Home

"user manual"

1. __ Minimum Maximum Probability Step Load Data gt 10 00 20 00 0 20 2000 30 00 0 10 4000 50 00 0 30 60 00 80 00 0 30 1 00 90 00 0 05 100 00 0 05 zi OK Cancel Enter Galley Conelate Help Figure A 38 Custom data from worksheet Crystal Ball User Manual 305 Appendix A Selecting and Using Probability Distributions Entering tables of data into custom distributions Follow the rules in this section for loading data Unweighted values Single values are values that don t define a range Each value stands alone For a series of single values with the same probabilities unweighted values use a one column format or more than five columns The values go in each cell and the relative probabilities are all assumed to be 1 0 Choose Parameters gt Unweighted Values to enter these TT oz pi Sheet1_ Sheet2 Z Sheet3 Figure A 39 Single values with the same probability G Define Assumption Cell A2 E jol x Edit View Parameters Preferences Help Name faz xj Custom Distribution 2 0904 3 a 0 60 3 0 30 cy oo 12 00 16 o0 20 bo 24 00 28 o0 32 00 Load Data ann zi OK Cancel _ Enter Galley _Conelte Help Figure A 40 Unweighted values loaded in a custom distribution 306 Crystal Ball User Manual Using the custom distribution Weighted values For a serie
2. Material 1 Strength Number of Coils Variable values Shearing Modulus of at forecast Elasticity eh Sa E Cox minimums and maximums Wire Diameter in Spring Deflection in Figure 8 13 Tornado chart The bars next to each variable represent the forecast value range across the variable tested as discussed above Next to the bars are the values of the variables that produced the greatest swing in the forecast values The bar colors indicate the direction of the relationship between the variables and the forecast For variables that have a positive effect on the forecast the upside of the variable shown in blue is to the right of the base case the initial value in the cell before running the simulation and the downside of the variable shown in red is to the left side of the base case For variables that have a reverse relationship with the forecast the bars are reversed When a variable s relationship with the forecast is not strictly increasing or decreasing it is called non monotonic In other words if the minimum or maximum values of the forecast range do not occur at the extreme endpoints of the testing range for the variable the variable has a non monotonic relationship with the forecast 212 Crystal Ball User Manual Tornado Chart tool maximum minimum Figure 8 14 A non monotonic variable If one or more variables are non monotonic all the variable bars are the same color al
3. Setting Effect View Sets the forecast window s display type For more information on the View settings see Changing the distribution view and interpreting statistics on page 115 Frequency Shows the number or frequency of values occurring in a given interval bin This is the default view Cumulative Shows the number or proportion percentage of values less than Frequency or equal to a given amount Reverse Shows the number or proportion percentage of values greater Cumulative than or equal to a given amount Frequency Statistics Shows a set of descriptive statistics for the forecast values Percentiles Shows percentile information in 10 increments Goodness Of Fit If distribution fitting is selected for the forecast chart shows goodness of fit statistics for the selected distribution and ranking method Capability If process capability metrics are set for display shows a table of Metrics process capability quality statistics for the simulation See Viewing capability metrics on page 344 for more information Split View When selected shows all selected views simultaneously For details see Using Split View on page 122 Window Sets whether to display the forecast window automatically while the simulation is running or when the simulation is stopped You can display one or more forecasts while the simulation is running If you decide not to display the forecast the simulation continues to ru
4. ccecee cece eee cee sss isesi siinses neresini 243 Appendix A Selecting and Using Probability Distributions Understanding probability distributions 0 0 c ccccccces cee ceteeseeteeteeeeeeens 246 A probability example siscccccsceccdccssskcceiasteescnsvevsntescsosteetneatersnsassessees sive 246 Discrete and continuous probability distributions 0ccceeee 249 Selecting a probability distribution cece eee eens ceeeeeee reese eeeeeneee 250 Using basi distributions i 2 cc scssssecsans esate desabteddengtbaaaestdavngetiveanesceass Soastes 252 Using continuous distributions 0cccceccccs cc cee cee ceeeeeeeeeeeseeeeenseeseens 253 Beta GistriD ution cise icspicesse ded dadteaisectviaee sas Segncassagocsstignevsteeres E 255 Exponential distribution 0 0 0 0 c cece eee ceeceseeeeteteteeteeceeceaeeseeaeeas 257 Gamma distribution also Erlang and chi square csceseeeee 259 LOPIStIG GIStrIDULLON iiss csvescsdes ios sscenssasedssdensessocestsnne nearr NEEE EN ESENE 262 Lognormal distribution ccc cece eeeeeeeaeeeeaeeecaeceseesneeeeeeeenaees 263 Maximum extreme distribution 0 0 0 ee cece eee eeeeceeeeceeee settee 265 Minimum extreme distribution eceeeceseescesseeseceneeeeeeeeeeseesneeeaeeees 267 Normal distribution 0 0 cece eece eee e reece eee seca aeteae ences seeeesneeeeeeeeees 268 Pareto Gis trib Win eicon AE EEE ERE 270 Crystal Ball User Manual Contents Student s tdistribution s
5. 1 Run a simulation in Crystal Ball To produce a meaningful trend chart the simulation should have more than one forecast 2 When the simulation stops choose Analyze gt Trend Charts The Trend Charts dialog appears If no trend charts have been created for the open workbooks or restored results files the dialog is blank as shown in Figure 6 10 lolx Click New or choose existing charts to Open Close or Delete Figure 6 10 Trend Charts dialog 3 To create a new trend chart click New Crystal Ball User Manual 155 Chapter 6 Analyzing Other Charts The Choose Forecasts dialog opens similar to Figure 6 11 chooserorecasts o Choose forecasts to display on the trend chart BLY Model Clu Ending Sales Year 1 Q1 Duly Ending Sales Year 1 Q2 Dll Ending Sales Year 1 Q3 Dolly Ending Sales Year 1 Q4 Duly Ending Sales Year 2 Q1 Dll Ending Sales Year 2 Q2 Dolly Ending Sales Year 2 Q3 Duly Ending Sales Year 2 Q4 Duly Ending Sales Year 3 Q1 Dolly Ending Sales Year 3 Q2 Dolly Ending Sales Year 3 Q3 Dll Ending Sales Year 3 Q4 Figure 6 11 Choose Forecasts dialog for trend charts Tree view 5 By default this dialog appears in a hierarchical Tree view If you prefer click the List button to change to List view 4 Select two or more forecasts to include in the trend chart To select a forecast check the box in front of its name For this example select all forecasts in the Sales Pr
6. Figure 8 23 Specify Target dialog Bootstrap tool Crystal Ball User Manual 221 Chapter 8 Crystal Ball Tools 222 Set the target by selecting Profit Or Loss from the forecast list Click Next The Specify Options Step 2 of 3 dialog appears as shown in Figure 8 24 Specify options step 2 of 3 Bootstrap Crystal Ball Tool Decisioneering 1998 2004 Figure 8 24 First Specify Options dialog Bootstrap tool Make sure the one simulation method and the statistics options are selected Click Next The Specify Options Step 3 of 3 dialog appears as shown in Figure 8 25 Specify options step 3 of 3 Bootstrap Crystal Ball Tool Decisioneering 1998 2004 Figure 8 25 Second Specify Options dialog Bootstrap tool Crystal Ball User Manual Bootstrap tool 7 Set the following options Number Of Bootstrap 200 e Number Of Trials Per Sample 500 Show Only Target Forecast is selected 8 Click Start The bootstrap tool displays a forecast chart of the distributions for each statistic and creates a workbook summarizing the data BES cS Tap 2 a 3 a F 1 E 5 s 2 Profit or Loss 3 4 Correlations 5 Mean 0 007 0 065 6 Median 0 049 0 368 7 Standard Deviation 1 000 1 000 0 053 8 Variance 1 000 0 053 9 Skewness 1 000 10 Kurtosis 44 naf anf Uariahilite C Forecast Mean Edit View Forecast Preferences H
7. To reset the original certainty range before beginning the next section 1 In the forecast chart either e Move the right certainty grabber until the certainty maximum field shows positive infinity or Type Infinity in the right field 2 Either e Move the left certainty grabber until the certainty minimum field shows negative infinity or Type Infinity in the left field The forecast chart displays a certainty level of 100 as displayed in the original forecast Crystal Ball User Manual 111 Chapter 5 Analyzing Forecast Charts Focusing on the display range With Crystal Ball you can focus on a particular range of the forecast results by changing axis settings in the Chart Preferences dialog For instructions see Customizing chart axes and axis labels on page 138 Using the Scale group of axis settings you can define the display range in the following ways Table 5 1 Chart Preferences gt Axis gt Scale gt Type settings Type Setting Effect Auto Crystal Ball uses a default display range of 2 6 standard deviations from the mean which includes about 99 of the forecast values See Standard Deviation following Fixed Sets the display range end points manually Using fixed end points for the display range lets you focus on particular value ranges For example you can focus on positive values only to look at the profit for a profit loss forecast Standard Deviation Set
8. c cece eee eee eee eens 33 Setting assumption Preferences 0 0 0 cscs eee cee ceteeeeteeeeeeeesseeeeaeenes 38 Additional Distribution Gallery features ccc eee cee ce cess ceseeeeeeeeeees 4 Displaying the Distribution Gallery 0 c cece eee cece cee eeeeeneeeneenees 42 The Distribution Gallery window 0ceccecec ce eee cee ce ceeeteeeeeeeeees 42 Managing distributions ccceeceee cee ceseeeeeeeseseseseeceeeeeeesseessseeeeeeeeenees 45 Crystal Ball User Manual i Contents Manacing Cate ories vasici irnn E TEE EE E E REEE 49 Chapter 3 Defining Other Model Elements Defining decision variable cells and forecast cells ieee ee eee 58 Defining decision variable Cells cceceeceeeesccsse cee eeseesseeseeceeeeeeeeaaes 58 Defining forecasts oiio esatean tebe aee aE e RE E E EEEE 59 Working with Crystal Ball data s sesnssessesenseessessssesstessneessessssenseesensensenne 69 Editing Crystal Ball data o iccctietiesecasieeacsanesstees sales innisensee sides 69 Selecting and reviewing your data oo cece eeeeeceeeee cee eeeeeeeeeeseeseeees 72 Setting cell preferences es ccsecistiescesstesaaeelitemtisdaveastieatitieeioehesvstessedens 74 Saving and restoring your models uu cece cesses cess eceeeseceeeseeeeseeeaeenseeees 76 Chapter 4 Running Simulations About Crystal Ball simulations 0 c cece ce cee cee cee ceneeeeeeesecseeseeeneseeeneee 78 How Crystal Ball uses Monte Carlo simul
9. precedents All cells within the active worksheet that are referenced as part of the formula or a sub formula of the target cell Crystal Ball Note Precedents for this tool are handled differently than standard Excel precedents in that they cannot trace beyond the active worksheet Therefore you can only use precedents on the active worksheet as input variables 5 Click Add Assumptions Available assumptions appear in the dialog as shown in Figure 8 17 Specify input variables step 2 of 3 Tornado Chart Crystal Ball Tool Decisioneering 1998 2004 Use the buttons to add or remove assumptions decision variables and precedents to the list Coil Diameter in Shearing Modulus of Elasticity Add Decision Variables Number of Coils Bde eine sateen Material 3 Strength addRenge Remove z Remove All Dee Tz All 8 assumptions added lt Back Next gt Figure 8 17 Specify Input Variables dialog Tornado Chart tool 6 Remove Material 2 Strength and Material 3 Strength a Select an assumption to remove b Click Remove c Repeat steps 6a and 6b for the second assumption to remove Crystal Ball User Manual 215 Chapter 8 Crystal Ball Tools The last two assumptions have no impact on the target forecast If you leave them in the list they will appear in the charts even though they are unrelated to the target forecast 7 Click Next The Specify Options Step 3 Of 3
10. 2 Type 10 in the Certainty field 3 Press Enter Crystal Ball moves the certainty grabbers to include 10 of the trials centered around the median statistic of the entire range Anchoring a grabber then entering certainty You can also anchor a certainty grabber type the certainty level and Crystal Ball moves the free grabber to correspond to the value range for the level 1 Move the left certainty grabber until the certainty minimum value is 0 0 The certainty grabber turns to a lighter color and is considered anchored 2 Type 75 in the Certainty field 3 Press Enter Crystal Ball moves the free certainty grabber toward the anchored certainty grabber to include 75 of the values as in Figure 5 4 Crystal Ball Note If both grabbers are free the distribution is centered on the median Crystal Ball User Manual Understanding and using forecast charts Forecast Net Profit MM E ioj x Edit View Forecast Preferences Help 500 Trials Frequency Chart 497 Displayed Net Profit MM 2 f 3 a 20 5 ine 10 0 0 0 10 0 Millions p gt soo Certainty 75 0 q 21 Figure 5 4 Certainty level 75 from left certainty grabber You also can cross over the certainty grabbers to determine the certainty level for the two tails ends You can determine the certainty level for specific value ranges at any time either during or after the simulation Resetting the certainty range
11. Forecast Window Precision Filter Auto Extract View Window IV Show automatically While running simulation When simulation stops Fit distribution 7 J Fit a continuous probability distribution to the forecast Choose distributions Cancel Apply To Defauts Help Figure 5 18 The Forecast Preferences dialog Click a tab and set preferences as required These tabs also appear in the expanded Define Forecast dialog For details see the following sections or click the Help button in the Forecast Preferences dialog e Forecast Window preferences on page 62 manage window display and distribution fitting for the forecast e Precision preferences on page 64 manage precision control settings e Filter preferences on page 66 let you discard values inside or outside a range for the current forecast e Auto Extract preferences on page 68 let you specify which statistics to extract automatically to Excel when a simulation stops Also see the previous section Setting forecast preferences on page 125 Crystal Ball Note For more information on setting the Precision Control option and settings see Precision control on page 318 See Confidence intervals in the Statistical Definitions chapter of the online Crystal Ball Reference Manual for more information about how absolute and relative precision relate to the
12. v ort Summary t E Report GA Forecasts FR All M Assumptions IV Run date time Choose Decision Variables C One v eferences Open M Overay Charts M Runpn p M Trend Charts IV Run statistics Sensitivity Charts Figure 7 4 Custom Report dialog Check one or more items in the Report Sections group to include in the report See Report sections on page 186 for more information Crystal Ball Note Ifyou have activated the process capability features and have generated capability metrics you can include them in your custom report For details see Including capability metrics in reports on page 351 For each item checked in Report Sections define the report further by checking settings in the Details group Crystal Ball User Manual 185 Chapter 7 Creating Reports and Extracting Data As each item is highlighted in the Report Sections group appropriate settings appear in the Details group For descriptions of these see Report details beginning on page 187 For each item checked in Report Sections choose whether to display all of that type of item only selected items or all open items As each item is highlighted in the Report Sections group appropriate display settings appear in the group at the upper right of the dialog This group is labeled with the name of the current selection in the Report Sections group If you choose Choose a dialog appears so you can check the box in fr
13. Each adjustment changes the characteristics of the probability distribution For example the truncated normal distribution in Figure A 47 will no longer have an actual mean of 100 000 and standard deviation of 15 000 Also statistics values will be approximate for truncated distributions When using alternate percentile parameters the actual percentiles calculated for a truncated distribution will differ from the specified parameter values For example a normal distribution specified with 10 90 percentiles and truncated on either side of the distribution will have actual 10 90 percentiles greater or less than the specified percentiles a Crystal Ball Note Showing the mean line ofthe distribution is useful when truncating distributions However the mean line value might differ from the Mean parameter field The mean line shows the actual mean of the truncated distribution while the Mean parameter field shows the mean of the complete distribution Crystal Ball User Manual Comparing the distributions Define Assumption Cell A1 E E fe o x Edit view Parameters Preferences Help Name fan xj Normal Distribution Mean 102 707 11 60 000 00 80 000 00 100 000 00 120 000 00 140 000 00 gt 80 000 00 xj xtnfinity xj Mean 100 000 00 4 J Std Dev s15 000 00 H3 OK Cancel Enter_ Gallery Correlate Help Figure A 47 Truncated distribution example Comparing the distributions
14. Index Correlation Matrix 207 Decision Table 227 Scenario Analysis 231 Tornado Chart 214 Two dimensional Simulation 237 Excel charts pasting to 141 Excel functions and distributions 315 expanded Define Forecast dialog 61 exponential distribution defined 257 extenders see tools 198 extract data 190 extreme distributions defined 265 267 F Filter preferences 66 fitting distributions overview 27 forecast cells defined 18 forecast charts distribution types 115 format 113 groups 127 forecast charts analyzing 104 Forecast Preference dialog 61 forecast preferences 61 Forecast Preferences dialog Auto Extract tab 68 Filter tab 66 Forecast Window tab 62 Precision tab 64 Forecast Window preferences 62 forecasts defining 59 fitting distributions 27 percentiles view 118 selecting and reviewing 72 statistics view 118 understanding 106 forecasts defined 12 formulas in parameters 25 freezing assumptions 88 gamma distribution defined 259 376 Crystal Ball User Manual geometric distribution defined 285 global macros 102 goodness of fit tests 27 grouped data 302 groups 127 guidelines analyzing simulation results 104 H K help online 5 hot keys for charts 129 how this manual is organized 2 hypergeometric distribution defined 286 insufficient memory warning 87 Kolmogorov Smirnov test 29 122 L Latin hypercube sampling accuracy 318 logistic distribution defined 262 lognormal distribution defined 2
15. The number of possible occurrences in any unit of measurement is not limited to a fixed number The occurrences are independent The number of occurrences in one unit of measurement does not affect the number of occurrences in other units The average number of occurrences must remain the same from unit to unit Description The gamma distribution applies to a wide range of physical quantities and is related to other distributions lognormal exponential Pascal Erlang Poisson and chi square It is used in meteorological processes to represent pollutant concentrations and precipitation quantities The gamma distribution is also used to measure the time between the occurrence of events when the event process is not completely random Other applications of the gamma distribution include inventory control economics theory and insurance risk theory Example one A computer dealership knows that the lead time for re ordering their most popular computer system is 4 weeks Based upon an average demand of 1 unit per day the dealership wants to model the number of business days it will take to sell 20 systems Crystal Ball User Manual 259 Appendix A Selecting and Using Probability Distributions 260 Checking the conditions of the gamma distribution The number of possible customers demanding to buy the computer system is unlimited e The decisions of customers to buy the system are independent The demand remains
16. e You can choose Overlay gt Fit Probability Distributions in the overlay chart menubar to do a quick fit with the default or currently selected distributions and ranking method You can also use this command to switch off distribution fitting that is set with either the Overlay menu or Preferences menu You can choose Preferences gt Overlay gt Overlay Window in the overlay chart menubar to specify particular distributions and to choose one of three fit ranking methods This way also lets you change the fit options or use Apply To to set these preferences for other overlay charts To fit a probability distribution to all forecasts in an overlay chart using the Preferences gt Overlay command 1 Follow the steps for forecast charts given in Fitting a distribution to a forecast on page 120 Wherever the instructions say Forecast as in Preferences gt Forecast substitute Overlay 2 Click OK to perform the fit Crystal Ball fits the distributions and then displays a probability distribution for each forecast as shown in Figure 6 8 Crystal Ball User Manual Understanding and using overlay charts atl W F d Figure 6 8 Overlay chart with forecasts and lines of best fit As the legend shows the forecast in front fits best to a normal distribution while the other two are lognormal fits The Series tab of the Chart Preferences dialog was used to change the colors of the middle distribution and the li
17. Appendix A Selecting and Using Probability Distributions Example A lawn mower company is testing its gas powered self propelled lawn mowers They run 20 mowers and keep track of how many hours each mower runs until its first breakdown They use a Weibull distribution to describe the number of hours until the first failure Using discrete distributions Discrete probability distributions describe distinct values usually integers with no intermediate values and are shown as a series of vertical bars such as the binomial distribution at the bottom of Figure A 4 on page 248 A discrete distribution for example might describe the number of heads in four flips of a coin as 0 1 2 3 or 4 The following discrete distributions are described later in this section in alphabetical order Page references appear below the names Table A 3 Summary of discrete distributions Shape Name Summary Binomial The binomial distribution describes the number of times a particular Ili page 281 event occurs in a fixed number of trials such as the number of heads in Binomial 10 flips of a coin or the number of defective items in 50 items Discrete uniform In the discrete uniform distribution all integer values between the WHI Basic minimum and maximum are equally likely to occur It is the discrete equivalent of the continuous uniform distribution Discrete Uniform page 283 Geometric The geometric distribution describe
18. Figure 2 15 Correlation chart The points on the chart represent the pairing of assumption values as they would actually occur when running a simulation The solid line running through the middle of the chart indicates the location where values of a perfect correlation 1 0 or 1 0 would fall The closer the points are to the solid line the stronger the correlation In the example above an Inflation Rate assumption and an Oil Price Barrel assumption have been correlated using a coefficient of 0 8 a strong positive correlation As the points on the chart show higher inflation values tend to be associated with higher oil prices and vice versa This chart can help you begin to understand how the two assumptions are related You can specify as many of these paired correlations as you want for each assumption up to the total number of assumptions defined in a workbook You can generally ignore correlations between variables if one or both variables do not impact the output or are not highly correlated Setting assumption preferences The Define Assumption dialog has a Preferences menu in the menubar This menu has the following main options 38 Crystal Ball User Manual Additional assumption features Table 2 1 Preferences menu Define Assumptions dialog Setting Effect Assumption Preferences Manage window display during simulations Chart Preferences Determine the appearance of the assumption chart The Chart Preferen
19. Smaller bands always appear on top of larger bands This obscures the larger bands Don t confuse the actual width of a band with the portion that is visible You can display the true size of a band using the certainty band preferences described in Setting certainty bands on page 162 These preferences let you display the bands one at a time on the trend chart if you wish To change the placement of the certainty bands 1 In the trend chart open the View menu 2 Choose one of the views listed in Table 6 1 following Crystal Ball User Manual 157 Chapter 6 Analyzing Other Charts Table 6 1 Trend chart views View Effect Centered On Median Cumulative The default displays forecasts centered around the median of each forecast value Displays the forecasts anchored at the lower end of the forecast range shows the certainty that the forecast values will be at or below a given value cumulative probability Reverse Cumulative Displays the forecasts anchored at the higher end of the forecast range shows the certainty that the forecast values will be at or above a given value reverse cumulative probability Alternately you can choose Preferences gt Trend and choose a view in the View list of the Trend Preferences dialog Crystal Ball Note Use a Crystal Ball hot key to bypass the View menu Each time you press Ctrl d the next trend chart view appears Spec
20. They are entered and stored in assumption cells For more information on assumptions and probability distributions see About assumptions and probability distributions beginning on page 18 Forecasts Since all those scenarios produce associated results Crystal Ball also keeps track of the forecasts for each scenario These are important outputs of the model such as totals net profit or gross expenses They are defined with formulas in spreadsheet forecast cells For each forecast Crystal Ball remembers the cell value for all the trials scenarios If you run a simulation at Demo speed you can watch histograms of the results calculated for each forecast cell and can see how the results stabilize toward a smooth frequency distribution as the simulation progresses After hundreds or thousands of trials you can view sets of values the statistics of the results such as the mean forecast value and the certainty of any particular value Chapter 5 gives more information about charts of forecast results and how to interpret them Certainty The forecast results show you not only the different result values for each forecast but also the probability of obtaining any value Crystal Ball normalizes these probabilities to calculate another important number the certainty The chance of any forecast value falling between Infinity and Infinity is always 100 However the chance or certainty of that same forecast being at le
21. To terminate the simulation click Cancel If small adjustments to the correlation coefficients are not possible the simulation stops with an error message prompting you to reexamine your coefficients To solve the problem you can usually make large coefficients smaller or change your spreadsheet model to use formulas to calculate assumptions with large correlation coefficients Creating a report containing just your assumptions might make it easier to spot problems 324 Crystal Ball User Manual Using the Extreme Speed Feature In this appendix Overview e Compatibility issues Other important differences e Numerical differences e Maximizing the benefits of Extreme Speed This appendix describes Extreme speed a Speed preference available in Crystal Ball Professional and Premium editions that runs simulations up to 100 times faster than Normal speed Crystal Ball User Manual 325 Appendix C Using the Extreme Speed Feature Overview Extreme speed available only in Crystal Ball Professional and Premium editions runs simulations up to 100 times faster than Normal speed With Extreme speed you won t need to wait for simulation results or compromise on the number of trials you run Extreme speed makes it more practical to use CB Tools that run multi simulations or to seek optimal solutions with OptQuest in a reasonable length of time The Extreme Speed feature uses PSI Technology a high speed Excel compatible Polym
22. compatibility issues Appendix D Using the Process Capability Features Discusses the process capability features that can be activated to support Six Sigma DFSS Lean principles and similar quality programs Crystal Ball User Manual 3 Introduction Bibliography Lists related publications including statistics textbooks e Glossary Defines terms specific to Crystal Ball and other statistical terms used in this manual Index Lists subjects alphabetically with corresponding page numbers Conventions used in this manual This manual uses the following conventions Text separated by gt symbols means that you select menu options in the sequence shown starting from the left The following example means that you select the Exit option from the File menu 1 Select File gt Exit Steps with attached icons mean that you can click the icon instead of manually selecting the menu options in the text For example La 2 Select Define gt Define Assumption Notes provide additional information expanding on the text There are four categories of notes Crystal Ball Note Notes that provide additional directions or information about using Crystal Ball Excel Note Notes that provide additional information about using the program with Microsoft Excel OptQuest Note Notes that provide additional directions or information about using OptQuest Statistical Note Notes that provide additional
23. decision variables and forecasts from the copied range into the range selected in step 1 Any existing Crystal Ball data in the range selected for pasting will be overwritten Crystal Ball User Manual Working with Crystal Ball data Crystal Ball Note If there are more Crystal Ball data cells in the copied range than there are appropriate cells in the range selected for pasting as many cells as possible of the appropriate type will be pasted the remaining cells are ignored If there are more cells in the range selected for pasting than in the range that was copied the copied cells are reused starting with the first one in the list Note that when you copy a range Crystal Ball remembers the area covered by the range but not the contents of the range When you paste you paste all Crystal Ball data of the copied type or types currently within that range regardless of what was there when you copied it For example suppose you have two assumption cells in a range You use the Crystal Ball Copy Data command to copy all types of data cells in the range Then you delete one assumption modify the other and add a forecast definition to another cell within the range When you use the Crystal Ball Paste Data command you then paste the current contents of the range the remaining modified assumption cell and the new forecast cell For best results use the Paste Data command immediately after the Copy Data command You are probably
24. dialog appears as shown in Figure 8 18 Specify options step 3 of 3 Tornado Chart Crystal Ball Tool Decisioneering 1998 2004 Tornado Input __ Tornado Method Testing range 10 to 90 x Percentiles of the variables Percentage deviations Testing points E m from the base case or Base Case Tornado Output ci I Tornado chart Use median values IV Spider chart Show jz top variables cme o Figure 8 18 Specify Options dialog Tornado Chart tool 8 Set the following options as shown in Figure 8 18 e Testing Range 10 to 90 e Testing Points 5 e For Base Case Use Existing Cell Values e Tornado Method Percentiles Of The Variables Tornado Output Tornado Chart and Spider Chart e Show __ Top Variables 20 9 Click Start The tool creates the tornado and spider charts in their own workbooks with data tables as shown in Figure 8 19 on page 217 and Figure 8 20 on page 218 CB Tools Note For more information about the Tornado Chart dialogs click Help in a dialog 216 Crystal Ball User Manual Tornado Chart tool Interpreting the results In this example six assumptions are listed in the tornado chart The first assumption Material 1 Strength has the highest sensitivity ranking and is the most important A researcher running this model would investigate this assumption further in the hopes of reducing its uncertainty and therefore its
25. effect on the target forecast Material 1 Reliability The last two assumptions Wire Diameter and Spring Deflection are the least influential assumptions Since their effects on the Material 1 Reliability are very small you might ignore their uncertainty or eliminate them from the spreadsheet Material 1 Reliability 1 05 1 10 wateals Sven 16 00 7 20 coi iameerin o reos A 22 Number of Coils Shearing Modulus of Elasticity 1 18E 07 1 12E 07 Wire Diameter in Spring Deflection in Material 1 Reliability Variable Downside Upside Range Downside Upside Base Case Material 1 Strength 46 800 53 200 Coil Diameter in 0 7808 0 8192 Number of Coils 19 41 20 59 Shearing Modulus of Elasticity 1 12E 07 1 18E 07 Wire Diameter in 0 1481 0 1519 Deflection in 0 94 0 96 Figure 8 19 Tornado chart with data The spider chart shows similar information Material 1 at the top of the tornado chart has the steepest positive slope in the spider chart Notice that more reliability values are given one for each of five levels within the testing range Crystal Ball User Manual 217 Chapter 8 Crystal Ball Tools e Material 1 Strength s Coil Diameter in Number of Coils Shearing Modulus of Elasticity Wire Diameter in 1 00 10 0 30 0 50 0 70 0 90 0 Percentiles of the variables e Spring Deflection in Mat
26. et al Numerical Recipes in C 2nd Ed Cambridge England Cambridge University Press 1993 Odeh R E and J O Evans Percentage Points of the Normal Distribution Applied Statistics London Royal Statistical Society 1974 Sedgewich Robert Algorithms Reading MA Addison Wesley Publishing Co 1983 Specific distributions Extreme value distribution Castillo Enrique Extreme Value Theory in Engineering London Academic Press 1988 Lognormal distribution Aitchison J and J A Brown The Lognormal Distribution New York Cambridge University Press 1973 Weibull distribution King James R Probability Charts for Decision Making Rev Ed New York Industrial Press Inc 1981 Henley Ernest J and Hiromitsu Kumamoto Reliability Engineering and Risk Assessment Englewood Cliffs NJ Prentice Hall Inc 1981 Tornado charts and sensitivity analysis Clemen Robert T Making Hard Decisions An Introduction to Decision Analysis 2nd Ed Belmont CA Duxbury Press 1997 360 Crystal Ball User Manual Morgan M Granger and Max Henrion with a chapter by Mitchell Small Uncertainty A Guide to Dealing with Uncertainty in Quantitative Risk and Policy Analysis New York Cambridge University Press 1990 Two Dimensional simulation Burmaster David E and Andrew M Wilson An Introduction to Second Order Random Variables in Human Health Risk Assessments Human and Ecological Risk Assessment 2 4
27. gt General gt 3D Chart Preferences gt Chart Preferences gt Chart Type gt Marker Lines gt lt central tendencies gt Preferences gt Chart Preferences gt General gt Legend Preferences gt Chart Preferences gt Chart Type gt Marker Lines gt Percentiles View menu Preferences gt lt chartname gt Preferences Cycles through chart views Frequency Cumulative Frequency Reverse Cumulative Frequency for assumption and forecast charts Cycles through bin or group interval values to adjust the number of columns or data points Cycles through gridline settings None Horizontal Vertical Both Cycles through chart types Area Line Column for sensitivity charts Bar direction Bar magnitude Pie in Contribution To Variance view Cycles between two dimensional and three dimensional chart display Cycles through central tendency marker lines None Mean Median Mode except for sensitivity and trend charts Toggles the legend display on and off Cycles through percentile marker lines None 10 20 90 Cycles through window views Chart Statistics Percentiles Goodness Of Fit if distribution fitting is selected except for trend charts Crystal Ball User Manual 129 Chapter 5 Analyzing Forecast Charts Basic customization instructions These instructions apply most specifically to forecast charts However many of them apply to other charts as well For that reas
28. random number A mathematically selected value which is generated by a formula or selected from a table to conform to a probability distribution Glossary Term random number generator A method implemented in a computer program that is capable of producing a series of independent random numbers Most real world problems involving elements of uncertainty are too complex to solve analytically There are simply too many combinations of input values to calculate every possible result Monte Carlo simulation is an efficient technique that requires only a random number table or a random number generator on a computer Crystal Ball implements Monte Carlo simulation in a repetitive three step process For each trial of a simulation Crystal Ball repeats the following three steps 1 For every assumption cell Crystal Ball generates a random number according to the probability distribution you defined and places it into the spreadsheet 2 Crystal Ball recalculates the spreadsheet 3 Crystal Ball then retrieves a value from every forecast cell and adds it to the chart in the forecast windows This is an iterative process that continues until either The simulation reaches a stopping criterion e You stop the simulation manually 78 Crystal Ball User Manual Generate random numbers for assumption cells Calculate entire spreadsheet Display results in a forecast chart About Crysta
29. 000 falls between 3 000 and 7 000 forming a triangle These conditions match those of the triangular distribution The triangular distribution has three parameters Minimum Likeliest and Maximum The conditions outlined in this example contain the values for these parameters 3 000 Minimum 5 000 Likeliest and 7 000 Maximum Crystal Ball User Manual Using continuous distributions You would enter these values as the parameters of the triangular distribution in Crystal Ball The following triangular distribution shows the probability of x number of gallons being sold per week oix Edit View Parameters Preferences Help Name Jan xj y Triangular Distribution 3000 3500 4000 4500 5000 5500 6000 6500 7 000 Minimum 3 000 Likeliest E 000 Maximum 7 000 xj oK Cancel Enter_ Galley Correlate Help Figure A 16 Triangular distribution Example two The triangular distribution also could be used to approximate a computer controlled inventory situation The computer is programmed to keep an ideal supply of 25 items on the shelf not to let inventory ever drop below 10 items and not to let it ever rise above 30 items Check the triangular distribution conditions The minimum inventory is 10 items The maximum inventory is 30 items The ideal level most frequently on the shelf is 25 items These conditions match those of the triangular distribution The result would be a dist
30. 115 Using Split View scenici steel an iE R ov E EE 122 Setting forecast preferences sssssssrseisisesssissirsrneistssrrsissiesrnneneensent 125 Setting forecast chart preferences sssssssssessirersessresirssrissrrersristent 127 Setting chart preference sssini ireas cesses iee inen aan iaaa KENS 128 Available settings lt issc sici2 cstiacreszecstaaseciticehiidssntbsnabeleasnaravancaineetieeaas 128 Setting preferences with hot keys cccccccccceeseeeeceeeee ete eneeeseeseesee 129 Basic Customization instructions cee eee cee cee eee eeeseseeceeeeeeeeeeeseees 130 Specific Customization INStrUCTIONS 2 0 eee eee cee cee e ces eeee cee eeeeeteeeeaeees 131 Managing existing Charts cccce eee eseeee cee eeeeeceseceeceeeeeeeseecasseesneeeeaeed 140 Opening Ghats ie is cuca teacs cassis don vansooeboveatveuatsantaiseasntegineDbventecestededovenss 140 Copying and pasting charts to other applications 0 ceeeeeees 141 Printing charts fc deccnstegevnics civ i nern E Aar Ena a a EE r tas 142 Closing charts ssm ensne o a EE E E E E O E 142 Deleting ChartSissraimesisisr ssnin eene eies ES En era a Arra FEENS 143 Chapter 6 Analyzing Other Charts CVETVIEW ara i AE AOE EEEE A ETORO 146 Understanding and using overlay charts snsssssesssenseresnstrnsrttsrsntseneenene 146 Creating an overlay chart s sssssssisssssrssisisstssisrinsrnristesrsrnsinsensentessent 147 Customizing overlay Charts cccccc cec
31. 2 17 The Distribution Gallery window Distribution Gallery Cell C5 Edit View Categories Help Menubar Category pane with category folders Aa A Lognormal Distribution pane Description pane 42 Crystal Ball User Manual Discrete Uniform The parameters for the nomal distribution are mean and standard deviation cma st eo Figure 2 17 The Distribution Gallery window 10l xl Additional Distribution Gallery features As shown in Figure 2 17 the Distribution Gallery has a menubar a category pane with folders containing distributions a distribution pane that displays all distributions in the selected category and a description pane that describes the selected distribution The following sections explain each part of the Distribution Gallery The Distribution Gallery menubar and buttons The Distribution Gallery menubar has the menus summarized in Table 2 3 Table 2 3 Distribution Gallery menus Menu Command Summary Edit Commands that let you copy paste modify and delete distributions You can copy from any category you are allowed to use but you can only paste modify and delete within the Favorites category or new categories you or others have created You cannot modify or delete distributions in the Basic and All categories these categories are reserved only for unmodified distributions shipped with Crystal Ball Categories Commands that le
32. 306 Changes from Crystal Ball 2000 x 5 x eeee eee ee eeeeeee crete eeseeneees 311 Other important custom distribution notes ccce eee eee eeeeeneeens 311 Truncating distrib tor ierra ei iisi 312 Beawatene srerriediotioii iroda aaa ested E E R A ALEE R e AAS 312 Comparing the distributions ccc cece eeecsseeeeceeeeeeeeeecneeeeeeeeeel 313 Using probability functions ve scisccccecseccsssacasivacesseaaesasiesnesseinasssaiaestsseeasaesseess 315 Limitations of probability functions 2 0 0 0 ee cece eeee cee ee etree eeees 316 Probability functions and random seeds ccceeeeeeese ees eeeeeeeeeeseens 316 Appendix B Maximizing Your Use of Crystal Ball SimUlation ACCULACY cscs ca cadiedd atesetivetesssacntecinths sanesbnatian devia device annaa 318 Precision contol esogeni oiia a E A Saadedededaess 318 Sampling Method siveccictaccdstiess cee ates aenitton dean densis 321 Simulation Speed shssiissccsiarniaactseaatiesttsstasdianerecterenstiacbeaacentaagenn TAAR 321 Sample sienna O E 323 Correlated assumption Ss essiri ssns sssrini iisen 323 Appendix C Using the Extreme Speed Feature OVERVIEW aieiaa EE A OTA E AA N ARA 326 Compatibility ISSUES sieesi veicceteaseett iatidaadeaensdesaieedaaided sania nansa ainean 326 Multiple workbook models ssssssssssssssiseiessitsisirtistsrersinrisrsinnsinensenent 327 Circular referentes cies daddencdcctsedacises dacdesuneadeessaseess sande aa a a aaa 327 Crystal Ball Excel functions
33. 4 seconds per recalculation cycle will take up to an hour to simulate 1 000 trials Greater numbers of assumptions and forecasts slow the simulation especially if the assumptions and forecasts are scattered across many spreadsheets in your model Start by examining the structure and nature of your model to locate possible efficiencies You can also use the sensitivity feature or the Tornado Chart tool to determine which assumptions contribute the least amount of uncertainty to your most important forecasts Freeze or eliminate the least important assumptions from the simulation Correlated assumptions can also consume a significant amount of processing time the time grows geometrically as the number of correlated assumptions increases Reduce the use of other applications Quitting other applications and closing or minimizing windows can be helpful in reducing overhead and increasing simulation speed Increase your system s RAM The amount of RAM in your computer has a large effect on the speed of simulations Modern operating systems give applications such as spreadsheets the appearance of additional RAM through the use of virtual memory Virtual memory lets you run a greater number of applications than would otherwise be possible but slows down overall processing speed because the system is frequently accessing the hard drive If you hear your hard disk being used during a simulation there might not be enough RAM to hold all par
34. 6 1 An overlay chart with 3D formatting After the simulation stops you can also use the overlay chart to fit probability distributions to the forecasts This process is similar to the distribution fitting feature described in Chapter 2 except that the fit is applied to forecast data not historical data Creating an overlay chart The example model used in this section Reliability xls shows the reliability of a design component when it is manufactured from different materials and is subjected to varying stresses Each forecast chart displays the design component s reliability as a ratio of two distributions strength and stress A reliable component has values greater than 1 because the component s strength exceeds the stress subjected to it An unreliable component has values less than 1 because its strength is less than the stress For this model the overlay chart can be used to compare component reliabilities To create an overlay chart 1 Runa simulation in Crystal Ball To produce a meaningful overlay chart the simulation should have more than one forecast Reliability xls is used in these instructions AM 2 Choose Analyze gt Overlay Charts Crystal Ball User Manual 147 Chapter 6 Analyzing Other Charts 148 The Overlay Charts dialog appears If you have not yet created any overlay charts for open workbooks or restored results files the dialog is blank as shown in Figure 6 2 Ps overaycharts o
35. Ball Note There is no absolute limit to the nwmber of assumptions you can define per worksheet In general you should define less than 1000 assumptions decision variables and forecasts per worksheet The Expanded Define Forecast dialog also has the Defaults button Clicking on Defaults restores the original default settings in place of any new settings you have made You can also click Apply To to use the settings in other charts and worksheets For more information click the Help button in the Apply To dialog Setting forecast preferences The forecast preference settings appear when you click the More button in the Define Forecast dialog box You can also display them by choosing Preferences gt Forecast in a forecast chart menubar The tabs in this dialog control several important aspects of Crystal Ball e Forecast Window preferences on page 62 window display and distribution fitting for the forecast e Precision preferences on page 64 precision control settings e Filter preferences on page 66 value filtering which discards values inside or outside a range for the current forecast e Auto Extract preferences on page 68 automatic data extraction to Excel when a simulation stops Crystal Ball Note For more information on setting the Precision Control option and settings see Precision control on page 318 See Confidence intervals in the Statistical Definitions chapter of th
36. Control Panel Statistics tab with frozen items X XXX at the end if items were frozen Second section is as follows User macros executed lt workbook1 sheet 1 gt CBBeforeSimulation lt workbook1 sheetl gt CBAfterTrial etc Forecasts Summary Chart And Size Default 100 for all charts Statistics Percentiles Selected in the Percentiles dialog the default is Deciles Assumptions Parameters Lists the parameters used to define the distribution Chart And Size Default 50 for all charts Statistics Theoretical and empirical if saved during the simulation Percentiles Selected in the Percentiles dialog the default is Deciles Correlations 188 Crystal Ball User Manual Creating reports Table 7 2 Custom report sections and details Continued Section Detail Description Decision Variables Type Continuous or Discrete Step Size Bounds Overlay Charts Chart And Size Default 100 for all charts Trend Charts Chart And Size Default 100 for all charts Sensitivity Charts Chart And Size Default 100 for all charts Crystal Ball Note When no details are selected for a custom report section only a single row is output with the Crystal Ball item name and the cell reference Report processing notes The following are special notes concerning Crystal Ball reports No section is created if there are no Crystal Ball items for that section Statistics follow the chart by default The Choose
37. Excel file in which you work and store your data A worksheet is part of a workbook Crystal Ball User Manual 371 Glossary 372 Crystal Ball User Manual In this index A comprehensive index designed to give you quick access to the information in this manual Crystal Ball User Manual 373 Index A absolute cell references 24 25 alternate parameters 25 255 analyzing simulation results guidelines 104 Anderson Darling test 29 121 Apply To 139 applying settings 139 assumption preferences 38 assumption cells defined 18 assumption parameters alternate 25 Assumption Preferences dialog 39 assumptions correlated 33 defining 19 entering 19 freezing 88 selecting and reviewing 72 Auto Extract preferences 68 B Batch Fit example 199 tool 199 beta distribution defined 255 bibliography 360 binomial distribution defined 281 Bootstrap example 221 tool 219 bootstrap methods 219 C capability metrics 340 categories creating distribution 50 deleting distribution 51 rearranging distribution 51 sharing distribution 52 53 54 55 viewing and editing distribution properties 51 categories managing distribution 49 Category pane Distribution Gallery 44 caveats tornado and spider charts 218 cell preferences 74 374 Crystal Ball User Manual cell references 24 absolute 25 dynamic and static 24 relative 25 certainty levels forecast chart 107 trend charts 162 certainty defined 12 chart preference
38. Ifyou try to create a sensitivity chart but Store Assumption Values For Sensitivity Analysis was not checked in the Run Preferences dialog check it and then reset the simulation and run the simulation again In this example there are four assumptions listed in the sensitivity chart The first assumption Volume Of Water Per Day accounts for approximately 65 of the variance in forecast values and can be considered the most important assumption in the model A researcher running this model would want to investigate this assumption further in the hopes of reducing its uncertainty and therefore its effect on the target forecast The last assumption Concentration Of Contaminant In Water contributes the least to forecast variance about 2 In fact this assumption has such a small Crystal Ball User Manual 169 Chapter 6 Analyzing Other Charts effect it could be ignored or altogether eliminated by clearing it from the spreadsheet Sensitivity charts like this one illustrate that one or two assumptions typically have a dominant effect on the uncertainty of a forecast How Crystal Ball calculates sensitivity Glossary Term rank correlation A method whereby assumption values are replaced with their ranking from lowest value to highest value using the integers 1 to N prior to computing the correlation coefficient Crystal Ball calculates sensitivity by computing rank correlation coefficients between every assumption and ev
39. Integer Your code is here End Function Public Function CBAfterSimulation _ aIterations As Long aTrials As Long aMaxTrials As Long aProgress As Double _ aWasAborted As Boolean _ aWasStopOnError As Boolean aWasStoppedOnPrecision As Boolean aHasNewBestSolution As Boolean As Integer Number of the simulation in a multi simulation run usually 1 Number of trials run Number of trials scheduled to run trials completed True if manually stopped True if stopped on error True if stopped on precision control New value of objective Your code is here End Function Public Function CBBeforeTrial aTrial As Long As Integer Your code is here End Function Public Function CBAfterTrial aTrial As Long As Integer Your code is here End Function Public Function CBAfterRecalc aTrial As Long As Integer Your code is here End Function Crystal Ball Note Returning anything other than a zero from a macro will stop the simulation Priority rules The following rules govern the running order of macros e If multiple workbooks are open Crystal Ball searches the workbooks in the order of their priority The priority number is set by the Crystal Ball User Manual 101 Chapter 4 Running Simulations 102 CB SetCBWorkbookPriority macro If this macro has not been called on the workbook no workbook order is guaranteed Crystal Ball Note CB SetCBWorkbookPriority sets the running order for macros when multip
40. Preferences Trials Sampling Speed Options m Calculate percentiles as Probability below a value Probability above a value 10 90 etc P10 P90 etc IV Calculate capability metrics Options Figure D 1 Statistics tab of the Run Preferences dialog Check Calculate Capability Metrics 4 Click the Options button to set the capability calculation options described in the next section 5 Click OK when setting changes are complete Setting capability calculation options Once you activate the process capability features you can set a variety of options to further customize these features for your situation To set the process capability options 1 Display the Statistics tab of the Run Preferences dialog 2 Click the Options button as described in step 5 of the previous section Crystal Ball User Manual 341 Appendix D Using the Process Capability Features 342 The Capability Options dialog appears Capability Options E xj M Short vs long term Use short term metrics C Use long term metrics Z score shift fi 5 Calculation method Calculate metrics from normal distribution Significance level 0 05 If normality test fails Calculate from forecast values Calculate from best fitting distribution Calculate metrics from forecast values always Cancel Help Figure D 2 Capability Options dialog 3 Indicate whether metrics should
41. Preferences Help Name Jan xj y Normal Distribution 8 00 10 00 Figure A 13 Normal distribution The distribution in Figure A 13 shows the probability of the inflation rate being a particular percentage Pareto distribution Parameters Location Shape Description The Pareto distribution is widely used for the investigation of distributions associated with such empirical phenomena as city population sizes the occurrence of natural resources the size of companies personal incomes stock price fluctuations and error clustering in communication circuits Pareto 270 Crystal Ball User Manual Using continuous distributions P Define Assumption CeMAL Lo Edit View Parameters Preferences Help Name Jan EF Pareto Distribution 2 F Figure A 14 Pareto distribution Calculating parameters There are two standard parameters for the Pareto distribution Location and Shape The Location parameter is the lower bound for the variable After you select the Location parameter you can estimate the Shape parameter The Shape parameter is a number greater than 0 usually greater than 1 The larger the Shape parameter the smaller the variance and the thicker the right tail of the distribution appears To calculate a more exact shape you can estimate the mean and use the equation for shapes greater than 1 mean B L L B 1 where B is the Shape parameter and L is the Locat
42. Statistical Definitions chapter of the online Crystal Ball Reference Manual Crystal Ball Note Crystal Ball uses Spearman rank correlation to calculate correlation coefficients For more information on how Crystal Ball calculates Spearman rank correlation coefficients see Rank correlation in the Statistical Definitions chapter of the online Crystal Ball Reference Manual When defining correlations the more correlations you define the greater the possibility that some correlations might be in conflict with each other preventing Crystal Ball from running a simulation Conflicts can arise when a group of assumptions are improperly related to each other by large positive and or large negative correlation coefficients When this condition occurs the correlations are said to be inconsistent For more information see Correlated assumptions on page 323 With Crystal Ball you use the Define Correlation dialog to specify a correlation coefficient for any pair of assumptions in the same workbook To define a correlation coefficient 1 Select the cell of one of the assumptions you want to correlate for example a cell describing the inflation rate Crystal Ball User Manual 33 Chapter 2 Defining Model Assumptions Statistical Note Which of the pair you select first is not important since the correlation coefficient is bidirectional La 2 Choose Define gt Define Assumption The distribution previo
43. Student s distribution has an additional parameter than controls the shape of the distribution Degrees of Freedom over the normal distribution the greater flexibility of the Student s distribution is sometimes preferred for more precise modeling of nearly normal quantities found in many econometric and financial applications The default parameters for the Student s distribution are Midpoint Scale and Degrees of Freedom Crystal Ball User Manual Using continuous distributions Define Assumption Cell Ai E lol x Edit View Parameters Preferences Help Student s t Distribution Midpoint 0 00 EF Scale 1 00 S Deg Freedom 5 EFI Ook Cancel Enter Gallery _Corelate Hep Figure A 15 Student s t distribution The Midpoint parameter is the central location of the distribution also mode the x axis value where you want to place the peak of the distribution The Degrees of Freedom parameter controls the shape of the distribution Smaller values result in thicker tails and less mass in the center The Scale parameter affects the width of the distribution by increasing the variance without affecting the overall shape and proportions of the curve Scale can be used to widen the curve for easier reading and interpretation For example if the midpoint were a large number say 5000 the scale could be proportionately larger than if the midpoint were 500 Example For examples see Normal dist
44. The Matrix The tool loads the correlation coefficients from the matrix into your Crystal Ball model Loading a full correlation matrix for all assumptions can take several minutes or even longer CB Tools Note Ifa Matrix Successfully Loaded message doesn t appear press Tab or Return to exit the current cell and then click Load The Matrix again 12 Reset the simulation 13 Rerun the simulation Va The forecast statistics for the correlated simulation are shown in Figure 8 12 iix Edit View Forecast Preferences Help 500 Trials Statistics View 495 Displayed Statistic Trials Mean _ Median Mode Standard Deviation Variance Skewness Kurtosis Coeff of Variability Minimum Maximum Mean Std Error Figure 8 12 Correlated simulation statistics 210 Crystal Ball User Manual Tornado Chart tool The standard deviation is now much higher than the original simulation due to the correlations The original model without the correlations ignored this risk factor and its effects CB Tools Note For details on the Correlation Matrix dialogs and their settings click Help in a dialog Tornado Chart tool The Tornado Chart tool measures the impact of each model variable one at a time on a target forecast The tool displays the results in two ways e Tornado chart e Spider chart This method differs from the correlation based method built into Crystal Ball in that this tool tests each assumption deci
45. a 0A Forecasts Ollu C8 C e j o Figure 4 7 Freeze dialog Tree view All Crystal Ball data cells defined for the current simulation appear in the dialog By default the dialog appears in a hierarchical Tree view You can e Click Show Assumptions Decision Variables or Forecasts to hide all assumptions decision variables or forecasts respectively Each button is a toggle when clicked again it shows that type of data cell Click Select All to select or check all cells currently showing in the tree Click Select None to select no cells that is to uncheck all data cells currently showing in the tree You can use these buttons together For example to select all assumptions you can click Show Decision Variables and Forecasts to hide all decision variables and forecasts Then click Select All Since only assumptions are visible only assumptions are selected for freezing Click on a workbook or worksheet box to select or deselect all data cells in that folder You can also click View List to change from Tree view to List view shown in Figure 4 8 Crystal Ball User Manual 89 Chapter 4 Running Simulations ioxi Choose items to freeze during the simulation view FYE rola Oea See IC Name Type Worksheet Assumptions Sheet1 Forecasts Sheet1 Decision Variables Sheet1 DLA Oil Price Barel Assumptions Sheet1 Figure 4 8 Freeze dialog List view 2 Check the boxes in front of cells yo
46. already familiar with this from using Excel if you copy in Excel and then do anything else but select a range for pasting the copied area is unselected and you must select it again before continuing Clearing Crystal Ball data To clear Crystal Ball data 1 Select a cell or a range of cells that contains the Crystal Ball data you want to clear 2 Select Define gt Clear Data If you select a range of cells with more than one kind of Crystal Ball data for example an assumption and a forecast Crystal Ball prompts for which data type you want to clear 3 Check the type or types to clear and click OK Crystal Ball clears the selected Crystal Ball data type assumptions decision variables and forecast definitions from each cell To clear all of one type of Crystal Ball data from all cells in the active worksheet 1 Select either Define gt Select All Assumptions e Define gt Select All Decisions Crystal Ball User Manual 71 Chapter 3 Defining Other Model Elements 72 e Define gt Select All Forecasts HF 2 Select Define gt Clear Data Crystal Ball clears the Crystal Ball data from all the selected cells in the active worksheet Excel Note Edit gt Clear gt All also clears all Crystal Ball data from selected cells Selecting and reviewing your data After you define assumptions decision variables and forecast cells and return to the spreadsheet you might want to check the cell definit
47. an example of the normal distribution A Define Assumption Cell C19 E 10 x Edit View Parameters Preferences Help Name Market Penetration xj M Normal Distribution gt 2a Oo 2 O a 6 00 7 00 8 00 9 00 10 00 OK Cancel _ Emer Gallery Correlate Help Figure 2 3 Normal distribution Crystal Ball Note Ifyou want to change the distribution type click Gallery to return to the Distribution Gallery and then select another distribution 4 In the dialog type a name for the assumption optional If the assumption cell already has a name next to it or above it on the spreadsheet or has been named in Excel the name appears in the dialog You can use that name or type a new name You can also use cell referencing to name the assumption Type the parameters for the distribution Default values appear for the distribution parameters You can type new values or cell references and formulas in any field For more about using cell references and formulas see Entering cell references and formulas on page 24 To see more information click the More button near the Name field More information appears in the Define Assumption dialog as shown in Figure 2 4 Crystal Ball User Manual 21 Chapter 2 Defining Model Assumptions C Define Assumption NCS O Edit View Parameters Preferences Help Name Market Penetration xj Yy Normal Distribution Less bu
48. and Beta parameters Description The beta distribution is a very flexible distribution commonly used to represent variability over a fixed range One of the more important applications of the beta distribution is its use as a conjugate distribution for the parameter of a Bernoulli distribution In this application the beta distribution is used to represent the uncertainty in the probability of occurrence of an event It is also used to describe empirical data and predict the random behavior of percentages and fractions The value of the beta distribution lies in the wide variety of shapes it can assume when you vary the two parameters alpha and beta If the parameters are equal the distribution is symmetrical If either parameter is 1 and the Crystal Ball User Manual 255 Appendix A Selecting and Using Probability Distributions 256 other parameter is greater than 1 the distribution is J shaped If alpha is less than beta the distribution is said to be positively skewed most of the values are near the minimum value If alpha is greater than beta the distribution is negatively skewed most of the values are near the maximum value Because the beta distribution is very complex the methods for determining the parameters of the distribution are beyond the scope of this manual For more information about the beta distribution and Bayesian statistics refer to the texts in the Bibliography Example A company that manufactures
49. appendix Crystal Ball Note The beta distribution changed from previous versions to Crystal Ball 7 Both the original and revised functions appear for compatibility CB Beta has three parameters but CB Beta2 is the Crystal Ball 7 version with Minimum and Maximum instead of Scale Crystal Ball User Manual 315 Appendix A Selecting and Using Probability Distributions Limitations of probability functions Distributions defined with probability functions differ from those entered with the Define Assumption command in these ways e You can t correlate them e You can t view charts or statistics on them e You can t extract data from them or include them in reports e They are not included in sensitivity analyses or charts Probability functions and random seeds Sampling preferences on page 83 describes how you can use the Sampling tab of the Run Preferences dialog to use the same sequence of random numbers for each simulation If you use Define gt Define Assumption or the Define Assumption toolbar button to define assumptions the same sequence of random numbers is used for each simulation even if you switch from Extreme speed to Normal speed or back to Extreme speed If you use the probability functions to define assumptions one sequence of random numbers is used for Extreme speed and a different sequence is used for Normal speed 316 Crystal Ball User Manual Maximizing Your Use of Crystal Ball In this ap
50. by sub dividing them into groups and displaying their frequency counts goodness of fit A set of mathematical tests performed to find the best fit between a standard probability distribution and a data set grabber also certainty grabber or truncation grabber A control that lets you use the mouse to change values and settings group interval A subrange of a distribution that allows similar values to be grouped together and given a frequency count 366 Crystal Ball User Manual iteration also trial A three step process in which Crystal Ball generates random numbers for assumption cells recalculates the spreadsheet model or models and displays the results in a forecast chart kurtosis The measure of the degree of peakedness of a curve The higher the kurtosis the closer the points of the curve lie to the mode of the curve A normal distribution curve has a kurtosis of 3 Latin hypercube sampling In Crystal Ball a sampling method that divides an assumption s probability distribution into intervals of equal probability The number of intervals corresponds to the Minimum Sample Size option available in the Run Preferences dialog A random number is then generated for each interval Compared with conventional Monte Carlo sampling Latin hypercube sampling is more precise because the entire range of the distribution is sampled in a more even consistent manner The increased accuracy of this method comes at the expense of
51. can use special commands to copy paste and clear Crystal Ball cell definitions from cells These are different from similar Excel commands and must be used to copy Crystal Ball cell definitions data Other Crystal Ball commands let you select and review your data The following sections describe these data commands ting Crystal Ball data Crystal Ball provides edit commands to let you copy paste or clear assumptions decision variables or forecasts from cells This feature lets you set up an entire row or column of assumptions decision variables or forecasts with just a few steps If you select a range of cells with more than one kind of Crystal Ball data for example an assumption and a forecast Crystal Ball prompts for which data type you want to copy or clear If you want to copy Crystal Ball cell definitions use only the Crystal Ball Copy Data Paste Data or Clear Data commands Using the Excel copy and paste commands only copies the cell value and attributes including cell color or pattern Crystal Ball User Manual 69 Chapter 3 Defining Other Model Elements 70 Excel Note The Excel commands do not copy the Crystal Ball data even though they might appear to since the cell colors are copied To copy Excel data and Crystal Ball data you must use the Excel edit commands and then use the Crystal Ball edit commands to copy the Crystal Ball data Copying Crystal Ball data To copy Crystal Ball assumptions
52. chart s menubar Before printing you can choose Edit gt Page Setup to format the chart on the page Then choose Edit gt Print Preview to view the chart as it will print on the selected paper size For example Figure 5 21 shows the Print Preview dialog for a forecast chart from Toxic Waste Site xls in Landscape orientation on Letter paper Notice the block of information about the chart that appears beneath it the chart title the number of trials and the certainty value for this forecast S Print preview a Ioj x 6 O088 B a Page 134 Risk Assessment 38888 a 8B Aouenbel4 Figure 5 21 Print preview dialog for a forecast chart Closing charts When you close a chart you remove it from memory but don t delete it permanently To close a chart 1 Open the Analyze menu and select the type of chart to close Assumption Charts Forecast Charts Overlay Charts Trend Charts or Sensitivity Charts 142 Crystal Ball User Manual Managing existing charts When the dialog for that chart appears check the box in front of each chart you want to close Click Close The selected chart or charts are closed without prompting Crystal Ball Note You can use Analyze gt Close All to close all chart windows from the current simulation and restored results Deleting charts You don t need to open a chart to delete it as long as the model or saved results file containin
53. chart frequencies and probabilities more easily Crystal Ball Note You can copy forecast charts and paste them into other applications For more information see page 141 Crystal Ball User Manual 127 Chapter 5 Analyzing Forecast Charts Setting chart preferences You can set a number of chart preferences to customize the appearance of your Crystal Ball charts Customization can help you analyze and present your data Available settings Table 5 4 summarizes how you can customize Crystal Ball charts and refers to the section that describes each customization Use the Help button on each tab of the Chart Preferences dialog for a description of each setting Also see the next section Setting preferences with hot keys on page 129 for ways to change the appearance of charts without using menu commands Table 5 4 Ways to customize Crystal Ball charts How Do1 Described in Add or edit and format a title Adding and formatting chart titles page 132 Change the chart type Setting the chart type page 132 Show more or fewer columns or data points Changing the chart density page 134 Show or hide grid lines Showing grid lines page 135 Show or hide the chart legend Showing the chart legend page 135 Set special chart effects such as transparency or 3D lines areas and columns Setting special chart effects page 136 Set chart color
54. choose Custom and enter a set of custom percentiles if the set you need is not already available in the dialog Data types will be extracted in the order they appear in the Select Data To Extract list You can use the up and down arrows to rearrange the data types In the Forecasts group select forecasts for data extraction All Choose or None To include the selected data and restored results for all forecasts in the current simulation choose All To include the selected data for only selected forecasts or to include forecasts in selected restored results files choose Choose Choose None if you don t want to extract any forecast data If you choose Choose the Choose Forecasts dialog appears Crystal Ball User Manual 191 Chapter 7 Creating Reports and Extracting Data 192 O Sales Projection xls Model lul Ending Sales Year 1 lly Ending Sales Year 1 lily Ending Sales Year 1 lul Ending Sales Year 1 lily Ending Sales Year 2 Lily Ending Sales Year 2 Lili Ending Sales Year 2 Q3 Lili Ending Sales Year 2 Q4 Lili Ending Sales Year 3 Q1 Lili Ending Sales Year 3 Q2 Lili Ending Sales Year 3 Q3 Lill Ending Sales Year 3 Q4 1G Reliability cbr 3 06 ability xls Model lul Material 1 Reliability ll Material 2 Reliability lul Material 3 Reliability o o oO oO oO oO oO m Figure 7 7 The Choose Forecasts d
55. contain a model depending on what type of information you have in your spreadsheet and how you use it Data vs analysis If you only use spreadsheets to hold data sales data inventory data account data etc then you don t have a model Even if you have formulas that total or subtotal the data you might not have a model that is useful for simulation A model is a spreadsheet that has taken the leap from being a data organizer to an analysis tool A model represents the relationships between input and output variables using a combination of functions formulas and data As you Crystal Ball User Manual 9 Chapter 1 Crystal Ball Overview 10 add more cells to the model your spreadsheet begins to portray the behavior of a real world system or situation Traditional spreadsheet analysis So now you have a model or you have created your first model For each variable in your model ask yourself How certain am I of its value Will it vary Is this a best estimate or a known fact You might notice that your model has some variables in it that aren t definitely certain Perhaps you don t have the actual data yet this month s sales figures or the variable behaves unpredictably individual item cost This lack of knowledge about particular values or how some variables behave contributes to the model s uncertainty Traditional spreadsheet analysis tries to capture this uncertainty in one of three ways Po
56. cost could be 5 8 or 10 In this example you will use the custom distribution to describe a series of single values Crystal Ball User Manual 295 Appendix A Selecting and Using Probability Distributions 296 To enter the parameters of this custom distribution 1 Type 5 in the Value field and click Enter Since you do not specify a probability Crystal Ball defaults to a relative probability of 1 00 for the value 5 A single value bar displays the value 5 00 Statistical Note Relative probability means that the sum of the probabilities does not have to add up to 1 So the probability for a given value is meaningless by itself it makes sense only in relation to the total relative probability For example if the total relative probability is 3 and the relative probability for a given value is 1 the value has a probability of 0 33 2 Type 8 in the Value field 3 Click Enter Since you did not specify a probability Crystal Ball defaults to a relative probability of 1 00 displayed on the scale to the left of the Custom Distribution dialog for the value 8 A second value bar represents the value 8 4 Type 10 in the Value field 5 Click Enter Crystal Ball displays a relative probability of 1 00 for the value 10 A third single value bar represents the value 10 Figure A 28 shows the value bars for the values 5 8 and 10 each with a relative probability of 1 00 Crystal Ball User Manual Using the custom di
57. different combinations of settings for special effects For example Figure 6 27 shows a sensitivity pie chart with 3D and Transparency chart effects Assumptions have similar values and ranks to the directional bar chart shown in Figure 6 21 on page 169 Sensitivity Risk Assessment Edit View Sensitivity Preferences Help 1 000 Trials Contribution to Variance Chart Sensitivity Risk Assessment Assumptions Volume of Water per Day A CPF Body Weight ion of Contaminant in Water Figure 6 27 Transparent three dimensional sensitivity pie chart Crystal Ball User Manual Understanding and using assumption charts Understanding and using assumption charts Assumption charts are created automatically when you run a simulation They cannot be deleted only opened or closed During or after a simulation runs you can view assumption charts to show trial values that is random numbers for the current simulation layered over the assumption s ideal probability distribution Assumption Al lol xj Edit View Preferences Help 1 000 Trials Normal Distribution 999 Displayed Book1 Sheet1 IA1 gt ta Oo 2 O gt infinity 4 Prinfinity Mean fp 00 Std Dev fi 00 Figure 6 28 Assumption chart Assumption charts are especially useful for checking and comparing Run Preferences settings For example you can look at charts for the same assumption before and after increasing th
58. distribution will not be used for fitting 6 Specify how the distributions should be ranked In ranking the distributions you can use any one of three standard goodness of fit tests Anderson Darling This method closely resembles the Kolmogorov Smirnov method except that it weights the differences between the two distributions at their tails greater than at their mid ranges This weighting of the tails helps to correct the Kolmogorov Smirnov Crystal Ball User Manual 121 Chapter 5 Analyzing Forecast Charts method s tendency to over emphasize discrepancies in the central region e Chi Square This test is the oldest and most common of the goodness of fit tests It gauges the general accuracy of the fit The test breaks down the distribution into areas of equal probability and compares the data points within each area to the number of expected data points Kolmogorov Smirnov The result of this test is essentially the largest vertical distance between the two cumulative distributions 7 Click OK to perform the fit Crystal Ball Note During a simulation Crystal Ball disables distribution fitting on forecast charts and overlay charts after 1 000 trials and until the simulation stops to enhance performance A final fit is performed at end of the simulation Using Split View Split View displays charts and statistics at the same time If you are using Crystal Ball s process capability features Split View is the d
59. distributions that are copied or modified by users For example if you want to copy a triangular distribution from Basic and modify it you could paste it to Favorites and change it there You can use the Categories menu to create new category folders for holding distributions Then you can use the Edit commands to add distributions to the new categories and modify them The Distribution pane The Distribution pane shows all distributions in the selected category You can use the View menu to change how they appear as shown in Table 2 4 Table 2 4 Distribution view examples View Command Example Thumbnails A A de Normal val Triangular Large Icons ZA A m iT Normal val Triangular Uniform Small Icons ake mm i af slit Triangular Uniform Lognormal Yes No Discrete Uniform Details Name Parameters Summary P Normal Mean Standard Deviation Familiar bell curve used to describe natural phenorr a Triangular Minimum Likeliest Maximum Used for rough estimation when data is limited 44 Crystal Ball User Manual Additional Distribution Gallery features The Description pane The Description pane appears at the bottom of the Distribution Gallery and provides a detailed description of the selected distribution You can turn off Show Description in the View menu to show more distributions in the Distribution pane Managing distributions Working with distributions is a k
60. enough This method isn t as accurate as the multiple simulation method but it takes significantly less time to run Crystal Ball User Manual 219 Chapter 8 Crystal Ball Tools Table 8 5 Supported bootstrap methods Continued Multiple simulation method Repeatedly simulates the model and then creates a distribution of the statistics from each simulation This method is more accurate than the one simulation method but it might take a prohibitive amount of time no CB Tools Note When you use the multiple simulation method the tool temporarily turns off the Use Same Sequence Of Random Numbers option Statistical Note In statistics literature the one simulation method is also called the non parametric bootstrap and the multi simulation method is also called the parametric bootstrap One simulation method Simulate forecast create original sample Create resample from original sample with replacement Compute statistics for resample ave yo reached the number of bootstrap amples Form distribution for all resample statistics Multiple simulation method eee Simulate forecast create sample Compute statistics for sample ave yo reached the number of bootstrap amplesg yes Form distribution for all sample statistics Figure 8 22 Comparison of one simulation and multiple simulation bootstrap
61. event taking place Who should use this program Crystal Ball is for decision makers from the businessperson analyzing the potential for new markets to the scientist evaluating experiments and hypotheses Crystal Ball and has been developed with a wide range of spreadsheet uses and users in mind You don t need highly advanced statistical or computer knowledge to use Crystal Ball to its full potential All you need is a basic working knowledge of your personal computer and the ability to create a spreadsheet model What you will need Crystal Ball runs on several versions of Microsoft Windows and Microsoft Excel For a complete list of required hardware and software see README htm in your Crystal Ball installation folder by default C Program Files Decisioneering Crystal Ball 7 Crystal Ball User Manual 1 Introduction Any changes to these requirements can be found at http Avww crystalball com About the Crystal Ball documentation set The Crystal Ball Installation and Licensing Guide describes how to install and license Crystal Ball For a brief introduction and tutorials that offer hands on experience with Crystal Ball see the Crystal Ball Getting Started Guide That guide also contains a summary of the task information included in this User Manual For information about distribution defaults and formulas plus other statistical information as well as keyboard shortcuts for commands see the Crystal Ball Reference Manu
62. explains how to set cell preferences Chapter 4 Running Simulations Provides step by step instructions for setting up and running a simulation in Crystal Ball Chapter 5 Analyzing Forecast Charts Explains how to use Crystal Ball s powerful analytical features to interpret the results of a simulation focusing on forecast charts Chapter 6 Analyzing Other Charts Provides additional information to help you analyze and present the results of your simulations using advanced charting features Chapter 7 Creating Reports and Extracting Data Provides additional information to help you share Crystal Ball data and graphics with other applications and describes how to prepare reports with charts and data Chapter 8 Crystal Ball Tools Describes tools that extend the functionality of Crystal Ball such as the Tornado Chart and Decision Table tools Appendix A Selecting and Using Probability Distributions Describes all the pre defined probability distributions used to define assumptions in Crystal Ball and includes suggestions on how to choose and use them Appendix B Maximizing Your Use of Crystal Ball Describes different aspects that enhance the performance of the program s features Appendix C Using the Extreme Speed Feature Discusses the Extreme Speed feature available with Crystal Ball Professional and Premium editions and describes its benefits and
63. factor governing simulation accuracy During a simulation Monte Carlo sampling generates natural what if type scenarios while Latin Hypercube s sampling is constrained but more accurate See page 321 for further discussion on sampling methods Precision control The precision control feature lets you set how precise you want the forecast statistics to be Crystal Ball runs the simulation until the selected forecast statistics reach the required precision as determined by calculating confidence intervals Crystal Ball Note See Confidence intervals in the Statistical Definitions chapter of the online Crystal Ball Reference Manual for more information about how Crystal Ball calculates confidence intervals Generally speaking as more trials are calculated the confidence intervals narrow and statistics become more precise The precision control feature in Crystal Ball uses this characteristic of confidence intervals to determine when a specified precision of a statistic has been reached It compares the specified precision to the confidence interval When the calculated confidence interval drops to less than the specified precision the simulation stops Crystal Ball User Manual Simulation accuracy For each forecast you can specify precision in either absolute terms in units of the forecast or in relative terms as percentages These settings are made on the Precision tab of the expanded Define Forecast dialog or th
64. failures of electronic equipment the time between arrivals at a service booth or repairs needed on a certain stretch of highway It is related to the Poisson distribution which describes the number of occurrences of an event in a given interval of time or space An important characteristic of the exponential distribution is the memoryless property which means that the future lifetime of a given object has the same distribution regardless of the time it existed In other words time has no effect on future outcomes Example one A travel agency wants to describe the time between incoming telephone calls when the calls are averaging about 35 every 10 minutes This same example was used for the Poisson distribution to describe the number of calls arriving every 10 minutes The first step in selecting a probability distribution is matching your data with a distribution s conditions Checking the exponential distribution The travel agency wants to describe the time between successive telephone calls The phone calls are not affected by previous history The probability of receiving 35 calls every 10 minutes remains the same Crystal Ball User Manual 257 Appendix A Selecting and Using Probability Distributions 258 The conditions in this example match those of the exponential distribution The exponential distribution has only one parameter rate The conditions outlined in this example include the value for this p
65. fastest simulations Crystal Ball User Manual 85 Chapter 4 Running Simulations Options preferences x Trials Sampling Speed m Run options IV Store assumption values for sensitivity analysis IV Enable correlations IV Run user defined macros IV Warn if insufficient memory IV Show control panel IV Leave control panel open on reset statistics Figure 4 5 Options tab Run Preferences dialog The Options tab sets a number of run preferences For general instructions see Setting run preferences beginning on page 80 Table 4 4 describes the settings available on this tab Table 4 4 Options tab Run Preferences dialog Setting Effect Store Assumption Values For Sensitivity Analysis Stores the randomly generated values used during the simulation for display while running To see those values during a simulation you must select this option and the appropriate Window preference for each assumption you want displayed Values can also be exported to a spreadsheet after the simulation using the Extract Data command This setting also allows Crystal Ball to generate sensitivity data during simulations This information appears in the sensitivity chart to display the influence each assumption has on a particular forecast For more information on the sensitivity chart see Understanding and using sensitivity charts on page 165 The sensitivity chart is not available unless
66. field 3 Next adjust the Scale settings By default Auto appears in the Type list and the chart appears with the most appropriate end points for that chart type To use another scale choose it from the Type drop down list and enter the minimum Min and maximum Max values to use for the scale Most chart axis combinations offer Fixed as an alternative The value axis for assumption forecast and overlay charts also offers Standard Deviation and Percentile 4 The Format Number settings control the format of the axis label numbers Choose appropriate settings for your chart For Format settings Cell Format uses the format of the underlying cell Most choices are similar to those used in Excel General Number Currency Scientific Percentage or Date Crystal Ball User Manual Setting chart preferences The Decimal settings control the number of decimal points e When checked Thousand Separator inserts a thousands separator symbol where appropriate except when Scientific formatting is set The thousands separator that appears is the one defined in Windows International or Regional Options settings Crystal Ball Note The Format Number settings also control the format of assumption parameters in the Define Assumption dialog and assumption charts 5 When all your settings are complete click OK The chart appears with your new settings Applying settings to the current chart and other charts If you would like
67. for the Vision Research model Crystal Ball Report Forecasts Simulation started on 12 15 2004 at 15 29 58 Simulation stopped on 12 15 2004 at 15 30 13 I Forecasts I I Cell C21 Summary Certainty level is 78 01 Certainty range is from 32 0 to Infinity Entire range is from 21 1 to 72 0 Base case is 39 2 After 5 000 trials the std error of the mean is 0 1 Gross Profit if Approved MM i ja 70 60 i oo ag 4 40 lad i ia a10 ono r soo Hoo go 80 1 Millons y me i a a ey l Worksheet Vision Research xls Model T Forecast Gross Profit if Approved MM Statistics Forecast values Trials 5 000 Mean 39 5 Median 38 9 Mode Standard Deviation 8 8 X M 4 gt HN Report Jal gt Figure 7 1 Sample forecast report 182 Crystal Ball User Manual Creating reports Basic steps E Crystal Ball Note If appears in your report instead of a numeric value try making the column wider to show the entire number To create a report 1 Choose Analyze gt Create Report The Create Report Preferences dialog appears as in Figure 7 2 Create Report Preferences a xi options Select a report A A ia BY Assumptions Decision Variables Forecasts ff gt E Zo 2 24a Figure 7 2 Create Report dialog Reports tab 2 Click an icon to choose a report You c
68. for user defined categories To view or edit properties 1 Click a category in the Category pane 2 Choose Categories gt Properties The Category Properties dialog appears 3 You can edit the information in any field that appears white Read only fields are gray 4 When you have finished viewing or editing category properties click OK The Category Properties dialog closes and any property changes are saved Deleting categories You can delete user defined categories and all the distributions within them but be careful not to delete a distribution you will need to use later To delete a category 1 Select the category in the Category pane 2 Choose Categories gt Delete A confirmation message appears 3 When you click OK the category and all its distributions are immediately deleted Crystal Ball Note Original built in distributions shipped with Crystal Ball are never deleted only copies can be deleted Rearranging category order Changing the order of folders in the Category pane can be helpful if you want to arrange categories in alphabetical order or put frequently used categories near the top Crystal Ball User Manual 51 Chapter 2 Defining Model Assumptions To move a category folder 1 Select the folder you want to move in the Category pane 2 Choose Categories gt Move Up or Categories gt Move Down The folder moves up or down one position each time you choose the menu command
69. gt All options always include restored results if they exist If scroll bars are present in a chart they appear in the report If an assumption has been truncated you might want to add marker lines to show where the distribution has been truncated To do this display the Chart Type tab in the Chart Preferences dialog and set an appropriate Value marker Crystal Ball User Manual 189 Chapter 7 Creating Reports and Extracting Data Extracting data Crystal Ball lets you extract assumption and forecast information generated during a simulation Crystal Ball places the extracted data in a worksheet location you select You can only extract data after you run a simulation or restore saved results To extract data Z 1 Choose Analyze gt Extract Data The Extract Data dialog appears as in Figure 7 6 x Data Options O Sensitivity Data O Trial values Figure 7 6 Extract Data dialog Data tab 2 To select the type of data to extract choose the appropriate setting in the Select Data To Extract list described in Table 7 3 Table 7 3 Extract Data dialog Select Data To Extract settings Option Extracts Statistics Descriptive statistics summarizing the assumption and forecast values Percentiles The probability of achieving values below a particular threshold in the selected increments DE Note You can reverse the meaning of the percentiles by changing the setting in the Run Preferences gt O
70. gt Define Forecast The forecast dialog for the first forecast cell appears Change the forecast definition optional Click OK If you have selected more than one forecast cell each forecast appears in turn Repeat steps 3 and 4 to check the forecast for each cell Reviewing selected cells To review Crystal Ball data cells on any open workbook 1 Select Define gt Select The Select dialog appears as shown in Figure 3 7 ET icix Choose items to select on the spreadsheets Forecast button 9 Sheet1 amp G Assumptions Decision Variable button OLA Oil Price Bare Ol 85 Assumption button OG Decision Variables D 05 1 Forecasts Dials C2 RT owt __ Figure 3 7 Select dialog Tree view Crystal Ball User Manual 73 Chapter 3 Defining Other Model Elements By default the dialog appears in hierarchical Tree view All the assumptions are listed first then all the decision variables and finally all the forecasts You can click the Assumption Decision Variable and Forecast buttons to show and hide assumptions decision variables and forecasts respectively If you prefer to view available cells in list format shown in Figure 3 8 S click the List button E o Choose items to select on the spreadsheets eee SS Sheet Forecasts Sheet1 Decision Variables Sheet1 OLA Oil Price Barel Assumptions Sheet1 Figure 3 8 Select dialog List view 2 Check the cells you w
71. high degree of confidence that the precision requirements have been met However if you have a large number of forecasts defined with precision control set you can adjust the confidence level up or down to globally change the accuracy of all forecasts together Crystal Ball User Manual Simulation speed Sampling method Choosing between Monte Carlo and Latin hypercube sampling affects how the random numbers are generated for the individual assumptions In almost all cases Latin hypercube will produce more accurate forecast statistics especially the mean given the same number of trials as Monte Carlo since it is a more consistent sampling method If you are primarily interested in the accuracy of your statistics you should select Latin hypercube as the sampling method in the Run Preferences gt Sampling dialog page 83 If you are primarily interested in evaluating how your spreadsheet model behaves under various what if scenarios you should select Monte Carlo as the sampling method Monte Carlo produces assumptions with the most randomness and hence will simulate real life situations the best Simulation speed Monte Carlo simulations can be very time consuming You can change a number of factors that affect the speed of simulations The factors are listed below in order of importance 1 Change the Speed preferences In the Run Preferences gt Speed dialog page 84 there are options that can dramatically increase t
72. how Crystal Ball runs a simulation K Crystal Ball Note You need to reset the simulation before changing run preferences To change any run preferences Ep 1 Choose Run gt Run Preferences to display the Run Preferences dialog Run Preferences Sampling Speed Options Statistics Number of trials to run 5000 JV Stop on calculation errors IV Stop when precision control limits are reached Confidence level o5 x Figure 4 1 Run Preferences dialog Trials tab 2 Click the tab with the preferences you want to change described in the following sections e Trials preferences on page 81 Specify when to stop a simulation namely number of trials calculation errors and precision control e Sampling preferences on page 83 Set the sampling seed value method and sample size e Speed preferences on page 84 Determine whether a simulation runs in Normal Demo or Extreme speed if available and set additional speed control options e Options preferences on page 86 Set a number of run preferences including whether sensitivity data and assumption values are stored whether assumption correlations are activated whether 80 Crystal Ball User Manual Setting run preferences user macros are run whether low memory triggers a warning and whether the Crystal Ball Control Panel appears e Statistics preferences on page 87 Determine how Crystal Ball displays per
73. information about statistics Sometimes you have to press two or more keys at the same time For example Ctrl c means that you hold down the Ctrl key and type c Capitalization is important Ctrl c and Ctrl C are two different key sequences 4 Crystal Ball User Manual Getting help e A key sequence without hyphens means you type the sequence in the order shown but not simultaneously For example Ctrl q N means that you press the Ctrl key and type q simultaneously and then type N Screen capture notes All the screen captures in this document were taken in Excel 2000 for Windows 2000 Professional and Excel 2003 for Windows XP using a Crystal Ball Run Preferences random seed setting of 999 Due to round off differences between various system configurations you might obtain slightly different calculated results than those shown in the examples Getting help As you work in Crystal Ball you can display online help in a variety of ways e Click the Help button in a dialog 2 e Click the Help tool in the Crystal Ball toolbar in Excel In the Excel menubar choose Help gt Crystal Ball gt Crystal Ball Help In the Distribution Gallery and other dialogs press F1 Additional resources Decisioneering Inc offers these additional resources to increase the effectiveness with which you can use our products Technical support Technical Support is available for all registered customers with a current maintenance agreement a
74. into the Define Forecast dialog the Capability Metrics view is available for the forecast chart For a definition of each statistic see Capability metrics list on page 353 zmizi Edit View Forecast Preferences Help 1 000 Trials Capability Metrics View 1 000 Displayed Calculated from forecast values Metric Value gt Mean 13 01 Standard Deviation 1 17 Pp 0 32 Ppk lower 0 31 Ppk upper 0 33 Ppk 0 31 _ Pom 0 28 Z LSL 0 86 _ Z USL 0 84 Zst 1 69 Zit total 0 19 p N C below 0 21 p N C above 0 22 p NIC total 042 PPM below 205 468 42 PPM above 218 167 04 PPN total 423 635 47 Figure 5 14 Capability Metrics view Split View Split View shows forecast charts and related statistics onscreen at the same time For more information see Using Split View on page 122 Fitting a distribution to a forecast You can fit distributions to forecast charts a feature that is similar to distribution fitting for assumptions described beginning on page 27 You can fit distributions to forecasts two ways e You can choose Forecast gt Fit Probability Distributions in the forecast chart menubar to do a quick fit with the default or currently selected distributions and ranking method You can also use this command to switch off distribution fitting that is set with either the Forecast menu or Preferences menu You can choose Preferences gt Forecast gt Forecast Window in the forecast chart menu
75. is checked on the Tools gt Options gt Calculation tab in Excel If Crystal Ball detects a circular reference and Iteration is not checked this error appears Example message Circular reference detected at cell Book 1 xls Sheet1 A1 Code 5523 e Workaround Stop the simulation and check Iteration on the Tools gt Options gt Calculation tab Crystal Ball Note In Extreme speed circular references with short Iteration setting may not match Excel s values because of differences in calculation algorithms For most consistent results set Iteration to at least 1 000 However if a circular reference is non converging its results can differ greatly when run at both Extreme and Normal speed regardless of the Iteration setting If a circular reference is non converging this error message appears Example message Unable to run in Extreme speed for the following reason Circular references do not converge results cannot be guaranteed to match Normal speed To bypass this message turn off Stop on calculation errors in the Run Preferences dialog Code 5545 e Workaround There is no workaround Check the formulas in the workbook that have created this circular reference and look for a problem that keeps this circular reference from converging to a single value Crystal Ball User Manual 327 Appendix C Using the Extreme Speed Feature Crystal Ball Note Simulations with circular references run in n
76. is a 50 probability that net profit will be at or below this value sini Edit View Forecast Preferences Help 500 Trials Frequency Chart 499 Displayed Net Profit MM 32 0 06 gt 28 3 0 05 24 S J O H20 9 a 5 2 pi H162 5 Ei D es 8 Ha 0 20 0 30 0 Millions D foo Certainty 75 0 4 s2610 Figure 5 8 Forecast chart frequency Cumulative frequency Figure 5 9 shows the Net Profit forecast chart as a cumulative distribution This chart shows the number or proportion percentage of values less than or equal to a given amount LO Forecast NetProft am ST Edit View Forecast Preferences Help 500 Trials Cumulative Frequency 499 Displayed Net Profit MM 1 00 0 90 0 80 070 Z 0604 Sos 2 0 a oao 0 30 4 020 4 Median 7 8 010 0 00 10 0 0 0 10 0 20 0 30 0 Millions gt soo Certainty 75 0 q s2610 Figure 5 9 Forecast chart cumulative frequency 116 Crystal Ball User Manual Understanding and using forecast charts To create this chart the frequencies are added cumulatively starting from the lower end of the range and then plotted as a cumulative frequency curve To understand the cumulative distribution look at a particular value 7 8 million in the example above The chart shows that the probability of 7 8 million is about 50 approximately 50 of the values are less than 7 8 million while approximately
77. methods 220 Crystal Ball User Manual Bootstrap tool Since the bootstrap technique doesn t assume that the sampling distribution is normally distributed you can use it to estimate the sampling distribution of any statistic even an unconventional one such as the minimum or maximum value of a forecast You can also easily estimate complex statistics such as the correlation coefficient of two data sets or combinations of statistics such as the ratio of a mean to a variance Statistical Note To estimate the accuracy of Latin hypercube statistics you must use the multiple simulation method Bootstrap example In the Crystal Ball Examples folder there is a Futura Apartments xls workbook you can use to experiment with the Bootstrap tool This spreadsheet model forecasts the profit and loss for an apartment complex To run the Bootstrap tool 1 In Excel with Crystal Ball loaded open the workbook Futura Apartments xls 2 Select Run gt Tools gt Bootstrap K Crystal Ball Note Ifthe Bootstrap command is not available reset the simulation and try again The Specify Target dialog appears as shown in Figure 8 23 Specify target step 1 of 3 Bootstrap Crystal Ball Tool Decisioneering 1998 2004 Determine the reliability and accuracy of your forecast statistics Profit or Loss Please select the forecast from the list above or specify data range Futura Apartments xis Model sCs9 Y ce o e
78. might be uncertain These types of assumptions are called second order assumptions also second order random variables see Burmaster and Wilson 1996 in the Bibliography You can model these types of assumptions in Crystal Ball by placing the uncertain parameters of the distribution in separate cells and defining these cells as assumptions You then link the parameters of the variability assumption to the uncertainty assumptions using cell references To illustrate this for the Toxic Waste Site xls spreadsheet 1 Enter the values 70 and 10 into cells G4 and G5 respectively These are the mean and standard deviation of the Body Weight assumption in cell C4 which is defined as a normal distribution 2 Define an assumption for cell G4 using a normal distribution with a mean of 70 and a standard deviation of 2 3 Define an assumption for cell G5 using a normal distribution with a mean of 10 and a standard deviation of 1 4 Enter references to these cells in the Body Weight assumption For an example see Figure 8 47 following Crystal Ball User Manual 243 Chapter 8 Crystal Ball Tools Pbetine Assumption Cee Edit View Parameters Preferences Help Name Body Weight EF y Normal Distribution 100 00 Figure 8 47 Assumption using cell references for the mean and standard deviation When you run the tool for second order assumptions the uncertainty of the assumptions parameters is modeled in the outer simu
79. million is about 05 of having a greater value while the probability for 0 is about 80 This is also correct since the probability of Net Profit falling between those two values is 75 80 05 75 or Certainty 75 Notice in this chart that the reverse cumulative frequency values are Crystal Ball User Manual 117 Chapter 5 Analyzing Forecast Charts 118 complements of the cumulative frequency values 20 80 1 00 and 95 05 1 0 the probability values for 0 0 and 26 1 million respectively Statistics You can display a full set of descriptive statistics for a simulation in the forecast window by choosing View gt Statistics Forecast Net Profit MM a loj x Edit View Forecast Preferences Help 500 Trials Statistics 499 Displayed Statistic Forecast gt Trials 500 kL Mean 6 9 __ Median 7 8 Mode Standard Deviation 10 5 Variance 110 7 Skewness 0 24944 Kurtosis 2 91 Coeff of Variability 1 51 Minimum 15 0 Maximum 35 3 _ Mean Std Error 0 5 Figure 5 11 Forecast window statistics The example in Figure 5 11 shows statistics for the entire range of values 100 of the forecast values including the outliers extreme values excluded from the default display range Statistical terms listed on this table are discussed in the Statistical Definitions chapter of the online Crystal Ball Reference Manual and the glossary Crystal Ball Note If the Precision
80. number Format Cel format Decimal 2 J Thousand separator Figure 5 6 The Format Number group on the Axis tab Choose a format from the drop down list shown in Figure 5 7 They are similar to Excel formats Format number Format Cell format Cell format Figure 5 7 Format settings For most formats you can also specify the number of decimal places and whether to use a thousands separator Click OK to apply the settings or use Apply To as described on page 139 114 Crystal Ball User Manual Understanding and using forecast charts Changing the distribution view and interpreting statistics The forecast settings related to distribution type determine the overall appearance of a forecast chart You can also choose to display a table of statistics or percentiles instead of or in addition to a chart To set the distribution type or display a data table 1 Open the View menu in the forecast window 2 Choose a distribution type or other view to display on the forecast chart as described in Table 5 2 Table 5 2 Forecast Preferences dialog Forecast Window gt View settings View Setting Effect Frequency Shows the number or frequency of values occurring in a given interval This is the default distribution type Cumulative Frequency Shows the number or proportion percentage of values less than or equal to a given amount Reverse Cumulative Shows the number or proporti
81. of the first dialog Enter the range of cells on your spreadsheet that contains the empirical values that Crystal Ball should use to calculate a correlation coefficient Enter the range of cells in the standard Al A2 format where A designates the column and 1 and 2 designate the first and last cell rows respectively For example if one set of values is in column Q rows 10 through 15 and the second set of values is in column R rows 10 through 15 enter the range in the left field as Q10 Q15 and the range in the right field as R10 R15 Crystal Ball calculates the correlation coefficient enters it in the field to the left of the correlation coefficient scale and moves the slider control to the correct position Crystal Ball Note The two cell ranges do not necessarily have to have the same dimensions but they must contain the same number of value cells and must be in the same workbook The cell ranges are read in a row by row fashion Each time you select a new assumption or correlation coefficient Crystal Ball displays a sample correlation of the correlated assumption values in the chart to the right Crystal Ball User Manual 37 Chapter 2 Defining Model Assumptions Define Correlation Cell D13 iol x Choose an assumption and enter a correlation Correlation Chart Example Choose oi Price Barre iO Hey ah AE Gey teLh N oil ef Assumption Oil Price Barrel Oil Price Barrel Normal Distribution
82. overlay chart window to switch among several graphic and numeric views described in Table 5 2 on page 115 e Use the Overlay menu to add additional forecasts to the chart or remove them all and toggle between the default view and Goodness Of Fit view e Use Preferences gt Overlay to choose a view described in Table 5 2 on page 115 determine when the overlay chart window should appear and specify whether to fit distributions to all forecasts described on page 152 Crystal Ball User Manual 151 Chapter 6 Analyzing Other Charts 152 e Use Preferences gt Chart Preferences to further customize the chart s appearance as described beginning on page 128 Customizing overlay charts helps you compare forecasts by viewing their differences in several ways For example the area and column chart types might obscure parts of some distributions behind other distributions but the outline and line chart types show virtually all of each distribution For best display of many types of data you can choose 3D view and then rotate the chart as shown in Figure 6 7 Using distribution fitting with overlay charts Similar to the distribution fitting with historical data described in Fitting distributions to data beginning on page 27 you can fit standard probability distributions to forecasts in an overlay chart in addition to comparing forecasts to each other You can fit distributions to forecasts in overlay charts two ways
83. probability distribution the rate is the only key factor If needed for modeling a situation information on the size of the interval must be encoded in your spreadsheet formulas Example two A travel agency wants to describe the number of calls it receives in 10 minutes The average number of calls in 10 minutes is about 35 Again you begin by identifying and entering the values to set the parameters of the Poisson distribution in Crystal Ball In this example the conditions show one important value 35 calls or a rate of 35 The result would be a distribution showing the probability of receiving x number of calls in 10 minutes Crystal Ball User Manual Using discrete distributions Yes no distribution Yes No Parameter Probability of Yes 1 Conditions The random variable can have only one of two values for example 0 and 1 The mean is p or probability 0 lt p lt 1 Description The yes no distribution is also called the Bernoulli distribution in statistical textbooks This distribution describes a set of observations that can have only one of two values such as yes or no success or failure true or false or heads or tails It is a discrete probability distribution The yes no distribution has one parameter Probability of Yes 1 Q penne assumptions cenar Te Edit View Parameters Preferences Help NomeT CCG Yes No Distribution Aaa a o o no Probability of Yes 1 0 5 xj O
84. range minimum and maximum 3 In the forecast chart menubar choose Preferences gt Forecast gt Filter Forecast Window Precision Fiter Auto Extract IV Seta filter on the forecast values Include values in the range Infinity to Infinity C Exclude values in the range infinity to finfinit Filtering removes values from charts and statistical calculations Values are not lost and can be restored by changing these settings Cancel Apply To Defaults Help Figure 5 5 Filter tab Forecast Preferences dialog 4 On the Filter tab of the Forecast Preferences dialog set a filter on the forecast values and include values in the range between the display range minimum and maximum 5 When the settings are complete click OK 6 Choose View gt Statistics in the forecast chart menubar to show statistics for the display range or in Split View look at the statistics table on the right side Formatting chart numbers By default the number format displayed on the forecast chart comes from the underlying format of the forecast cell You can select another cell format using the Chart Preferences dialog Crystal Ball User Manual 113 Chapter 5 Analyzing Forecast Charts To change the format of numbers in a forecast chart 1 In the forecast window choose Preferences gt Chart The Chart Preferences dialog appears Click the Axis tab Find the Format Number group as shown in Figure 5 6 Format
85. resulting from life and fatigue tests It is commonly used to describe failure time in reliability studies and the Waha breaking strengths of materials in reliability and quality control tests Weibull distributions are also used to represent various physical quantities such as wind speed The Weibull distribution is a family of distributions that can assume the properties of several other distributions For example depending on the shape parameter you define the Weibull distribution can be used to model the exponential and Rayleigh distributions among others The Weibull distribution is very flexible When the Weibull Shape parameter is equal to 1 0 the Weibull distribution is identical to the exponential distribution The Weibull Location parameter lets you set up an exponential distribution to start at a location other than 0 0 When the Shape parameter is less than 1 0 the Weibull distribution becomes a steeply declining curve A manufacturer might find this effect useful in describing part failures during a burn in period Define Assumptions cetar inixi Edit View Parameters preferences Help Name Jan E3 Weibull Distribution Location 0 00 EF Scale 1 00 EFI Shape 2 EF Ok Cancel Enter Gallery _ Correlate Help Figure A 18 Weibull distribution 278 Crystal Ball User Manual Using continuous distributions When the Shape parameter is equal to 2 0 as in Figure A 18 a special form of the Wei
86. run a simulation set the chart settings the way you want them save your simulation results and go on to other work Then at a moment s notice you can call up the results open existing charts and reports and create new charts and reports without running the simulation again You can also extract data from restored results Saving Crystal Ball simulation results To save Crystal Ball simulation results select Analyze gt Save Results You can only save results after a simulation stops Crystal Ball uses the name of the active workbook for the default name of the saved results file The Save Results dialog appears as in Figure 4 12 It lets you determine the name and location of the saved results file Te zhxt Save in Examples odre Reliability cbr File name Book 1 Save as type Crystal Ball Result Files cbr x Cancel Figure 4 12 Save Results dialog Crystal Ball User Manual 97 Chapter 4 Running Simulations 98 Crystal Ball saves the simulation results including charts and reports in a file with extension cbr Once you have saved the simulation results you can continue working with Crystal Ball Crystal Ball Note Ifyou choose Analyze gt Save Results only results from the current simulation are saved Previously restored results are not saved This can be an issue where both current and saved results are used for example in overlay charts Suppose current and restored forecasts are both us
87. the left pane of the dialog as shown in Figure 2 14 Define Correlation Cell D13 Coefficient m Entry field y l o o Coefficient C ial Pl Oi PricefBamrel o Correlation list Calc button Figure 2 14 Define Correlation dialog with second assumption After you select the second assumption the cursor moves to the field below the Choose button the Coefficient Entry field The chosen assumptions appear in the list of correlations The currently selected assumption also appears immediately next to the Choose button 36 Crystal Ball User Manual Additional assumption features 6 Enter a correlation coefficient using one of the following methods Enter a value between 1 and 1 inclusive in the Coefficient Entry field Type the number that you want to use in the field to the left of the slider control After you type the number the slider control on the correlation coefficient scale moves to the selected value e Choose a cell that contains the correlation coefficient Crystal Ball Note If you choose a cell with values that change during the simulation it is the initial value of the cell that is used for the coefficient Drag the slider control along the correlation coefficient scale The value you select appears in the field to the left of the scale Type the desired correlation coefficient in the Coefficient field in the correlation list e Click Calc A small dialog appears at the bottom
88. the chart Crystal Ball Note To clear all forecast selections choose Trend gt Remove All in step 1 above 3 Click OK to accept the settings and modify the chart Crystal Ball User Manual 159 Chapter 6 Analyzing Other Charts 160 Once you select forecasts and create a trend chart you can change the order that forecasts appear in the forecast axis To change the order of forecasts 1 In the trend chart menubar choose Preferences gt Chart gt Chart Type The Chart Type dialog appears as shown in Figure 6 14 on page 161 All charted forecasts appear in the Series list at the top of the dialog The forecast at the top of the list appears as the first forecast at the left of the forecast axis Select a forecast in the list and use the up and down arrow keys to the right of the list to move the forecast up or down the list Optional step You can choose Defaults at any time to restore all settings to their original default values Optional step If you want to apply the settings to more than one chart click Apply To Then specify how they should be applied see page 139 for details and click OK Otherwise go to step 5 When all forecasts are positioned in order click OK to accept the current Chart Preferences settings and modify the chart Changing the overall appearance of trend charts When you first choose Preferences gt Chart in the trend chart menubar the General tab of the Chart Preferences d
89. then modify it they can publish their own versions and overwrite each others changes If you are publishing a category you might want to make the shared folder read only to avoid this problem To edit or delete a path or rearrange the path order 1 Open the Subscribe To A Category dialog as described in step 2 above 2 Select the path to a target category 3 Click an action button Edit Delete Move Up or Move Down 4 When you are finished click OK If you delete a path to a subscribed category that category disappears from the Category pane of the Distribution Gallery You can resubscribe to it at any time to use it again Sharing categories through email You can use email to share a user defined category with a Crystal Ball user who is not on your network To share categories using email 1 Create and address an email message 2 Use the file attachment command and browse to the category folder on your machine By default this is C Documents and Settings your username Application Data Decisioneering Crystal Ball 7 x Store 3 Select the user defined category to send Crystal Ball category files have this format CBCategory_name cbc For example the Favorites category appears as CBCategory_Favorites cbc 4 Send the email To use a category received in an email 1 Open the email 2 Save the attachment in the Crystal Ball category folder by default C Documents and Settings your username Appli
90. to describe the number of dry wells you would drill before the next producing well you would use the geometric distribution Assume that in the past you have hit oil about 10 of the time The first step in selecting a probability distribution is matching your data with a distribution s conditions Checking the geometric distribution The number of trials dry wells is unlimited e You continue to drill wells until you hit the next producing well The probability of success 10 is the same each time you drill a well These conditions match those of the geometric distribution The geometric distribution has only one parameter Probability In this example the value for this parameter is 0 10 representing the 10 probability of discovering oil You would enter this value as the parameter of the geometric distribution in Crystal Ball The distribution illustrated in Figure A 21 shows the probability of x number of wells drilled before the next producing well Crystal Ball User Manual 285 Appendix A Selecting and Using Probability Distributions 286 inixi Edit View Parameters Preferences Help Name Jan x y Geometric Distribution Probability o o B Figure A 21 Geometric distribution Example two An insurance company wants to describe the number of claims received until a major claim arrives Records show that 6 of the submitted claims are equal in dollar amount to all the other claims c
91. to zero and are nearly independent remove the correlation For coefficients that are close to one replace one of the assumptions with a formula in your spreadsheet If Crystal Ball detects inconsistently correlated assumptions when a simulation is running it first determines whether small adjustments to the correlation coefficients are possible This process might take a long time depending on the number of correlated assumptions Crystal Ball displays the message Examining the Correlation Coefficients If you get this message you should probably stop and redefine your correlations If small adjustments to the correlation coefficients are possible a dialog appears letting you decide whether to cancel the simulation or continue with the adjusted coefficients Select one of the following responses Crystal Ball User Manual 323 Appendix B Maximizing Your Use of Crystal Ball e Click Adjust Coefficients This Time Only to continue the simulation with the adjusted coefficients Since adjusted coefficients are not saved permanently a dialog appears again if you stop and restart the simulation e To continue the simulation replacing your original correlation coefficients with the adjusted ones click Adjust Coefficients Permanently Crystal Ball Note Correlation coefficients that were specified using a cell reference in place of an actual value are replaced with a permanent value the cell references are removed
92. understand start review the menus and toolbars and close Crystal Ball Now spend a few moments learning how Crystal Ball can help you make better decisions under conditions of uncertainty Chapters 2 through 7 of this Crystal Ball User Manual describe how to build run analyze and present your own Crystal Ball model simulations Crystal Ball User Manual 7 Chapter 1 Crystal Ball Overview Model building and risk analysis overview Glossary Term simulation Any analytical method that is meant to imitate a real life system especially when other analyses are too mathematically complex or too difficult to reproduce Glossary Term spreadsheet model Any spreadsheet that represents an actual or hypothetical system or set of relationships Glossary Term forecast A statistical summary of the simulation results in a spreadsheet model displayed graphically or numerically Glossary Term risk The possibility of loss damage or other undesirable event and the severity associated with the event Risk and Crystal Ball is an analytical tool that helps executives analysts and others make decisions by performing simulations on spreadsheet models The forecasts that result from these simulations help quantify areas of risk so decision makers can have as much information as possible to support wise decisions The basic process for using Crystal Ball then is to 1 Build a spreadsheet model that
93. use the goodness of fit test and distribution selection that is set in the Forecast Window tab of the Forecast Preferences dialog opened by choosing Preferences gt Forecast in the forecast window Before you choose to calculate from the best fitting distribution if the distribution is not normal consider that e You are not guaranteed of attaining a good fit to the forecast values and The fitting process might take a long time depending on how many simulation trials you are running Crystal Ball Note In odd or rare circumstances it is possible for the normality test to fail and the best fitting distribution still be a normal distribution or for the normality test to pass and the best fitting distribution be non normal Alternatively you can choose the second main setting Calculate Metrics From Forecast Values to bypass the normality test and always calculate the metrics directly from the forecast data Setting specification limits and targets The capability metrics only appear if you specify either an upper or lower specification limit or both for the forecast You can also specify an optional target To specify these limits ily 1 Either define a new forecast or select an existing forecast and choose Define gt Define Forecast The Define Forecast dialog appears With process capability features activated it looks similar to the following figure Crystal Ball User Manual 343 Appendix D Using the Pro
94. variable is continuous or discrete e Step specifies the interval between values for discrete variables For example Step 1 could be used to specify whole dollars while Step 5 could specify 50 cent increments For details about the fields and options in this dialog click the Help button in the dialog Crystal Ball Note You can use cell referencing to name a decision variable define the lower and upper limits and set the step size For more information see Entering cell references and formulas on page 24 Click OK Repeat steps 1 4 for each decision variable in your model Crystal Ball Note There is no absolute limit to the nwmber of assumptions you can define per worksheet In general you should define less than 1000 assumptions decision variables and forecasts per worksheet Defining forecasts After you define the assumption cells and decision variables you are ready to select and define forecast cells Forecast cells usually contain formulas that refer to one or more assumption and decision variable cells The forecast cells combine cells in your model to produce the results you need When you define a forecast cell you Name the forecast Specify the units for the forecast Indicate whether you want to automatically display the forecast window during the simulation Set precision control settings for the forecast Set filtering settings for the forecast Indicate whether to automatically extract stat
95. when using one of these tools All of the compatibility and other differences apply to these tools as well 332 Crystal Ball User Manual Other important differences Precision control and cell error checking Extreme speed uses a bursting feature to optimize the balance between speed and system responsiveness Since bursts are usually several hundreds or thousands of trials at a time this feature affects a number of features like precision control confidence testing and checking for cell errors Crystal Ball tests for precision control and performs confidence testing in OptQuest after each burst of several hundred or thousands of trials In Normal speed these checks are performed every 50 trials by default As a result you may see your simulations run more trials than usual when these features are turned on Generally speaking these features are not as critical to simulation performance in Extreme speed as they are in Normal speed Because of bursting Crystal Ball cannot check cell errors after every trial as it does in Normal speed only at the end of a burst If cell errors occur in forecast cells such as Divide by zero or VALUE and the Run Preferences gt Stop On Calculation Errors option is selected the simulation stops at the end of the next burst Spreadsheet updating While running simulations in Extreme speed cells in your spreadsheet model appear to update every second or so after each burst This is nor
96. wide and skewed distribution Crystal Ball User Manual Bootstrap tool E Forecast Mean ij E H al Figure 8 27 Bootstrap forecast chart of mean The Statistics view further lets you analyze the statistics sampling distribution If the mean standard error or coefficient of variability is very large the statistic might not be reliable and might require more samples or more trials This example has a relatively low standard error and coefficient of variability so the forecast mean is an accurate estimate of the population mean a Forecast Mean Figure 8 28 Bootstrap forecast statistics for the mean Crystal Ball User Manual 225 Chapter 8 Crystal Ball Tools The results workbook has a correlation matrix showing the correlations between the various statistics High correlation between certain statistics such as between the mean and the standard deviation usually indicates a highly skewed distribution You can also use the Bootstrap tool to analyze the distribution of percentiles but you should run at least 1 000 bootstrap samples and 1 000 trials per sample to obtain good sampling distributions for these statistics according to Efron and Tibshirani see Bibliography Decision Table tool 226 Decision variables are values that you can control such as how much to charge for a product or how many wells to drill But in situations with uncertainty it is not alway
97. your data Adding and removing assumptions By default the sensitivity chart includes all of the assumptions from the simulation To remove or change the assumptions included follow the instructions in this section Crystal Ball Note The total number of assumptions included in the chart affects the calculation of the Contribution to Variance percentages To select just which assumptions are displayed in the chart see the text and Figure 6 24 on page 175 To remove or change assumptions included in the sensitivity chart 1 In the Sensitivity Chart window choose Sensitivity gt Choose Assumptions The Choose Assumptions dialog appears Crystal Ball User Manual 3 Understanding and using sensitivity charts F choose Assumptions ox Choose assumptions to display on the sensitivity chart eR A ooo a 00 Toxic Waste Site xis 9 Model MLA Body Weight LALA Volume of Water per Day ALA Concentration of Contaminant in Water Mla CPF Figure 6 22 Choose Assumptions dialog Check the assumptions to add to the sensitivity chart and uncheck those to remove from the chart Click OK to activate the selected assumptions Changing the target forecast To change the target forecast in the sensitivity analysis 1 2 3 In the Sensitivity Chart window choose Sensitivity gt Choose Target Forecast The Choose Forecasts dialog appears similar to Figure 6 20 on page 168 Check a new target forecast Click
98. 0 are independent of each other What happens to the first patient does not affect the second patient The probability of curing a patient 0 25 25 remains the same each time a patient is tested Since the conditions of the variable match the conditions of the binomial distribution the binomial distribution would be the correct distribution type for the variable in question e If historical data are available use distribution fitting to select the distribution that best describes your data Crystal Ball can automatically select the probability distribution that most closely approximates your data s distribution The feature is described in detail in Fitting distributions to data beginning on page 27 You can also populate a custom distribution with your historical data After you select a distribution type determine the parameter values for the distribution Each distribution type has its own set of parameters For example there are two parameters for the binomial distribution trials and probability The conditions of a variable contain the values for the parameters In the example used the conditions show 100 trials and 0 25 25 probability of success In addition to the standard parameter set each continuous distribution except uniform also lets you choose from alternate parameter sets which substitute percentiles for one or more of the standard parameters For more information on alternate parameters see Alter
99. 00 Hourly Wage Ranges in Dollars Figure A 3 Probability distribution of wages Compare the probability distribution in the example above to the probability distributions in Crystal Ball Figure A 4 A Distribution Gallery Cell C5 Edit View Categories Help Triangular Weibull Logistic Student s t ill Binomial Poisson INN Neg Binomial Geometric Discrete Uniform Figure A 4 Distribution Gallery dialog 248 Crystal Ball User Manual Understanding probability distributions The probability distribution in the example in Figure A 3 has a shape similar to many of the distributions in the Distribution Gallery This process of plotting data as a frequency distribution and converting it to a probability distribution provides one starting point for selecting a Crystal Ball distribution Select the distributions in the gallery that appear similar to your probability distribution then read about those distributions in this chapter to find the correct distribution For information about the similarities between distributions see Comparing the distributions on page 313 For a complete discussion of probability distributions refer to the sources listed in the bibliography Discrete and continuous probability distributions Notice that the Distribution Gallery shows whether the probability distributions are discrete or continuous Discrete probability distributi
100. 08 Retail Price 4 95 7 95 3 95 5 95 Cost of Goods 2 20 3 80 1 95 2 40 Gross Profit 1 546 1 502 1 292 1 805 Total Gross Profit 6 144 Data Series ISales Volume A Sales Volume B Sales Volume C Sales Volume D Chi Square 3 322 7 333 4 694 2 478 Distribution Best fit Lognormal Beta Beta Student s t Normal 38 683 32 772 65 600 3 217 Triangular 179 072 11 028 102 439 64 650 Lognormal 3 322 27 283 56 733 6 700 Uniform 498 589 94 733 16 200 280 933 M 4 gt mi Description Model Sales Data ial Figure 8 5 Batch Fit results below the model Using Batch Fit results in a model Now that assumptions have been generated from historic sales data the next steps are to convert Sales Volume formulas to values and replace them with the new assumptions Then you can run a simulation with more accurate results To use Batch Fit results in a model 1 Select the formulas in cells C5 through F5 2 Use Excel s Copy and Paste Special commands to copy the formula cells and paste them back into the same locations When the Paste Special dialog appears choose Values as the Paste setting Leave the other default settings and click OK 3 Copy assumption data into the worksheet Select cells C15 through F15 Select Define gt Copy Data Notice this is Crystal Ball s Copy Data command not Excel s Copy command Crystal Ball Note This function copies Crystal Ball data only not the cell value Crystal Ball User Manua
101. 257 extreme maximum 265 extreme minimum 267 gamma 259 geometric 285 hypergeometric 286 logistic 262 lognormal 263 negative binomial 289 normal 268 Pareto 270 Poisson 291 selecting 250 Student s t 272 triangular 274 uniform 276 yes no 293 process capability features 340 properties distribution category 51 publishing distribution categories 52 Q quality features 340 R range names in parameters 25 ranking methods 32 rearranging distribution categories 51 references 360 absolute cell 25 dynamic and static 24 relative cell 25 references cell 24 referrals consulting 6 relative cell references 24 relative cell referencing 25 relative probability 296 reports overview 182 requirements 1l resetting a simulation 90 restoring simulation results 98 reversed percentiles 88 Crystal Ball User Manual 377 Index run preferences 80 running a simulation 90 S saving and restoring simulation results 97 saving simulation results 97 Scenario Analysis example 231 tool 231 screen capture notes 5 second order variables 243 Select dialog 73 selecting some data 73 sending distribution categories 54 sensitivity analysis accuracy 318 calculation 170 sensitivity charts creating 167 customizing 172 overview 165 setting defaults 139 setting forecast preferences 61 sharing distribution categories 52 53 54 55 simulation results restoring 98 simulation results saving 97 simulation results saving and restoring 97
102. 3 Continue moving folders up and down until they are arranged as preferred Sharing categories over networks The ability to publish and subscribe to categories is very powerful You can create large numbers of specific distributions add them to a category and then publish it on the network for other users in your organization Likewise you can easily subscribe to categories published by others You can also email categories to others and use categories sent to you To publish a category to a shared folder on your computer or to a location on a network 1 Select a category in the Distribution Gallery 2 Choose Categories gt Publish The Browse For Folder dialog appears as shown in Figure 2 19 Select a Folder to publish this category E My Documents Ej My Computer amp ee My Network Places fa Recycle Bin Internet Explorer Cancel New Folder Z Figure 2 19 The Browse For Folder dialog 3 Browse to locate the folder where you want to place the category You can click to expand a folder and view its subfolders If you need to create a new folder select the folder that should include the new folder 52 Crystal Ball User Manual Additional Distribution Gallery features The new folder appears and you can rename it For example if you want to create a folder under My Documents select My Documents Then when you click New Folder the new folder appears under My Documents at the same level as My
103. 50 are greater This would be correct for a median value Note also that the chart shows that the probability for 26 1 million is about 95 while the probability for 0 is about 20 This is also correct since the probability of Net Profit falling between those two values is 75 95 20 75 or Certainty 75 Reverse cumulative frequency Figure 5 10 shows the Net Profit forecast chart as a reverse cumulative distribution This chart shows the number or proportion percentage of values greater than or equal to a given amount mix Edit View Forecast Preferences Help 500 Trials Reverse Cumulative Frequency 499 Displayed Net Profit MM 1 00 0 90 0 80 4 gt 0704 a 5 0 60 4 2 0504 3 0 40 2 0 30 4 0 20 4 0104 00 10 0 0 0 10 0 20 0 30 0 Millions po Ceteiny 50 x aso Figure 5 10 Forecast chart reverse cumulative frequency To create this chart the frequencies are added cumulatively starting at the higher end of the range and then plotted as a declining cumulative frequency curve To understand the reverse cumulative distribution look at a particular value 7 8 million in the example above The chart shows that the probability of 7 8 million is about 50 approximately 50 of the values are less than 7 8 million while approximately 50 are greater This would be correct for a median value Note also that the chart shows that the probability for 26 1
104. 6 4 Overlay chart for selected forecasts The overlay chart appears with the frequency distributions for the selected forecasts superimposed over each other By default they are displayed as columns You can change them to lines or areas in two or three dimensions and can change the number of data points or intervals bins with the Chart Preferences settings page 128 A default name is assigned you can also change it with the Chart Preferences settings Follow the steps in Customizing overlay charts on page 151 and Setting chart preferences beginning on page 128 to change a variety of chart features and highlight those of greatest interest Crystal Ball Note You can also use hot keys or keyboard equivalents for commands to quickly change the chart preferences For a list of these see Table 5 5 on page 129 7 In this example press Ctrl d until the reverse cumulative chart view appears Then press Ctrl t until the outline chart type appears As shown in Figure 6 5 the chart view changes to show all three distributions completely This view in outline view most clearly suggests that Material 3 has superior reliability and is dominant since a greater proportion of its distribution is to the right of 1 00 and its values for all probability levels are higher than the others Crystal Ball User Manual 149 Chapter 6 Analyzing Other Charts 150 Overlay Chart 1 Figure 6 5 Overlay chart with three
105. 63 parameter sets 265 M macros global 102 user defined 99 managing categories Distribution Gallery 49 managing charts 140 managing distirbutions Distribution Gallery 45 manual conventions 4 maximum extreme distribution 265 memory warning of insufficient 87 methods bootstrap 219 minimum extreme distribution 267 model sensitivity 165 modifying distributions 47 48 monotonic relationships 212 Monte Carlo simulation benefits 13 defined 11 in Crystal Ball 78 N negative binomial distribution 289 New Category dialog 50 next distribution prev distribution 30 normal distribution defined 268 O online help 5 outliers 107 outlying values 107 overlay chart with Two dimensional Simulation 241 overlay charts customizing 151 distribution fitting 152 153 overview 146 Overlay Example 149 P parameters alternate 25 formulas 25 range names 25 Parameters menu 255 parameters alternate 255 Pareto distribution defined 270 pasting charts 141 pasting distributions 46 percentiles reversing 88 view 118 Poisson distribution defined 291 precision control defining with forecast 59 displayed percentiles 118 displayed statistics 118 increasing accuracy 318 Precision preferences 64 Preferences menu distributions 38 preferences chart 128 printing charts 142 probability 246 probability distributions 12 beta 255 binomial 281 chi square 261 comparing 313 custom 294 discrete uniform 283 erlang 261 exponential
106. 892 919 1996 Frey H Christopher and David S Rhodes Characterizing Simulating and Analyzing Variability and Uncertainty An Illustration of Methods Using an Air Toxics Emissions Example Human and Ecological Risk Assessment 2 4 762 797 1996 Hoffman F O and J S Hammonds Propagation of Uncertainty in Risk Assessments The Need to Distinguish Between Uncertainty Due to Lack of Knowledge and Uncertainty Due to Variability Risk Analysis 14 5 707 712 1994 Rai S N D Drewski and S Bartlett A General Framework for the Analysis of Uncertainty and Variability in Risk Assessment Human and Ecological Risk Assessment 2 4 972 989 1996 Uncertainty analysis Morgan M Granger and Max Henrion with a chapter by Mitchell Small Uncertainty A Guide to Dealing with Uncertainty in Quantitative Risk and Policy Analysis New York Cambridge University Press 1990 Hoffman F O and J S Hammonds Propagation of Uncertainty in Risk Assessments The Need to Distinguish Between Uncertainty Due to Lack of Knowledge and Uncertainty Due to Variability Risk Analysis 14 5 707 712 1994 International Atomic Energy Agency Evaluating the Reliability of Predictions Made Using Environmental Transfer Models AEA Safety Series No 100 1 106 STI PUB 835 Vienna Austria 1989 Kelton W David and Averill M Law Simulation Modeling amp Analysis 3rd Ed New York McGraw Hill Inc 2000 Cryst
107. 9 MCR TI MA Probability 0 8 xj Shape 50 xj OK Cancel C Ete Gallery Correlate Help Figure A 23 Negative binomial distribution Some characteristics of the negative binomial distribution When Shape 1 the negative binomial distribution becomes the geometric distribution The sum of any two negative binomial distributed variables is a negative binomial variable e Another form of the negative binomial distribution sometimes found in textbooks considers only the total number of failures until the r th successful occurrence not the total number of trials To model this form of the distribution subtract out r the value of the shape parameter from the assumption value using a formula in your worksheet 290 Crystal Ball User Manual Using discrete distributions Poisson distribution alh Poisson Parameter Rate Conditions The number of possible occurrences in any interval is unlimited The occurrences are independent The number of occurrences in one interval does not affect the number of occurrences in other intervals The average number of occurrences must remain the same from interval to interval Description The Poisson distribution describes the number of times an event occurs in a given interval such as the number of telephone calls per minute or the number of errors per page in a document Example one An aerospace company wants to determine the number of defects per 100 squar
108. A mH A Normal Triangular Uniform Lognormal Figure 2 1 Common distribution types 18 Crystal Ball User Manual Defining assumptions During a simulation Crystal Ball calculates numerous scenarios of a model by repeatedly picking values from the probability distribution for the uncertain variables and using those values for each assumption cell Commonly a Crystal Ball simulation calculates hundreds or thousands of scenarios or trials in just a few seconds The value to use for each assumption for each trial is selected randomly from the defined possibilities Because distributions for independent variables are so important to simulations selecting and applying the appropriate distribution is the main part of defining an assumption cell For more information on probability distributions see Understanding probability distributions beginning on page 246 Defining an assumption To define an assumption you must 1 Identify a distribution type as described in Selecting a probability distribution beginning on page 250 Crystal Ball uses probability distributions to describe the uncertainty in your assumption cells From a gallery of distribution types you choose the ones that best describe the uncertain variables in the problem you are trying to solve Appendix A describes each distribution type in detail You can also select a distribution type by fitting a distribution to data For details see Fitting dis
109. Click New or choose existing charts to Open Close or Delete Close Delete Figure 6 2 Overlay Charts dialog To create a new overlay chart click New The Choose Forecasts dialog appears as shown in Figure 6 3 chooserorecasts o Choose forecasts to display on the overlay chart View Fa seee IEF 5 06 a 0 Model lal Material 1 Reliability lal Material 2 Reliability Dll Material 3 Reliability Figure 6 3 Choose Forecasts dialog Tree view By default this dialog appears in a hierarchical Tree view If you prefer click the List button to change the view from a tree to a list For more information about this dialog see page 156 or click the Help button Check the boxes in front of the forecasts to include in the overlay chart Figure 6 3 above shows the Choose Forecasts dialog in Tree view with the three forecasts selected Click OK to create a new overlay chart with the selected forecasts as shown in Figure 6 4 Crystal Ball User Manual Menubar Chart view Chart title Grid lines y axis title Edit View Overlay Preferences Help Understanding and using overlay charts Q Overlay Chart 1 E 5 x 1 000 Trials Frequency View Overlay Chart o o E Wi Material 1 Reliability o Q a Material 2 Reliability I Material 3 Reliability Probability Auanbaly 3 N Figure
110. Control feature is checked in the Run Preferences dialog and the forecast has Precision Control options set the Precision column appears in the Statistics view Percentiles You can display percentile information in 10 increments in the forecast window by choosing View gt Percentiles A percentile is the percent chance or probability of a forecast value being less than or equal to the value that corresponds to the percentile the default For example Figure 5 12 displays the percentile view of the Net Profit forecast where the 90th percentile corresponds to 19 3 million meaning that there is a 90 chance of a forecast value being equal to or less than 19 3 million Another interpretation is that 90 of the forecast values are equal to or less than 19 3 million Notice that the Median in Statistics view is the same as the 50th percentile in Percentiles view in this case 7 8 million Crystal Ball User Manual Understanding and using forecast charts zix Edit Vew Forecast Preferences Help 500 Trials Percentiles 499 Displayed Figure 5 12 Forecast Percentiles view Crystal Ball Note Ifthe Precision Control feature is checked in the Run Preferences dialog and the forecast has Precision Control options set the Precision column appears in the Percentiles view Goodness Of Fit If you have selected distribution fitting described in the next section you can choose the Goodness Of Fit view to display comparativ
111. Create Report to display the Create Report dialog 2 Click a report type Full Forecast or Custom If you choose Full or Forecast the capability metrics appear in a block for each forecast following the statistics and percentiles for that forecast Additional process capability information appears in the summary and all selected marker lines appear in forecast and overlay charts If you choose Custom the Custom Report dialog appears Custom Report a J xj Report sections m Forecasts Details Forecasts M Repot Summary Sp em Forecasts y O Assumptions gt IV Chart fioo C Choose D1 Decision Variables ae Co O Overay Charts I Siaistics ea O Trend Charts I Percentiles O Sensitivity Charts 7 Capability Figure D 11 Custom Report dialog with Forecasts Details Crystal Ball User Manual 351 Appendix D Using the Process Capability Features 3 Ifit is not already highlighted highlight Forecasts in the Report Sections list The Forecasts Details appear With the process capability features activated Capability Metrics is checked by default 4 If you don t want to include the capability metrics for some reason uncheck that setting in the Forecasts Details list Otherwise leave it checked and follow the instructions in online help or Defining custom reports on page 185 to finish defining the custom report 5 When all settings are complete click OK to generate the repor
112. Crystal Ball 7 2 User Manual N Crystal Ball This manual and the software described in it are furnished under license and may only be used or copied in accordance with the terms of the license agreement Information in this document is provided for informational purposes only is subject to change without notice and does not represent a commitment as to merchantability or fitness for a particular purpose by Decisioneering Inc No part of this manual may be reproduced or transmitted in any form or by any means electronic or mechanical including photocopying and recording for any purpose without the express written permission of Decisioneering Inc Written designed and published in the United States of America To purchase additional copies of this document contact the Technical Services or Sales Department at the address below Decisioneering Inc 1515 Arapahoe St Suite 1311 Denver Colorado USA 80202 Phone 1 303 534 1515 Toll free sales 1 800 289 2550 Fax 1 303 534 4818 1988 2006 Decisioneering Inc Decisioneering is a registered trademark of Decisioneering Inc Crystal Ball is a registered trademark of Decisioneering Inc CB Predictor is a trademark of Decisioneering Inc OptQuest is a registered trademark of Optimization Technologies Inc Microsoft is a registered trademark of Microsoft Corporation in the U S and other countries FLEXIm is a trademark of Macrovision Cor
113. Extreme Speed simulation multiple times on a large or complex model you might find that you receive the Sufficient memory may not warning dialog This dialog lets you disable the Store Assumption Values run preference and change the number of trials to run Notice that these changes in the warning dialog reset those preferences in the Run Preferences dialog on the Trials and Options tabs They are then applied to other models unless you reset them after the large model has run Spreadsheets with no Crystal Ball data Numerical Spreadsheets without Crystal Ball data always run in Normal speed even if Extreme speed is selected in the Run Preferences dialog Crystal Ball Note This could happen if you have Crystal Ball spreadsheet functions entered in the model but have not defined assumptions forecasts or decision variables differences Ordinarily there may be slight numerical differences in simulation results between Extreme and Normal speed modes These differences are primarily due to roundoff errors and are usually in the relative range of le 12 to le 15 In certain circumstances as for extremely large models containing thousands of formulas these slight differences can compound and grow in magnitude 334 Crystal Ball User Manual Numerical differences If you are concerned about possible differences in your own model Crystal Ball provides a special tool that you can use to compare the results between the two run modes
114. Gallery For instructions see the referenced pages Creating new categories You can create a category of distributions to help with model development or to share with other users To create a category 1 Choose Categories gt New The New Category dialog box appears as shown in Figure 2 18 ST 2x Name limit 15 chars Description a Created by fo Version OO Location Path o Status flocal Edtable tti S Cancel Help Figure 2 18 The New Category dialog 2 Enter the name of your category Optionally you can enter a description your name and a version number helpful for shared categories Note If you enter an ampersand amp before a letter in the category name that letter becomes a shortcut key You can then use it to select the category from the keyboard by holding down the Alt key and pressing the shortcut key Shortcut keys appear underlined when you press the Alt key The new category should use a different shortcut key from other categories 3 Click OK The new folder appears in the Category pane and the category can be selected and used like Favorites or any other user defined category Crystal Ball User Manual Additional Distribution Gallery features Viewing and editing category properties Category properties include the information entered in the New Category dialog when a category is created You can view properties for any category but you can only edit properties
115. K Cancel Ete Galley Correlate Help Figure A 25 Yes no distribution Example A machine shop produces complex high tolerance parts with a 02 probability of failure and a 98 probability of success If a single part is pulled from the line Figure A 26 shows the probability that the part is good Crystal Ball User Manual 293 Appendix A Selecting and Using Probability Distributions Hi Define Assumption Cell A1 4 lolx Edit View Parameters Preferences Help Name fan xj v Yes No Distribution 0 80 0 60 0 40 i 0 1 2 Probability of Yes 1 xj Ok Cancel Enter Galley Help Probability Figure A 26 Probability of pulling a good part Using the custom distribution 294 If none of the provided distributions fits your data you can use the custom distribution to define your own For example a custom distribution can be especially helpful if different ranges of values have specific probabilities You can create a distribution of one shape for one range of values and a different distribution for another range The following sections explain how to use the custom distribution and provide examples of its use Custom distribution An Custom With Crystal Ball you can use the custom distribution to represent a unique situation that cannot be described using other distribution types you can describe a series of single values discrete ranges or continuou
116. Many of the distributions discussed in this chapter are related to one another in various ways For example the geometric distribution is related to the binomial distribution The geometric distribution represents the number of trials until the next success while the binomial represents the number of successes in a fixed number of trials Similarly the Poisson distribution is related to the exponential distribution The exponential distribution represents the amount of time until the next occurrence of an event while the Poisson distribution represents the number of times an event occurs within a given period of time In some situations as when the number of trials for the binomial distribution becomes very large the normal and binomial distributions become very similar For these two distributions as the number of binomial trials approaches infinity the probabilities become identical for any given interval For this reason you can use the normal distribution to approximate the binomial distribution when the number of trials becomes too large for Crystal Ball to handle more than 1000 trials You also can use the Poisson distribution to approximate the binomial distribution when the number of trials is large but there is little advantage to this since Crystal Ball takes a comparable amount of time to compute both distributions Crystal Ball User Manual 313 Appendix A Selecting and Using Probability Distributions 314 Likewise the nor
117. NV ENVI CW ics sacs sds Seen 25a a dnd cvodadeanndadesiinda sabi E EAA A 198 Batch Fitto leerse aa aeaa eanna deecdeedcscvesSheavancteterpenddoasendtonaeoevte 199 Batchi Fit example soimisiriisisiinisngnornaria naiiai 199 Correlation Matrix tool cccccccccccccccccccecseseececeeeccssesecceceseeuseessssuenseseusuauesss 206 Correlations erea Shoaccads ra a eaea e a Oa EE EEA OTERO 206 Correlation Matrix cccccccccccecccccssscecccceeeseeceseueecessesuececeseeeaeeeseueeeseesesaaes 206 Correlation Matrix example wicasnscisssnacinanctstaintiadasucteiainneriasivulesiundintis 207 Tornado Chart tOol sarsie T EE E EE OE 211 Tornado hart mesye a e a a a a a 211 Spider Chart ssori an en a a EE E AE E R E 213 Tornado Chart example ssossrrrsisiss seias inniinn inean 214 LMA ONS aaae aA E AA A A A E 218 Bo tstrap tool eias peaareen rne E A atan AEE an Ee ASE EEEE 219 Bootstrap example sronirocnseinen oen niniin E EEEN EE E S 221 Decision Table tool cccccccccceececccceeecccceeueseccecueceessueuuecccseeueescssuueesesesauaasss 226 Decision Table example 0 0 cece en eeeteeceecneeneetaetas 227 Scenario ANALYSIS TOO isscstsccserisciandoncestgoesuvesaassetessapebevsgnadonnenbawigwasseoneeatoays 231 Scenario Analysis CxaMple sss amaieria 231 Two Dimensional Simulation tool ccccccccccceeececcccueeceeseueueeceseeeeueesseeuens 236 Two Dimensional Simulation example 0 c cece cece eee ceneeeeeeees 237 Second order assumptions
118. OK to activate the selected forecast Setting sensitivity preferences You can set a number of preferences that determine The sensitivity view that appears Whether the sensitivity chart appears automatically and whether it appears while the simulation is running or after it stops How many assumptions are shown in the chart starting with the most sensitive Whether sensitivities are limited to a certain sensitivity value or higher To set sensitivity preferences 1 Choose Preferences gt Sensitivity Crystal Ball User Manual 173 Chapter 6 Analyzing Other Charts 174 The Sensitivity Preferences dialog appears as shown in Figure 6 23 Sensitivity Preferences J xj View Contribution to Variance x Window IV Show automatically While running simulation When simulation stops Figure 6 23 Sensitivity Preferences dialog Sensitivity Window tab To change how the sensitivities are presented use the View drop down list e Contribution To Variance shows sensitivities as values that range from 0 to 100 and indicate relative importance by showing the percentage of the forecast variance contributed by each assumption e Rank Correlation shows sensitivities as rank correlations that range from 1 to 1 and indicate both magnitude and direction of the correlation of each assumption with the forecast e Sensitivity Data shows a table of contributions to variance and rank correlations for each
119. Pictures Click OK The dialog closes and a copy of your category is saved there Using shared categories If other users have published categories in a shared folder on their computers or the network you can access them for use with the Distribution Gallery This is called subscribing to categories To subscribe to a category learn its name and location then 1 2 Open the Distribution Gallery and choose Categories gt Subscribe The Subscribe To A Category dialog appears as shown in Figure 2 20 Subscribe to a Category E Enter the paths to search for categories If duplicate categories are found the first one in the list will be used Cancel Help Figure 2 20 The Subscribe To A Category dialog x Di a po SIN Click Add The Browse For Folder dialog appears as shown in Figure 2 19 on page 52 Locate the target folder and click OK to add the new path to the Subscribe To A Category dialog box Click OK to load all categories in the listed paths and close the Subscribe To A Category dialog box All loaded categories are available for use as if they were on your local computer Crystal Ball User Manual 53 Chapter 2 Defining Model Assumptions 54 Crystal Ball Note Shared categories can be used like local categories in most ways However they cannot be modified unless they also exist in a folder on your local computer If several users copy a published category locally and
120. Previous Scenario Reset Original Values 95 00 95 10 2 52E 04 54 02111213 95 20 2 22E 04 53 56886657 95 30 2 11E 04 70 34865898 95 40 2 35E 04 62 99855485 95 50 2 22E 04 60 73774051 95 60 2 36E 04 57 42244167 95 70 2 66E 04 62 4104222 95 80 2 25E 04 61 96721601 95 90 2 50E 04 60 08780074 96 00 2 63E 04 5683218803 96 10 2 21E 04 77 02164987 96 20 3 14E 04 72 73861511 96 30 2 15E 04 58 43870157 wss ssy 4SIY feg jad 134LAA jo auinjoA biaa Apog u Figure 8 37 Scenario Analysis results for Toxic Waste Site Windows Note For the next few steps you might want to arrange the windows in Excel vertically so that you can see the results next to the original workbook 9 Select the row with the 98 00 assuming default percentile settings 10 Click Paste Selected Scenario The scenario of assumption values that produced the 98th percentile value of the target forecast appears in the Toxic Waste workbook Crystal Ball recalculates the workbook and updates the forecast cells 11 Click Paste Next Scenario In the workbook the assumption and forecast values change to the values for the next scenario for 98 10 12 Click Reset Original Values The original assumption and forecast values appear in the workbook CB Tools Note For more information about the Scenario Analysis dialogs click Help in a dialog 234 Crystal Ball User Manual Scenario Analysis tool Interp
121. Sales xds Fe Magazine Checkout Sales E Magazine A Magazine B Magazine C Magazine D Sales Volume 562 362 646 508 Retail Price 4 95 7 95 3 95 5 95 Cost of Goods 2 20 3 80 1 95 2 40 Gross Profit 1 546 1 502 1 292 1 805 Total Gross Profit 6 144 M 4 gt r Description Model Sales Data ial Oy Figure 8 1 Magazine Sales workbook Crystal Ball User Manual 199 Chapter 8 Crystal Ball Tools 200 In this model cells C5 through F5 are formulas that refer to the first row of data on the Sales Data worksheet However the model would be more accurate if you replaced these formulas with assumptions based on historical data To do this you can use the Batch Fit tool to generate an assumption for each data column of the Sales Data worksheet Then you can use Crystal Ball commands to copy and paste those assumptions from the Batch Fit data to the first data row of the Magazine Sales model Because Crystal Ball cannot define assumptions in formula cells you need to convert the formulas in C5 to F5 to values before pasting the assumption cells Generating Batch Fit results To run the Batch Fit tool 1 In Excel with Crystal Ball loaded open the workbook Magazine Sales xls 2 Select Run gt Tools gt Batch Fit The Select Distributions dialog appears as shown in Figure 8 2 Select distributions step 1 of 3 Batch Fit Crystal Ball Tool Decisioneering 1998 2004 Fit probability distrib
122. Sheet ted Dp Figure 8 42 Two Dimensional Simulation results table The tool also graphs the results of the two dimensional simulations on an overlay chart and a trend chart The overlay chart preferences can be set to show the risk curves for the simulations for different sets of uncertainty assumption values To do this set the Chart Type for each series to Line and choose Cumulative Frequency view It is convenient to use the chart hotkeys Ctrl t for the chart type and Ctrl d for view Optionally use Ctrl n to move or remove the legend and Ctrl m to cycle through central tendency marker lines For this example Figure 8 43 shows that most of the risk curves are clustered densely toward the center while a few outlier curves are scattered to the right showing the small probability of having a much greater risk Statistical Note In risk analysis literature the curves are often called the alternate realizations of the population risk assessment 240 Crystal Ball User Manual Two Dimensional Simulation tool Overlay Risk Assessment 000 Trials Figure 8 43 Overlay chart of risk curves The trend chart shows certainty bands for the percentiles of the risk curves The band width shows the amount of uncertainty at each percentile level for all the distributions Trend Risk Assessment all Z E Figure 8 44 Trend chart of certainty bands Crystal Ball User Manual 241 Chapter 8 Cryst
123. To delete a single range of data without using the data table click the range to select it and either e Set the Probability or Height of Min and Height of Max to 0 or e Choose Edit gt Delete Row or right click and choose Delete Row Statistics for custom distributions are approximate Crystal Ball User Manual 311 Appendix A Selecting and Using Probability Distributions Truncating distributions 312 You can change the bounds or limits of each distribution except the custom distribution by dragging the truncation grabbers or by entering different numeric endpoints for the truncation grabbers This truncates the distribution You can also exclude a middle area of a distribution by crossing over the truncation grabbers to white out the portion you want to exclude Crystal Ball Note To display the truncation grabbers open an assumption in the Define Assumption dialog and click the More button For example suppose you want to describe the selling price of a house up for auction after foreclosure The bank that holds the mortgage will not sell for less than 80 000 They expect the bids to be normally distributed around 100 000 with a standard deviation of 15 000 In Crystal Ball you would specify the mean as 100 000 and the standard deviation as 15 000 and then move the left grabber to set the limit of 80 000 The grabber whites out the portion you want to exclude as shown in Figure A 47 Be aware
124. To use this tool 1 Open and click the model you want to test 2 Launch the comparison tool CompareRunModes xla Locate and double click that file in the Tools folder under the Crystal Ball installation directory by default C Program Files Decisioneering Crystal Ball 7 Tools The Compare Run Modes dialog appears Compare Run Modes E x Compare Run Modes Crystal Ball Tool Decisioneering 1598 2005 Compare Extreme speed and Normal speed simulation modes for the active workbook This tool will create a new workbook with a summary of the differences Detect differences above 0 0000000000001 Relative X Number of trials to run 1000 I Compare assumptions in addition to forecasts Figure C 2 Compare Run Modes dialog 3 Indicate the amount of difference to detect whether that difference is absolute or relative and the number of trials to run Check the box if you want to compare assumptions as well as forecasts Depending on the size of your model differences might not become obvious until after a fairly large number of trials have run You might want to run 5 000 or more trials for your test 4 When you are ready click Run Comparison The simulation runs once in Extreme speed and once in Normal speed Results appear in a new workbook The comparison Summary tab appears when the comparison is complete Crystal Ball User Manual 335 Appendix C Using the Extreme Speed Feature Figure C 3 shows co
125. USL if there is only an upper specification limit Zlt total For long term metrics when both specification limits are defined the number of standard deviations between the long term forecast mean and the lower boundary of combining all defects onto the upper tail of the normal curve Also equal to Zst total minus the Z score shift value if a short term index is available 354 Crystal Ball User Manual Capability metrics list Table D 1 Capability statistics calculated by Crystal Ball Continued Long Short Metric term term Description p N C v v Probability of a defect below the lower specification limit DPU BELOW below p N C v v Probability of a defect above the upper specification limit DPU4gove above p N C v v Probability of a defect outside the lower and upper specification limits total DPUToraL PPM v v Defects below the lower specification limit per million units below PPM v y Defects above the upper specification limit per million units above PPM v v Defects outside both specification limits per million units total LSL v v Lower specification limit the lowest acceptable value of a forecast involved in process capability or quality analysis USL v v Upper specification limit the highest acceptable value of a forecast involved in process capability analysis Target s v The ideal target value of a forecast involved in process capability analysis Z scor
126. a simulation Crystal Ball Note In previous versions of Crystal Ball you could choose whether to use static or dynamic cell referencing in parameters With static referencing all cell references are resolved at the start of a simulation and then frozen while a simulation is running If you open a model from a previous version any static references are converted into dynamic references If you don t want parameter values to change when a simulation is running be sure cell references in parameters do not reference Crystal Ball data cells assumptions decision variables and forecasts directly or indirectly through formulas Crystal Ball User Manual Additional assumption features Relative references Relative references remember the position of a cell relative to the cell containing the assumption For example suppose an assumption in cell C6 refers to cell C5 If the assumption in C6 is copied to cell C9 the relative reference to C5 will then refer to the value in cell C8 This lets you easily set up a whole row or column of assumptions each having similar distributions but slightly different parameters by performing just a few steps An absolute reference on the other hand always refers back to the originally referenced cell in this case C5 Absolute references To indicate an absolute reference you must use a dollar sign before the row and the column For example to copy the exact contents of cell C5 into an assumpti
127. ab Add Comment To Cell Adds an Excel comment that provides more information about the Crystal Ball data within each cell Crystal Ball Note Crystal Ball only updates cell comments when you define or redefine an assumption decision variable or forecast Set Cell Value To Changes assumption cell values to the selected value Mean or Median when no simulation is running Set To Range Midpoint Changes decision variable cell values to the selected value range Midpoint Minimum or Maximum when no simulation is running For more information click the Help button in the dialog Click Apply To The Apply To dialog appears as shown in Figure 3 10 Crystal Ball User Manual 75 Chapter 3 Defining Other Model Elements Cn x M Apply preferences for C This tab To all cells on This sheet This workbook All open and new workbooks Cancel Hep Figure 3 10 The Apply To dialog for cell preferences 5 Choose whether to apply settings from only the current tab or all tabs in the Cell Preferences dialog 6 Choose whether to apply the settings to all cell preferences of the chosen type s on the current Excel worksheet all worksheets in the current workbook or all open workbooks and any new workbooks to be created later Crystal Ball Note The default is All Workbooks Open And New 7 Click OK to close the Apply To dialog and apply the settings to the chosen cell types and worksheet
128. added memory requirements to hold the full Latin hypercube sample for each assumption See Sampling preferences on page 83 mean The familiar arithmetic average of a set of numerical observations the sum of the observations divided by the number of observations mean standard error The Standard Deviation of the distribution of possible sample means This statistic gives one indication of how accurate the simulation is median The value midway in terms of order between the smallest possible value and the largest possible value mode That value which if it exists occurs most often in a data set model sensitivity The overall effect that a change in an assumption cell produces in a forecast cell This effect is solely determined by the formulas in the spreadsheet model Crystal Ball User Manual 367 Glossary Monte Carlo simulation A system which uses random numbers to measure the effects of uncertainty in a spreadsheet model outliers also outlying values Values generated during a simulation on the extreme end of a distribution and are excluded from the display range PDF Probability density function that represents the probability that an infinitely small variable interval will fall at a given value probabilistic model A system whose output is a distribution of possible values In Crystal Ball this system includes a spreadsheet model containing mathematical relationships probability distrib
129. age 92 Analyze your results See Understanding and using forecast charts beginning on page 106 for suggestions If you have the Professional edition of Crystal Ball consider using CB Predictor or OptQuest for further analysis Take advantage of the many resources Decisioneering offers to help you get the most out of Crystal Ball If you are new to Crystal Ball the Crystal Ball Getting Started Guide offers tutorials to quickly introduce Crystal Ball s features and workflow Consider completing the tutorials before you continue on with the more detailed chapters that follow in this User Manual You can choose Start gt Programs gt Crystal Ball 7 gt Crystal Ball Tutorial to run through a brief online tutorial that teaches Crystal Ball basics Or choose Start gt Programs gt Crystal Ball 7 gt Training CD Demo to learn about a more extensive online tutorial available for purchase from Decisioneering Inc For a list of support training and referral services see Additional resources on page 5 Papers user group information conference schedules newsletter subscriptions and more are available on our Web Site http www crystalball com Starting and closing Crystal Ball 14 You can start Crystal Ball manually or you can set up Crystal Ball to start automatically whenever you start Excel Starting Crystal Ball manually To start Crystal Ball manually 1 In Windows choose Start gt Programs g
130. ain the conditions fit the binomial distribution with two important values 100 trials and 0 6 60 probability of success These values specify the parameters of the binomial distribution in Crystal Ball The result would be a distribution of the probability that x number of people prefer the company s product Discrete uniform distribution Discrete Uniform Parameters Minimum Maximum Conditions The minimum value is fixed The maximum value is fixed All integer values between the minimum and maximum are equally likely to occur Description In the discrete uniform distribution all integer values between the minimum and maximum are equally likely to occur It is a discrete probability distribution The discrete uniform distribution is very similar to the uniform distribution page 276 except it is discrete instead of continuous all its values must be integers The discrete uniform distribution can be used to model rolling a six sided die In that case the minimum value would be 1 and the maximum 6 Crystal Ball User Manual 283 Appendix A Selecting and Using Probability Distributions Define Assumption Cell A1 E Ioj x Edit View Parameters Preferences Help Discrete Uniform Distnbution 5 6 7 8 9 10 11 12 13 14 15 16 Minimum 5 y Maximum 15 Y ox Enter Gallery Correlate z Figure A 20 Discrete uniform distribution Example A manufacturer determines that he must
131. al available in Adobe Acrobat pdf format If you have Crystal Ball Professional or Premium Edition the CB Predictor User Manual OptQuest User Manual and Crystal Ball Developer Kit User Manual offer additional information about those Crystal Ball products For users of Six Sigma DFSS Lean principles and similar quality methodologies the Process Capability Guide offers tutorials and other information to help you use Crystal Ball s process capability features All of these Crystal Ball documents are available in Adobe Acrobat pdf format To view them choose Start gt Programs gt Crystal Ball 7 gt Documentation or start Crystal Ball and choose Help gt Crystal Ball gt Crystal Ball Manuals To view and download the Crystal Ball Developer Kit User Manual available with a Crystal Ball Professional or Premium Edition license choose Help gt Crystal Ball gt Crystal Ball Developer Kit This Crystal Ball User Manual includes the following e Chapter 1 Crystal Ball Overview Introduces Crystal Ball and explains how it uses spreadsheet models to help with risk analysis and many types of decision making Chapter 2 Defining Model Assumptions Describes how to define assumption cells in models and how to use the Crystal Ball Distribution Gallery 2 Crystal Ball User Manual Chapter 3 Defining Other Model Elements Describes how to define decision variable cells and forecast cells in models It also
132. al Ball Tools You can focus in on a particular percentile level such as the 95th percentile by viewing the statistics of the 95th percentile forecast shown in Figure 8 45 Figure 8 45 95th percentiles forecast statistics Crystal Ball Note Ifyou have the Probability Above A Value option selected in the Run Preferences gt Options dialog the percentiles will be reversed in meaning so that the Ist percentile represents the uppermost 1 and the 99th percentile represents the lowest 1 For more information see Statistics preferences on page 87 Compare the results of the two dimensional simulation to a one dimensional simulation with both uncertainty and variability co mingling together of the same risk model as in Figure 8 46 Forecast Risk Assessment Figure 8 46 Forecast chart for one dimensional simulation 242 Crystal Ball User Manual Two Dimensional Simulation tool The mean of the 95th percentiles 1 77E 4 is lower than the 95th percentile risk of the one dimensional simulation shown above at 2 06E 4 This indicates the tendency of the one dimensional simulation results to overestimate the population risk especially for highly skewed distributions Second order assumptions Some assumptions contain elements of both uncertainty and variability For instance an assumption might describe the distribution of body weights in a population but the parameters of the distribution
133. al Ball User Manual 361 Bibliography 362 Crystal Ball User Manual In this glossary A compilation of terms specific to Crystal Ball as well as statistical terms used in this manual Additional definitions and details are included in the Crystal Ball Reference Manual accessed online by choosing Start gt Programs gt Crystal Ball 7 gt Documentation Crystal Ball User Manual 363 Glossary algorithm A rule that specifies how to solve a particular problem assumption An estimated value or input to a spreadsheet model assumption cell A value cell in a spreadsheet model that has been defined as a probability distribution base case The value in a Crystal Ball assumption decision variable or forecast cell at the start of a simulation CDF Cumulative distribution function that represents the probability that a variable will fall at or below a given value certainty bands In a trend chart a graphic depiction of a particular certainty range for each forecast certainty level The percentage of values in the certainty range compared to the number of values in the entire range certainty range The linear distance for the set of values between the certainty grabbers on the forecast chart coefficient of variability also coefficient of variance or coefficient of variation A measure of relative variation that relates the standard deviation to the mean Results can be represented in percentages for
134. al world situation that matches or closely approximates the normal distribution conditions A more detailed discussion of calculating standard deviation follows this example Example The normal distribution can be used to describe future inflation You believe that 4 is the most likely rate You are willing to bet that the inflation rate could as likely be above 4 as it could be below You are also willing to bet that the inflation rate has a 68 chance of falling somewhere within 2 of the 4 rate That is you estimate there is approximately a two thirds chance that the rate of inflation will be between 2 and 6 The first step in selecting a probability distribution is matching your data with a distribution s conditions Checking the normal distribution The mean inflation rate is 4 e The inflation rate could as likely be above or below 4 The inflation rate is more likely to be close to 4 than far away In fact there is approximately a 68 chance that the rate will lie within 2 of the mean rate of 4 These conditions match those of the normal distribution The normal distribution uses two parameters Mean and Standard Deviation Figure A 13 shows the values from the example entered as parameters of the normal distribution in Crystal Ball a mean of 0 04 4 and a standard deviation of 0 02 2 Crystal Ball User Manual 269 Appendix A Selecting and Using Probability Distributions olx Edit View Parameters
135. alysis tools The following steps can help guide your simulation analysis by focusing on details as well as general trends 1 Look at the big picture Consider each forecast chart from a high level viewpoint Look at the shape of the distribution e Is it distributed normally or skewed negatively or positively e Is it flat spread out on both sides of the mean or peaked with most values clustered closely around the mean e Does it have a single mode most likely value or is it bimodal with several peaks or humps e Is it continuous or are there groups of values separated from the rest maybe even extreme outliers that fall outside of the display range The statistical concepts in the Statistical Definitions chapter of the online Crystal Ball Reference Manual can help with this part of your analysis 2 Look at the certainty level the probability of achieving values within a certain range You can enter a range such as all values greater than 0 dollars if you re analyzing profits and view the certainty of falling within the range 0 to Infinity in this case You can also enter a certainty say 75 and see what range of values would be required to meet that level Determining the certainty level beginning on page 107 explains these and other ways of analyzing certainty 3 Focus on the display range You can change the display range to focus on different sections o
136. an choose one of five pre defined reports or click the Custom button to define your own The choices are as follows e Assumptions report summary plus assumption parameters charts and correlations e Decision Variables decision variable bounds variable types and step size if discrete e Forecasts report summary plus forecast summaries charts statistics percentiles and capability metrics if generated e Full the default all sections and details except assumption statistics and percentiles Index only forecast assumption and decision variable summaries Crystal Ball User Manual 183 Chapter 7 Creating Reports and Extracting Data 184 Custom displays the Custom Report dialog for report definition If you choose Custom click the Custom button and complete the Custom Report dialog as described in the next section Defining custom reports on page 185 Otherwise continue with step 4 Click the Options tab to set a location and format for the report Create Report Preferences xi Reports M Location New workbook Sheet name R t C Current workbook ciad M Formatting IV Include cell locations IV Include cell comments M Chart format Image Excel chart Figure 7 3 Create Report dialog Options tab In the Location group of the Options tab choose whether to create the report in a new Excel workbook or the current workbook If you c
137. and less than 100 to use as a precision control statistic Must Be Within Selects which range to use for precision control absolute units or relative Plus Or Minus percentage Units Determines the size of the confidence interval in actual forecast units used to test the precision of the forecast statistics Percent Determines the size of the confidence interval in percentage terms used to test the precision of the forecast statistics Crystal Ball User Manual 65 Chapter 3 Defining Other Model Elements Filter preferences DefnerorecastCoAL Name an EF Al Units EF Forecast Window Precision Fiter Auto Extract Include values in the range Finfinity to infinity Exclude values in the range infinity to Infinity I When filtering values for this forecast remove values for the same trial from other forecasts too Filtering removes values from charts and statistical calculations Values are not lost and can be restored by changing these settings Cancel Apply To Defaults Help Figure 3 5 Filter tab forecast preferences This tab lets you discard values inside or outside a range for the current forecast or globally for all forecasts in a model The values are not permanently deleted only discarded for the purposes of your current analysis You can choose among the settings listed in Table 3 3 Then click OK to apply settings on the current tab to the active forec
138. and margins Click OK to accept the settings Printing distribution information To print a distribution 1 Double click a distribution in the Distribution Gallery to display it in the Define Assumption dialog Choose Edit gt Print in the Define Assumption menubar Confirm the printer and other settings Click OK to print the distribution Printed information includes the chart with current parameter settings The distribution name and parameters appear in a box beneath the graphic Crystal Ball Note You can preview before printing In step 2 choose Edit gt Print Preview before you choose Edit gt Print Managing categories You can use distribution categories to organize distributions and share them with other users in your organization For example you can Create a category to hold a few distributions you use most often Create a category to hold all distributions related to a certain model or series of models Create a category with distributions you want to share with other users Delete categories you no longer need Rearrange category order within the Distribution Gallery Share or publish categories with other users over networks Crystal Ball User Manual 49 Chapter 2 Defining Model Assumptions 50 e Use categories that others have published e Unpublish categories that are no longer needed Table 2 2 on page 41 summarizes category management tasks you can perform in the Distribution
139. ant to select 3 Click OK to highlight all the selected cells so you can change their preferences or perform other actions on them You can use the Select dialog to select cells on more than one worksheet but you need to activate each worksheet in turn to review and if desired apply a command to all the selected cells Setting cell preferences 74 Crystal Ball lets you change the appearance of assumption forecast and decision variable cells so you can quickly identify them on your spreadsheets You can set Crystal Ball to change the appearance of these cells as you define them or you can change the appearance of predefined cells To set cell preferences 1 Choose Define gt Cell Preferences The Cell Preferences dialog appears as shown in Figure 3 9 Crystal Ball User Manual Setting cell preferences Cell Preferences xi Assumption Decision Variables Forecast Cell shading Coo hE X Pattern aal J Add comment to cell J Set cell value to distribution Mean z Cancel Apply To Defauts Help Figure 3 9 The Cell Preferences dialog Click the tab for the kind of cell to format Assumption Decision Variables or Forecast Make appropriate settings for the chosen cell type Color Changes the color of each Crystal Ball data cell of the type modified by the selected tab Pattern Changes the pattern of each Crystal Ball data cell of the type modified by the selected t
140. arameter 35 calls every minute or a rate of 35 Enter this value to set the parameter of the exponential distribution in Crystal Ball Define Assumption Cell A1 R E lol x Edit View Parameters Preferences Help Name A1 E7 y Exponential Distribution 2 E amp Figure A 7 Exponential distribution The distribution in Figure A 7 shows the probability that x number of time units 10 minutes in this case will pass between calls Example two Acar dealer needs to know the amount of time between customer drop ins at his dealership so that he can staff the sales floor more efficiently The car dealer knows an average of 6 customers visit the dealership every hour Checking the exponential distribution The car dealer wants to describe the time between successive customer drop ins The probabilities of customer drop ins remain the same from hour to hour These conditions fit the exponential distribution Crystal Ball User Manual Using continuous distributions The resulting distribution would show the probability that x number of hours will pass between customer visits Gamma distribution also Erlang and chi square Gamma Parameters Location Scale Shape Conditions The gamma distribution is most often used as the distribution of the amount of time until the r th occurrence of an event in a Poisson process When used in this fashion the conditions underlying the gamma distribution are
141. assumption Use settings in the Windows group to determine whether the chart appears automatically If Show Automatically is checked you can choose whether to display the chart while the simulation is running or after it stops To limit sensitivities by rank or value click the Criteria tab to display it Crystal Ball User Manual Understanding and using sensitivity charts Sensitivity Preferences E x Sensitivity Window Display only those assumptions matching these criteria IV Show the 25 assumptions with the highest sensitivity values I Show assumptions with sensitivity values Assumptions not meeting these criteria are grouped into an Other category on the Contribution to Variance chart Figure 6 24 Sensitivity Preferences dialog Criteria tab If you have a model with a lot of assumptions you can check either or both boxes to limit the number of assumptions shown in the chart to a fixed number or to assumptions above a certain sensitivity value If you check both the more restrictive of the two criteria is used 5 You can click Defaults at any time to restore original default settings for the Sensitivity Preferences dialog 6 When all settings are complete click OK to activate them and close the dialog Crystal Ball Note You can copy sensitivity charts and paste them into other applications For more information see page 141 Setting sensitivity chart preferences Settings in the Sensitivity Pr
142. ast Or you can click Apply To to apply settings on the active tab to the active worksheet the active workbook or all workbooks For more information click Apply To and then setting without using Apply To For more information see Table 3 3 on page 67 At any time you can click Defaults to restore the original default settings on the active tab in the dialog 66 Crystal Ball User Manual Defining decision variable cells and forecast cells Table 3 3 Filter tab forecast preferences Setting Effect Set A Filter On The Forecast Values Activates the filter settings for the forecast Include Values In The Range Discards values from the forecast if they fall above or below the two values in the range fields Endpoints are included not excluded Exclude Values In The Range Discards values from the forecast if they fall between the two values in the range fields The range is inclusive Crystal Ball discards values inside the range as well as values equal to the range endpoints When filtering values for this forecast remove values for the same trial from other forecasts too For each trial that a value is not included or is excluded removes the value for that trial from all other forecasts in the model and assumptions too For example if the filter for the current forecast is set to include values from 4 through 10 and the value for the third trial is 12 the value for the third trial will b
143. ast zero which you might want to calculate to make sure that you make a profit might be only 45 For any range you define Crystal Ball calculates the resulting certainty This way not only do you know that your company has a 12 Crystal Ball User Manual Steps for using Glossary Term assumption cell A value cell in a spreadsheet model that has been defined as a probability distribution Glossary Term forecast cell Cells that contain formulas that refer to one or more assumption and decision variable cells and combine the values in the assumption decision variable and other cells to calculate a result Glossary Term decision variable cell Cells that contain the values or variables that are within your control to change The decision variable cells must contain simple numeric values not formulas or text Steps for using Crystal Ball chance to make a profit but you can also quantify that chance by saying that the company has a 45 chance of making a profit on a venture a venture you might therefore decide to skip Benefits of Monte Carlo analysis Crystal Ball uses Monte Carlo simulation to overcome both of the spreadsheet limitations listed earlier e You can describe a range of possible values for each uncertain cell in your spreadsheet Everything you know about each assumption is expressed all at once For example you can define your business phone bill for future months as any value b
144. ation Stock prices are usually positively skewed rather than normally symmetrically distributed Stock prices exhibit this trend because they cannot fall below the lower limit of zero but might increase to any price without limit Similarly real estate prices illustrate positive skewness since property values cannot become negative Example The lognormal distribution can be used to model the price of a particular stock You purchase a stock today at 50 You expect that the stock will be worth 70 at the end of the year If the stock price Crystal Ball User Manual 263 Appendix A Selecting and Using Probability Distributions 264 drops at the end of the year rather than appreciating you know that the lowest value it can drop to is 0 On the other hand the stock could end up with a price much higher than expected thus implying no upper limit on the rate of return In summary your losses are limited to your original investment but your gains are unlimited Using historical data you can determine that the standard deviation of the stock s price is 12 aa Statistical Note If you have historical data available with which to define a lognormal distribution it is important to calculate the mean and standard deviation of the logarithms of the data and then enter these log parameters using the Parameters menu Log Mean and Log Standard Deviation Calculating the mean and standard deviation directly on the raw data wi
145. ation 0 0 0 0 cece cece 78 Steps for running simulations c ccc cece ce ceee cee eeeeeseeeeseceeeeeseeaaes 79 Setting TUN PLELETEM CES 6 i sca sasesnsvsssacsosdevnneadtssodellveartdetannsiascraeavynnvateeesestes 80 Trials Preferences sorce Yel igskdiadat aacestedin beawtasecdelasncaeun E besmaunnions 81 Sampling preferences c sccsecsessseecsscestecceetdeceneecasensesssegedescscseseuceesies 83 Speed PrefEren Ces oirro E cea eadidh ovis E LEE RERI A E treaties 84 OPtlONs preference lt 0 crscseneiseirsaesesseterivocesseasicersnrseneturtonnssiteaesectaes 86 Statistics preferences lt 5 s sesdersiGedavs egecavunes hav egeieesabosessnseessesrcostastldesteniees 87 Freezing Crystal Ball data cells ccc cece cee ceceeeeeeeecseneceeeeseeeseeeeenees 88 Running simulations s cccicc ccsiceceatsess cach uakina ii Eai 90 About running simulations 0 0 0 ce eee ee eee ceseeeeeeeeeeeceeeceeeneeeeaeees 90 Running a similation ssnsdin ana aidai 92 Stopping a simulation sis cicie testes cesies cssenescsnseesaneeesaaveasieesnestenseensteannas 92 Continuing a simulation 2 0 00 eee eee cee cee eee eeceecneeeeeeteeeseeesaeeneeeses 92 Resetting and rerunning a simulation eee eee eeeeeeeeeseeeeeeeeeeees 92 SINGlE StEP PING ss iisncsocdssed iets eavcedieasei demesne tidom AAEE A Ea E 93 The Crystal Ball Control Panel 0 0 cece eee eee ceeeee cee eeeeeesseneceenes 93 Managing chart windows si s 0cssscedesers
146. ation speed Extreme Speed Available only in Crystal Ball Professional Edition and Crystal Ball Premium Edition This option runs simulations up to 100 times faster than Normal mode but is not suitable for some models For details see Appendix C Using the Extreme Speed Feature Normal Speed The standard simulation option for general model processing Demo Speed Runs simulations in slow motion to make it easier to watch values change in spreadsheet cells and charts Options Sets update rules for the active worksheet Normal and Demo speeds only For Normal Update Every Updates Crystal Ball data in Excel after each speed Trial simulation trial Dynamic references are still updated internally if another setting is chosen Update Every _ Defines the update rate in terms of time The Seconds default value is 1 Minimize Minimizes the Excel window This option Workbooks produces the fastest simulations Fastest For Demo Maximum With optimal processing what is the greatest speed Number of number of trials to run each second The Trials Second default value is 10 Chart Windows Sets the redraw rate for any charts open during a simulation Redraw Every _ Seconds Defines the redraw rate in terms of time The default value is 0 5 Suppress Chart Windows Fastest Closes all charts during simulation Selecting this option overrides the Show Window preferences set for any charts This option produces the
147. atistics and Percentiles views page 344 Use Crystal Ball s Split View feature to display the Capability Metrics view to the right of one of the forecast chart views page 345 This is the default View marker lines for specification limits and the target value on forecast charts page 347 Extract capability metrics to a worksheet using Analyze gt Extract Data or the Auto Extract forecast preferences page 349 Include capability metrics in reports page 351 Include several capability metrics in OptQuest if you have Crystal Ball Professional or Premium editions and specify them as objectives or requirements Access capability metrics in the Crystal Ball Developer Kit if available to you using these calls CB SetFore CB GetFore CB DefineForeND CB SetView CB GetForeStat CB RunPrefsND CB GetRunPrefs CB CreateRptND and CB ExtractDataND see the Crystal Ball Developer Kit User Manual for more information Crystal Ball Note Several of these calls have equivalent spreadsheet functions and can be used in worksheet cells instead of VBA macros The following sections provide details 340 Crystal Ball User Manual Activating the process capability features Activating the process capability features Activating the process capability features To activate Crystal Ball s process capability features Ep 1 Choose Run gt Run Preferences to display the Run Preferences dialog 2 Click the Statistics tab Run
148. ave Open On Reset on the Options tab of the Run Preferences dialog For more information on the Run Preferences Options tab see Options preferences on page 86 Crystal Ball User Manual 93 Chapter 4 Running Simulations ox Menubar Run Analyze Help More button Progress bar EARRARRREAERAENANNENNATENENNE 0 Total trials 1 000 1 000 room o Figure 4 10 The Crystal Ball Control Panel When you click the More button simulation statistics appear below the controls They show how fast the simulation ran and how many assumptions decision variables and forecasts were included in it Statistics Run statistics Total running time seconds Trials second average Random numbers generated second Crystal Ball data Assumptions Correlations Correlation groups Decision variables Forecasts Figure 4 11 Control Panel with simulation statistics crystal Ball Crystal Ball Note Clicking on the Excel window title bar brings Excel forward and makes the forecast windows disappear Ifthis happens you can quickly bring the forecast windows back to the front by clicking the Crystal Ball and Excel icons in the Windows task bar to display the chart windows again 94 Crystal Ball User Manual Managing chart windows The Crystal Ball Control Panel menubar The following table describes the Control Panel menus and their commands Table 4 6 Crystal Ball Control Panel menus Menu Commands Run Start Simula
149. bar to specify particular distributions and to choose one of 120 Crystal Ball User Manual Understanding and using forecast charts three fit ranking methods This way also lets you change the fit options or use Apply To to set these preferences for other forecasts To fit a probability distribution to a forecast chart using the Preferences gt Forecast command 1 Create a model and run a simulation 2 Select a forecast chart 3 In the forecast chart menubar choose Preferences gt Forecast 4 In the Forecast Window tab of the Forecast Preferences dialog check Fit A Continuous Probability Distributions To The Forecast and click Fit Options The Fit Options dialog appears as shown in Figure 5 15 Fit Options E x Which distributions 7 Ranking method AII continuous Anderson Darling C Choose Chi Square C Kolmogorov Smirnov Cancel Heip Figure 5 15 Fit Distribution dialog for forecast charts 5 Specify which distributions are to be fitted e All Continuous fits the data to all those distributions in which ever y value in the distribution s range is possible these distributions appear as solid shapes on the Distribution Gallery e Choose displays another dialog from which you can select a subset of the continuous distributions to include in the fitting Note that if you are trying to fit negative data to a distribution that requires positive data that
150. bull distribution called the Rayleigh distribution results A researcher might find the Rayleigh distribution useful for analyzing noise problems in communication systems or for use in reliability studies Calculating parameters There are three standard parameters for the Weibull distribution Location Scale and Shape The Location parameter is the lower bound for the variable The Shape parameter is a number greater than 0 usually a small number less than 10 When the Shape parameter is less than 3 the distribution becomes more and more positively skewed until it starts to resemble an exponential distribution shape lt 1 At a shape of 3 25 the distribution is symmetrical and above that value the distribution becomes more narrow and negatively skewed After you select the Location and Shape parameter you can estimate the Scale parameter The larger the scale the larger the width of the distribution To calculate a more exact scale estimate the mean and use the equation _ mean L i where ois the scale B is the shape L is the location and T is the gamma function You can use the Excel GAMMALN function and Excel Solver to help you calculate this parameter Statistical Note For this distribution there is a 63 probability that x falls between aand O L Or estimate the mode and use the equation _ mode L 1 0 where is the scale P is the shape and Z is the location Crystal Ball User Manual 279
151. bution Gallery click a category in the Category pane 2 Click a distribution in the Distribution pane 3 Choose Edit gt Copy or right click and choose Copy from the right click menu 4 Click Favorites or another user defined category in the Category pane 5 Choose Edit gt Paste or right click and choose Paste from the right click menu Crystal Ball User Manual Additional Distribution Gallery features The distribution is pasted into the selected category If another distribution in that category had the same name the new distribution appears with the next larger number at the end For example if the category already had Normal a new normal distribution would be named Normal2 To rename the copied distribution click its name and wait a moment The name is highlighted for editing Type the new name and press Enter Modify the new distribution as needed Crystal Ball Note To easily copy a distribution from one category to another select the category with the distribution you want to copy Then drag the distribution into the closed category folder where you want to copy it When you select that category the copied distribution appears and you can rename it or modify it as described in the next section However the distribution you copied remains in its original folder unchanged Modifying distributions Distributions in the Basic or All categories are read only and cannot be changed However once you have copie
152. cal error such as division by zero occurs in any forecast cell If a calculation error occurs to let you find the error Crystal Ball doesn t restore the cell values If no calculation errors occur the simulation continues until it reaches the Number Of Trials To Run or if set when the specified precision is reached Crystal Ball Note In Extreme Speed the simulation stops at the end of a burst of trials when an error occurs and not immediately upon detecting the error Stop When Precision Control Limits Are Reached When checked stops the simulation when certain statistics reach a specified level of precision You choose the statistics and define the precision that triggers this option in each Define Forecast dialog For instructions see Precision preferences on page 64 Any forecasts set to use precision control must all reach their specified precision within the confidence level to stop the simulation If all the forecasts set to use precision control don t meet the specified precision the simulation stops when it reaches the Number Of Trials To Run By default precision control is on Confidence Sets the precision level confidence level that Level indicates when to stop a simulation 82 Crystal Ball User Manual Setting run preferences Sampling preferences a x Sampling Speed Options Statistics Trals M Random number generation IV Use same sequence of random numbe
153. can choose from among several chart display types such as column line area bar or pie To change the chart display type 1 Display the Chart Type tab of the Chart Preferences dialog The Chart Type group appears in the middle of the Chart Type page If more than one series appears in the list box at the top of the tab select one to work with The settings on the tab apply to the selected series 2 To change the chart display type open the Type drop down list and select a display type Depending on the basic chart and series types you can choose from among these display types Table 5 7 Chart types Example Type Description Column Displays data as vertical columns that correspond to the group intervals chart bins of the data The column chart is the default chart type for generated data in assumption forecast and overlay charts 132 Crystal Ball User Manual Setting chart preferences Table 5 7 Chart types Continued Example Type Description ssl Line Shows data as an outline of the peaks and pi a as valleys My F f TAANA i af N l ki nd 1M emma Di gt a ss Area Shows data as darkened peaks and valleys E f cS we me Bar Shows sensitivity data as horizontal bars to the Sex directional right and left of the 0 line showing magnitude 255 and direction of sensitivity E 2 Bar Shows sensitivity data as horizontal bar
154. can help you discover if any of your assumptions have non monotonic relationships with the target forecast For more information see Tornado Chart tool on page 211 Discrete distributions Sensitivity calculations might be inaccurate for assumptions or forecasts that have a small set of discrete values Because the sensitivity calculation relies on rank correlation a slight loss of information occurs when the assumption or forecast values are replaced by ranks This loss of information is generally offset by the advantage of Crystal Ball User Manual 171 Chapter 6 Analyzing Other Charts 172 being able to measure sensitivity between dissimilar types of distributions However when a large percentage of assumption or forecast values are very similar or identical this loss of information grows and can significantly distort the calculation of correlations Be aware of this problem for example for assumptions when using distributions such as Binomial with a small Trials parameter for example lt 10 e forecasts when formulas in your spreadsheet result in identical values for example if then logic INT functions and so on Customizing sensitivity charts You can customize sensitivity charts by changing the assumptions forecasts and sensitivity preferences As you become more familiar with the sensitivity chart practice choosing the options that help you get the answers you seek and that are appropriate for
155. cast charts in Split View which shows charts and statistics side by side This section focuses on each type of chart separately For more information about Split View and its features see Using Split View on page 122 The highest value on the probability scale to the left of the chart is the probability for the mode You can estimate the probability for other points on the forecast chart using this probability scale In the example below the mode in the middle of the distribution has a frequency of 32 meaning that there are 32 values in the interval expressed by that column The mode has a probability of 0 065 or 6 5 meaning that there is a 6 5 chance of a value falling within this interval Figure 5 1 shows the elements of the forecast chart Forecast Net Profit MM a o x Number of trials Edit View Forecast Preferences Help 500 Trials Frequency View 498 Displayed Number of trials displayed Certainty grabbers Certainty level Certainty range Net Profit MM 32 fF 28 24 N c o z Aou nbaJy H16 Probability 10 0 0 0 10 0 20 0 Millions gt Certainty fioo Z q Hnfinity Figure 5 1 Forecast chart Crystal Ball User Manual Understanding and using forecast charts Crystal Ball forecasts the entire range of results for a given situation However the forecast chart shows only the display range which is a subrange of
156. cation Data Decisioneering Crystal Ball 7 x Store Crystal Ball User Manual Additional Distribution Gallery features The next time you start Crystal Ball the new category appears in the Category pane of the Distribution Gallery Alternately you can save the attachment in another folder then follow the steps in Using shared categories on page 53 to subscribe to it Editing shared categories If you have published a user defined category and want to change it 1 Make the changes in your Distribution Gallery 2 Follow the steps in Sharing categories over networks on page 52 to publish the category to the same place 3 When you are prompted to overwrite the existing file click OK This publishes the modified category Now the next time a subscriber loads Crystal Ball the modified category appears in the Distribution Gallery Unpublishing shared categories If you have published a user defined category and want to remove it from publication or unpublish it 1 Use My Computer or Windows Explorer to browse to the folder where the category was published The category file appears as CBCategory_name cbc For example the Favorites category would appear as CBCategory_Favorites cbc 2 Select the category file and delete it Now the next time a subscriber loads Crystal Ball the deleted category will not appear in the Distribution Gallery Existing assumptions based on distributions from that category w
157. cell at a time As a result exploring the entire range of possible outcomes is next to impossible you cannot realistically determine the amount of risk that is impacting your bottom line e What if analysis always results in single point estimates which do not indicate the likelihood of achieving any particular outcome While single point estimates might tell you what is possible they do not tell you what is probable This is where simulation with Crystal Ball comes in Monte Carlo simulation and Crystal Ball Glossary Term spreadsheet model any spreadsheet that represents an actual or hypothetical system or set of relationships Glossary Term simulation Any analytical method that is meant to imitate a real life system especially when other analyses are too mathematically complex or too difficult to reproduce Spreadsheet risk analysis uses both a spreadsheet model and simulation to analyze the effect of varying inputs on outputs of the modeled system One type of spreadsheet simulation is Monte Carlo simulation which randomly generates values for uncertain variables over and over to simulate a model History Monte Carlo simulation was named for Monte Carlo Monaco where the primary attractions are casinos containing games of chance Games of chance such as roulette wheels dice and slot machines exhibit random behavior The random behavior in games of chance is similar to how Monte Carlo simulation s
158. centiles and activates Crystal Ball s process capability features For details on each tab see the referenced pages or click the Help button in the Run Preferences dialog Change any preferences on any tab 4 Click OK 5 To reset settings on the active tab to the original defaults click Defaults Trials preferences Run Preferences Sampling Speed Options Statistics Number of trials to run 5000 IV Stop on calculation errors IV Stop when precision control limits are reached Confidence level e5 Figure 4 2 Trials tab Run Preferences dialog The Trials tab sets preferences that stop a simulation number of trials calculation errors and precision control For general instructions see Setting run preferences beginning on page 80 Table 4 1 describes the settings available on this tab Crystal Ball Note The current simulation must be reset before precision control settings will take effect Crystal Ball User Manual 81 Chapter 4 Running Simulations Table 4 1 Trials tab Run Preferences dialog Setting Effect Number Of Trials To Run Defines the maximum number of trials that Crystal Ball runs before it stops the simulation If you check either of the checkboxes on this dialog Crystal Ball only uses the maximum number of trials if forecast results do not meet the other stop criteria first Stop On Calculation Errors When checked stops the simulation when a mathemati
159. ces settings are discussed in Setting chart preferences beginning on page 128 If you choose Assumption Preferences the Assumption Preferences dialog appears as shown in Figure 2 16 Assumption Preferences x Assumption Window View Probability v Window IV Show automatically While running simulation When simulation stops To show the generated random values in the window turn on the Store assumption values run preference Run Preferences Cancel Apply To Defaults Help Figure 2 16 The Assumption Preferences dialog This dialog lets you Choose a view for the assumption chart e Probability shows a graph of all possible values for the assumption variable and the probability of their occurrence e Cumulative Probability shows a graph of the probability that the assumption variable will fall at or below a given value e Reverse Cumulative Probability shows a graph of the probability that the assumption variable will fall at or above a given value Statistics shows a table of measures of central tendency variability minimum and maximum values and other statistics for the assumption variable Crystal Ball User Manual 39 Chapter 2 Defining Model Assumptions e Percentiles shows a table of percentiles and their associated values for the assumption variable Note For examples of each view see Changing the distribution view and interpr
160. cess Capability Features Define Forecast CNB me A w A w aa 2 Target EF Figure D 3 Define Forecast dialog with limits and target fields 2 Enter specification limits and target values for this forecast into the appropriate fields LSL lower specification limit USL upper specification limit and Target target value for this forecast If you prefer you can enter cell references by typing or browsing All of these fields are optional but Crystal Ball only calculates capability metrics if a value is entered for one or both of the specification limits w 3 To set forecast preferences at the same time click the More button 4 When all settings are complete click OK Crystal Ball Note For information about the relationship of the LSL and USL to the certainty range see the next section Viewing LSL USL and Target marker lines Viewing capability metrics 344 Once you have defined a forecast with at least one specification limit and optionally a target you can run a simulation and display capability metrics for the forecast To display capability metrics 1 Define a forecast with LSL USL and Target values as described in Setting specification limits and targets on page 343 2 Run the simulation and display the chart for that forecast 3 In the forecast window choose View gt Capability Metrics A table of metrics appears similar to the following figure Crystal Ball Us
161. charts Crystal Ball User Manual 99 Chapter 4 Running Simulations Table 4 8 User defined macros Continued Macro Name Runs CBAfterSimulation When the simulation is complete or stopped with a Stop command by single stepping after each step or for any other reason The following figure shows where each macro fits into the Crystal Ball simulation cycle E te ke CBBeforeSimulation gt DS te a ee Ges see CBBeforeTrial Set Values Recalculate CBAfterRecalc CBAfterTrial CBAfterSimulation Figure 4 14 When user defined macros run Crystal Ball searches for macros with these names in open workbooks and runs them at the appropriate points during the simulation The macros must be contained within workbooks and worksheets not VBA modules and Run User Defined Macros must checked on the Options tab of the Run Preferences dialog Crystal Ball Note Crystal Ball checks for the presence of any CBBeforeTrial CBAfterTrial and CBAfterRecalc macros before the simulation runs by attempting to execute them If this poses a problem you may skip the logic in your macro by first testing for aTrial 0 100 Crystal Ball User Manual Running user defined macros For information on running user defined macros in Extreme speed see Running user defined macros on page 330 Interfaces You must use these VBA interfaces for the macros Public Function CBBeforeSimulation As
162. chical Tree view If you prefer click the List box to change the view from a tree to a list For more information on the Choose Forecast dialog click the Help button Check the box in front of the forecast name to include it in the sensitivity chart Toxic Waste xls has only one forecast Risk Assessment Click OK to create a new sensitivity chart as shown in Figure 6 21 The illustrated chart has a transparency effect applied using the chart preferences to make all the numbers easier to read For instructions see Setting special chart effects beginning on page 136 Crystal Ball User Manual Menubar Chart view Chart title Understanding and using sensitivity charts Q Riske assessment TST Edit View Sensitivity Preferences Help 1 000 Trials Contribution to Variance View Sensitivity Risk Assessment 0 0 21 0 42 0 63 0 Volume of Water per Day z S 65 2 f nace CPF 25 5 Body Weight 7 2 Concentration of Contaminan Figure 6 21 Sensitivity chart for the selected forecast The assumptions are listed on the left side starting with the assumption with the highest sensitivity If necessary use the scroll bar to view the entire bar chart it is at the left of the chart You can drag the edges of the chart to resize it make it narrower wider taller or shorter This can also change the tick labels along the top of the chart Crystal Ball Note
163. column or bin check Show Column Gaps Crystal Ball Note Even if you uncheck Show Column Gaps gaps will always appear in a discrete distribution 5 When all your settings are complete click OK The chart appears with your new settings Crystal Ball User Manual Setting chart preferences Showing grid lines Grid lines are vertical or horizontal lines that can appear on charts to help you compare the charted data with the values along the horizontal and vertical axes of the chart Crystal Ball Note For basic chart customization instructions see page 130 To hide or show grid lines 1 Display the General tab of the Chart Preferences dialog page 130 2 Locate the Options group in the middle of the page 3 Choose a setting from the Gridlines drop down list to show only horizontal grid lines Horizontal only vertical grid lines Vertical both horizontal and vertical grid lines Both or choose None to hide both horizontal and vertical grid lines 4 When all your settings are complete click OK The chart appears with the new settings Crystal Ball Note You can press Ctrl l to toggle the horizontal grid lines on and off Showing the chart legend The legend shows the name and chart color for each series in the chart Crystal Ball Note For basic chart customization instructions see page 130 To hide or show a chart legend 1 Display the General tab of the Chart Preferences dialog pa
164. comparison purposes continuous probability distribution A probability distribution that describes a set of uninterrupted values over a range In contrast to the discrete distribution the continuous distribution assumes there is an infinite number of possible values correlation In Crystal Ball a dependency that exists between assumption cells 364 Crystal Ball User Manual correlation coefficient A number between 1 and 1 that specifies mathematically the degree of positive or negative correlation between assumption cells A correlation of 1 indicates a perfect positive correlation minus 1 indicates a perfect negative correlation and 0 indicates there is no correlation cumulative frequency distribution A chart that shows the number or proportion or percentage of values less than or equal to a given amount decision variable cell Cells that contain the values or variables that are within your control to change The decision variable cells must contain simple numeric values not formulas or text decision variable A Crystal Ball variable in your model that you can control deterministic model Another name for a spreadsheet model which yields single valued results discrete probability distribution A probability distribution that describes distinct values usually integers with no intermediate values In contrast the continuous distribution assumes there is an infinite number of possible values display ran
165. confidence interval If you want to reset forecast preference defaults to the original settings shipped with Crystal Ball click Defaults To copy preferences to other forecasts click Apply To Crystal Ball User Manual Understanding and using forecast charts You can specify whether to apply all forecast preferences or just the current tab and whether to apply them to the current Excel sheet all sheets in the workbook or all open and new workbooks When all settings are complete click OK to apply them Setting forecast chart preferences The chart preferences can help you customize the appearance of forecast charts For a list of available settings see Table 5 3 on page 125 To make these settings choose Preferences gt Chart Preferences in the forecast chart menubar The following chart preference settings are especially helpful for chart interpretation Chart type lets you show forecast charts as columns areas or lines in two or three dimensions so you can view the data plot from different perspectives and grasp the big picture more easily Chart density increases and decreases the number of bars or data points so you can spot trends more easily Grid lines make it easier to determine frequencies and probabilities Marker lines make it easier to locate means medians modes percentiles and other important values Axis scaling and rounding let you show more or fewer axis values to read
166. constant from week to week These conditions match those of the gamma distribution Note that in this example the dealership has made several simplifying assumptions about the conditions In reality the total number of computer purchasers is finite and some might have influenced the purchasing decisions of others The shape parameter is used to specify the r th occurrence of the Poisson event In this example you would enter 20 for the shape parameter 5 units per week times 4 weeks The result is a distribution showing the probability that x number of business days will pass until the 20th system is sold Figure A 8 illustrates the gamma distribution loxi Edit View Parameters preferences Help Name ja xj Gamma Distribution So AD Aaa a T a E E 2s ate E Location 0 zj Scalefi Shape 20 x OK Cancel C Ete Gallery Correlate Help Figure A 8 Gamma distribution Example two Suppose a particular mechanical system fails after receiving exactly 5 shocks to it from an external source The total time to system failure defined as the random time occurrence of the 5th shock follows a gamma distribution with a shape parameter of 5 Crystal Ball User Manual Using continuous distributions Some characteristics of the gamma distribution e When shape 1 gamma becomes a scalable exponential distribution The sum of any two gamma distributed variables is a gamma variable e Ifyou have historical
167. ct from the list of forecasts above toi Field Development xis Model C 30 j Figure 8 29 Specify Target dialog Decision Table tool 5 Select the NPV forecast 6 Click Next The Select One Or Two Decisions dialog appears as shown in Figure 8 30 Select one or two decisions step 2 of 3 Decision Table Crystal Ball Tool Decisioneering 1998 2004 Select one or two decisions to evaluate Figure 8 30 Select One or Two Decisions dialog Decision Table tool 228 Crystal Ball User Manual 10 Decision Table tool Move Wells To Drill and Facility Size to the Chosen Decision Variables list a Select Wells To Drill in the Available Decision Variables field b Click gt gt c Repeat steps 7a and 7b for the Facility Size Click Next The Specify Options dialog appears as shown in Figure 8 31 Decision Table Crystal Ball Tool Decisioneering 1998 2004 rSimulation Control Test fe values for Wells to drill Test 7 values for Facility size Run each simulation for 500 trials maximum While Running Show forecasts as defined Show only target forecast Hide all forecasts lt Back Cancel Help Figure 8 31 Specify Options dialog Decision Table tool Set the following options as shown above Number of values to test for Wells To Drill 6 Number of values to test for Facility Size 7 e Maximum trials per simulation 500 Show Only Target Forecast is s
168. d Value of 999 e Monte Carlo simulation When using this tool use these options to make the resulting simulations comparable 3 Select Run gt Tools gt 2D Simulation K Crystal Ball Note Ifthe 2D Simulation command is not available reset the simulation and try again The Specify Target dialog appears as shown in Figure 8 39 Specify target step 1 of 3 2D Simulation Crystal Ball Tool Decisioneering 1993 2004 Independently analyze uncertainty and variability using two dimensional simulation Risk Assessment Please enter the target cell or select from the list of forecasts above Toxic Waste Site xls Model scs9 Y Figure 8 39 Specify Target dialog 2D Simulation tool Crystal Ball User Manual 237 Chapter 8 Crystal Ball Tools 4 Select the Risk Assessment forecast 5 Click Next The Specify Assumptions dialog appears as shown in Figure 8 40 Specify assumptions step 2 of 3 2D Simulation Crystal Ball Tool Decisioneering 1998 2004 Identify the types of your assumptions Uncertainty Variability Body Weight Volume of Water per Day Concentration of Contaminant in CPF ped KN Cancel Help Figure 8 40 Specify Assumptions dialog 2D Simulation tool 6 Move Body Weight and Volume Of Water Per Day to the Variability list a Select Body Weight b Click gt gt c Repeat steps 6a and 6b for Volume Of Water Per Day This separates the assump
169. d a distribution or used it to define an assumption you can modify it in a number of ways To modify a distribution 1 Create a distribution as described in Creating new distributions on page 45 and Copying and pasting distributions on page 46 Right click the user defined distribution and choose Modify or click the distribution and choose Edit gt Modify in the Distribution Gallery menubar A prompt appears to remind you that you are modifying a distribution and not defining an assumption Click OK to display the distribution in a window that looks like the Define Assumption dialog Modify the distribution as if it were an assumption following the instructions in Defining assumptions beginning on page 18 to modify the distribution When you are finished click OK The modifications are saved with the distribution in the category and no assumption is created in the spreadsheet Crystal Ball User Manual 47 Chapter 2 Defining Model Assumptions 48 Modifying distribution summaries and descriptions You can edit the distribution Summary and Description text for any distribution in Favorites or another user defined category To modify Summary text for a distribution 1 In the Distribution Gallery set the View menu to Details 2 Inthe Distribution pane right click the Summary text you want to edit 3 In the right click menu choose Modify Summary For distribution name 4 When the Modify S
170. data that you believe fits the conditions of a gamma distribution computing the parameters of the distribution is easy First compute the mean x and variance s of your historical data Then compute the distribution s parameters shape parameter x7 s scale parameter s x Chi square and Erlang distributions You can model two additional probability distributions the chi square and Erlang distributions by adjusting the parameters entered in the Gamma Distribution dialog To model these distributions enter the parameters as described below Chi square distribution With parameters N and S where N number of degrees of freedom and S scale set your parameters as follows shape A scale 2S The chi square distribution is the sum of the squares of N normal variates Erlang distribution The Erlang distribution is identical to the gamma distribution except the shape parameter is restricted to integer values Mathematically the Erlang distribution is a summation of N exponential distributions Crystal Ball User Manual 261 Appendix A Selecting and Using Probability Distributions 262 Logistic distribution Logistic Parameters Mean Scale Description The logistic distribution is commonly used to describe growth i e the size of a population expressed as a function of a time variable It can also be used to describe chemical reactions and the course of growth for a population or i
171. decision variables or forecasts from one area of the spreadsheet to another area in the same workbook or a different workbook 1 Select a cell or a range of cells that contains the Crystal Ball data you want to copy 2 Select Define gt Copy Data If you select a range of cells with more than one kind of Crystal Ball data for example an assumption and a forecast Crystal Ball prompts for which data type you want to copy 3 Check the type or types to copy and click OK Crystal Ball remembers the description of the selected cell range Pasting Crystal Ball data To paste Crystal Ball data 1 Select a cell or a range of cells that you want to paste into Usually you will want it to be at least as large as the range you copied It should contain cells with values if you are pasting assumptions or decision variables unless the range is all blank see note below and cells with formulas if you are pasting forecasts It can be on a different worksheet in the same or a different workbook Crystal Ball Note If you are pasting assumptions and or decision variables into a completely blank range of cells Crystal Ball pastes them along with the underlying cell value from each copied cell This does not work for forecasts which must be pastsed into a cell with a formula or if there is already data in any cell within the range ie 2 Choose Define gt Paste Data Crystal Ball pastes all selected data types assumptions
172. defined in Windows Regional and Language Options dialog If actual values in the file contain commas or the designated list separator those values must be enclosed in quotes Allowable formats for values are identical to those allowed within the assumption parameter dialog including date time currency and numbers Specify which distributions are to be fitted e All Continuous fits the data to all of the built in continuous distributions these distributions appear as solid shapes on the Distribution Gallery e Choose displays another dialog from which you can select a subset of the continuous distributions to include in the fitting The third option selects the continuous distribution that was highlighted on the Distribution Gallery when you clicked the Fit button Crystal Ball Note If you try to fit negative data to a distribution that can only accept positive data that distribution will not be fitted to the data Specify how the distributions should be ranked In ranking the distributions you can use any one of three standard goodness of fit tests Anderson Darling This method closely resembles the Kolmogorov Smirnov method except that it weights the differences between the two distributions at their tails greater than at their mid ranges This weighting of the tails helps to correct the Kolmogorov Smirnov method s tendency to over emphasize discrepancies in the central region e Chi Square This test is th
173. describes an uncertain situation 2 Runa simulation on it 3 Analyze the results This User Manual is structured to match those main tasks The guidelines on page 13 fill in some details and indicate where each task is discussed If you are new to Crystal Ball and risk analysis tools you might not be familiar with or know what is meant by models or risk analysis Or if you do know you might want a better understanding of how Crystal Ball performs a risk analysis The following sections give a brief overview of risk analysis and modeling They build a foundation for understanding the many ways Crystal Ball and related Decisioneering products can help you minimize risk and maximize success in virtually any decision making environment risk analysis Uncertainty is usually associated with risk where risk includes the possibility of an undesirable event coupled with severity For example if sales for next month are above a certain amount a desirable event then orders will reduce the inventory If the reduction in inventory is large enough there will be a delay in shipping orders an undesirable event Ifa delay in shipping means losing orders severity then that possibility presents a risk As uncertainty and risk increase decision making becomes more difficult 8 Crystal Ball User Manual Model building and risk analysis overview There are two points to keep in mind when analyzing risk e Where is the risk How sig
174. distributions Now experiment with chart hot keys Press Ctrl d again until the frequency distribution appears Press Ctrl t to display the column chart Try pressing Ctrl b to change the number of frequency bins columns in this view Then press Ctrl 3 to make the chart three dimensional Overlay Chart 1 Figure 6 6 Overlay chart 3D view Crystal Ball User Manual Understanding and using overlay charts If you want you can drag the right or left edge of the chart to give it a taller narrower look as shown in Figure 6 6 or a stretched look as in Figure 6 7 In 3D view the Enable Rotation checkbox appears in the upper right corner of the chart When it is checked you can click inside the chart and drag to rotate it This can enhance the data display for both analysis and presentation Figure 6 7 shows a rotated overlay chart Q overlay Chart 1 Edit View Overlay Preferences Help 1 000 Trials Frequency View IV Enable Rotation sl B xd Probability Overlay Chart Wi Material 1 Reliability Material 2 Reliability Wi Material 3 Reliability AOuenbel4 Figure 6 7 A rotated overlay chart Crystal Ball Note Note that rotation settings are for the current session only and are not saved with the chart For more customization suggestions see the next section Customizing overlay charts You can customize overlay charts in a variety of ways Use the View menu in the
175. dom numbers A given seed value produces the same sequence of random numbers every time you run a simulation sensitivity The amount of uncertainty in a forecast cell that is a result of both the uncertainty probability distribution and model sensitivity of an assumption cell sensitivity analysis The computation of a forecast cell s sensitivity with respect to the assumption cells skewed An asymmetrical distribution skewed negatively A distribution in which most of the values occur at the upper end of the range skewed positively A distribution in which most of the values occur at the lower end of the range skewness The amount a curve differs from a normal symmetrical distribution The greater the degree of skewness the more points of the curve lie to either side of the peak of the curve A normal distribution curve having no skewness is symmetrical Skewness is computed by finding the third moment about the mean and dividing by the cube of the standard deviation Crystal Ball User Manual 369 Glossary spreadsheet model Any spreadsheet that represents an actual or hypothetical system or set of relationships standard deviation The square root of the variance for a distribution A measurement of the variability of a distribution i e the dispersion of values around the mean See formulas in Standard deviation in the Statistical Definitions chapter of the online Crystal Ball Reference Manua
176. dow and hit Esc A dialog will appear and you can choose to End or Continue 198 Crystal Ball User Manual Batch Fit tool Batch Fit tool The Batch Fit tool fits probability distributions to multiple data series You can select any or all of the continuous probability distributions normal triangular uniform etc to fit to any number of series limited only by the size of your spreadsheet Batch Fit is intended to help you create assumptions when you have historical data for several variables It selects which distribution best fits each series of historical data and gives you the distribution and its associated parameters for you to use in your model This tool also gives you a table of goodness of fit statistics for the best fitting distribution and provides a matrix of correlations calculated between multiple data series so you can easily see which series are related and to what degree To use the Batch Fit tool your data series must be contiguous in adjacent rows or columns in either rows or columns You can select any combination of the continuous probability distributions to fit to all your data series Batch Fit example In the Crystal Ball Examples folder there is a Magazine Sales xls spreadsheet you can use to experiment with the Batch Fit tool This spreadsheet model shows the estimated gross profit resulting from newsstand sales of four of the company s most popular magazines Ej Magazine
177. e the band which represents the 90 certainty range shows the range of values into which a forecast has a 90 chance of falling By default the bands are centered around the median of each forecast The trend chart in Figure 6 9 displays certainty ranges on a quarterly basis over a three year period Because the model contains quarterly forecast formulas dependent on the previous quarter s results the bands widen in the future This occurs because the standard deviation of the forecasts increases or widens for each quarter Trend charts like this one demonstrate the compounding of uncertainty that occurs as predictions are made farther and farther into the future 154 Crystal Ball User Manual Understanding and using trend charts You can customize trend charts to display the probability that given forecasts will fall in a particular part of a value range For example if a model contains forecasts related through time you can use trend charts to view the certainty ranges for each forecast side by side You can also compare at a glance the certainty ranges for an early time period and a later time period Crystal Ball Note Trend charts are only meaningful when you have multiple forecasts that are related to each other Creating trend charts k The following are basic steps for creating trend charts The steps are illustrating using an example spreadsheet Sales Projection xls shipped with Crystal Ball To create a trend chart
178. e 3D display on and off Setting chart colors This preference sets the color of the current chart series This is the color that appears for the series in the chart legend if visible Crystal Ball Note For basic chart customization instructions see page 130 To change the chart colors 1 Display the Chart Type tab of the Chart Preferences dialog The Chart group appears in the middle of the Chart Type page If more than one series appears in the list box at the top of the tab select one to work with The settings on the page apply to the selected series 2 Open the Color drop down list and choose a color 3 While you have a series selected consider adjusting the chart type page 132 and marker line settings page 137 too 4 When settings for the current series are complete follow steps 2 through 3 to customize settings for any other series in the chart Crystal Ball User Manual Setting chart preferences 5 When all your settings are complete click OK The chart appears with your new settings Showing the mean and other marker lines You can display mean mode median standard deviation and other marker lines on assumption forecast and overlay charts These lines help you locate various values in the charted distribution Crystal Ball Note Ifyou have activated the process capability features and have entered an LSL USL or Target value you can include marker lines for them in your forecast chart Fo
179. e A 37 Location of data 42 D7 EF IV Keep linked to spreadsheet When loading data unlinked Replace existing distribution Append to easting distribution Tl Probabilities are cumulative a a Figure A 37 Load Data dialog Custom Distribution The default settings are appropriate for most purposes but the following other options are available Crystal Ball User Manual Using the custom distribution e When loading unlinked data you can choose to replace the current distribution with the new data or append new data to the existing distribution e Ifprobabilities are entered cumulatively into the spreadsheet you are loading you can check Probabilities Are Cumulative Then Crystal Ball determines the probabilities for each range by subtracting the previous probability from the one entered for the current range You will need to choose View gt Cumulative Probability to display the data cumulatively in the assumption chart 3 Enter a location range for your data When all settings are correct click OK Crystal Ball enters the values from the specified range into the custom distribution and plots the specified ranges as shown in Figure A 38 G Define Assumption Cell C14 Edit View Parameters Preferences Help Name C14 xi Custom Distribution 005 D 004 0 03 a 3 002 Oo P NU 500 10 00 20 00 30 00 40 00 50 00 60 00 70 00 80 00 90 00 100 00
180. e Distribution Gallery Crystal Ball User Manual 45 Chapter 2 Defining Model Assumptions 46 The following sections explain both methods Creating distributions by copying To create a distribution by copying another distribution follow the steps in the next section Copying and pasting distributions Then you can rename and modify the new distribution as needed Creating distributions by saving assumptions To create a distribution by saving an assumption 1 With Crystal Ball running click a cell in the spreadsheet It can be a value cell where you want to define an assumption or a blank cell 2 Click the Define Assumption tool or choose Define gt Define Assumption 3 Follow the steps beginning on page 18 to select a distribution and enter parameters 4 Choose Edit gt Add To Gallery in the Define Assumption dialog 5 The Add To Gallery dialog appears where you can name the new distribution and select a category for it 6 Click OK to save the distribution in the selected category If you created correlation data it is not saved although the distribution type and any parameter settings are saved Now the new distribution is available for use just as another other distribution either supplied with Crystal Ball or user defined Copying and pasting distributions To copy and paste a distribution within Favorites or another user defined category or from one category to another 1 In the Distri
181. e Forecast Preferences dialog Each method absolute or relative has its own benefits and drawbacks Specifying precision in absolute terms can give you greater control of the simulation when the shape and scale of the forecast distribution is roughly known For example for a Gross Profit forecast from the Vision Research model that ranges from 25 5 to 64 0 million dollars you can require the precision of the mean to be within plus or minus 100 000 or some other convenient measure of accuracy However with the same forecast range an absolute accuracy of 1000 might require an unreasonably large number of trials to reach So the drawback of using absolute precision is that it might require experimentation to determine reasonable accuracy values Q Forecast Gross Profit if Approved MM loj x Edit View forecast Preferences Help 1 000 Trials Frequency Chart 999 Displayed Gross Profit if Approved MM Aouenbel4 Probability 25 0 300 350 400 45 0 500 550 600 Millions D infinity Certainty 100 0 q Infinity Figure B 1 Absolute precision control example Specifying precision in relative terms can give you greater control of the simulation when the shape and scale of the forecast distribution is largely unknown and you are interested in the accuracy only as it relates to the overall distribution itself In the Gross Profit example above you might not know or care if the distribution
182. e Weibull distribution called the Rayleigh distribution results A researcher might find the Rayleigh distribution useful for analyzing noise problems in communication systems or for use in reliability studies When the shape parameter is set to 3 25 the Weibull distribution approximates the shape of the normal distribution however for applications when the normal distribution is appropriate us it instead of the Weibull distribution The gamma distribution is also a very flexible family of distributions When the shape parameter is 1 0 the gamma distribution is identical to the exponential distribution When the shape parameter is an integer greater than one a special form of the gamma distribution called the Erlang distribution results The Erlang distribution is especially useful in the areas of inventory control and queueing theory where events tend to follow Poisson processes Finally when the shape parameter is an integer plus one half e g 1 5 2 5 etc the result is a chi squared distribution useful for modeling the effects between the observed and expected outcomes of a random sampling When no other distribution seems to fit your historical data or accurately describes an uncertain variable you can use the custom distribution to simulate almost any distribution The Load Data button on the Custom Distribution dialog lets you read a series of data points or ranges from value cells in your worksheet If you like you can use
183. e eeeceeceeceeseene esses seeseseenseees 151 Using distribution fitting with overlay charts 0c ccc cece eee 152 Understanding and using trend Charts c cece eee ee eee eee cee eens eneeeaee 154 Creating trend Charts cciccciaccstisessleteesiesteetessasecisesusaeata cusssueeastasiaaas 155 Customizing trend charts wo ccc cece ceeeceeeecesaecnsaeceseeseeeesnseeseaees 157 Understanding and using sensitivity Charts cccccseeesces cesses entrene tenne 165 AbDOUt SEnSitIVIty Charts sserrep innn EEE AS 166 Creating sensitivity Charts 0 c cece cee ece cree cee eeeeeseeesaecneeseeesssecaeees 167 How Crystal Ball calculates sensitivity cccccsecceeeeetereeeeeeeeeee 170 TMM ALOIS ee sanoa aan aaa aa o EEA ara aane 171 Customizing sensitivity Charts 0 0 cece cece nttttstnttrtnnnen reneste ent 172 Understanding and using assumption charts 0 0 0 0 cee eee eee ceeeeeeeeeees 179 Customizing assumption Charts cecee eee eeeeseeeeeeseeceeseeceeeeeeeeseees 179 Chapter 7 Creating Reports and Extracting Data Creating reports nosier a sSaedssnovededevensoaniisnn lost vveetivedesstaucnddatesdneae 182 BASIC STEPS cores shat cutie eona e E NEE NE A E E T EREE ES 183 Crystal Ball User Manual iii Contents Defining CustOm PEPOrts wiisscsstncssectes netini iE E ASER EE 185 Extracting ata isc ssicssitsceticesetsscuelssttelvasteas a snaecdetiaws tages antes AE ENSS ESA 190 Chapter 8 Crystal Ball Tools CO
184. e filtered from the current forecast and all other forecasts and assumptions in the model regardless of the values in the other forecasts If this setting is checked and you run a Forecasts report globally filtered appears for the forecast in the Summary data following the filter description Crystal Ball Note This setting can be checked in the Define Forecast dialog for several forecasts with different filter settings In that case filtering for each selected forecast will be applied across all forecasts Crystal Ball User Manual 67 Chapter 3 Defining Other Model Elements Auto Extract preferences Define Forecast Cell A2 i x Name e3 A Units Forecast Window Precision Filter Auto Extract Extract forecast statistics automatically to your spreadsheet Iv SEAN when the simulation stops O Forecast Name O Trials O Mean O Median O Mode a Starting cell E g Al gt Formatting _ JV Include labels O Std Deviation IV AutoFormat O variance O Skewness RAN m Direction O Kurtosis O Coeff of Variability Fill downward C Minimum Fillto the ri ight CO Maximum xi Cancel Appl To Defauts Help Figure 3 6 Auto Extract tab forecast preferences This tab lets you specify which statistics to extract automatically to Excel after the simulation stops Crystal Ball Note The Auto Extract settings create tables of unforma
185. e fit statistics for each of the selected distribution types The distributions are ordered according to the selected ranking method Figure 5 13 shows statistics for each ranking method and each continuous distribution type Notice that Beta is ranked highest for all three methods for this forecast Q Forecast material vRelabity TE Edit View Forecast Preferences Help 1 000 Trials Goodness of Fit View 1 000 Displayed Selected test for ranking distributions Chi Square Distribution AD Chi Square KS Parameters Beta Pa Z Minimum0 86 Maximum 1 39 Alpha S 25 Beta Weibull 4 17 78 02 Location 0 91 Scale 0 24 Shape 3 01 Normal 18 19 02 Mean 1 13 Std Dev 0 08 Lognormal 23 99 03 Mean 1 13 Std Dev 0 08 Student s t 26 25 03 Midpoint 1 13 Scale 0 08 Deg Freedom 30 00 Gamma 27 01 03 Location 0 49 Scale 0 01 Shape 66 51 Logistic f 49 28 03 Mean 1 13 Scale 0 05 Triangular 90 34 09 Minimum 0 93 Likeliest 1 11 Maximum 1 33 Min Extreme 114 93 06 Likeliest 1 17 Scale 0 07 Max Extreme 119 86 07 Likeliest 1 09 Scale 0 07 Uniform 557 53 23 Minimum 0 93 Maximum 1 37 Pareto 1116 36 30 Location 0 93 Shape 5 42 Exponential 9252 02 56 Rate 0 89 Figure 5 13 Forecast Goodness Of Fit view Crystal Ball User Manual 119 Chapter 5 Analyzing Forecast Charts Capability Metrics If the process capability features are activated on the Statistics tab of the Run Preferences dialog and if either an LSL USL or both are entered
186. e intermediate value between two numbers That is a continuous distribution assumes there is an infinite number of values between any two points in the distribution In many situations you can effectively use a continuous distribution to approximate a discrete distribution even though the continuous model does not necessarily describe the situation exactly For a comparison of continuous and discrete distributions see page 249 The continuous distributions listed in Table A 2 are described later in this section in alphabetical order Page references appear below the names Table A 2 Summary of continuous distributions Shape Name Summary A Beta Beta page 255 The beta distribution is a very flexible distribution commonly used to represent variability over a fixed range It can represent uncertainty in the probability of occurrence of an event It is also used to describe empirical data and predict the random behavior of percentages and fractions Exponential Exponential page 257 The exponential distribution is widely used to describe events recurring at random points in time or space such as the time between failures of electronic equipment the time between arrivals at a service booth or repairs needed on a certain stretch of highway It is related to the Poisson distribution which describes the number of occurrences of an event in a given interval of time or space Crystal Ball User Man
187. e likelihood of achieving a given result As you run the simulation observe the shape of the distribution to get a feel for the problem you are trying to solve If you are new to Crystal Ball try running simulations in Demo speed to watch the forecast values change from trial to trial In Figure 4 9 frequency distribution for Net Profit for example you see that most of the values are clustered around central group intervals You also see the variation around these group intervals and a clump of values at the left end Chapter 5 also explains how to focus on a particular range of forecast values Crystal Ball Note Clicking on the Excel menu bar brings Excel forward and makes the forecast windows disappear If this happens you can quickly bring the forecast windows back to the front by clicking the Crystal Ball and Excel icons in the Windows task bar During the simulation Crystal Ball saves the forecast values in a list that grows as the simulation progresses Beginning on page 190 you will learn how to export these forecast values to other programs such as a statistical analysis program Crystal Ball User Manual 91 Chapter 4 Running Simulations Running a simulation To run a simulation a gt 1 Select Run gt Start Simulation or click the Start Simulation button in the Crystal Ball toolbar If you set the forecast to appear in the Define Forecast dialog a forecast window appears As the simulation proceeds the for
188. e number of trials and switching between Monte Carlo and Latin hypercube sampling If certain settings yield a noticeably better match between the random numbers and the ideal distribution you might want to use those settings even if the simulation runs longer Customizing assumption charts Because assumption charts look so similar to forecast charts many of their menu commands and settings are the same Crystal Ball User Manual 179 Chapter 6 Analyzing Other Charts 180 Setting assumption chart views You can use the View menu to choose five views Probability Cumulative Probability Reverse Cumulative Probability Statistics and Percentiles For a description of these views and how to select them see Changing the distribution view and interpreting statistics beginning on page 115 Setting assumption preferences Assumption preferences set with Preferences gt Assumptions are similar to those for forecasts described in Setting forecast preferences beginning on page 125 By default assumption charts are not shown when you run a simulation You can change the Show Automatically setting to automatically show assumption charts while a simulation is running or when it stops Assumption Preferences q x Assumption Window View Probability x Window IV Show automatically While running simulation When simulation stops To show the generated random values in the window turn on the Store assu
189. e oldest and most common of the goodness of fit tests It gauges the general accuracy of the fit The test breaks down the distribution into areas of equal probability and compares the data points within each area to the number of expected data points The chi square test in Crystal Ball does not use the associated p value the way other statistical tests e g or F do e Kolmogorov Smirnov The result of this test is essentially the largest vertical distance between the two cumulative distributions Crystal Ball User Manual 29 Chapter 2 Defining Model Assumptions 30 7 Click OK to fit the distributions to your data Crystal Ball successively fits the selected distributions to your data The fitted distributions appear in the Comparison Chart dialog starting with the highest ranked distribution down through to the lowest You can use the Next and Previous buttons to scroll through the fitted probability distributions Each probability distribution is shown superimposed over the data as shown in Figure 2 7 Edit View Preferences Help 500 Values Frequency View 494 Displayed Comparison Chart ag un o n T ea N al a Fel 5 psg SEa T oO 2 a Q o 2 D i ihik ihin 200 00 400 00 600 00 800 00 Fit 1 Weibull distribution o Data values lt lt Previous Figure 2 7 Comparison Chart dialog Use the Comparison Chart dialog to visually com
190. e online Crystal Ball Reference Manual for more information about how absolute and relative precision relate to the confidence interval When you expand the Define Forecast dialog or open the Forecast Preferences dialog the Forecast Window tab appears by default as shown in Figure 3 3 Crystal Ball User Manual 61 Chapter 3 Defining Other Model Elements Forecast Window preferences Define Forecast Cell A3 xl w S Forecast Window Precision Fiter Auto Extract View Frequency Window lV Show automatically While running simulation When simulation stops Fit distribution Fit a continuous probability distribution to the forecast Fit Options Figure 3 3 Forecast Window tab forecast preferences This tab manages window display and distribution fitting for the forecast You can choose among the settings listed in Table 3 1 Then click OK to apply settings on the current tab to the active forecast Or you can click Apply To to apply settings on the active tab to the active worksheet the active workbook or all workbooks For more information click Apply To and then click Help in the Apply To dialog At any time you can click Defaults to restore the original default settings on the active tab in the dialog 62 Crystal Ball User Manual Defining decision variable cells and forecast cells Table 3 1 Forecast Window tab forecast preferences
191. e s v An optional shift value to use when calculating long term capability shift metrics The default set in the Capability Options dialog is 1 5 Crystal Ball User Manual 355 Appendix D Using the Process Capability Features 356 Crystal Ball User Manual In this bibliography Bootstrap Monte Carlo simulation Probability theory and statistics Random variate generation methods Specific distributions Extreme value distribution Lognormal distribution Weibull distribution Tornado charts and sensitivity analysis Two Dimensional simulation Weibull distribution Uncertainty analysis Crystal Ball User Manual 357 Bibliography Bootstrap Efron Bradley and Robert J Tibshirani An Introduction to the Bootstrap Monographs on Statistics and Applied Probability vol 57 New York Chapman and Hall 1993 Mooney C Z and R D Duval Bootstrapping A Nonparametric Approach to Statistical Inference Sage University Paper Series on Quantitative Applications in the Social Sciences series no 07 095 Newbury Park CA Sage 1993 Monte Carlo simulation Burmaster D E and P D Anderson Principles of Good Practice for the Use of Monte Carlo Techniques in Human Health and Ecological Risk Assessments Risk Analysis 14 4 477 481 1994 Hammersley J M and D C Handscomb Monte Carlo Methods New York Chapman and Hall 1965 Kalos Malvin H and Paula A Whitlock Monte Carlo Method
192. e values that are within your control to change The decision variable cells must contain simple numeric values not formulas or text These are used by some of the Crystal Ball tools and by OptQuest Forecast cells dependent variables contain formulas that refer to one or more assumption and decision variable cells The forecast cells combine the values in the assumption decision variable and other cells to calculate a result A forecast cell for example might contain the formula C17 C20 C21 Crystal Ball Note For previous versions of Crystal Ball it might have been necessary to define forecasts in the same cells as assumptions or decision variables to capture that data for later extraction Now assumption and decision variable data can be extracted as well as forecast data For this reason Crystal Ball 7 no longer supports two types of cell definition in the same cell If an assumption or a decision variable is defined in the same cell as a forecast in a Crystal Ball 4 x or 5 x 2000 x workbook the forecast will be deleted when the workbook is converted to Crystal Ball 7 format Defining assumptions About assumptions and probability distributions For each uncertain variable in a simulation or assumption you define the possible values with a probability distribution The type of distribution you select depends on the conditions surrounding the variable For example some common distribution types are shown in Figure 2 1 A
193. e yards of carbon fiber material when the defects occur an average of 8 times per 100 square yards The first step in selecting a probability distribution is matching your data with a distribution s conditions Checking the Poisson distribution e Any number of defects is possible within 100 square yards The occurrences are independent of one another The number of defects in the first 100 square yards does not affect the number of defects in the second 100 square yards The average number of defects 8 remains the same for each 100 square yards These conditions match those of the Poisson distribution The Poisson distribution has only one parameter Rate In this example the value for this parameter is 8 defects You would enter this value to specify the parameter of the Poisson distribution in Crystal Ball Crystal Ball User Manual 291 Appendix A Selecting and Using Probability Distributions 292 olx Edit View Parameters Preferences Help Name Jar xJ y Poisson Distribution 014 4 012 4 2040 5 008 4 2 0 06 0 04 0 02 l 0 00 a E E __ 0 2 4 6 8 10 12 14 16 18 Figure A 24 Poisson distribution The distribution illustrated in Figure A 24 shows the probability of observing x number of defects in 100 square yards of the carbon fiber material Statistical Note The size of the interval to which the rate applies 100 square yards in this example has no bearing on the
194. ecast chart updates to reflect the changing values in the forecast cell Crystal Ball Note When you start Crystal Ball the Crystal Ball Control Panel appears as shown in Figure 4 10 on page 94 You can use the Crystal Ball Control Panel to perform many of the procedures described in this section For details see The Crystal Ball Control Panel on page 93 Stopping a simulation F To stop a simulation 1 Select Run gt Stop Simulation or click the Stop button on the Crystal Ball toolbar or Control Panel Continuing a simulation To continue a simulation i 1 Select Run gt Continue Simulation or click the Continue button on the Crystal Ball toolbar or Control Panel Resetting and rerunning a simulation To reset the simulation and start over again 1 Select Run gt Reset Simulation or click the Reset button on the Crystal K Ball toolbar or Control Panel A dialog appears with a message asking you to confirm your request to reset the simulation 2 Click OK Crystal Ball resets the number of trials to 0 and clears the list of values and statistics for each assumption and forecast However the assumption and forecast definitions remain 92 Crystal Ball User Manual j gt Running simulations Change any assumptions or forecasts as needed 4 Select Run gt Start Simulation or click the Run button on the Crystal Ball toolbar The simulation starts over again Single stepping Before you run a simula
195. ed Excel functions Crystal Ball supports LOOKUP and OFFSET functions in both Normal and Extreme speed However if you use more than a thousand in a model this can slow down the initial parsing time and cause the simulation to run slowly 338 Crystal Ball User Manual Using the Process Capability Features In this appendix e Overview e Activating the process capability features e Setting specification limits and targets e Viewing capability metrics e Extracting capability metrics Including capability metrics in reports e Capability metrics list This appendix describes Crystal Ball s process capability features which support quality improvement methodologies such as Six Sigma DFSS Design for Six Sigma and Lean principles Crystal Ball User Manual 339 Appendix D Using the Process Capability Features Overview The Crystal Ball process capability features are provided to support quality improvement methodologies such as Six Sigma DFSS Design for Six Sigma and Lean principles For additional information and tutorials see the Process Capability Guide Use the Statistics tab of the Run Preferences dialog to activate these features Once they are activated you can Add upper and lower specification limits and target values to forecast definitions page 343 Display the Capability Metrics view in the forecast window in addition to the basic Frequency Cumulative Frequency Reverse Cumulative Frequency St
196. ed in an overlay chart If the overlay chart is then saved as part of the current simulation results when it is restored it will only contain forecasts from the saved current results set Forecasts from the previously restored results will no longer be included in the chart Restoring Crystal Ball simulation results To restore Crystal Ball simulation results that you saved earlier select Analyze gt Restore Results The Restore Results dialog appears as shown in Figure 4 13 Look in D Eamples z elsi a mE Reliability cbr wwe ooo Files of type Cystal Ball Result Files cbr gt Cancel A Figure 4 13 Restore Results dialog The file list contains only files that have the cbr file type Select the results file you want to restore and click Open If you try to open a file that is not Crystal Ball saved results or if Crystal Ball recognizes that a saved results file has somehow been damaged you will not be able to open the file and an error message appears Crystal Ball User Manual Running user defined macros Because you are restoring results and not simulation cell definitions or data you do not need to reset the simulation before restoring results Crystal Ball Note Results files can be restored at any time regardless of whether the original workbooks are open or whether another simulation has run or not You can open as many results files as you want not just one as in Crystal Ball 2000 or earlier
197. ed on page 344 Options button When Calculate Capability Metrics is checked displays the Capability Options dialog for specifying short term or long term capability metrics formulas a Z score shift value and other capability metrics calculation settings For more information see Setting capability calculation options on page 341 Freezing Crystal Ball data cells The Freeze command excludes or freezes certain assumptions decision variables or forecasts from a simulation The Freeze command lets you Temporarily disable cells that have been defined but are not wanted for the current simulation See the effect of changing certain assumptions or decision variables while holding other cells to their spreadsheet values Crystal Ball Note The Freeze command can be especially useful when you have multiple workbooks open and don t want to include all of their data cells in a simulation You can freeze any unwanted cells instead of closing the workbooks that contain them 88 Crystal Ball User Manual Ol Freezing Crystal Ball data cells To freeze cells in a Crystal Ball model 1 Choose Run gt Freeze The Freeze dialog appears as shown in Figure 4 7 2 Freeze R Choose items to freeze during the simulation pec Eel ajea see RADY E On Sheet1 G4 Assumptions OA Oil Prce Bare Ol4 85 Assumption button 308 on Variables D5 Forecast button Decision Variable button oO
198. eeeee eee eeees 11 Steps for using Crystal Ball 0 ccc eee cee eee eee ceeteeeneeseeesseceeeeseecseenes 13 Starting and closing Crystal Ball oo eee cece eee eeesseeeceeeeneeesaenes 14 Starting Crystal Ball manually 2 0 0 cece cece cee eee cree eee sees ceeeneeeeeenaes 14 Starting Crystal Ball automatically cccccccccee cee cesene ee eeeteeeecneenees 15 Closing Crystal Ball sssini derea 15 The Crystal Ball menus and toolbar s ssssssssnsssesssissisesseississrsensersrnninsessen 15 The Crystal Ball menus 0 0 ccc eee eee eee ce ceeeeeeeeeeceseseeceeeeeeessaeeneenes 15 The Crystal Ball toolbar sissi issenensis e a E E E 16 Chapter 2 Defining Model Assumptions QVELVIEW aiia E OETA ONE O AEAEE 18 Defining assumptioNSsgoiseryss an e n E EE EE EEE E 18 About assumptions and probability distributions 0 0c0 eee ees 18 Defining an assumption 2 0 0 cece cee cceeeeceeeeee eee eeceeceeeceeeseeecssesaeeneeeaes 19 Entering an aSSUMPtiON 0 0 eee cece ee cece eceeceeeceeeeeeeceseseeceeeeeeeesaecseeees 19 Additional assumption features cccccccccces cece eeteetecnecnseceecseeseeneeeecnee 23 Entering cell references and formulas 0 c cece eesceeeceeeeeeeseens 24 Alternate parameter sets isciecssceine sevsdiserssnssicssesdsnevsspseonssuvcuneghetneodstenss 25 Fitting distributions to data cc cece cece ese cseceseeeeeeseecsessesseeeaeeneenees 27 Specifying correlations between assumptions
199. eens 334 Numerical differences cccccccessssssescoeccceccccceusecsseescecccscceeeeeeesessesseeceess 334 Maximizing the benefits of Extreme Speed 0 cece cece cece eeeeeeeees 337 String intermediate results in formulas 00 0 eee cece ceeee ener eeee 337 Calls to user defined functions cccccccccccsseseeseececeecceeeceseeaaeneeseeeeeeees 337 Dynamic aSSUMPptiONS 2 0 0 eee ce cece E N EEEE NEE ERTEN 338 Excel nC Onis resi Gus fectecc nets a a a a a A 338 Appendix D Using the Process Capability Features OUTA E E E E tuevesiapur dees eeseeetbaognntd 340 Activating the process capability features cece eee eeeeee ce eeeeteeereeeee 341 Activating the process capability features ccccceeeeeeeeeeseeteenes 341 Setting capability calculation Options ccccceseeeee cee cete sensere eneee 341 Setting specification limits and targets 0 ccecece eee eeeeee cee eeeeeteeeeeeees 343 Viewing capability Metrics sssssssssssisiisrssiiessys isidici aneda ii 344 Viewing forecast charts and capability metrics together 004 345 Viewing LSL USL and Target marker lines c cece eee tees 347 Extracting capability Metrics 0 cece ee cece cee cee eeeeeeeecesessecneeseeeeseeeaeees 349 Extracting capability metrics automatically ce eee eee eeeeeees 349 Extracting capability metrics manually 0 0 0 0 cece eee eee eseeeeeeeeees 350 Including capability metrics in re
200. efault Otherwise you can activate Split View through the View menu or the Preferences gt Forecast menu in the forecast chart window To activate Split View 1 In the forecast window choose View to open the View menu 2 Choose Split View at the bottom of the menu The Frequency chart and statistics both appear similar to the next figure 122 Crystal Ball User Manual Understanding and using forecast charts Forecast Profit or Loss vertical pane splitter Figure 5 16 Frequency chart and statistics in Split view You can resize the window and use the vertical pane splitter to adjust the size of the chart and the statistics pane 3 If you want continue opening the View menu and choosing charts or data The following figure shows a Frequency chart a Cumulative Frequency chart plus Statistics and Percentiles tables Crystal Ball User Manual 123 Chapter 5 Analyzing Forecast Charts Forecast Profit or Loss horizontal pane splitter Figure 5 17 Frequency and Cumulative Frequency charts with statistics and percentiles in Split View You can click in any of the Split View panes and use the chart hot keys to modify them without using the View or Preferences menus For a list see Table 5 5 on page 129 You can also resize the chart window and drag the horizontal and vertical pane splitters to resize each part of the Split View window To clear Split View or remo
201. eferences dialog determine how sensitivity is viewed and set the criteria for showing assumptions Additional settings in the Chart Preferences dialog determine the appearance of the chart To set sensitivity chart preferences 1 In the sensitivity chart window choose Preferences gt Chart The chart preferences dialog appears as shown in Figure 6 25 Crystal Ball User Manual 175 Chapter 6 Analyzing Other Charts Chart Preferences y xi Chart title IV Auto Sensttivity Risk Assessment Chart bins Options Density Medium z Gridlines Vertical x I Show column gaps Legend Right x Effects F 3D chart I Transparency 50 Figure 6 25 Sensitivity chart Chart Preferences General tab 2 On the General tab you can set the following features described on the pages in parentheses e Chart title page 132 e Gridlines page 135 e Legend page 135 e Chart effects page 136 Except for the disabled Chart Bins preferences the General tab settings are the same as those for forecast charts For descriptions of each setting while you are viewing the dialog click Help 3 Click the Chart Type tab for more settings shown in Figure 6 26 176 Crystal Ball User Manual Understanding and using sensitivity charts x Chart type Bar directional X Bar chart options IV Show multiple colors Figure 6 26 Sensitivity chart Chart Preferences Chart Type tab Y
202. elected Click Start The tool runs a simulation for each combination of decision variable values It compiles the results in a table of forecast cells indexed by the decision variables CB Tools Note For more information on the Decision Table dialogs click Help in a dialog Crystal Ball User Manual 229 Chapter 8 Crystal Ball Tools Interpreting the results For this example the Decision Table tool ran 42 simulations one for each combination of wells to drill and facility sizes The simulation that resulted in the best mean NPV was the combination of 12 wells and a facility size of 150 mbd B iS s 2 ao 7 o a a N N B N LZ UP O STA O LE IMP 0 SI3AA m OP IHP 9 SIIBAA 0S UP 9 SIA O Facility size 300 00 ility size 350 00 Figure 8 32 Decision table for Oil Field Development results To view one or more of the forecasts in the decision table select the cells and click Forecast Charts To compare one or more forecasts on the same chart select the cells and click Trend Chart or Overlay Chart Trend NPV Trend E 15 x Edit View Trend Preferences Help 500 Trials Centered on Median View NPV Trend Figure 8 33 Trend chart of 150 mbd forecasts 230 Crystal Ball User Manual Scenario Analysis tool You can create the above trend chart by selecting all the forecast cells in the Facility Size 150 00 row of t
203. electrical devices for custom orders wants to model the reliability of devices it produces After analyzing the empirical data the company knows that it can use the beta distribution to describe the reliability of the devices if the parameters are alpha 10 and beta 2 The first step in selecting a probability distribution is matching your data with a distribution s conditions Checking the beta distribution e The reliability rate is a random value somewhere between 0 and 1 The shape of the distribution can be specified using two positive values 10 and 2 These conditions match those of the beta distribution P befine Assumption CoAT Edit View Parameters Preferences Help Name Jan EF Y Beta Distribution Probability T T T y 010 020 030 040 050 060 070 O80 090 1 00 Figure A 6 Beta distribution Crystal Ball User Manual Using continuous distributions Figure A 6 shows the beta distribution with the alpha parameter set to 10 the beta parameter set to 2 and Minimum and Maximum set to 0 and 1 The reliability rate of the devices will be x Exponential distribution b Exponential Parameter Rate Conditions The exponential distribution describes the amount of time between occurrences The distribution is not affected by previous events Description The exponential distribution is widely used to describe events recurring at random points in time or space such as the time between
204. elects variable values at random to simulate a model When you roll a die you know that either a 1 2 3 4 5 or 6 will come up but you don t know which for any particular trial It is the same with the variables that have a known range of values but an uncertain value for any particular time or event for example interest rates staffing needs stock prices inventory phone calls per minute Crystal Ball User Manual 11 Chapter 1 Crystal Ball Overview Glossary Term probability distribution A set of all possible events and their associated probabilities Glossary Term assumption An estimated value or input to a spreadsheet model Glossary Term forecast A statistical summary of the simulation results in a spreadsheet model displayed graphically or numerically Glossary Term certainty The percent chance that a particular forecast value will fall within a specified range Probability distributions and assumptions For each uncertain variable in a simulation you define the possible values with a probability distribution A simulation calculates numerous scenarios of a model by repeatedly picking values from the probability distribution for the uncertain variables and using those values for the cell Commonly a Crystal Ball simulation calculates hundreds or thousands of scenarios in just a few seconds In Crystal Ball distributions and associated scenario input values are called assumptions
205. elp 200 Trials Frequency View 199 Displayed Mean 2 3 2 e a Certainty 100 0 3 q iny Figure 8 26 Bootstrap example results CB Tools Note For more information about the Bootstrap dialogs click Help in a dialog Crystal Ball User Manual 223 Chapter 8 Crystal Ball Tools 224 Interpreting the results The Bootstrap tool displays sampling distributions in forecast charts for the following statistics e Mean e Median e Standard deviation e Variance e Skewness e Kurtosis Coefficient of variability one simulation When you use the multiple simulation method the tool also displays sampling distributions for these statistics e Range minimum e Range maximum e Range width For percentiles the Bootstrap tool displays the percentile sampling distributions on the overlay and trend charts To display the individual percentile forecast charts select Analyze gt Forecast Charts Crystal Ball Note Ifyou have the Probability Above A Value option selected in the Run Preferences gt Options dialog the percentiles are reversed in meaning so that the Ist percentile represents the uppermost 1 and the 99th percentile represents the lowest 1 For more information on this reversal See Statistics preferences on page 87 The forecast charts visually indicate the accuracy of each statistic A narrow and symmetrical distribution yields more precise statistics estimates than a
206. empt wages within the department A glance at the chart illustrated in Figure A 2 reveals that the most common wage range is 12 00 to 15 00 Approximately 60 employees out of a total of 180 earn from 12 to 15 00 per hour Crystal Ball User Manual Understanding probability distributions 60 50 Number of 40 Employees 20 6 00 9 00 12 00 15 00 18 00 Hourly Wage Ranges in Dollars Figure A 2 Raw frequency data for a probability distribution You can chart this data as a probability distribution A probability distribution shows the number of employees in each interval as a fraction of the total number of employees To create a probability distribution you divide the number of employees in each interval by the total number of employees and list the results on the chart s vertical axis The chart illustrated in Figure A 3 shows you the number of employees in each wage group as a fraction of all employees you can estimate the likelihood or probability that an employee drawn at random from the whole group earns a wage within a given interval For example assuming the same conditions exist at the time the sample was taken the probability is 0 33 a 1 in 3 chance that an employee drawn at random from the whole group earns between 12 and 15 an hour Crystal Ball User Manual 247 Appendix A Selecting and Using Probability Distributions 0 33 Probability 0 22 6 00 9 00 12 00 15 00 18
207. en all windows for that chart type or e Choose from the commands listed in Table 4 7 Table 4 7 Crystal Ball window management commands Button Command Effect Analyze gt Open Selected Cells Opens charts for all assumption and forecast cells in the selected range 1 vilh Analyze gt Assumption Charts Analyze gt Forecast Charts Opens a dialog so you can choose which assumption charts to open or close Opens a dialog so you can choose which forecast charts to open or close ai Analyze gt Overlay Charts Opens a dialog so you can choose which overlay charts to open or close create or delete E Analyze gt Trend Charts Opens a dialog so you can choose which trend charts to open or close create or delete F Analyze gt Sensitivity Charts Opens a dialog so you can choose which sensitivity charts to open or close create or delete 96 Crystal Ball User Manual Button Saving and restoring simulation results Table 4 7 Crystal Ball window management commands Command Effect Analyze gt Close All Closes all chart windows and clears the current simulation results and stored results from memory Analyze gt Cascade Neatly stacks all windows in front of Excel Saving and restoring simulation results You can save Crystal Ball simulation results to disk and later restore them This is useful in several ways For example you can
208. ening the menu and choosing charts or data The following figure shows a Frequency chart a Cumulative Frequency chart plus Statistics and Capability Metrics tables 346 Crystal Ball User Manual Viewing capability metrics Forecast FLOW RATE FORECAST ml sec E j 5 x Edit View forecast Preferences Help 1 000 Trials Split View FLOW RATE FORECAST ml sec 0 05 USL 53 9283 5 004 3 M 49 2453 0 03 i mEn Std Dev 48 0331 0 02 h 47 2617 0 00 45 6617 46 7236 47 7855 48 8474 49 9093 50 9712 52 0331 53 0950 1 00 1 000 0 80 USL 53 9283 600 D Target 50 5950 F 0 60 Std Dev 50 4574 od 3 oao PAE 0 20 200 o lt 46 6452 477053 487654 498256 50 8857 51 9458 53 0059 54 0660 ml sec p 47 26 Certainty 94 92 zA q 333 Figure D 6 Frequency and Cumulative Frequency charts with capability metrics and statistics in Split view You can click in any of the Split View panes and use the chart hot keys to modify them without using the View or Preferences menus For a list see Table 5 5 on page 129 You can also resize the chart window and drag the horizontal and vertical pane splitters to resize each part of the Split View window Viewing LSL USL and Target marker lines By default after you add specification limits and a target to a forecast markers for these values appear on the forecast char
209. er Manual Viewing capability metrics loxi Edit View Forecast Preferences Help 1 000 Trials Capability Metrics View 1 000 Displayed Standard Deviation 1 17 Pp 0 32 Ppk lower 0 31 Ppk upper 0 33 Ppk 0 31 Ppm 0 28 Z LSL 0 86 Z USL 0 84 Zst 1 69 Zit total 0 19 p N C below 0 21 p N C above 0 22 p N C total 0 42 PPM below 205 468 42 PPM above 218 167 04 PPM total 423 635 47 Figure D 4 Capability Metrics view For a description of each statistic see Capability metrics list on page 353 Viewing forecast charts and capability metrics together It can be very helpful to view a forecast chart and its capability metrics side by side This is now the default view when capability metrics are activated To do this manually 1 Follow the instructions in the previous section to display capability metrics 2 In the forecast window choose View to open the View menu 3 Choose Split View at the bottom of the menu 4 Open the View menu again and choose Frequency The Frequency chart and capability metrics both appear similar to the next figure Crystal Ball User Manual 345 Appendix D Using the Process Capability Features Forecast FLOW RATE FORECAST ml sec ema i HT arget 50 5950 l Std Dev 50 4574 1 Std Dev 48 0331 JER if tz TH Figure D 5 Frequency chart and capability metrics in Split view 5 If you want continue op
210. erial 1 Reliability Variable 10 0 30 0 50 0 70 0 90 0 Material 1 Strength 1 05 109 112 1 16 1 19 Coil Diameter in 1 07 1 10 1 12 1 14 Number of Coils 1 09 1 11 1 12 1 14 1 15 Shearing Modulus of Elasticity 1 15 4 13 1 12 1 11 Wire Diameter in 1 44 1 13 1 12 1 12 Spring Deflection in 1 14 1 13 1 12 1 12 Figure 8 20 Spider chart with data Limitations While tornado and spider charts are very useful there are some limitations e Since the tool tests each variable independently of the others the tool doesn t consider correlations defined between the variables The results shown in the tornado and spider charts depend significantly on the particular base case used for the variables To confirm the accuracy of the results run the tool multiple times with different base cases This characteristic makes the one at a time perturbation method less robust than the correlation based method built into Crystal Ball s sensitivity chart Hence the sensitivity chart is preferable since it computes sensitivity by sampling the variables all together while a simulation is running 218 Crystal Ball User Manual Bootstrap tool Bootstrap tool Bootstrap is a simple technique that estimates the reliability or accuracy of forecast statistics or other sample data Classical methods rely on mathematical formulas to describe the accuracy of sample statistics These methods assume that the distrib
211. ery forecast while the simulation is running Correlation coefficients provide a meaningful measure of the degree to which assumptions and forecasts change together If an assumption and a forecast have a high correlation coefficient it means that the assumption has a significant impact on the forecast both through its uncertainty and its model sensitivity Positive coefficients indicate that an increase in the assumption is associated with an increase in the forecast Negative coefficients imply the opposite situation The larger the absolute value of the correlation coefficient the stronger the relationship To help interpret the rank correlations Crystal Ball provides a default chart view called the Contribution To Variance view This view makes it easier to answer questions such as What percentage of the variance or uncertainty in the target forecast is due to assumption X It is important to note that the Contribution To Variance method is only an approximation and is not precisely a variance decomposition Crystal Ball calculates Contribution To Variance by squaring the rank correlation coefficients and normalizing them to 100 Both the alternate Rank Correlation View and the Contribution To Variance view display the direction of each assumption s relationship to the target forecast Assumptions with a positive relationship have bars on the right side of the zero line Assumptions with a negative relationship have bars on the le
212. es a particular event occurs in a fixed number of trials The difference is that binomial distribution trials are independent while hypergeometric distribution trials change the success rate for each subsequent trial and are called trials without replacement For example suppose a box of manufactured parts is known to contain some defective parts You choose a part from the box find it is defective and remove the part from the box If you choose another part from the box the probability that it is defective is somewhat lower than for the first part because you have removed a defective part If you had replaced the defective part the probabilities would have remained the same and the process would have satisfied the conditions for a binomial distribution Example one You want to describe the number of consumers in a fixed population who prefer Brand X You are dealing with a total population of 40 consumers of which 30 prefer Brand X and 10 prefer Brand Y You survey 20 of those consumers The first step in selecting a probability distribution is matching your data with a distribution s conditions Checking the hypergeometric distribution The population size 40 is fixed The sample size 20 consumers represents a portion of the population Initially 30 of 40 consumers preferred Brand X so the initial success rate is 30 This rate changes each time you question one of the 20 consumers depending on the preference o
213. es depend on each other in any way you should correlate them to increase the accuracy of your simulation s forecast results There are two types of correlations Table 8 3 Types of correlations positive correlation Indicates that two assumptions increase or decrease together For example the price of gasoline and shipping costs increase and decrease together negative correlation Indicates that an increase in one assumption results in a decrease in the other assumption For example the more items you buy from a particular vendor the lower the unit cost The correlation coefficient range is 1 to 1 where 0 indicates no correlation The closer the coefficient is to 1 the stronger the relationship between the assumptions You should never use a coefficient of 1 represent relationships this closely correlated with formulas in the spreadsheet model Correlation matrix In Crystal Ball you enter correlations one at a time using the Correlation dialog Instead of manually entering the correlations this way you can use the Correlation Matrix tool to define a matrix of correlations between assumptions in one simple step This saves time and effort when building your spreadsheet model especially for models with many correlated assumptions The correlation matrix is either an upper or lower triangular matrix with ones along the diagonal When entering coefficients think of the matrix as a multiplication table If yo
214. et Sampling Method to Monte Carlo gt 3 Runa simulation by selecting Run gt Start Simulation Choose View gt Statistics in the forecast chart The forecast statistics for the simulation are similar to those shown below Crystal Ball User Manual 207 Chapter 8 Crystal Ball Tools ixi Edit View Forecast Preferences Help 500 Trials Statistics View 500 Displayed Statisti Trials 500 Mean 6 579 Median 6 517 Mode Standard Deviation 5 467 Variance 29 889 397 Skewness 0 03295 Kurtosis 3 12 Coeff of Variability 0 83103 Minimum 11 010 Maximum 26 042 Mean Std Error 244 Figure 8 9 Uncorrelated simulation statistics Choose Run gt Reset to reset the simulation Select Run gt Tools gt Correlation Matrix The Select Assumptions dialog appears as shown in Figure 8 10 Select assumptions step 1 of 2 E E x Correlation Matrix Crystal Ball Tool amp Decisioneering 1998 2004 View or edit the full correlation matrix for a selected set of assumptions Income fund Growth and Income fund Aggressive Growth fund Figure 8 10 Select Assumptions dialog Correlation Matrix tool Include all the assumptions in the correlation matrix by moving all the assumptions from the Available Assumptions list to the Selected Assumptions list by either e Double clicking on each assumption to move 208 Crystal Ball User Manual Correlation Matrix tool e Selecting each assumption to move and clickin
215. et assumption and chart preferences and display help as described in Additional assumption features beginning on page 23 7 When you have finished entering parameters to define the assumption click Enter The distribution changes to reflect the values you entered Crystal Ball Note Ifyou click OK instead of Enter Crystal Ball accepts the parameters and closes the dialog 8 Click OK If you selected a range of cells repeat steps 3 8 to define the assumption for each cell assumption features As you enter assumption parameters you can use cell references and alternate parameters If you have historical data available you can use Crystal Ball s distribution fitting feature to help simplify the process of selecting a probability distribution You can also specify correlations between assumptions or freeze assumptions to exclude them from a simulation The following sections discuss advanced features that help you refine assumption definitions and use assumptions more effectively e Entering cell references and formulas below e Alternate parameter sets on page 25 e Setting assumption preferences on page 38 e Fitting distributions to data on page 27 e Specifying correlations between assumptions on page 33 e Freezing Crystal Ball data cells on page 88 Crystal Ball User Manual 23 Chapter 2 Defining Model Assumptions 24 Entering cell references and formulas In addi
216. et model predicts how best to develop a new oil field by selecting the optimal number of wells to drill rate of oil production and size of the refinery to build that will maximize the net present value of the field Crystal Ball Note Before you run the Decision Table tool you might want to choose Run gt Run Preferences gt Sampling and check Use Same Sequence Of Random Numbers This starts each simulation with the same random number for the sake of consistency and makes the effect of the decision variable more obvious To run the Decision Table tool 1 D In Excel with Crystal Ball loaded open the workbook Oil Field Development xls In the Run gt Run Preferences gt Sampling dialog set e Random Number Generation to use the Same Sequence Of Random Numbers and An Initial Seed Value of 999 e Monte Carlo simulation When using this tool use these options to make the resulting simulations comparable In the Run Preferences dialog click OK Select Run gt Tools gt Decision Table Crystal Ball User Manual 227 Chapter 8 Crystal Ball Tools K Crystal Ball Note Ifthe Decision Table command is not available reset the simulation and try again The Specify Target dialog appears as shown in Figure 8 29 Specify target step 1 of 3 Decision Table Crystal Ball Tool Decisioneering 1998 2004 Evaluate the effects of alternate decisions in your simulation model Please enter the target cell or sele
217. eting statistics beginning on page 115 Determine if and when the assumption chart window appears when a simulation runs Crystal Ball Note To show the generated values in the window turn on the Store Assumption Values For Sensitivity Analysis run preference To do so click the Run Preferences button then click the Options tab For more information about the Assumption Preferences settings click the Help button in the Assumption Preferences dialog You can click Apply To to copy these settings to other assumptions If necessary you can click Defaults to restore original default settings When the settings are complete click OK 40 Crystal Ball User Manual Additional Distribution Gallery features Additional Distribution Gallery features You can use the Distribution Gallery to add manage and share libraries of distributions This powerful feature lets work groups modify and share customized distributions over their local networks for collaborative use in custom models They can also email these to other Crystal Ball users for use with their models The following table summarizes how you can use these additional features For instructions see the referenced pages Table 2 2 Distribution Gallery tasks How Dol Command Reference Display the Distribution Gallery Define gt Define page 42 Assumption Manage distributions Create new distributions Edit gt Add To Gallery in page 45 Define Assumption dial
218. etween 2500 and 3750 instead of using a single point estimate of 3000 Crystal Ball then uses the defined range in a simulation e With Monte Carlo simulation Crystal Ball displays results in a forecast chart that shows the entire range of possible outcomes and the likelihood of achieving each of them In addition Crystal Ball keeps track of the results of each scenario for you Crystal Ball Follow these general steps to create and interpret simulations with Crystal Ball the remaining chapters provide detailed instructions 1 Create a spreadsheet model in Microsoft Excel format with data and formula cells that represent the situation you want to analyze What is a model on page 9 discusses spreadsheet models Also see the references in the Bibliography beginning on page 357 2 Start Crystal Ball If you haven t set up Crystal Ball to load automatically with Microsoft Excel start Crystal Ball as described on page 14 3 Load your spreadsheet model 4 Using Crystal Ball define assumption cells and forecast cells If appropriate for your situation you can also define decision variable cells For more information see Entering an assumption beginning on page 19 and continue on with Chapter 3 Crystal Ball User Manual 13 Chapter 1 Crystal Ball Overview Set run preferences for your simulation as described beginning on page 80 Run the simulation following the instructions beginning on p
219. ey part of defining a model in Crystal Ball Assumptions are the main source of simulation input and distributions are the core component of assumption definitions Several Crystal Ball features help you modify and save distributions for future use and share them with other Crystal Ball users You can e Modify distribution parameters when defining an assumption and then save the modified distribution directly to the Distribution Gallery for later use Cut and paste default and modified distributions to create custom categories of distributions that you use frequently or use with certain models For example you can copy a small number of frequently used distributions from the All cateogry into Favorites to instantly access your favorite distributions without scrolling through All Or you can create a custom category and add them there Modify distribution summaries and descriptions to accompany particular models or reflect parameter changes e Delete unwanted distributions from Favorites and custom categories e Print distributions for reference and presentation Table 2 2 on page 41 summarizes distribution management tasks you can perform in the Distribution Gallery For instructions see the referenced pages Creating new distributions You can create new distributions by Copying and renaming a distribution within the Distribution Gallery Defining a distribution in the Define Assumption dialog and adding it to th
220. f around 0 25 25 These facts are the conditions surrounding the variable Review the descriptions of the probability distributions This chapter describes each distribution in detail outlining the conditions underlying the distribution and providing real world examples of each distribution type As you review the descriptions look for a distribution that features the conditions you have listed for this variable Select the distribution that characterizes this variable A distribution characterizes a variable when the conditions of the distribution match those of the variable The conditions of the variable describe the values for the parameters of the distribution in Crystal Ball Each distribution type has its own set of parameters which are explained in the following descriptions For example look at the conditions of the binomial distribution as described on page 281 e For each trial only two outcomes are possible success or failure Crystal Ball User Manual Selecting a probability distribution The trials are independent What happens on the first trial does not affect the second trial and so on The probability of success remains the same from trial to trial Now check the patients cured variable in Tutorial 2 in the Crystal Ball Getting Started Guide against the conditions of the binomial distribution e There are two possible outcomes the patient is either cured or not cured e The trials 10
221. f the forecast chart For example you can set the display range to focus on just the upper or lower tail of the forecast Focusing on the display range on page 112 discusses this analysis technique Crystal Ball User Manual Guidelines for analyzing simulation results Look at different views of the forecast Use the View menu to switch among different ways of viewing the forecast distribution graphically frequency cumulative frequency or reverse cumulative frequency or numerically statistics percentiles goodness of fit data or capability metrics You can also choose whether to show charts and statistics simultaneously or separately For instructions see Changing the distribution view and interpreting statistics on page 115 Customize the forecast chart Use the chart preferences to change graphic presentations from bars to areas or lines or experiment with different colors 2D versus 3D more or fewer plotted intervals or data points and other display variations Setting chart preferences beginning on page 128 describes the many ways you can format charts for analysis and presentation of results Create other kinds of charts You can display overlay charts page 146 trend charts page 154 and sensitivity charts page 165 As with forecast charts you can customize these charts as well Selecting different views of the data can help you analyze it and present it to others Create reports Create re
222. f the first range beside the maximum in the second column Be sure to check Probabilities Are Cumulative in the Load Data dialog Minimum of First Range Figure A 45 Connected continuous uniform ranges P Define Assumption CAMEO Lo Edit View Parameters Preferences Help Name E10 xi A Custom Distribution slative Probabil o o o N e AR i 9 00 1200 1500 1800 2100 2400 2700 30 00 Figure A 46 Connected continuous uniform ranges after loading Other data load notes You can load each type of range separately or you can specify the range type with the greatest number of parameters and load all types together Other rules are e Cumulative probabilities are supported for all but sloping ranges e Blank probabilities are interpreted as a relative probability of 1 0 e Ranges or values with 0 probabilities are removed Sloping ranges with Height of Min and Height of Max equal to 0 are also removed e For continuous connected ranges for either endpoint values or probabilities if the starting cell is blank the previous end value is used as the start for this range 310 Crystal Ball User Manual Using the custom distribution e When you load a discrete value that exists in the table already its probability is incremented by 1 For continuous ranges this is not allowed an error message about overlapping ranges appears Changes from Crystal Ball 2000 x 5 x In previous versions of Cry
223. f the previous consumer The conditions in this example match those of the hypergeometric distribution Crystal Ball User Manual 287 Appendix A Selecting and Using Probability Distributions 288 Statistical Note If you have a probability from a different sized sample instead of a success rate you can estimate initial success by multiplying the population size by the probability of success In this example the probability of success is 75 75 x 40 30 and 30 40 75 The three parameters of this distribution are initial Success number of Trials and Population size The conditions outlined in this example contain the values for these parameters a population Size of 40 sample size Trials of 20 and initial Success of 30 30 of 40 consumers will prefer Brand X You would enter these values as the parameters of the hypergeometric distribution in Crystal Ball C Define Assumption cenar TT Edit View Parameters Preferences Help Name Jan xj y Hypergeometric Distribution 020 4 3 oe 4 2 012 4 a 008 4 0 04 oo lil m__ GQ A ep fa Sa Re ay a Success 30 Thy Trials 20 S Populationfjso S OK Cancel Enter Gallery Correlate Help Figure A 22 Hypergeometric distribution The distribution illustrated in Figure A 22 shows the probability that x number of consumers prefer Brand X Example two The U S Department of the Interior wants to describe the movemen
224. ferences gt Chart Click the Chart Type tab shown in Figure 6 14 on page 161 On the Chart Type tab click the Certainty Bands button The Certainty Bands dialog appears as shown in the next figure Crystal Ball User Manual Understanding and using trend charts Certainty ands m Choose certainty bands 25 50 75 100 C Quartiles 25 increments Quintiles 20 increments C Deciles 10 increments Icosatiles 5 increments Figure 6 15 Certainty Bands dialog 5 Choose a set of certainty bands to display on the trend chart 6 To create your own set choose Custom and enter a series of certainty bands separated by commas 7 Click OK The trend chart appears with the selected certainty bands A legend appears on the right side of the trend chart indicating which levels are represented by which bands Crystal Ball Note If the legend doesn t include all bands drag the top or the bottom of the trend chart to increase its height until all bands appear Changing value axis preferences You can change several trend chart axis preferences as described below You can add a name for the value axis set number formats and set value rounding You can also change Scale settings from Auto to Fixed and specify a range minimum and maximum value These scale settings let you display the probability that given forecasts will fall in a particular part of a value range To change value axis setting
225. forecast means the uncertainty assumption values and the statistics including percentiles of the forecast distribution for each simulation Crystal Ball User Manual 239 Chapter 8 Crystal Ball Tools P O gt m n o juawssassy ysy juaWssassly ySIy g uawissassyy ysiy pjuawissassy ysy g luawissassy ysIy g juawissassy ysIy Juaussassy Sly ighuawssassy ysIy iz te Assumptions Concentration of Contaminant in Water Model 105 77 94 53 82 84 82 06 94 71 102 68 94 75 102 93 CPF Model 14E 02 1 6E 02 1 9E 02 2 0E 02 1 8E 02 1 8E 02 2 0E 02 1 9E 02 Statistics Mean 4 51E 05 4 86E 05 4 99E 05 5 16E 05 5 57E 05 5 64E 05 5 89E 05 Median 4 35E 05 4 69E 05 4 81E 05 4 98E 05 5 36E 05 5 44E 05 5 68E 05 Standard Deviation 2 11E 05 2 28E 05 2 34E 05 2 41E 05 2 60E 05 2 64E 05 2 76E 05 Variance 4 45E 10 5 18E 10 5 46E 10 5 83E 10 6 78E 10 6 97E 10 7 60E 10 Skewness 0 56 0 56 0 56 0 56 0 56 0 56 0 56 Kurtosis 3 44 3 44 3 44 3 44 3 44 3 44 3 44 Coeff of Variability 0 47 0 47 0 47 0 47 0 47 0 47 0 47 Minimum 4 78E 09 5 16E 09 5 30E 09 5 47E 09 5 90E 09 5 98E 09 6 25E 09 Maximum 1 29E 04 1 39E 04 1 43E 04 1 48E 04 1 59E 04 1 61E 04 1 69E 04 Range 1 29E 04 1 39E 04 1 43E 04 1 48E 04 1 59E 04 1 61E 04 1 69E 04 Percentiles 1 35E 05 1 46E 05 1 50E 05 1 55E 05 1 67E 05 1 69E 05 1 77E 05 1 85E 05 1 99E 05 2 04E 05 2 11E 05 2 28E 05 2 31E 05 241E 05 M4 MN Sheeti Sheet2
226. ft size of the zero line If you want to show just the absolute magnitude of the relationship you can change the Chart Type preference setting described on page 177 to Bar Magnitude Crystal Ball Note To ensure appropriate accuracy in Contribution To Variance view consider running at least 10 000 trials 170 Crystal Ball User Manual Understanding and using sensitivity charts Limitations Sensitivity charts have several limitations you should be aware of Correlated assumptions Sensitivity calculations might be inaccurate for correlated assumptions For example if an important assumption were highly correlated with an unimportant one the unimportant assumption would likely have a high sensitivity with respect to the target forecast Assumptions that are correlated are flagged as such on the sensitivity chart In some circumstances turning off correlations in the Run Preferences dialog might help you to gain more accurate sensitivity information e Non monotonic relationships Sensitivity calculations might be inaccurate for assumptions whose relationships with the target forecast are not monotonic A monotonic relationship means that an increase in the assumption tends to be accompanied by a strict increase in the forecast or an increase in the assumption tends to be accompanied by a strict decrease in the forecast For example the relationship y Log x is monotonic While y Sin x is not The Tornado Chart tool
227. g capability metrics Extracting capability metrics automatically You can automatically extract capability metrics whenever a simulation runs To do this 1 Ina forecast window choose Preferences gt Forecast and display the Auto Extract tab of the Forecast Preferences dialog Forecast Preferences E x Forecast Window Precision Fiter Auto Extract v Extract forecast statistics automatically to your spreadsheet when the simulation stops Starting cell O Mode 4 it e g Al O Std Deviation O Variance al Formatting O Skewness I Include labels O Kurtosis O Coet of Variability M AutoFormat O Minimum O Maximum Direction O Mean Standard Eror Fill j O Percentile Capability Metrics Fill to the right Figure D 8 Choosing to Auto Extract Capability Metrics 2 Check Extract Forecast Statistics Automatically and scroll down to the bottom of the data selection list 3 Check Capability Metrics enter a Starting Cell and then click OK Now when you run a simulation capability metrics are written to the specified area of your worksheet along with any other data you have requested For more information about the Auto Extract feature see Auto Extract preferences on page 68 Be sure to choose an open area of your worksheet as the Starting Cell to avoid overwriting your model Crystal Ball User Manual 349 Appendix D Using the Process Capabil
228. g it is open To delete an overlay trend or sensitivity chart 1 2 Open the model containing the chart Open the Analyze menu and select the type of chart to delete Overlay Charts Trend Charts or Sensitivity Charts When the dialog for that chart appears check the box in front of each chart you want to delete Click Delete The selected chart or charts are deleted without prompting Crystal Ball Note Assumption and forecast charts cannot be deleted in this way Crystal Ball User Manual 143 Chapter 5 Analyzing Forecast Charts 144 Crystal Ball User Manual Analyzing Other Charts In this chapter e Overview Understanding and using overlay charts e Understanding and using trend charts e Understanding and using sensitivity charts e Understanding and using assumption charts This chapter extends the information on analyzing simulation results provided in Chapter 5 Analyzing Forecast Charts You will learn how to use additional charts to interpret and present data For information on customizing charts managing chart windows and printing charts see Chapter 5 Crystal Ball User Manual 145 Chapter 6 Analyzing Other Charts Overview The previous chapter gave guidelines for analyzing and customizing simulation results with a focus on forecast charts This chapter describes how to create and use a variety of additional charts Overlay charts to view forecasts superimposed over eac
229. g on gt gt to move it e Making an extended selection using the Shift or Ctrl keys 7 Click Next The Specify Options dialog appears Specify options step 2 of 2 S j x Correlation Matrix Crystal Ball Tool Decisioneering 1998 2004 rLocation of Matrix Create a temporary correlation matrix on a new worksheet Create and link to a correlation matrix on the active worksheet Specify upper left corner of the matrix VV Format matrix rOrientation Upper triangular matrix C Lower triangular matrix Figure 8 11 Specify Options dialog Correlation Matrix tool 8 Choose the following settings e Create A Temporary Correlation Matrix On A New Worksheet e Upper Triangular Matrix 9 Click Start The tool creates a temporary matrix in a new workbook 10 Enter the following correlation coefficients into the matrix Crystal Ball User Manual 209 Chapter 8 Crystal Ball Tools QoQ gt S z 3 Ss 2 o m a 8 Load the matrix lt i z 2 8 9 2 t f o pagi 2 2 amp y e Qa Qa a a Money Market fund 1 000 0 200 0 100 0 100 Income fund 1 000 0 300 0 200 Growth and Income fund Aggressive Growth fund 1 000 0 500 CB Tools Note Leaving a cell blank is not the same as entering a zero Values that are not specified in the matrix will be filled in with estimates of appropriate values when the simulation runs 11 Click Load
230. ge The linear distance for the set of values displayed on the forecast chart dominant A relationship between distributions in which one distribution s values for all percentile levels are higher than another s see also subordinate entire range The linear distance from the minimum forecast value to the maximum forecast value forecast A statistical summary of the assumptions in a spreadsheet model output graphically or numerically Crystal Ball User Manual 365 Glossary forecast cell Cells that contain formulas that refer to one or more assumption and decision variable cells and combine the values in the assumption decision and other cells to calculate a result forecast definition The forecast name and parameters assigned to a cell in a Crystal Ball dialog forecast filtering A process by which Crystal Ball discards forecast values outside or inside a specified range forecast formula A formula that has been defined as a forecast cell forecast value also trial A value calculated by the forecast formula during an iteration These values are kept in a list for each forecast and are summarized graphically in the forecast chart and numerically in the descriptive statistics formula cell A cell that contains a mathematical formula frequency also frequency count The number of times a value recurs in a group interval frequency distribution A chart that graphically summarizes a list of values
231. ge 130 2 Locate the Options group in the middle of the page 3 Choose a setting from the Legend drop down list to show the legend at the right side of the chart Right the left side of the chart Left or the bottom of the chart Bottom To hide the legend choose None 4 When all your settings are complete click OK The chart appears with the new settings Crystal Ball Note You can press Ctrl n to toggle the legend on and off Crystal Ball User Manual 135 Chapter 5 Analyzing Forecast Charts 136 Setting special chart effects You can use special effects to help present your data most effectively Transparency ensures that all chart series and values are visible and 3D effects add graphic depth that can be so useful when many series are charted for example bars become blocks as shown in the chart density figure on page 134 Crystal Ball Note For basic chart customization instructions see page 130 To set special chart effects 1 Display the General tab of the Chart Preferences dialog page 130 2 Locate the Effects group at the bottom of the page 3 You can check any or all of the available effects to see how they enhance your chart If you check Transparency you can also choose a percent 0 is completely opaque and 100 is completely transparent 4 When all your settings are complete click OK The chart appears with the new settings Crystal Ball Note You can press Ctrl 3 to toggl
232. h adjusted probabilities Example two Before beginning example two clear the values entered in example one as follows 1 Right click in the chart and choose Clear Distribution from the right click menu In this example you will use the custom distribution to describe a continuous range of values since the unit cost can take on any value within the specified intervals 1 Choose Parameters gt Continuous Ranges to enter value ranges 2 Enter the first range of values Type 5 in the Minimum field e Type 15 in the Maximum field Type 75 in the Probability field This represents the total probability of all values within the range 3 Click Enter Crystal Ball displays a continuous value bar for the range 5 00 to 15 00 as in Figure A 30 and returns the cursor to the Minimum field Notice that the height of the range is 0 075 This represents the total probability divided by the width number of units in the range 10 Crystal Ball User Manual Using the custom distribution Define Assumption Cell D11 Ky iol x Edit View Parameters Preferences Help Name D11 xj M Custom Distribution 5 00 6 00 7 00 8 00 9 00 10 00 1100 1200 13 00 1400 15 00 Enter one or more continuous ranges with probabilities Minimum fT O O M i lity f OK _ Cancel _ Enter Gallery _ Correlate Hele Figure A 30 A continuous custom distribution Enter the second range of values e Type 16 in the Mini
233. h distribution into the specified number of intervals bins A higher number increases the evenness of the sampling method while reducing the randomness For more information see Simulation accuracy on page 318 Speed preferences Run Preferences J Options Statistics Trials Sampling E Run mode Extreme speed C Normal speed Demo speed M Chart windows Redraw every fo 5 seconds Suppress chart windows fastest Figure 4 4 Speed tab Run Preferences dialog The Speed tab sets run mode and adjusts how fast a simulation runs Extreme speed is the default simulation speed for Crystal Ball Professional Edition and Crystal Ball Premium Edition Crystal Ball Standard Edition can only run in Normal or Demo speed If you choose Normal or Demo speed the Options button is active and you can make additional settings described in Table 4 3 A Crystal Ball Note Ifyou are using Crystal Ball Professional Edition or Crystal Ball Premium Edition be sure to read Appendix C Using the Extreme Speed Feature for important information about Extreme speed For general instructions see Setting run preferences beginning on page 80 Table 4 3 describes the settings available on this tab 84 Crystal Ball User Manual Setting run preferences Table 4 3 Speed tab Run Preferences dialog Group Setting Effect Run Mode Determines overall simul
234. h other Trend charts to view the certainty ranges of all your forecasts on a single chart Sensitivity charts to view the impact your assumptions have on forecast results Assumption charts to confirm random number sampling For most effective analysis and presentation consider this chapter as an extension of Chapter 5 and use them both together Understanding and using overlay charts 146 After completing a simulation with multiple related forecasts you can create an overlay chart to view the relative characteristics of those forecasts on one chart Then you can compare differences or similarities that otherwise might not be apparent There is no limit to the number of forecasts you can view at one time on an overlay chart For example if a model has several forecasts based on slightly different interest rate assumptions an overlay chart can be used to show how these slight differences manifest themselves in the variability of the forecasts You can customize the overlay chart to accentuate these differences or similarities The following overlay chart shows the relative reliabilities of three manufacturing materials Crystal Ball User Manual Understanding and using overlay charts Goveraychart Te Edit View Overlay Preferences Help 1 000 Trials Frequency View IV Enable Rotation Overlay Chart Wi Material 1 Reliability C Material 2 Reliability Probability Aouenbel4 i Material 3 Reliability Figure
235. hart as shown in Figure 5 2 D T a 2 Z m i T ss ange Min 12 8 Click the certainty grabber and drag it when the cursor changes to a hand Figure 5 2 Moving the left certainty grabber The certainty minimum field shows the value that corresponds to the position of the left certainty grabber the certainty maximum field shows the value that corresponds to the position of the right certainty grabber The Certainty field shows the certainty level for the area between the certainty grabbers Crystal Ball shades the columns outside the certainty grabbers a different color to show that those values have been excluded as illustrated in Figure 5 3 Crystal Ball User Manual Understanding and using forecast charts A Forecast Net Profit MM E lol o x Edit View Forecast Preferences Help 500 Trials Frequency Chart 497 Displayed Net Profit MM o 3 3 a E p cy oc 10 0 20 0 Millions Figure 5 3 Certainty level values 0 to 20 in Millions The Net Profit forecast chart in Figure 5 3 is the same as the example preceding it except that the certainty grabbers have been moved The certainty minimum shows 0 0 and the certainty maximum shows 20 0 The key statistic is the certainty level of 69 4 in the Certainty field By moving the certainty grabbers you have changed the certainty range Crystal Ball compares the number of values lying within the certainty range to the
236. he menubar For more information on these alternate parameters see Lognormal distribution in the Equations and Methods chapter of the online Crystal Ball Reference Manual For more information about this menu see Alternate parameter sets on page 25 Maximum extreme distribution Parameters Likeliest Scale Crystal Ball User Manual 265 Appendix A Selecting and Using Probability Distributions 266 A Max Extreme Description The maximum extreme distribution is commonly used to describe the largest value of a response over a period of time for example in flood flows rainfall and earthquakes Other applications include the breaking strengths of materials construction design and aircraft loads and tolerances The maximum extreme distribution is also known as the Gumbel distribution This distribution is closely related to the minimum extreme distribution described beginning on page 267 Define Assumption Cell A1 o x Edit View Parameters Preferences Help Maximum Extreme Distribution Likeliest 0 00 Scale 1 00 EF OK Cancel Cete Gallery Correlate Help Figure A 11 Maximum extreme distribution Calculating parameters There are two standard parameters for the maximum extreme value distribution Likeliest and Scale The Likeliest parameter is the most likely value for the variable the highest point on the probability distribution or mode After yo
237. he results table and clicking on Trend Chart This chart shows that the forecast with the highest mean NPV also has the largest uncertainty compared to other forecasts with smaller NPVs of the same facility size This indicates a higher risk that you could avoid with a different number of wells although the lower risk is accompanied with a lower NPV Crystal Ball Note If you have the Probability Above A Value option selected in the Run Preferences gt Options dialog the percentiles will be reversed in meaning so that the Ist percentile represents the uppermost 1 and the 99th percentile represents the lowest 1 For more information on this reversal See Statistics preferences on page 87 Scenario Analysis tool The Scenario Analysis tool runs a simulation and then sorts and matches all the resulting values of a target forecast with their corresponding assumption values This lets you investigate which combination of assumption values gives you a particular result You can run the Scenario Analysis tool on any Crystal Ball model You simply select a target forecast to analyze and then the forecast s percentile or value range you want to examine The resulting table shows all the values for the target forecast in the designated range sorted along with the corresponding assumption values Crystal Ball Note To help ensure the most accurate results select Run gt Run Preferences gt Trials and be sure the Stop On Calculation E
238. he speed of your simulation Of course if you have Crystal Ball Professional Edition or Premium Edition you can run in Extreme speed with dramatic cuts in run times If your model is incompatible with Extreme speed or you have Crystal Ball Standard Edition you can in increasing order of helpfulness e Redraw forecasts less often available in Extreme speed Update worksheets only every few seconds or so e Suppress chart windows during simulations available in Extreme speed e Minimize Excel workbooks during simulations e Minimize Excel workbooks and suppress chart windows the fastest combination of settings 2 Use the precision control feature The precision control feature page 81 can be used to either increase the precision of your simulations or increase the speed of your simulations If you set the confidence level to a high number your simulations will be Crystal Ball User Manual 321 Appendix B Maximizing Your Use of Crystal Ball 322 more precise but might run significantly longer However if you do not need as precise a result you can set the confidence level to a lower number and the simulation speed will increase Using this feature to speed up your model will require you to experiment with different confidence levels Reduce the size of the model by reducing the number of assumptions forecasts and correlations Large models require more time per trial For example a model that takes 3 or
239. hem first because the tool gathers all Crystal Ball definitions from all open workbooks 2 Select Run gt Tools gt Tornado Chart K Crystal Ball Note Ifthe Tornado Chart command is not available reset the simulation and try again The Specify Target Step 1 Of 3 dialog appears All of the forecasts from Reliability xls appear in the list Specify target step 1 of 3 E xj Tornado Chart Crystal Ball Tool Decisioneering 1998 2004 Measure the sensitivity of each assumption or variable individually to a specific target cell Material 1 Reliabilit Material 2 Reliability Material 3 Reliability Please enter the target cell or select from the list of forecasts above Reliability xls Model sD 18 Figure 8 16 Specify Target dialog Tornado Chart tool Select the Material 1 Reliability forecast 4 Click Next The Specify Input Variables Step 2 Of 3 dialog appears You can include any value cell in your tornado chart calculations using cell names if available However the cells are usually Crystal Ball User Manual Tornado Chart tool Table 8 4 Tornado chart input cells assumptions Cells defined as assumptions in Crystal Ball For more information about assumptions see Defining assumptions on page 18 decision variables Cells defined as decision variables in Crystal Ball For more information about defining decision variables see Defining decision variable cells on page 58
240. hoose Current Workbook a new sheet is created right after the current sheet If you want enter a descriptive name for the new sheet in the Sheet Name field In the Formatting group indicate whether to include the cell location workbook worksheet and cell address in the rightmost column of report headers and whether to include cell comments By default these settings are checked Crystal Ball Note If you choose to include cell comments only non Crystal Ball comments are included Crystal Ball cell comments are redundant and are filtered out In the Chart Format group choose Image to create a Crystal Ball chart or choose Excel to create an Excel chart Crystal Ball User Manual Creating reports If you choose Image you can format charts using the Crystal Ball Chart Preference settings Image is the default chart format When all settings are made click OK Crystal Ball creates the report as an Excel worksheet You can modify print or save the report in the same way as any other worksheet For example you can choose the File gt Print option for your spreadsheet model as you would for a normal spreadsheet Defining custom reports To define a custom report 1 Follow the steps in the previous section Basic steps beginning on page 183 Choose Custom in step 2 and click the Custom button The Custom Report dialog appears Custom Report TA Details Report Summary
241. ialog appears Chart Preferences Bex chat Type Axis Chart title IV Auto M Chart bins Ee LE E Options err ere Density Medium Gridlines Both x Figure 6 13 General tab Chart Preferences dialog for trend charts Crystal Ball User Manual Understanding and using trend charts Except for the disabled Chart Bins preferences the General tab settings are the same as those for forecast and other charts You can set the following features described on the pages in parentheses e Chart title page 132 e Gridlines page 135 Legend page 135 e Chart effects page 136 For descriptions of each setting while you are viewing the dialog click Help Setting certainty band type and colors By default the trend chart certainty bands appear as solid colored areas You can change the bands from solid areas to uncolored areas bordered by lines You can also change the color of each certainty band To change the chart type or color settings 1 Choose Preferences gt Chart gt Chart Type The Chart Type tab of the Chart Preferences dialog appears Chart Preferences xj Ending Sales Year 1 Q2 E Ending Sales Year 1 Q3 Ending Sales Year 1 Q4 al Ending Sales Year 2 Q1 Ending Sales Year 2 Q2 Ending Sales Year 2 Q3 Ending Sales Year 2 Q4 x Chart type ares 7 Band color 7 Select a band on the right to change the color gt Figure 6 14 Cha
242. ialog for data extraction Tree view Only forecasts for which data was generated or restored are included in the list Results for the current simulation appear first followed by any restored results that might be loaded These results are identified as file names with the cbr extension By default the Choose Forecasts dialog is in a hierarchical Tree view You can click the List button at the top to change it to List view For more information on this dialog click the Help button In the Assumptions group select assumptions for data extraction All Choose or None To include the selected data for all assumptions and restored results in the current simulation choose All To include the selected data for only selected assumptions or to include assumptions in selected restored results files choose Choose The Choose Assumptions dialog appears so you can choose from list of available assumptions It works similarly to the Choose Forecasts dialog described previously For more information on the Choose Assumptions dialog click the Help button e Choose None if you don t want to extract any assumption data Click the Options tab to specify a location or formatting for the extracted data Crystal Ball User Manual Extracting data Extract Data Location New workbook Sheet name Data Current workbook New sheet Starting cell Current sheet Al m Formatting IV Include labels IV AutoF
243. ics tab Run Preferences dialog The Statistics tab settings determine how Crystal Ball displays percentiles They also activate capability metrics to support Six Sigma and other quality programs as shown in Figure 4 6 For general instructions see Setting run preferences beginning on page 80 Table 4 5 describes the settings available on this tab Crystal Ball User Manual 87 Chapter 4 Running Simulations Table 4 5 Statistics tab Run Preferences dialog Setting Effect Calculate Percentiles As Sets how Crystal Ball defines percentiles Selecting either of these options also affects the percentiles used for the assumption alternate parameters Probability Below A Value Defines percentiles as the percent chance probability that the associated variable value is at or below a particular value This setting is the default Probability Above A Value Defines percentiles as the percent chance probability that the associated variable value is at or above a particular value Format Percentiles As Sets how Crystal Ball displays percentiles in charts and reports with a percent sign or the percentile preceded by P Calculate Capability Metrics Activates the process capability features in Crystal Ball when checked Crystal Ball displays capability metrics indicating process quality provided at least an upper or lower specification limit is entered in the Define Forecast dialog as describ
244. iew e Choose Preferences gt Chart to change chart features so that similarities or differences are more clearly accentuated e Click Cancel to return to the Fit Distribution dialog 9 To use the currently selected distribution either the best fit or another of your choice click Accept The Assumption dialog appears with the best fitting parameters taken from the chosen distribution as shown in Figure 2 9 You can change the distribution parameters before you click OK Crystal Ball User Manual 31 Chapter 2 Defining Model Assumptions Define Assumption Cell A1 _ O xj Edit View Parameters Preferences Help Name 21 EF y Weibull Distribution 1 4 Probability 2 8 200 00 400 00 600 00 800 00 Location IEE Scale 294 59 Ss Shape 1 9336863 S OK Cancel _ Enter Gallery Correlate Help Figure 2 9 The best fitting distribution after acceptance A distribution fitting example The following steps offer a specific example of fitting a distribution to data in a file 1 Create a new spreadsheet and select a cell a 2 Choose Define gt Define Assumption The Distribution Gallery appears 3 Click Fit to select the source of the fitted data The Kit Distribution dialog appears as shown in Figure 2 6 on page 28 4 Choose the location of historical data For this example click Text File and Browse and locate the TESTDATA txt file in the Examples f
245. ifying when trend charts appear You can set preferences that determine whether the trend chart appears automatically and if so whether it appears while the simulation is running or after it stops To set these display preferences 1 Choose Preferences gt Trend The Trend Preferences dialog appears as shown in Figure 6 12 158 Crystal Ball User Manual Understanding and using trend charts x View Centered on Medan f Window IV Show automatically While running simulation When simulation stops Figure 6 12 Trend Preferences dialog Trend Window tab 2 To change the trend chart view use the drop down View list The views are discussed in the previous section Changing trend chart A 2 views 3 Use settings in the Windows group to determine whether the chart appears automatically If Show Automatically is checked you can choose whether to display the chart while the simulation is running or after it stops 4 Youcan click Defaults at any time to restore original default settings for the Trend Preferences dialog 5 When all settings are complete click OK to activate them and close the dialog Adding removing and ordering forecasts To add and remove forecasts from a trend chart 1 In the trend chart menubar choose Trend gt Choose Forecasts The Choose Forecasts dialog appears similar to Figure 6 11 on page 156 2 Check and uncheck forecasts to add and remove them from
246. ight of Max is the height of the range Maximum 4 Click Enter The range returns to its original color and its height appears unchanged 300 Crystal Ball User Manual Using the custom distribution 5 Click in the range again to select it and set the Height of Max to 0 025 Then click Enter The right side of the range drops to half the height of the left as shown in Figure A 33 The range is selected to show its parameters after the change C Define assumptione cetona ioixi Edit view Parameters Preferences Help Name pr x Custom Distribution Enter one or more continuous or discrete sloping ranges Minimum fis o0 Maximum 21 00 Min Height fo os Max Height Jo o2s Step OK ca Gallery Correlate Dee Figure A 33 Sloping continuous value range 6 You can change the range from continuous to discrete values by adding a step value Type 5 in the Step field and click Enter The sloped range is now discrete Separate bars appear at the beginning and end of the range and every half unit in between 16 16 5 17 17 5 and so on until 21 as shown in Figure A 34 on page 302 If the discrete range represented money it could only include whole dollars and 50 cent increments Crystal Ball Note You can enter any positive number in the Step field If you entered 1 in this example the steps would fall on consecutive integers such as whole dollars Leave the Step parameter blank for continuous range
247. ility distributions 246 For each uncertain variable in a simulation you define the possible values with a probability distribution The type of distribution you select depends on the conditions surrounding the variable For example some common distribution types are A A mH A Normal Triangular Uniform Lognormal Figure A 1 Common distribution types During a simulation the value to use for each variable is selected randomly from the defined possibilities A simulation calculates numerous scenarios of a model by repeatedly picking values from the probability distribution for the uncertain variables and using those values for the cell Commonly a Crystal Ball simulation calculates hundreds or thousands of scenarios in just a few seconds A probability example To begin to understand probability consider this example You want to look at the distribution of non exempt wages within one department of a large company First you gather raw data in this case the wages of each non exempt employee in the department Second you organize the data into a meaningful format and plot the data as a frequency distribution on a chart To create a frequency distribution you divide the wages into groups also called intervals or bins and list these intervals on the chart s horizontal axis Then you list the number or frequency of employees in each interval on the chart s vertical axis Now you can easily see the distribution of non ex
248. ill still work but those distributions won t be available for creating new assumptions A note about modifying shared categories If a shared category is moved deleted or otherwise modified subscribers to that category won t notice the changes until they close Crystal Ball and open it again It is wise to notify subscribers when a change of that type occurs so they can start a new Crystal Ball session and refresh their view of the Distribution Gallery Crystal Ball User Manual 55 Chapter 2 Defining Model Assumptions 56 Crystal Ball User Manual Defining Other Model Elements In this chapter Defining decision variable cells Defining forecasts e Working with Crystal Ball data Setting cell preferences Saving and restoring your models Chapter 2 describes how to start building a spreadsheet model by defining assumption cells This chapter provides step by step instructions for completing models in Crystal Ball so you can run simulations against them As you work through these instructions you will learn to define decision variable cells and forecasts and how to cut copy and paste Crystal Ball data After you complete this chapter read Chapter 4 for information on setting run preferences and running simulations Crystal Ball User Manual 57 Chapter 3 Defining Other Model Elements Defining decision variable cells and forecast cells Chapter 2 describes how to define assumption cells and use the Distrib
249. ined or third party functions are allowed The functions may be written in Visual Basic or they may be in XLL or COM Automation DLL libraries that have been opened in Excel 328 Crystal Ball User Manual Compatibility issues Pure functions To be compatible with Extreme Speed user defined functions must be pure A pure function is one that computes its value solely on the basis of values passed to it as arguments A function that is not pure might reference global data not passed as an argument For example it might get the value of a worksheet cell or a defined name and use this as an input to compute the function value If the global data depends on the assumptions if for example it is a worksheet cell with a formula computed from the assumptions it will have a distribution of values in Normal Speed but it will appear deterministic have a single value in Extreme Speed This is because the worksheet cells change on every trial in Normal Speed but they don t change in Extreme Speed Range arguments Range arguments in user defined functions are only compatible with Extreme Speed when they are handled as Variant types For example for a function called on the worksheet as MyFunc A1 E4 5 4 Function MyFunc MyData As Variant Rows As Long Cols As Long As Double For I 1 to Rows For J 1 to Cols MsgBox MyData I J or otherwise work with the cell range as an array Next J Next I End Function Volati
250. ining 58 defaults setting 139 Define Correlation dialog 35 Define Forecast dialog expanded 61 defining assumptions 19 deleting charts 143 deleting distribution categories 51 deleting distributions 48 Description pane Distribution Gallery 45 dialogs Choose Forecasts 192 Select 73 discrete distributions 249 ranges 302 discrete uniform distribution 283 display range preferences 112 displaying charts 140 distribution preferences 38 distribution categories 49 creating 50 deleting 51 rearranging 51 sharing 52 53 54 55 distribution category properties editing and viewing 51 distribution fitting overlay charts 152 153 overview 27 Distribution Gallery Category pane 44 Description pane 45 displaying 42 Distribution pane 44 managing categories 49 managing distributions 45 menubar and buttons 43 using 41 window 42 Distribution Gallery tasks 41 Distribution pane Distribution Gallery 44 distributions chi squared 259 copying 46 creating 45 deleting 48 discrete or continuous 249 Erlang 259 managing 45 modifying 47 48 pasting 46 truncating 312 distributions as Excel functions 315 distributions parameters 255 distributions probability 12 dynamic cell referencing 24 E editing distribution category properties 51 emailing distribution categories 54 entering custom data 306 Erlang distribution 259 erlang distribution 261 examples Batch Fit 199 Bootstrap 221 Crystal Ball User Manual 375
251. int estimates e Range estimates e What if scenarios Point estimates Point estimates are when you use what you think are the most likely values technically referred to as the mode for the uncertain variables These estimates are the easiest but can return very misleading results For example try crossing a river with an average depth of three feet Or if it takes you an average of 25 minutes to get to the airport leave 25 minutes before your flight takes off You will miss your plane 50 of the time Range estimates Range estimates typically calculate three scenarios the best case the worst case and the most likely case These types of estimates can show you the range of outcomes but not the probability of any of these outcomes What if scenarios What if scenarios are usually based on range estimates and are often constructed informally What is the worst case for sales What if sales are best case but expenses are the worst case What if sales are average but expenses are the best case What if sales are average expenses are average but sales for the next month are flat Crystal Ball User Manual Model building and risk analysis overview As you can see this is extremely time consuming and results in lots of data but still doesn t give you the probability of achieving different outcomes You are still faced with these two fundamental limitations of ordinary spreadsheets e You can change only one spreadsheet
252. ion parameter You can use Excel Solver to help you calculate this parameter setting the constraint of B gt 1 Or estimate the variance and use the equation for shapes greater than 2 B z 26 17 variance Crystal Ball User Manual 271 Appendix A Selecting and Using Probability Distributions 272 where B is the Shape parameter and Z is the Location parameter You can use Excel Solver to help you calculate this parameter setting the constraint of B gt 2 Student s f distribution A Student s t Parameters Midpoint Scale Degrees of Freedom Conditions The values are distributed symmetrically about the midpoint The likelihood of values at the extreme ends is greater than those of the normal distribution Description In classical statistics the Student s distribution is used to describe the mean statistic for small sets of empirical data when the population variance is unknown Classically degrees of freedom is typically defined as the sample size minus 1 For purposes of simulation the Student s distribution resembles a normal curve but with thicker tails more outliers and greater peakedness high kurtosis in the central region As degrees of freedom increase at around 30 the distribution approximates the normal distribution For degrees of freedom larger than 30 you should use the normal distribution instead The Student s is a continuous probability distribution Since the
253. ions to make sure you have defined them as you intended With Crystal Ball you can select all the assumption forecast and decision variable cells and check their corresponding dialogs in sequence Reviewing assumption cells To review all assumption cells 1 Select Define gt Select All Assumptions Crystal Ball selects each cell defined as an assumption Select Define gt Define Assumption gt The distribution dialog for the first assumption cell appears Change the assumption parameters optional 4 Click OK If you have more than one assumption cell each assumption appears in turn Repeat steps 3 and 4 to check the assumptions for each cell Reviewing decision variable cells To review all decision variable cells ea 1 Select Define gt Select All Decisions I Crystal Ball selects each cell defined as a decision variable ep 2 Select Define gt Define Decision The Define Decision Variable dialog for the first decision variable cell appears 3 Change the decision variable parameters optional Crystal Ball User Manual Working with Crystal Ball data Click OK If you have more than one decision variable cell each decision variable appears in turn Repeat steps 3 and 4 to check the decision variables for each cell Reviewing forecast cells To review all forecast cells 1 Select Define gt Select All Forecasts Crystal Ball selects each cell defined as a forecast Select Define
254. ions used to calculate each of these statistics see the online Crystal Ball Reference Manual Crystal Ball Note Z scores are typically reported only for normal data Crystal Ball always displays Z scores It is up to the user to determine if the values are appropriate Table D 1 Capability statistics calculated by Crystal Ball Long Short Metric term term Description Mean v v Mean of the forecast values Standard s v Standard deviation of the forecast values deviation Cp v Short term capability index indicating what quality level the forecast output is potentially capable of producing It is defined as the ratio of the specification width to the forecast width If a Cp is equal to or greater than 1 then a short term 3 sigma quality level is possible Pp v Long term capability index indicating what quality level the forecast output is potentially capable of producing It is defined as the ratio of the specification width to the forecast width If a Cp is equal to or greater than 1 then a long term 3 sigma quality level is possible Cpk v One sided short term capability index for normally distributed lower forecasts the ratio of the difference between the forecast mean and lower specification limit over three times the forecast short term standard deviation often used to calculate process capability indices with only a lower specification limit Ppk v One sided long term capability index for normally di
255. istics Crystal Ball User Manual 59 Chapter 3 Defining Other Model Elements 60 Crystal Ball Note Ifyou have activated the Crystal Ball process capability features you can also enter a lower specification limit LSL an upper specification limit USL and a Target value For more information see Setting specification limits and targets on page 343 To define forecast cells 1 Select a formula cell or a range of formula cells lilt 2 Select Define gt Define Forecast The Define Forecast dialog appears Figure 3 2 Define Forecast dialog Crystal Ball Note If you have activated Crystal Ball s process capability features additional fields appear These are described in Setting specification limits and targets on page 343 3 Complete the Define Forecast dialog e Name is the name of the forecast e Units is the name of the units that will appear at the bottom of the forecast chart such as Millions For details about the fields and options in this dialog click the Help button in the dialog y 4 Toset additional forecast preferences click the More button to expand the Define Forecast dialog The expanded Define Forecast dialog has four tabs of additional options and fields listed in Setting forecast preferences on page 61 5 Click OK 6 Repeat steps 1 4 for each forecast in your model Crystal Ball User Manual Defining decision variable cells and forecast cells Crystal
256. ith the highest ranking fit is chosen to represent your data You can review the distributions sorted in order of their fit tests using the comparison chart This chart shows the fitted distributions superimposed over your data so you can visually check the quality of the fits Several chart preferences make it easier to pinpoint discrepancies in the fits If desired you can override the highest ranking probability distribution with another one of your choice Crystal Ball Note Only continuous distributions are considered for distribution fitting Continuous and discrete distributions are defined on page 249 Distribution fitting can also be used to check the characteristics of a forecast chart See Fitting a distribution to a forecast on page 120 for more information Use the Fit Distribution dialog to specify the source of your data the distributions to be fitted and the goodness of fit test to use Each goodness of fit test is calculated for every distribution but only the selected test determines how the distributions are ranked Crystal Ball Note Difficulties can occur when distribution fitting is selected for one or more forecasts with large numbers of trials This is true for Normal Crystal Ball User Manual 27 Chapter 2 Defining Model Assumptions as well as Extreme speed To avoid these difficulties fitting is disabled for all run modes after 1 000 trials have been run A final fit is performed when the simulati
257. ity Distributions New York John Wiley amp Sons Inc 1986 Spurr William and Charles P Bonini Statistical Analysis for Business Decisions Homewood IL Richard D Irwin Inc 1973 Wadsworth George P and Joseph G Bryan Introduction to Probability and Random Variables New York McGraw Hill Book Co 1960 Winkler Robert L and William L Hays Statistics Probability Inference and Decision New York Holt Rinehart and Winston 1975 Random variate generation methods Devroye Luc Non Uniform Random Variate Generation New York Springer Verlag 1986 Iman Ronald L and W J Conover A Distribution Free Approach to Inducing Rank Correlation Among Input Variables Communications in Statistics B11 3 1982 Iman Ronald L and J M Davenport Rank Correlation Plots for Use With Correlated Input Variables Communications in Statistics B11 3 1982 Iman Ronald L and J M Davenport Latin Hypercube Sampling A Program User s Guide Technical Report SAND79 1473 Sandia National Laboratories 1980 Kennedy William J Jr and James E Gentle Statistical Computing New York Marcel Dekker Inc 1980 Crystal Ball User Manual 359 Bibliography Knuth Donald E The Art of Computer Programming Vols I III Reading MA Addison Wesley Publishing Co 1969 Newman Thomas G and Patrick L Odell The Generation of Random Variates New York Hafner Publishing Co 1971 Press William H
258. ity Features 350 Extracting capability metrics manually You can also extract capability metrics manually after a simulation runs To do this l 4 Choose Analyze gt Extract Data to display the Extract Data Preferences dialog Extract Data Preferences 7 xj Capability metrics Figure D 9 Choosing to manually extract capability metrics Check Capability Metrics at the bottom of the Select Data To Extract list Choose appropriate Forecasts and Assumptions settings and specify locations and other preferences on the Options tab For more information see Extracting data on page 190 or click the Help button in the Extract Data Preferences dialog Click OK to extract the data Capability metrics are written to the specified location along with any other data you have requested See Figure D 10 for an example Crystal Ball User Manual Including capability metrics in reports B c U FLOW RATE FORECAST ml sec TOTAL COST FORECAST A Capability metrics Mean Standard Deviation Zst total 1 64 _ Zit 0 14 _ p N C below 0 05 _ p N C above 0 00 Ld p N C total 0 05 PPM below 50 879 49 _ PPM above 55 92 Description Model Data id Figure D 10 Manually extracted capability metrics Including capability metrics in reports You can include capability metrics in full forecast or custom reports To do this 1 Choose Analyze gt
259. k Create output on the active worksheet Specify upper left corner of the output IV Format output 8 13 a Output Orientation m IV Show table of goodness of fit statistics f Data in rows Data in columns Define correlations above fo lt Back cancel bep Figure 8 4 Select Output Options dialog Batch Fit tool Select the Create Output On The Active Worksheet option Click the Select Cells icon to the right of the Specify Upper Left Corner Of The Output field Select the Sales Model tab Select cell B13 Click the return icon to return to the Batch Fit dialog 18 19 20 21 Make sure the Format Output option is checked Set the Output Orientation to Data In Columns Check the Show Table Of Goodness of fit Statistics option To include a correlation analysis of the data Check the Calculate Correlations Between Series option b Enter a threshold value in the Define Correlations Above field The threshold must be between 0 and 1 inclusive The tool displays the correlation matrix at the end of the results section Click Start The tool fits each selected distribution to each data series The results appear on the Sales Model worksheet below the existing table similar to Figure 8 5 Crystal Ball User Manual i Magazine Sales xds Batch Fit tool Magazine Checkout Sales Magazine A Magazine B Magazine C Magazine D Sales Volume 562 362 646 5
260. kbooks click the Help button for details and click OK Otherwise go to step 5 5 Click OK to apply the settings on all tabs to the active chart Specific customization instructions This section explains how you can customize each part of your Crystal Ball charts to present the information most effectively Topics include e Adding and formatting chart titles page 132 Setting the chart type page 132 Changing the chart density page 134 Showing grid lines page 135 Showing the chart legend page 135 Setting special chart effects page 136 Setting chart colors page 136 Showing the mean and other marker lines page 137 Customizing chart axes and axis labels page 138 Applying settings to the current chart and other charts page 139 Crystal Ball Note For basic chart customization instructions see page 130 Crystal Ball User Manual 131 Chapter 5 Analyzing Forecast Charts Adding and formatting chart titles To add or change a title 1 Display the General tab of the Chart Preferences dialog page 130 The Chart Title group appears at the top of the General page By default Auto is checked and a default title appears 2 To edit the title uncheck Auto and type the new title in the text box 3 When all your settings are complete click OK The chart appears with your new settings Setting the chart type Depending on the basic chart type assumption forecast trend overlay or sensitivity you
261. l subordinate A relationship between distributions in which one distribution s values for all percentile levels are lower than another s see also dominant trial also iteration A three step process in which Crystal Ball generates random numbers for assumption cells recalculates the spreadsheet model or models and displays the results in a forecast chart trial as used to describe a parameter in certain probability distributions The number of times a given experiment is repeated value cell A cell that contains a simple numeric value variable A quantity that can assume any one of a set of values and is usually referenced by a formula variance The square of the standard deviation i e the average of the squares of the deviations of a number of observations from their mean value Variance can also be defined as a measure of the dispersion or spread of a set of values about a mean When values are close to the mean the variance is small When values are widely scattered about the mean the variance is large See formulas in Variance in the Statistical Definitions chapter of the online Crystal Ball Reference Manual virtual memory Memory which uses your hard drive space to store information after you run out of random access memory Virtual memory supplements your random access memory 370 Crystal Ball User Manual workbook An Excel file composed of at least one worksheet worksheet An
262. l it is possible to construct formulas that produce an intermediate character string result which is used in later formulas to produce a numeric result For maximum speed gains you should avoid the use of strings in your formulas Strings in formulas will not be flagged as incompatible but will cause simulations to run much slower depending on the number of string formulas in your model Workaround Some simple strings like Yes and No can be replaced with and 0 values or True and False Boolean values Calls to user defined functions Calls to user defined or third party functions cannot be vectorized and will cause simulations to run slower Crystal Ball User Manual 337 Appendix C Using the Extreme Speed Feature e Workaround There is no workaround for this case other than minimizing the use of these functions Dynamic assumptions Assumptions that contain cell references to parameter values that can change during a simulation are called dynamic assumptions Crystal Ball needs to perform additional validation checks on the parameters for these assumption types during a simulation causing simulations to run more slowly e Workaround If the dynamic assumption feature is not needed make sure that parameter cell references do not point to other assumption cells in your model Crystal Ball Note Cell references in custom distributions are treated as static instead of dynamic when a model is run at Extreme spe
263. l 203 Chapter 8 Crystal Ball Tools c Select cells C5 through F5 Ee d Select Define gt Paste Data The assumptions copy to the first row of the table leaving the current values Ej Magazine Sales xds 4 _ 0 x F E 3 4 Magazine A Magazine B Magazine C Magazine D 5 Sales Volume 6 Retail Price 4 95 7 95 3 95 5 95 7 Cost of Goods 2 20 3 80 1 95 2 40 8 Gross Profit 1 546 1 502 1 292 1 805 9 10 11 Total Gross Profit 36 144 19 Figure 8 6 Crystal Ball assumptions in cells C5 through F5 e Select cells C15 through F15 a f Select Define gt Clear Data This deletes the original assumptions generated by Batch Fit to eliminate redundant assumptions and speed up the simulation 4 Inthe Run gt Run Preferences gt Sampling dialog set e Random Number Generation to use the Same Sequence Of Random Numbers and a seed value of 999 e Monte Carlo simulation When using this tool use these options to make the resulting simulations comparable 5 Inthe Run gt Run Preferences gt Trials dialog set the Maximum Number Of Trials to 500 6 Click OK 7 Select Run gt Start Simulation The simulation runs until it stops Interpreting the results In this example you had historical data with no trend or seasonality for all your variables 204 Crystal Ball User Manual Batch Fit tool Crystal Ball Note If your historical data have a time series element trend or sea
264. l Ball simulations The final forecast chart reflects the combined uncertainty of the assumption cells on the model s output Keep in mind that Monte Carlo simulation can only approximate a real world situation When you build and simulate your own spreadsheet models you need to carefully examine the nature of the problem and continually refine the models until they approximate your situation as closely as possible For more information about Monte Carlo simulations and Crystal Ball see page 11 Crystal Ball provides statistics that describe the forecast results For more information on forecast results and statistics see Chapter 5 Analyzing Forecast Charts and the Statistical Definitions chapter of the online Crystal Ball Reference Manual Steps for running simulations To run simulations in Crystal Ball follow these basic steps 1 Define assumptions page 19 forecasts page 59 and decision variable cells if appropriate page 58 2 Ifyou want customize the appearance of each cell as described in Setting cell preferences beginning on page 74 3 Set run preferences as described in the next section beginning on page 80 4 Run the simulation as described in Running simulations beginning on page 90 The following sections explain each step Crystal Ball User Manual 79 Chapter 4 Running Simulations Setting run preferences You can change a number of run preferences factors that determine
265. l the way across Spider chart The spider chart illustrates the differences between the minimum and maximum forecast values by graphing a curve through all the variable values tested Curves with steep slopes positive or negative indicate that those variables have a large effect on the forecast while curves that are almost horizontal have little or no effect on the forecast The slopes of the lines also indicate whether a positive change in the variable has a positive or negative effect on the forecast Target forecast Material 1 Reliability e Material 1 Strength s Coil Diameter in Forecast axis values 1 10 Number of Coils Assumptions and other Shearing Modulus of variables Elasticity Wire Diameter in 1 00 10 0 30 0 50 0 70 0 90 0 Percentiles of the variables e Spring Deflection in Figure 8 15 Spider chart CB Tools Note A maximum of 250 variables can be displayed in these charts Crystal Ball User Manual 213 Chapter 8 Crystal Ball Tools 214 Tornado Chart example In the Crystal Ball Examples folder there is a Reliability xls workbook you can use to experiment with the Tornado Chart tool This spreadsheet model predicts the reliability of a spring using three different construction materials To run Tornado Chart 1 In Excel with Crystal Ball loaded open the workbook Reliability xls If you have any other spreadsheets open close t
266. lated Ppk lower and Ppk upper that takes into account the centering of the forecast with respect to the midpoint of the specified limits a Ppk equal to or greater than indicates a quality level of 3 sigmas or better Cpm Short term Taguchi capability index similar to Cpk but considers a target value which may not necessarily be centered between the upper and lower specification limits Ppm Long term Taguchi capability index similar to Ppk but considers a target value which may not necessarily be centered between the upper and lower specification limits Z LSL The number of standard deviations between the forecast mean and the lower specification limit Z USL The number of standard deviations between the forecast mean and the upper specification limit Zst For short term metrics when only one specification limit is defined equal to Z LSL if there is only a lower specification limit or Z USL if there is only an upper specification limit Zst total For short term metrics when both specification limits are defined the number of standard deviations between the short term forecast mean and the lower boundary of combining all defects onto the upper tail of the normal curve Also equal to Zlt total plus the Z score shift value if a long term index is available Zit For long term metrics when only one specification limit is defined equal to Z LSL if there is only a lower specification limit or Z
267. lation and the distribution of the assumption itself is modeled for different sets of parameters in the inner simulation Crystal Ball Note Often the parameters of assumptions are correlated For example you would correlate a higher mean with a higher standard deviation or a lower mean with a lower standard deviation Defining correlation coefficients between parameter distributions can increase the accuracy of your two dimensional simulation With data available as in sample body weights of a population you can use the Bootstrap tool to estimate the sampling distributions of the parameters and the correlations between them 244 Crystal Ball User Manual selecting and Using Probability Distributions In this appendix Understanding probability e Using discrete distributions distributions we Hs e Using the custom distribution e Selecti bability distributi Batic sa cic conan E Truncating distributions e Using basic distributions HoHo Comparing the distributions e Using continuous distributions This appendix explains probability and probability distributions Understanding these concepts will help you select the right probability distribution for your spreadsheet model This section describes in detail the distribution types Crystal Ball uses and demonstrates their use with real world examples Crystal Ball User Manual 245 Appendix A Selecting and Using Probability Distributions Understanding probab
268. le functions and array arguments User defined functions whose arguments are static their values do not change during a simulation are not called by Extreme Speed unless the Volatile property of the function has been set When a user defined function is encountered in Extreme Speed Crystal Ball first checks to see if the function is Volatile If Crystal Ball is denied access to the VBA project and the user defined function is not passing array or cell range arguments Crystal Ball treats the function as Volatile If access to the project is denied and the call is passing array or cell range arguments the following message appears Cannot interpret a user defined function with array arguments You must first check the Trust Access to Visual Basic Project checkbox in Excel s macro security settings dialog See Appendix C in the user manual for more information about this error Crystal Ball User Manual 329 Appendix C Using the Extreme Speed Feature 330 In that case in order for Extreme Speed to determine the state of the Volatile property in Excel 2002 XP and later you must first make the following setting in Excel Go to the Tools gt Options dialog Click the Security tab Click the Macro Security button Click the Trusted Publishers tab eee a E a Check Trust Access To Visual Basic Project Crystal Ball Note Note that all user defined functions are treated as Volatile in Excel 2000 Running
269. le workbooks are open For example a workbook with order 10 will run before a workbook with order 20 CB SetCBWorkbookPriority runs only at the workbook level and not at the worksheet level The definition of this macro is Public Function CB SetCBWorkbookPriority aPriority As Long A syntax example used in a subroutine at the module level is Application Workbooks Book1 xls Activate CB SetCBWorkbookPriority 10 This example assigns a priority of 10 to the workbook Book 1 xls Within each workbook macros on worksheets run in left to right sheet order in the workbook provided there are no macros at the workbook level If there are macros at the workbook level macros in that workbook at the worksheet level do not run If there are no Crystal Ball data cells in a workbook no macros run against it e Macros in modules are not supported Global macros If you want to run certain macros with any or all of your Crystal Ball models you need to make sure that the workbook containing those macros is open in Excel Then you can use calls within each model to run those macros as required Previous versions of Crystal Ball handled global macros differently Now you can run specific macros only for certain models and are no longer required to turn the macros on and off depending on the model that is being run Toolbar macros If you want to use custom Excel toolbar macros defined in a separate workbook while Crystal Ball is
270. level of sensitivity can be considered as the most influential assumption in the model To create a sensitivity chart 1 2 7 Close any spreadsheets that are currently open Open the spreadsheet you want to analyze Toxic Waste Site xls is used in these instructions You can choose Help gt Crystal Ball gt Examples Guide to choose the model from a list Choose Run gt Run Preferences gt Options Make sure Store Assumption Values For Sensitivity Analysis is checked and click OK Sensitivity analysis requires that all of the random numbers from the simulation be kept for comparison with forecast values Run a simulation until it stops Choose Analyze gt Sensitivity Charts The Sensitivity Charts dialog appears If you have not yet created any sensitivity charts for the active spreadsheet or restored results file the dialog is blank as shown in Figure 6 19 15 x Click New or choose existing charts to Open Close or Delete view TE set MIE Figure 6 19 The Sensitivity Charts dialog blank Click the New button Crystal Ball User Manual 167 Chapter 6 Analyzing Other Charts 168 The Choose Forecast dialog appears similar to Figure 6 20 Choose Forecast TST Choose the target forecast for the sensitivity chart vele see a ooo EE m oic Waste Site xis a 0 Model Duly Risk Assessment Figure 6 20 Choose Forecast dialog Tree view By default this dialog appears in a hierar
271. lick Apply To Then specify how they should be applied see page 139 for details and click OK Otherwise go to step 7 7 When settings are complete click OK The trend chart appears with the value axis changed Crystal Ball Note You can copy trend charts and paste them into other applications For more information see page 141 164 Crystal Ball User Manual Understanding and using sensitivity charts Understanding and using sensitivity charts Glossary Term sensitivity The amount of uncertainty in a forecast cell that is caused by both the uncertainty and model sensitivity of an assumption cell Glossary Term model sensitivity The overall effect that a change in an assumption cell produces in a forecast cell This effect is solely determined by the formulas in the spreadsheet model Assumption Cell Assumption Cell Measure of Model Sensitivity As you become more proficient at building spreadsheet models you will want to know how much a given assumption affects your result In other words you want to determine the sensitivity of the forecast to each assumption The overall sensitivity of a forecast to an assumption is a combination of two factors The model sensitivity of the forecast to the assumption The assumption s uncertainty To determine the model sensitivity yourself you would need to algebraically analyze the various relationships between your forecast cell a
272. ll not give you the correct lognormal distribution Alternatively use the distribution fiting feature described on page 27 The first step in selecting a probability distribution is matching your data with a distribution s conditions Checking the lognormal distribution e The price of the stock is unlimited at the upper end but cannot drop below 0 The distribution of the stock price is positively skewed The natural logarithm of the stock price yields a normal distribution These conditions match those of the lognormal distribution Figure A 10 Crystal Ball User Manual Using continuous distributions inixi Edit View Parameters preferences Help Name Jai xj y Lognormal Distribution 44 00 55 00 66 00 77 00 88 00 99 00 110 00 Mean 70 00 S Sid Dev 512 00 xj oK Cancel Enter_ Galley Correlate Help Figure A 10 Lognormal distribution In the lognormal distribution the mean parameter is set at 70 00 and the standard deviation set at 12 00 This distribution shows the probability that the stock price will be x Lognormal parameter sets By default the lognormal distribution uses the arithmetic mean and standard deviation For applications where historical data are available it is more appropriate to use the logarithmic mean and logarithmic standard deviation or the geometric mean and geometric standard deviation These options are available from the Parameters menu in t
273. loaded be sure to save that workbook to the folder C Documents and Settings yourname Application Data Microsoft Excel XLSTART Crystal Ball User Manual Analyzing Forecast Charts In this chapter Guidelines for analyzing simulation results Understanding and using forecast charts e Setting chart preferences e Managing existing charts This chapter describes how to use forecast charts to analyze simulation results As you read through the chapter you will learn to focus on a particular range of simulation results analyze forecast charts and interpret relevant descriptive statistics You will also learn how to customize forecast charts work with chart windows and share charts with other applications The next chapter introduces additional charts Chapter 7 describes how to extract Crystal Ball simulation data for use in Excel and other applications and how to prepare reports Crystal Ball User Manual 103 Chapter 5 Analyzing Forecast Charts Guidelines for analyzing simulation results 104 While a simulation runs Crystal Ball creates a forecast chart for each forecast cell Forecast charts condense much information into a small space You can display that information in several different ways both graphically and numerically When you analyze a simulation you can focus on forecast charts You can also display other kinds of charts generate reports and extract data for further processing using Excel or other an
274. lp _ Figure A 44 Mixed ranges loaded in a custom distribution Connected series of ranges sloping For a connected series of sloping continuous ranges choose Parameters gt Sloping Ranges to use a five column format The first column contains the lowest Minimum value of the right most range the second column contains the Maximum value of each connected range the third column contains the Height of Min relative probability of the Minimum value if it differs from the previous Height of Max otherwise it is left empty and the fourth column contains Height of Max relative probability of the Maximum value for that range The fifth column is left blank for continuous ranges but a fifth column is necessary to indicate that these are sloping ranges For example row 20 in Figure A 43 shows a connected continuous sloping range The Minimum cell is blank because the Minimum value is equal to 7 the previous Maximum The Height of Min is blank because it is equal to 6 the previous Height of Max Connected series of continuous uniform ranges cumulative For a connected series of continuous uniform ranges specified using cumulative probabilities use a three column format with the common endpoints of the ranges in the second column and the cumulative probabilities in the third column The first column is left blank except for the Crystal Ball User Manual 309 Appendix A Selecting and Using Probability Distributions minimum value o
275. lues such as yes or no success or failure true or false or heads or tails Binomial distribution all Binomial Parameters Probability Trials Statistical Note The word trials as used to describe a parameter of the binomial distribution is different from trials as it is used when running a simulation in Crystal Ball Binomial distribution trials describe the number of times a given experiment is repeated flipping a coin 50 times would be 50 binomial trials A simulation trial describes a set of 50 coin flips 10 simulation trials would simulate flipping 50 coins 10 times Conditions For each trial only two outcomes are possible The trials are independent What happens in the first trial does not affect the second trial and so on The probability of an event occurring remains the same from trial to trial Description The binomial distribution describes the number of times a particular event occurs in a fixed number of trials such as the number of heads in 10 flips of a coin or the number of defective items in 50 items Example one You want to describe the number of defective items in a total of 50 manufactured items 7 of which on the average were found to be defective during preliminary testing Crystal Ball User Manual 281 Appendix A Selecting and Using Probability Distributions 282 The first step in selecting a probability distribution is matching your data with a distribu
276. mal and Student s distributions are related With Degrees of Freedom gt 30 Student s t closely approximates the normal distribution The binomial and hypergeometric distributions are also closely related As the number of trials and the population size increase the hypergeometric trials tend to become independent like the binomial trials the outcome of a single trial has a negligible effect on the probabilities of successive observations The differences between these two types of distributions become important only when you are analyzing samples from relatively small populations As with the Poisson and binomial distributions Crystal Ball requires a similar amount of time to compute both the binomial and hypergeometric distributions The yes no distribution is simply the binomial distribution with Trials 1 The Weibull distribution is very flexible Actually it consists of a family of distributions that can assume the properties of several distributions When the Weibull shape parameter is 1 0 the Weibull distribution is identical to the exponential distribution The Weibull location parameter lets you set up an exponential distribution to start at a location other than 0 0 When the shape parameter is less than 1 0 the Weibull distribution becomes a steeply declining curve A manufacturer might find this effect useful in describing part failures during a burn in period When the shape parameter is equal to 2 0 a special form of th
277. mal and does not affect simulation results Single stepping mode still updates cell values for every step Very large models Very large models with greater than 10 000 formulas for example might take Crystal Ball many seconds or minutes in Extreme speed to analyze the model and convert it into a form for high speed simulation When this happens the following progress dialog appears J Preparing to run in Extreme speed Press Ctrl Break to cancel Figure C 1 Extreme speed Progress dialog Crystal Ball User Manual 333 Appendix C Using the Extreme Speed Feature If you choose you can cancel the simulation by pressing the Ctrl Break keys or wait for the process to finish Memory usage Running simulations in Extreme speed is extremely memory intensive Depending on your machine you may encounter issues with memory usage when running long simulations or working with large models During a long simulation greater than 100 000 trials you might also see that your machine s virtual memory is being used For example you might notice that your hard drive is being accessed continually and the simulation starts to slow down No message is given when this condition occurs If such a slow down occurs and it is absolutely necessary that additional trials be run you can try to disable the saving of assumption values see the Run Preferences gt Options tab and then restart the simulation Crystal Ball Note Ifyou run an
278. mparison results for the Tolerance Analysis xls example file with 5 000 trials In this case there were no differences in results and the model ran 28 8 times faster in Extreme speed EJ Microsoft Excel Book5 4 rea File Edit View Insert Format Tools Data RoboPDF Window Define Run Analyze Help x DAE BIG RIF OG sm FiO Z 4h 5 E Bis 8 Oli AG BA 5 P MAD thane Fie OP J13 Ke fe Comparison of Run Modes Workbook Tolerance Analysis xls Results comparison Total differences detected 0 Total cells checked 20096 Percentage differences to total 0 00 Average difference RMSE 0 Relative Maximum difference o Speed comparison Extreme speed time 3 27 seconds Normal speed time 94 27 seconds Speed difference 28 8 x Speed Comparison 0 00 20 00 40 00 60 00 80 00 100 00 Seconds Figure C 3 Comparison results for Tolerance Analysis xls 5 000 trials When you launch CompareRunModes xla a new menu command appears Run gt Compare Run Modes It stays there as long as Excel is running so you can use the command to run additional comparisons Crystal Ball Note Because of variations in the random number seeds you might see differences in comparision results if you use Excel s RAND or Crystal Ball s probability functions such as CB Uniform in your model 336 Crystal Ball User Manual Maximizing the benefits of Extreme Speed Maximizing the benefits of Extreme Speed To get maximum benefit fr
279. mption values run preference Run Preferences Cancel Appl To Defaults Help Figure 6 29 Assumption Preferences dialog While the Forecast Preferences dialog has a button for fitting distributions to forecasts distribution fitting isn t available in the Assumption Preferences dialog Instead there is a Run Preferences button so you can easily change the Store Assumption Values setting on the Options tab of the Run Preferences dialog For more information on the Assumption Preferences dialog click the Help button in the dialog Setting assumption chart preferences Assumption chart preferences are virtually identical to forecast chart preferences To review or change them choose Preferences gt Chart and follow the instructions beginning on page 128 Crystal Ball User Manual Creating Reports and Extracting Data In this chapter e Creating reports Extracting data This chapter describes how to create reports with charts and data and how to extract data for use in other applications Crystal Ball User Manual 181 Chapter 7 Creating Reports and Extracting Data Creating reports You can generate pre defined reports for your simulation or you can create a custom report with any or all of the following items e Report summary e Forecasts e Assumptions e Decision variables e Overlay charts Trend charts e Sensitivity charts Figure 7 1 shows part of a forecast report
280. mum field e Type 21 in the Maximum field Type 25 in the Probability field e Click Enter Crystal Ball displays a continuous value bar for the range 16 00 to 21 00 Its height is 050 equal to 25 divided by 5 the number of units in the range Both ranges now appear in the Custom Distribution dialog Figure A 31 Crystal Ball User Manual 299 Appendix A Selecting and Using Probability Distributions C Define assumption cetona aT Edit View Parameters Preferences Help Name pr xi Custom Distribution 16 00 18 00 20 00 Enter one or more continuous ranges with probabilities Minimum O M i ty Ok Cancel Enter Gallery _ Correlate Hele Figure A 31 Custom distribution with two continuous ranges You can change the probability and slope of a continuous range as described in the following steps 1 Click anywhere on the value bar for the range 16 to 21 The value bar changes to a lighter shade 2 Choose Parameters gt Sloping Ranges Additional parameters appear in the Custom Distribution dialog Enter one or more continuous or discrete sloping ranges Minimum fis 00 Maximum 21 00 Height of Min 0 25 Height of Max Step Figure A 32 Sloping Range parameters Custom Distribution dialog 3 Set the Height of Min and Height of Max equal to what currently appears in the chart 0 05 This can be an approximate value The Height of Min is the height of the range Minimum and the He
281. n You can override this option and close all the forecast windows using the Suppress Forecast Windows option in the Run Preferences dialog For more information see Setting run preferences on page 80 While Running Displays the forecast window automatically during simulations Simulation This slows down the simulation When Displays the forecast window automatically after the simulation Simulation Stops stops This is faster than displaying the window during simulations Fit Fits a continuous probability distribution to the forecast After selecting this checkbox Distribution in this group you can click Fit Options to select the distribution and goodness of fit tests you want Crystal Ball User Manual 63 Chapter 3 Defining Other Model Elements 64 Precision preferences Define Forecast Cell G11 i xi Name Gross Profit MM _al Units milions SS Forecast Window Precision Fiter Auto Extract IV Specify the desired precision for forecast statistics V Mean I Standard deviation I Percentile 95 Must be within plus or minus Units fo 0s e g 100 dollars C Percent E e g 5 percent Crystal Ball will run the simulation until the specified precision limits for all forecasts have been reached or until the maximum number of trials whichever is first Cancel Appl To Defauts Help Figure 3 4 Precision tab forecast preferences This tab manages the precision con
282. nate parameter sets on page 25 Crystal Ball User Manual 251 Appendix A Selecting and Using Probability Distributions Using basic distributions This section describes distributions in the Basic category of the Distribution Gallery Distribution Gallery Cell C5 Edit View Categories Help za A Hm lol x Triangular Uniform A ool Il px Lognormal Yes No Discrete Uniform Favontes Normal Description The normal distribution describes many naturel phenomena such as IQs people s heights the inflation rate or errors in measurements It is a continuous probability distribution The parameters for the nomal distribution are mean and standard deviation E cma e ee Figure A 5 Distribution Gallery Basic category Basic distributions are listed below in the same order they appear above For details see the page references below the names Table A 1 Summary of basic distributions Shape Name Summary Normal The normal distribution is the most important distribution in probability page 268 theory because it describes many natural phenomena such as people s IQs or heights Decision makers can use the normal distribution to describe uncertain variables such as the inflation rate or the future price of gasoline Normal Triangular The triangular distribution describes a situation where you know the gu ne S i y page 274 minimum maximum and most likely values to
283. nd a valid license authorization code There are a number of ways to reach Technical Support described in the README file in the Crystal Ball installation folder Online see http support crystalball com Training Decisioneering s Training group offers a variety of courses throughout the year to help improve how you make decisions For more information about Crystal Ball User Manual 5 Introduction Decisioneering courses call one of these numbers Monday through Friday between 8 00 a m and 5 00 p m Mountain Time 1 800 289 2550 toll free in US or 1 303 534 1515 or visit the Decisioneering Web site http www crystalball com training Consulting Decisioneering s Services group provides consulting services including the full range of risk analysis techniques from simulation optimization advanced statistical analysis and exact probability calculations to strategic thinking training expert elicitation and results communication to management To learn more about these consulting services call 1 800 289 2550 Monday through Friday between 8 00 A M and 5 00 P M Mountain Time or see our Web site at http www crystalball com consulting 6 Crystal Ball User Manual Crystal Ball Overview In this chapter Model building and risk analysis overview Steps for using Crystal Ball Starting and closing Crystal Ball The Crystal Ball menus and toolbar This chapter presents the basics you need to
284. nd your assumption cells These relationships include all of the formulas in the spreadsheet that link the assumption cells to the forecast cell Figure 6 17 illustrates two scenarios where model sensitivity and assumption uncertainty combine to affect two different forecasts In each diagram the distance between the assumption and the leverage point represents the model sensitivity The larger the distance the less model sensitivity Uncertainty Forecast Cell A Leverage Point Determined by the spreadsheet model It represents the extent to which the assumption cell affects the forecast cell Uncertainty Forecast Cell A Figure 6 17 Relationships between model sensitivity and assumption uncertainty Crystal Ball User Manual 165 Chapter 6 Analyzing Other Charts 166 In the top model the forecast is not as sensitive to the assumption because the model sensitivity is lower even though there is high uncertainty in the assumption In the bottom model the forecast is highly sensitive to even the small assumption uncertainty because of the high model sensitivity Calculating these complex relationships could be a difficult and time consuming task without a program like Crystal Ball About sensitivity charts Sensitivity charts show the influence each assumption cell has on a particular forecast cell During a simulation Crystal Ball ranks the assumptions according to their importance to each foreca
285. ndividual Define Assumption Cell A1 W lol x Edit View Parameters Preferences Help Logistic Distribution Figure A 9 Logistic distribution Calculating parameters There are two standard parameters for the logistic distribution mean and scale The mean parameter is the average value which for this distribution is the same as the mode since this is a symmetrical distribution After you select the mean parameter you can estimate the scale parameter The scale parameter is a number greater than 0 The larger the scale parameter the greater the variance Crystal Ball User Manual Using continuous distributions To calculate a more exact scale you can estimate the variance and use the equation ge 3 variance 2 T where is the scale parameter Lognormal distribution A Lognormal Glossary Term skewed positively A distribution in which most of the values occur at the lower end of the range Parameters Mean Standard Deviation Conditions The uncertain variable can increase without limits but cannot fall below zero The uncertain variable is positively skewed with most of the values near the lower limit The natural logarithm of the uncertain variable yields a normal distribution Description The lognormal distribution is widely used in situations where values are positively skewed for example in financial analysis for security valuation or in real estate for property valu
286. ne which assumption values produce specific forecast results Hw Please enter the target cell or select from the list of forecasts above Toxic Waste Site xls Model sCs9 Figure 8 35 Specify Target dialog Scenario Analysis tool 232 Crystal Ball User Manual Scenario Analysis tool 3 Select the Risk Assessment forecast 4 Click Next The Specify Options dialog appears Specify options step 2 of 2 Scenario Analysis Crystal Ball Tool Decisioneering 1998 2004 rRange of Forecast Results Select percentiles from fo to 100 C Select forecast values from to MWhile Running ulation Control Show forecasts as defined Run z trials maximum Show only target forecast Hide all forecasts pak eae cancel Hep Figure 8 36 Specify Options dialog Scenario Analysis tool 5 In the Range Of Forecast Results section specify a percentile range from 95 to 100 percent 6 Inthe While Running section select Show Only Target Forecast 7 Inthe Simulation Control section enter 1000 as the maximum number of trials to run 8 Click Start The tool creates a table of all the forecast values within the range specified in step 5 along with the corresponding value of each assumption as shown in Figure 8 37 Crystal Ball User Manual 233 Chapter 8 Crystal Ball Tools Pa Paste Selected Scenario Paste Next Scenario Paste
287. nes of best fit for greater contrast in the figure Crystal Ball Note You can copy overlay charts and paste them into other applications For more information see page 141 Crystal Ball User Manual 153 Chapter 6 Analyzing Other Charts Understanding and using trend charts After completing a simulation with multiple related forecasts you can create a trend chart to view the certainty ranges of all the forecasts on a single chart A trend chart summarizes and displays information from multiple forecasts making it easy to discover and analyze trends that might exist between related forecasts CI nx Edit View Trend Preferences Help 500 Trials Centered on Median View Trend Chart 1 22 000 000 20 000 000 Certainty Bands Hi0 18 000 000 H25 50 16 000 000 90 14 000 000 12 000 000 LO 48a sajes GuIpug ZO 48a sajes BuIpug 0 4a sajes Burpug tO 48a sajes Bupug LO z 488A sajes BuIpug ZO z 4834 sajes Buipug 0 z 434 sajes Buipug tO za sajes Bulpuz LO 483A sajes Buipug zo g 1834 sajes Bulpug E g 1834 sajes Bupug tO g 1234 sajes Bupu Figure 6 9 Upward trending sales figures by quarter Trend charts display certainty ranges for multiple forecasts in a series of colored bands Each band represents the certainty ranges into which the actual values of your forecasts fall For exampl
288. newest versions of Excel and so on Also small differences in the last few decimal places of certain built in function values should be expected due to minor algorithmic differences in the way formulas might be computed It is important to note that incompatibilities in functions and formula constructs only concern the cells involved in the calculation of a forecast cell If there are incompatibilities in cells that are not part of that calculation path these are not detected and the simulation is allowed to run 326 Crystal Ball User Manual Compatibility issues Multiple workbook models Extreme speed can now run simulations on multiple workbooks If you are running in Extreme speed and the workbook contains external references to cells in other closed workbooks Crystal Ball obtains the current value from those workbooks References to cells in other open workbooks are dynamically updated if those cells depend on one or more assumptions If the external reference is part of a formula not a simple external reference this is not compatible with Extreme speed e Example message Unable to interpret formula at cell location Book1 xls Sheet1 A1 Code 5524 Complex external reference e Workaround If possible consolidate all of the variables and formulas from a multi workbook model containing Crystal Ball data into a single workbook Circular references Circular references within a model are supported as long as Iteration
289. nificant is the risk Almost any change good or bad poses some risk Your own analysis will usually reveal numerous potential risk areas overtime costs inventory shortages future sales geological survey results personnel fluctuations unpredictable demand changing labor costs government approvals potential mergers pending legislation Once you identify your risks a model can help you quantify them Quantifying risk means determining the chances that the risk will occur and the cost if it does to help you decide whether a risk is worth taking For example if there is a 25 chance of running over schedule costing you 100 out of your own pocket that might be a risk you are willing to take But if you have a 5 chance of running over schedule knowing that there is a 10 000 penalty you might be less willing to take that risk Finding the certainty of achieving a particular result is often the goal of a model analysis Risk analysis takes a model and sees what effect changing different values has on the bottom line Risk analysis can Help end analysis paralysis and contribute to better decision making by quickly examining all possible scenarios e Identify which variables most affect the bottom line forecast Expose the uncertainty in a model leading to a better communication of risk What is a model Crystal Ball works with spreadsheet models specifically Excel spreadsheet models Your spreadsheet might already
290. nion operator A defined name whose value is a multiple area reference is not accepted Example message Multi area reference not supported at cell location Book1 xls Sheet1 A1 Code 5525 e Workaround For best results avoid using multiple area references when defining names or as arguments to any functions except those that accept a variable length comma separated list of cell ranges 3 D references Extreme speed does not support 3 D references where a cell range used as an argument in a function call for example spans multiple workbooks e Example message Unable to interpret formula at cell location Book 1 xls Sheet1 A1 Code 5514 Undefined identifier e Workaround For best results avoid using 3 D cell references Data Tables The one variable and two variable data table feature in Excel is not supported in Extreme speed This feature uses a TABLE function to perform an automatic what if analysis on one or two variables e Workaround Avoid the use of one and two variable data tables in models that will run in Extreme speed Other important differences There are several other differences between Extreme speed and Normal speed that you should be aware of OptQuest and CB Tools OptQuest and certain CB Tools run multiple simulations of your spreadsheet model back to back Since Extreme speed enhances the performance of each simulation you will see an overall speed improvement
291. nsitivity charts to other applications such as Microsoft Word Powerpoint and Excel Copying charts To copy charts for use in other applications 1 2 Select the chart to copy Open its View menu and choose the view to be copied If you choose a data view such as Percentiles Statistics or Goodness Of Fit the data will be pasted into many applications as alphanumeric data ready to edit add and so on This is true for Microsoft Excel and Word but not Powerpoint Data is pasted into Powerpoint as a graphic Graphic views such as Frequency are pasted as bitmap images In the chart s menubar choose Edit gt Copy The chart is copied to the Clipboard ready to paste into another application Pasting charts from the Clipboard To paste a chart into another application using its own Paste commands 1 2 Copy the Crystal Ball chart as described in the previous section Open a document spreadsheet slide and so on in the application to receive the chart Within that application choose Edit gt Paste or Edit gt Paste Special As described above if you copied a data view such as Percentiles Statistics or Goodness of Fit the data is pasted into many applications as editable numbers or text Graphic views such as Frequency are pasted as bitmap images Crystal Ball User Manual 141 Chapter 5 Analyzing Forecast Charts Printing charts To print a chart display it and choose Edit gt Print in the
292. number of values in the entire range from negative 14 7 million to positive 34 4 million to recalculate the certainty level With a certainty level of 69 4 you can be 69 4 confident of making a net profit between 0 and 20 million Changing the certainty minimum and maximum fields To determine the certainty level for a specific value range using the certainty minimum and maximum fields 1 In the forecast chart type a value in the certainty minimum field 2 Type a value in the certainty maximum field 3 Press Enter The certainty grabbers move to correspond to the values you entered Crystal Ball User Manual 109 Chapter 5 Analyzing Forecast Charts 110 Entering certainty directly Alternately you can type the certainty level on the forecast chart and let Crystal Ball calculate the minimum and maximum for you Make sure that the certainty grabbers are free whenever you click an certainty grabber it turns to a lighter color and is considered anchored To free the lighter color certainty grabber click it Then when you type in a certainty level you will center the certainty range on the median statistic Crystal Ball Note To free or anchor both certainty grabbers press Ctrl click or Shift click In the following procedure you will enter a certainty level of 10 and watch Crystal Ball move the certainty grabbers 1 Press Ctrl click or Shift click until both certainty grabbers are free a lighter color
293. occur For example you could describe the number of cars sold per week when past sales show the minimum maximum and usual number of cars sold Triangular Uniform In the uniform distribution all values between the minimum and maximum page 276 occur with equal likelihood Uniform 252 Crystal Ball User Manual Using continuous distributions Table A 1 Summary of basic distributions Continued Shape Name Summary A Lognormal Lognormal page 263 The lognormal distribution is widely used in situations where values are positively skewed for example in financial analysis for security valuation or in real estate for property valuation m Yes no The yes no distribution is a discrete distribution that describes a set of all page 293 observations that can have only one of two values such as yes or no success Sie or failure true or false or heads or tails Discrete In the discrete uniform distribution all integer values between the minimum uniform and maximum are equally likely to occur It is the discrete equivalent of the Discrete Uniform page 283 continuous uniform distribution Using continuous distributions Continuous probability distributions describe values over a range or scale and are shown as solid figures in the Distribution Gallery Continuous distributions are actually mathematical abstractions because they assume the existence of every possibl
294. og Copy distributions Edit gt Copy page 46 Paste distributions Edit gt Paste page 46 Modify distributions Edit gt Modify page 47 Modify distribution summaries and right click menu page 48 descriptions Delete distributions Edit gt Delete page 48 Set up pages for printing Edit gt Page Setup page 49 Print distribution information Edit gt Print page 49 Manage categories Create new categories Categories gt New page 50 View and edit category properties Categories gt Properties page 51 Delete categories Categories gt Delete page 51 Rearrange category order Categories gt Move Up page 51 Share categories over networks Categories gt Move Down Categories gt Publish page 52 Categories gt Subscribe Crystal Ball User Manual 41 Chapter 2 Defining Model Assumptions Table 2 2 Distribution Gallery tasks Continued How Dol Command Reference Use categories shared by others Categories gt Subscribe page 53 Send and receive categories page 54 Edi shared categories Categories gt Modify page 55 Unpublish shared categories page 55 Change distribution display within categories View menu page 44 View online help Help menu page 5 Displaying the Distribution Gallery To display the Distribution Gallery 1 With Crystal Ball open within Microsoft Excel click in a cell La 2 Click the Define Assumption tool or choose Define gt Define Assumption The Distribution Gallery appears as shown in Figure
295. ojection worksheet To select all forecasts in this example you can click the Select All button or check the box in front of Sales Projection xls or the Model tab name 5 Click OK The trend chart appears as shown in Figure 6 9 on page 154 As with overlay charts you can change the scale and proportions of the chart by dragging its edges For other customization possibilities see the next section 156 Crystal Ball User Manual Understanding and using trend charts Customizing trend charts You can customize trend charts by e Changing trend chart views Specifying when trend charts appear e Adding removing and ordering forecasts e Changing the overall appearance of trend charts e Setting certainty band type and colors Setting certainty bands e Changing value axis preferences The following sections describe how to make these changes For some of the options you can use hot keys to bypass the Trend Preferences dialog These hot keys are listed in Table 5 5 on page 129 and the Menus and Keyboard Commands chapter of the online Crystal Ball Reference Manual Changing trend chart views Use the trend chart View menu to change the placement of the certainty bands within the trend chart The default setting centers the bands around the median of each forecast You can change the location of the bands so that they are anchored at either the high end or the low end of the projected forecast ranges Crystal Ball Note
296. older under the Crystal Ball installation folder by default C Program Files Decisioneering Crystal Ball 7 Examples 5 Choose the distribution fitting characteristics For this example use these settings For Which Distributions choose All Continuous For Ranking Method choose Anderson Darling 6 Click OK to display the Comparison Chart dialog 32 Crystal Ball User Manual Additional assumption features For this Test Data example the Weibull distribution had the best fit of any distribution using the Anderson Darling fit test The parameters that were calculated for the Weibull distribution are displayed in Figure 2 8 on page 31 7 Use the Next and Previous buttons to view comparison charts for the other distributions 8 Return to Weibull and click Accept The Define Assumption dialog appears with the accepted Weibull distribution as shown in Figure 2 9 on page 32 Specifying correlations between assumptions In Crystal Ball assumption values are usually calculated independently of each other Crystal Ball generates random numbers for each assumption without regard to how random numbers are generated for other assumptions However dependencies often do exist between variables in a system being modeled The Correlated Assumptions feature in Crystal Ball lets you enter correlation coefficients to describe the dependencies between assumptions Correlation coefficients and correlation in general are explained in the
297. om running Crystal Ball in Extreme speed you should be aware of several issues in your models that could potentially cause simulations to run more slowly than expected When you first start a simulation in Extreme speed the model is analyzed and converted into a special form for high speed processing This form involves the use of a technique called vectorization to process batches of formula results together for maximum efficiency Certain functions and formula constructs in your model can interfere with this vectorization process Some of the more common cases are discussed below If one or more cases are detected Crystal Ball displays the message non vectored on the Excel status bar while the simulation is running Also additional information is usually written to the Crystal Ball log file which can be found with the suffix CB Log and extension txt in C Documents and Settings username Application Data Decisioneering Crystal Ball 7 0 Logs CB Log_wx txt Crystal Ball Note Ifyou have a very large number of trials the virtual memory feature of Microsoft Windows is activated This can slow down processing Too if you have large numbers of forecasts and assumptions managing these large arrays of values can take extra time If you think array management could be an issue try unchecking Store Assumption Values For Sensitivity Analysis on the Options tab of the Run Preferences dialog String intermediate results in formulas In Exce
298. ombined Again identify and enter the parameter values for the geometric distribution in Crystal Ball In this example the conditions show one important value a 0 06 6 probability of receiving that major claim The result would be a distribution showing the probability of x number of claims occurring between major claims Hypergeometric distribution Parameters Success Trials Population Statistical Note The word trials as used to describe a parameter of the hypergeometric distribution is different from trials as it is used when running a simulation in Crystal Ball Hypergeometric distribution trials describe the number of times a given experiment is repeated removing 20 manufactured parts from a box would be 20 hypergeometric trials A simulation trial describes the removing of 20 parts 10 simulation trials would simulate removing 20 manufactured parts 10 times Crystal Ball User Manual lll Hypergeometric Using discrete distributions Conditions The total number of items or elements the population size is a fixed number a finite population The population size must be less than or equal to 1000 The sample size the number of trials represents a portion of the population The known initial success rate in the population changes slightly after each trial Description The hypergeometric distribution is similar to the binomial distribution in that both describe the number of tim
299. on percentage of values greater than or Frequency equal to a given amount Statistics Shows a full set of descriptive statistics for a simulation in the forecast window Percentiles Shows percentile information in 10 increments where a percentile is the percent chance or probability of a forecast value being less than or equal to the value that corresponds to the percentile by default Goodness Of Fit If distribution fitting is selected in the Forecast or Preferences gt Forecast menus shows goodness of fit statistics for the selected distribution and ranking method Capability Metrics If process capability metrics are set for display shows a table of process capability quality statistics for the simulation See Viewing capability metrics on page 344 for more information Split View When selected shows all selected views simultaneously For details see Using Split View on page 122 The following sections show examples of each View type Crystal Ball User Manual 115 Chapter 5 Analyzing Forecast Charts Frequency Frequency the default forecast view shows a simple count of values the frequency for each interval on the x axis Figure 5 8 shows a frequency chart of net profit values for a simulation where there is a 75 probability of net profit falling between 0 00 and 26 1 million The chart has a median of 7 8 million This is also the 50th percentile By default there
300. on they are as general as possible although not all settings will apply to every type of chart To customize a chart 1 Create or display a chart and be sure it is the active chart window 2 Either double click the chart or choose Preferences gt Chart Preferences from the chart s menubar The Chart Preferences dialog appears similar to Figure 5 19 Chart Preferences M Chart bins Options Density Medium 7 Gridlines Horizontal x J Show column gaps Legend None 7 m Effects T 3D chart J Transparency Ea Figure 5 19 The Chart Preferences dialog General tab 3 Click each tab and make appropriate settings Tab contents are summarized in Table 5 6 For details about the settings on each tab click the Help button 130 Crystal Ball User Manual Setting chart preferences Table 5 6 Chart Preferences dialog tabs Tab Settings General Chart title overall appearance of the chart Chart Data sets series to display in the chart chart type and color of Type plotted series marker lines to display optional Axis Vertical and horizontal axes to display axis labels axis scales and axis number formats 4 Optional step If you want to apply the settings to more than one chart click Apply To Then specify whether to apply all chart preferences or just the current tab and whether to apply them to the current Excel sheet all sheets in the workbook or all open and new wor
301. on parameter field you would enter the cell reference C 5 This causes the value in cell C5 to be used in the assumption cell parameter field Later if you decide to copy and paste this assumption in the worksheet the cell references in the parameter field will refer to the contents of cell C5 Range names You can also enter cell references in the form of range names such as cellname Then the referenced cell can be located anywhere within a worksheet as long as its name doesn t change Formulas You can enter Excel formulas to calculate parameter values as long as the formula resolves to the type of data acceptable for that parameter For example if a formula returns a string it wouldn t be acceptable in a parameter that requires a numeric value such as Minimum or Maximum Alternate parameter sets For all the continuous probability distributions except uniform you can define the distributions using percentiles for parameters This option gives you added flexibility to set up assumptions when only percentile information is available or when specific attributes such as the mean and standard deviation of the variable in your model are unknown For example if you are defining a triangular distribution but are unsure of the absolute minimum and maximum values of the variable you could instead define the distribution using the 10th and 90th percentiles along with Crystal Ball User Manual 25 Chapter 2 Defining Model A
302. on stops a progress dialog appears during that fit so you can cancel the fit if necessary Using distribution fitting To use distribution fitting 1 Select the cell where you want to create an assumption It can be blank or contain a simple value not a formula Ay 2 Choose Define gt Define Assumption The Distribution Gallery appears 3 Click Fit to select the source of the fitted data The Fit Distribution dialog appears as shown in Figure 2 6 LI x Crystal Ball will fit a probability distribution to your data m Location of data G Range is Text file E Golumn fi m Which distributions _ m Ranking method All continuous Anderson Darling C Choose Chi Square C Normal Kolmogorov Smirnov Figure 2 6 Fit Distribution dialog 4 Choose one of the following two options e Ifthe historical data is in a worksheet in the active workbook choose Range and then enter the data s cell range e Ifthe historical data is in a separate text file click Text File and then either enter the path and name of the file or click Browse to search for the file If you want you can check Column and enter the number of columns in the text file 28 Crystal Ball User Manual Additional assumption features Crystal Ball Note When you use a file as your source of data each data value in the file must be separated by either a comma a tab character a space character or a list separator
303. on vectorized mode For this reason they will probably run more slowly than simulations without circular references Crystal Ball Excel functions The following Crystal Ball spreadsheet functions are handled normally CB IterationsFN distribution functions such as CB Binomial These functions are not supported in Extreme speed during a simulation CB GetForeStatFN CB GetForePercentFN CB GetRunPrefsFN CB GetAssumPercentFN CB GetCertaintyFN While running in Extreme speed all values for these functions return VALUE At the end of the simulation Crystal Ball performs a final recalculation on the model so these functions are evaluated properly Normally this should not present a problem unless one of these functions was defined as a forecast and you are expecting a valid value to be computed during the simulation If one of these Get functions feeds into a forecast during a simulation this is not compatible with Extreme speed Example message Unsupported Excel or Crystal Ball function at cell location Book1 xls Sheet1 A1 Code 5539 e Workaround Defining forecasts on statistical functions that are dependent on other forecasts is generally not a good modeling practice If you need to have a forecast cell defined on a statistical result from another forecast use the Auto Extract feature for the dependent forecast instead of using one of the above Crystal Ball functions User defined functions Calls to user def
304. one dimensional simulation for characterizing risk The Two Dimensional Simulation tool runs an outer loop to simulate the uncertainty values and then freezes the uncertainty values while it runs an inner loop of the whole model to simulate the variability This process repeats for some number of outer simulations providing a portrait of how the forecast distribution varies due to the uncertainty The primary output of this process is a chart depicting a series of cumulative frequency distributions You can interpret this chart as the range of possible risk curves associated with a population CB Tools Note When using this tool set the Seed Value option in the Crystal Ball Run Preferences dialog so that the resulting simulations are more comparable Crystal Ball User Manual Two Dimensional Simulation tool Two Dimensional Simulation example In the Crystal Ball Examples folder there is a Toxic Waste Site xls workbook you can use to experiment with the Two Dimensional Simulation tool This spreadsheet model predicts the cancer risk to the population from a toxic waste site This spreadsheet has two variability assumptions and two uncertainty assumptions To run the Two Dimensional Simulation tool 1 In Excel with Crystal Ball loaded open the workbook Toxic Waste Site xls Ep 2 Inthe Run gt Run Preferences gt Sampling dialog set Random Number Generation to use the Same Sequence Of Random Numbers and An Initial See
305. ons describe distinct values usually integers with no intermediate values and are shown as a series of vertical columns such as the binomial distribution at the bottom of Figure A 4 on page 248 A discrete distribution for example might describe the number of heads in four flips of a coin as 0 1 2 3 or 4 Continuous probability distributions such as the normal distribution describe values over a range or scale and are shown as solid figures in the Distribution Gallery Continuous distributions are actually mathematical abstractions because they assume the existence of every possible intermediate value between two numbers That is a continuous distribution assumes there is an infinite number of values between any two points in the distribution However in many situations you can effectively use a continuous distribution to approximate a discrete distribution even though the continuous model does not necessarily describe the situation exactly In the dialogs for the discrete distributions Crystal Ball displays the values of the variable on the horizontal axis and the associated probabilities on the vertical axis For the continuous distributions Crystal Ball does not display values on the vertical axis since in this case probability can only be associated with areas under the curve and not with single values For more information on the separate probability distributions and how to select them see these sections e Continuous di
306. ont of each item to display When these settings are complete click OK to close the dialog and return to the Create Report dialog In the Create Report dialog click the Options tab to display it and continue with step 5 on page 184 When all the report options are set click OK to close the dialog and create the report Report sections The Crystal Ball Custom report offers the following sections You can include any or all of them Report sections Report Summary Forecasts Assumptions Decision Variables Overlay Charts Trend Charts Sensitivity Charts Figure 7 5 Custom report sections Table 7 1 Create Report dialog Report Sections settings Setting Effect Report Summary Includes the report title date and time Run Preferences settings and run statistics Forecasts Includes various types of forecast information including the name charts percentiles statistics and more Assumptions Includes various types of assumption information including parameters charts percentiles statistics and correlations 186 Crystal Ball User Manual Creating reports Table 7 1 Create Report dialog Report Sections settings Continued Setting Effect Decision Variables Includes various types of decision variable information including type continuous or discrete with step if discrete plus lower and upper bounds Overlay Charts Includes overlay charts in the report You can scale
307. ormat T Include cell locations Figure 7 8 Extract Data dialog Options tab In the Location area To extract data to a new workbook choose New Workbook To extract data to a new worksheet in the active workbook choose Current Workbook gt New Sheet To extract data to the current sheet choose Current Workbook gt Current Sheet Specify the name of the sheet and the first cell of the range where the extracted data will be stored Check settings in the Formatting group to indicate how you want to format the extracted data When checked Include Labels adds row and column headers to the data table Otherwise just the numeric values are extracted When checked Include Cell Locations adds the workbook worksheet and cell address above the object name in the column header Otherwise only the object name appears Book1 D E Cell Location labels Sheet1 A2 Statistics A2 Trials 1000 Mean 0 02 Crystal Ball User Manual 193 Chapter 7 Creating Reports and Extracting Data 194 When checked AutoFormat applies the following formats to extracted data Bold font for column headers Border beside row labels Border beneath column headers Border before first assumption Numeric formatting to values AutoFit width to columns 9 Ifyou want you can click the Defaults button at any time to restore the original settings to both tabs of the Extract Data dialog 10 When both the Data and Options tab set
308. orphic Spreadsheet Interpreter for running simulations on workbooks This technology was developed by Frontline Systems the maker of Excel s Solver add in PSI Technology supports nearly all of the 320 standard Excel functions including the financial statistical and engineering functions that are part of the Analysis Toolpak By default Crystal Ball is set to use Extreme speed when initially loaded Ifa model is not compatible with Extreme speed a dialog offers the opportunity to downshift to Normal speed temporarily for that simulation Compatibility issues below explains conditions that cause a model to be incompatible with Extreme speed The simulation speed can be changed using the Speed tab in the Run Preferences dialogs see Speed preferences on page 84 for details Compatibility issues While Extreme speed can greatly decrease simulation run times not all models are compatible with Extreme speed When you start a simulation Crystal Ball detects if your spreadsheet is compatible with Extreme speed and warns you if it is incompatible If you choose you can run the simulation in Normal speed using standard Excel or you can change your spreadsheet model to correct the incompatibility This section lists functions and formula constructs that are not compatible with Extreme speed and suggests workarounds Besides the issues listed below some differences may also exist due to undocumented Excel behaviors changes in the
309. ossen iieii ai SE ERE i iS 272 Triangular distribution oo eee eee cece eee ceeceeeee iS inana AE 274 Unitorin distributions s cciscissmusivonscnconssoessonssdincsvecessebeavediuvesesestaseneass 276 Weibull distribution also Rayleigh distribution 0 0 cee 278 Using discrete distributions i cectecesseccicnecsiiseeesiesaeaineveseaseenastoiedeedssseasiesseass 280 Binomial distributii sc cdcceersisessenoenssnsaserdverasvaucesoeatedacanvdeveanvset nerves 281 Discrete uniform distribution 20 0 0 ccc ee eee cee eeeeeeeeseseeeceeeeteeeeeees 283 Geometric GISHIDUTION itis eskcdasesseiedes need anieendaceapstinnessnaatbeneesnanaaaees 285 Hypergeometric Cistribution 0 0 cece cceeeeeeceeeeteeececneeeseessenee 286 Negative binomial distribution 0 0c cece cece cess eeeesee cee eeeeeseeenaeens 289 Poisson distributo s icei cdeoi sestcondiestbaadeaendavcansadsatasedbesdtadseddcadeandeaasneaales 291 Y 682110 CUSCrIDUU ON ves css et eccalecsosuesossans innnan ennaa 293 Using the custom distribution cece cece ce eres eeee cee ceeeeeeesecseeeeeenees 294 Custom distributiom sisas cc selsaviesadh autedeaestennapsdadesiadanenosnagsbiaseeduieaes 294 Example ONG sinasin onnenn ii E AAE Ea A RR 295 EXAMPLE CWO i ciees sectevis ines ase caettecncsaseuees ussen eenean d asean eSlas EaR 298 Example three s csscit ississacssssasteaiddetaosiieteonsdeauiid Gans AEAEE 302 Entering tables of data into custom distributions 0 0ccee eee
310. ou can use the Chart Type drop down list to choose one of these chart types Table 6 2 Sensitivity chart types Bar directional The default horizontal f bars to the right and left of sa 255 the 0 line showing magnitude and direction of sensitivity Bar magnitude Horizontal bars to the right of the 0 line showing faa jal 55 1 de magnitude of sensitivity but not direction Pie A circle divided into proportional wedges 2 0 showing the magnitude of ae sensitivity only available 25 5 for Contribution To 65 2 Variance view For bar charts choose whether to use a different color for each assumption the default or whether to use the same color for all assumptions Crystal Ball User Manual 177 Chapter 6 Analyzing Other Charts 178 If you uncheck Show Multiple Colors you can choose a specific color to use for all assumptions 6 Choose whether to show value labels on the chart the default or uncheck Show Values On Chart to show only graphics but no values 7 Optional step You can choose Defaults at any time to restore all settings to their original default values 8 Optional step If you want to apply the settings to more than one chart click Apply To Then specify how they should be applied see page 139 for details and click OK Otherwise go to step 9 9 Click OK to apply the settings on all tabs to the active chart You can apply
311. p displays Crystal Ball online help To hide or display the Crystal Ball toolbar for the current session choose View gt Toolbars gt Crystal Ball 7 16 Crystal Ball User Manual Defining Model Assumptions In this chapter e Overview Defining assumptions e Additional assumption features e Additional Distribution Gallery features This chapter provides step by step instructions for setting up assumption cells in Crystal Ball models so simulations can be run against them This chapter also describes all the ways you can use the Distribution Gallery to organize your favorite distributions and define categories of distributions to share with others The next chapter describes how to define decision variable and forecast cells and to cut copy and paste data If you are a new user you should start by working through the tutorials in the Crystal Ball Getting Started Guide and then read this chapter After you complete this chapter and Chapter 3 of this User Manual read Chapter 4 for information on setting preferences and running simulations Crystal Ball User Manual 17 Chapter 2 Defining Model Assumptions Overview Crystal Ball lets you define three types of cells e Assumption cells contain the values that you are unsure of the uncertain independent variables in the problem you are trying to solve The assumption cells must contain simple numeric values not formulas or text Decision variable cells contain th
312. pare the quality of the fits or to view the goodness of fit statistics Use the Comparison Chart features as described below e Click the Next and Previous buttons to scroll through the fitted distributions You can view the quality of each fit graphically and statistically in decreasing order e Use the View menu to change the chart view Choose Goodness of Fit to display results of the different goodness of fit tests for each distribution type Crystal Ball User Manual Additional assumption features Edit View Preferences Help 500 Values Gamma Max Extreme Logistic Normal Student s t Lognormal Min Extreme Triangular Exponential Pareto Uniform lt lt Previous Goodness of Fit View 494 Displayed Parameters 11 19 02 Location 11 43 Scale 394 59 Shape 1 9336 26 26 Lean EES Minimum 58 33 Maximum 2203 96 Alpha 3 1 5 2 30 2 89 4 18 5 92 16 11 27 59 49 21 130 43 131 75 18 76 30 11 56 42 50 26 80 71 61 88 164 66 113 89 251 61 1287 10 386 95 04 Location 58 37 Scale 88 59 Shape 4 73836 04 Likeliest 272 98 Scale 155 80 05 Mean 349 48 Scale 107 63 05 Mean 361 39 Std Dev 188 57 09 Midpoint 344 52 Scale 160 65 Deg Freedo 09 Mean 372 73 Std Dev 256 35 13 Likeliest 460 61 Scale 216 89 19 Minimum 10 88 Likeliest 137 44 Maximum 22 Rate 0 00 39 Location 25 51 Shape 0 40194 39 Minimum 23 40 Maximum 1 148 07 Figure 2 8 The Comparison Chart Goodness of Fit v
313. pendix e Simulation accuracy e Simulation speed e Sample size Correlated assumptions This chapter contains information that you can use to improve the overall performance of Crystal Ball These improvements occur in terms of the accuracy of your model or speed of the results Crystal Ball User Manual 317 Appendix B Maximizing Your Use of Crystal Ball Simulation accuracy 318 The accuracy of your simulation is primarily governed by two factors e The number of trials or length of the simulation The sampling method Generally speaking the more trials you run in a simulation the greater the accuracy of the statistics and percentiles information This greater accuracy comes at the expense of lengthier simulation times and higher memory usages see later sections on simulation speed and memory usage Also for a given number of trials the accuracy of the statistics and percentiles greatly depends on the shape and nature of the forecast distribution If you re not sure how many trials to run for a specific level of accuracy you can use Crystal Ball s precision control feature described below to automatically determine the appropriate number of trials to run For a detailed picture of simulation s statistical accuracy or simply if you re curious you can run the Bootstrap tool under the Run menu to generate a forecast chart for each statistic or percentile of interest The sampling method is the other primary
314. ple describes a special feature on the Custom Distribution dialog the Load Data button which lets you pull in numbers from a specified cell range grouped data on the worksheet This example is not a hands on exercise but the illustrations will guide you through the procedure After you read this section you can experiment with your own data by pulling in numbers from specified cell ranges on your worksheet Crystal Ball User Manual M Using the custom distribution In this example the same company decides that the unit cost of the new product can vary widely The company feels it has a 20 chance of being any number between 10 and 20 a 10 chance of being any number between 20 and 30 a 30 chance of being any number between 40 and 50 a 30 chance of being a whole dollar amount between 60 and 80 and there is a 5 chance the value will be either 90 or 100 All the values have been entered on the worksheet in this order range minimum value range maximum value for all but Single Value ranges total probability and step for the Discrete Range only Minimum Maximum Prob 10 20 0 2 Continuous Range 20 30 0 1 Continuous Range 40 50 0 3 Continuous Range 560 80 0 3 1 Discrete Range 90 0 05 Single Value 100 0 05 Single Value Figure A 35 Four column custom data range In this case discrete ranges have the most parameters So you can create an assumption choose Custom Distribution and then choose Paramete
315. poration Chart FX Chart FX is a registered trademark of Software FX Inc Q Powered b f PSI Technology is a registered trademark of Frontline Systems Inc Other product names mentioned herein may be trademarks and or registered trademarks of the respective holders MAN CBUM 070202 2 6 15 06 Contents Welcome to Crystal Ball Who should use this program cccecceeceeeceeesceseeseecseececeseeseesseeeeenees 1 What you Will Need sessist a AS 1 About the Crystal Ball documentation set 0 0 0 ccc cesses teerseeneeees 2 Conventions used in this manual cccccccessesseesesceceecececeeceaaeaaeseseeeeeeeeeeees 4 Getting hel piciascisieciatisendctsd let dg a R A A 5 Additonal TESGULCES vesicsesssecowsssdeubecesavcedarsasusasendversendvesepodedeougecussavavevedseseawesaces 5 Technical SUpport ronnen iaa EAE A 5 Training isos se socinethettucagyetersdonsedtaysdesetdeds sanusextatunsensvad coeglseedusssnrectadta dest saces 5 COTSULEIIN Secale nents egosisaseotusasiects caghaas detassastevsnesendSynivaesaalsaadors sate tee abet 6 Chapter 1 Crystal Ball Overview Model building and risk analysis OVErVieW cccceeseeee cee eee eeee cee ceeeeeeeeneee 8 Risk and risk analysisciseanie ap n EN A 8 Whatisamodel dsr AeereienncvsinaresDtesu aE A E a A O 9 Traditional spreadsheet analysis ccccccccesceceteese ee eeeeeeeeceeeeseneeseens 10 Monte Carlo simulation and Crystal Ball 0 cccccececete cece e
316. ports eee ee cece eens eee eeeecneeeeeeteee 351 Capability metrics list cice cctsasiccsssecidascescadedeetten idevsabeanaaseaateatincnnescensscoantet 353 Elele LET e h A 357 Glossa eere a E A E EE EA 363 DMR SH a cata E E 373 Crystal Ball User Manual Welcome to Crystal Ball Crystal Ball is a user friendly graphically oriented forecasting and risk analysis program that takes the uncertainty out of decision making Through the power of simulation Crystal Ball becomes an effective tool in the hands of the decision maker You can answer questions such as Will we stay under budget if we build this facility or What are the chances this project will finish on time or How likely are we to achieve this level of profitability With Crystal Ball you will become a more confident efficient and accurate decision maker Crystal Ball is easy to learn and easy to use Unlike other forecasting and risk analysis programs you do not have to learn unfamiliar formats or special modeling languages To get started all you have to do is create a spreadsheet From there this manual guides you step by step explaining Crystal Ball terms procedures and results And you do get results from Crystal Ball Through a technique known as Monte Carlo simulation Crystal Ball forecasts the entire range of results possible for a given situation It also shows you confidence levels so you will know the likelihood of any specific
317. ports with charts and data page 182 You can quickly produce pre defined or custom reports for analysis and presentation Extract data for numeric analysis and presentation You can extract simulation results to Excel for further analysis or for further export into other analytical tools See Extracting data on page 190 for details Use other Crystal Ball tools for different types of analyses Depending on the types of problems you are analyzing the Crystal Ball Batch Fit Bootstrap Correlation Matrix Decision Table Scenario Analysis Tornado Chart and 2D Simulation tools offer addition analysis techniques These tools are described in Chapter 8 If you have Crystal Ball Professional or Premium Edition you can also use OptQuest and CB Predictor for optimization and time series analysis Crystal Ball User Manual 105 Chapter 5 Analyzing Forecast Charts Understanding and using forecast charts 106 As discussed in the section About Crystal Ball simulations on page 78 forecast charts use frequency distributions to show the number frequency of values occurring in a given interval The highest value on the frequency scale to the right of the chart is the frequency for the interval that contains the greatest number of forecast values the mode of the graphed distribution You can estimate the frequency for other points on the forecast chart using this frequency scale Crystal Ball Note You can now display fore
318. ptions dialog For more information see Statistics preferences on page 87 190 Crystal Ball User Manual Extracting data Table 7 3 Extract Data dialog Select Data To Extract settings Continued Option Extracts Chart Bins For each group interval or bin the interval range as well as the probability and frequency of occurrences within the interval for the forecast This setting is independent of the Chart Preferences density setting that controls how many bins or data points appear in a graphic chart Sensitivities Sensitivity data such as the rank correlation coefficient for all pairs of assumptions and forecasts indicating the strength of the relationship If you plan to extract sensitivity data be sure to check Store Assumption Values For Sensitivity Analysis on the Options tab of the Run Preferences dialog page 86 Note Data is extracted for all assumptions regardless of assumptions selected for extraction Trial Values The generated assumption and forecast values for each simulation trial Crystal Ball Note Ifyou have activated the process capability features and have generated capability metrics you can extract them For details see Extracting capability metrics on page 349 If you choose Percentiles or Chart Bins a dialog appears so you can choose which percentiles or how many bins you want to use For more information click the Help button in the dialog You can
319. r for example 10th percentile likeliest and 99th percentile To select a parameter set to use as the default when defining new assumptions of this type choose Set Default from the Parameters menu Crystal Ball User Manual Additional assumption features Several special parameter sets are available with the lognormal distribution including geometric and logarithmic sets For more information see the Equations and Methods chapter in the online Crystal Ball Reference Manual Fitting distributions to data Glossary Term goodness of fit A set of mathematical tests performed to find the best fit between a standard probability distribution and a data set s distribution If you have historical data available Crystal Ball s distribution fitting feature can substantially simplify the process of selecting a probability distribution Not only is the process simplified but the resulting distribution more accurately reflects the nature of your data than if the shape and parameters of the distribution were estimated How distribution fitting works In distribution fitting Crystal Ball automatically matches your data against each continuous probability distribution A mathematical fit is performed to determine the set of parameters for each distribution that best describe the characteristics of your data The quality or goodness of each fit is judged using one of several standard goodness of fit tests The distribution w
320. r details see Viewing LSL USL and Target marker lines on page 347 Base Case is the value in an assumption decision variable or forecast cell prior to running the simulation For forecasts Certainty Range shows lines at the certainty range endpoints Marker lines are shown with labels such as Mean 125 Crystal Ball Note You can press Ctrl m to cycle through the median mean and base case or mode depending on chart type Press Ctrl p to cycle through every 10th percentile To display marker lines 1 Display the Chart Type tab of the Chart Preferences dialog The Marker Lines group appears in the lower right corner of the Chart Type page If more than one series appears in the list box at the top of the tab select one to work with The settings on the page apply to the selected series 2 To display a marker line check the box for that item If you choose Standard Deviation Percentile or Value another dialog appears e For Standard Deviation enter the standard deviation s where you want a marker to appear If you enter more than one separate them with commas Then choose whether you want the marker s to appear below the mean technically indicating negative standard deviations above the mean or both above and below e For Percentile select the group of percentiles where you want markers to appear or choose Custom and create your own group of percentile points separated by commas For Value enter
321. r greater than r Description The negative binomial distribution is useful for modeling the distribution of the number of trials until the r th successful occurrence such as the number of sales calls you need to make to close a total of 10 orders It is essentially a super distribution of the geometric distribution Example A manufacturer of jet engine turbines has an order to produce 50 turbines Since about 20 of the turbines do not make it past the high velocity spin test the manufacturer will actually have to produce more than 50 turbines Matching these conditions with the negative binomial distribution The number of turbines to produce trials is not fixed Crystal Ball User Manual 289 Appendix A Selecting and Using Probability Distributions The manufacturer will continue to produce turbines until the 50th one has passed the spin test e The probability of success 80 is the same for each test These conditions match those of the negative binomial distribution The negative binomial distribution has two parameters Probability and Shape The Shape parameter specifies the r th successful occurrence In this example you would enter 0 8 for the Probability parameter 80 success rate of the spin test and 50 for the Shape parameter Figure A 23 Coeine Assumpvor coal Lo Edit View Parameters Preferences Help Name fan x Negative Binomial Distribution SI S S S So o0 Bie bebe bon bon 668 7 6
322. ranges from 25 500 to 64 000 dollars or from 25 5 to 64 0 million dollars You might require only that the simulation s estimate of the mean fall within plus or minus 5 of itself Crystal Ball User Manual 319 Appendix B Maximizing Your Use of Crystal Ball 320 You might encounter the drawback of using relative precision when the forecast statistic is close to zero In the Net Profit example below also from the Vision Research model the forecast s distribution straddles the break even point of zero A relative precision of 5 of the mean or roughly 0 5 million results in a very small confidence interval relative to the full range width of 49 1 million that might take an unexpectedly large number of trials to satisfy Forecast Net Profit MM E iol x Edit View forecast Preferences Help 500 Trials Frequency View 498 Displayed Net Profit MM T wo N T ho oO T T Y N NR a Aouenbel4 2 k amp 2 o o D a E 1 1 1 T N o o d 10 0 0 0 10 0 20 0 30 0 Millions gt iim Certainty 100 q intinity Figure B 2 Absolute precision control example Finally Crystal Ball combines the individual forecast precision options page 64 with the confidence level value found in the Run Preferences gt Trials dialog page 81 to calculate confidence intervals Generally it is a good idea to leave this value at 95 or 90 so that you can have a
323. rap tool e Decision Table tool e Scenario Analysis tool e Two Dimensional Simulation tool For each tool there is a general description an introduction tutorial and references to descriptions of all dialogs fields and options Crystal Ball User Manual 197 Chapter 8 Crystal Ball Tools Overview Crystal Ball tools are Visual Basic programs in the Run gt Tools menu that extend the functionality of Crystal Ball They cover two aspects of Crystal Ball modeling setup and analysis Table 8 1 Setup tools Tool Action Batch Fit Automatically fits selected continuous probability distributions to multiple data series Correlation Matrix Rapidly defines and automates correlations of assumptions Tornado Chart Individually analyzes the impact of each model variable on a target outcome Table 8 2 Analysis tools Tool Action Bootstrap Addresses the reliability and accuracy of forecast statistics Decision Table Evaluates the effects of alternate decisions in a simulation model Scenario Analysis Displays what inputs created particular outputs Two Dimensional Simulation Independently addresses uncertainty and variability using a two dimensional simulation This chapter describes each tool and provides a step by step example for using it CB Tools Note Some of the tools can take a long time to run Should you need to cancel a tool before it stops by itself click in the Excel win
324. receive 10 over production costs or a minimum of 5 per unit to make the manufacturing effort worthwhile He also wants to set the maximum price for the product at 15 per unit so that he can gain a sales advantage by offering the product for less than his nearest competitor All values between 5 and 15 per unit have the same likelihood of being the actual product price however he wants to limit the price to whole dollars The first step in selecting a probability distribution is matching your data with a distribution s conditions Checking the uniform distribution The minimum value is 5 per unit The maximum value is 15 per unit All integer values between 5 and 15 are equally possible You would enter these values in Crystal Ball to produce a discrete uniform distribution showing that all whole dollar values from 5 to 15 occur with equal likelihood Figure A 20 illustrates this scenario 284 Crystal Ball User Manual Geometric Using discrete distributions Geometric distribution Parameter Probability Conditions The number of trials is not fixed The trials continue until the first success The probability of success is the same from trial to trial Description The geometric distribution describes the number of trials until the first successful occurrence such as the number of times you need to spin a roulette wheel before you win Example one If you are drilling for oil and want
325. reting the results In this example the simulation generated 1 000 forecast values Since you selected to analyze the percentiles from 95 to 100 the resulting table lists 51 forecast values or the top 5 of the forecast range including the endpoints The table sorts the values from lowest to highest value and also includes the assumption values that Crystal Ball generated for each trial Crystal Ball Note If you have the Probability Above A Value option selected in the Run Preferences gt Options dialog the percentiles will be reversed in meaning so that the Ist percentile represents the uppermost 1 and the 99th percentile represents the lowest 1 One way to analyze the Scenario Analysis results is to identify a particular forecast value and see what assumption values created that forecast value Another way to analyze the Scenario Analysis results is to generate an Excel chart with the data For example you might create a scatter chart comparing the risk assessment with the CPF as shown below Risk Assessment Volume of Water per Day 6 00 4 00 2 00 0 00 0 00E 00 200E 04 400E 04 6 00E 04 Figure 8 38 Scatter chart of Risk Assessment and CPF Crystal Ball User Manual 235 Chapter 8 Crystal Ball Tools Two Dimensional Simulation tool 236 Risk analysts must often consider two sources of variation in their models Table 8 7 Sources of model variation uncertainty Assumptions that are unce
326. ribution on page 268 The uses are the same except that the sample degrees of freedom will be lt 30 for the Student s t distribution Crystal Ball User Manual 273 Appendix A Selecting and Using Probability Distributions 274 Triangular distribution A Triangular Parameters Minimum Likeliest Maximum Conditions The minimum number of items is fixed The maximum number of items is fixed The most likely number of items falls between the minimum and maximum values forming a triangular shaped distribution which shows that values near the minimum and maximum are less likely to occur than those near the most likely value Description The triangular distribution describes a situation where you know the minimum maximum and most likely values to occur For example you could describe the number of cars sold per week when past sales show the minimum maximum and usual number of cars sold Example one An owner needs to describe the amount of gasoline sold per week by his filling station Past sales records show that a minimum of 3 000 gallons to a maximum of 7 000 gallons are sold per week with most weeks showing sales of 5 000 gallons The first step in selecting a probability distribution is matching your data with a distribution s conditions Checking the triangular distribution The minimum number of gallons is 3 000 e The maximum number of gallons is 7 000 The most likely number of gallons 5
327. ribution showing the probability of x number of items in inventory Crystal Ball User Manual 275 Appendix A Selecting and Using Probability Distributions 276 Uniform Uniform distribution Parameters Minimum Maximum Conditions The minimum value is fixed The maximum value is fixed All values between the minimum and maximum occur with equal likelihood Description In the uniform distribution all values between the minimum and maximum occur with equal likelihood Example one An investment company interested in purchasing a parcel of prime commercial real estate wants to describe the appraised value of the property The company expects an appraisal of at least 500 000 but not more than 900 000 They believe that all values between 500 000 and 900 000 have the same likelihood of being the actual appraised value The first step in selecting a probability distribution is matching your data with a distribution s conditions In this case The minimum value is 500 000 The maximum value is 900 000 All values between 500 000 and 900 000 are equally possible These conditions match those of the uniform distribution The uniform distribution has two parameters the Minimum 500 000 and the Maximum 900 000 You would enter these values as the parameters of the uniform distribution in Crystal Ball Crystal Ball User Manual Using continuous distributions Define Assumption CeMAZ Lo Edit Vie
328. rrors checkbox is checked before you use the Scenario Analysis tool Because of the way it creates workbooks while it is running the Scenario Analysis tool cannot be run in Extreme speed When this tool runs the simulation part of the tool runs in Normal speed even if the Run Mode preference is set to Extreme Speed Scenario Analysis example In the Crystal Ball Examples folder there is a Toxic Waste Site xls workbook you can use to experiment with the Scenario Analysis tool This spreadsheet model predicts the cancer risk to the population from a toxic waste site This spreadsheet has four assumptions and one forecast Crystal Ball User Manual 231 Chapter 8 Crystal Ball Tools E Toxic Waste Site ods i 0 x Risk Assessment at a Toxic Waste Site iil Body Weight kilograms Volume of Water per Day liters day Concentration of Contaminant micrograms iter Cancer Potency Factor inverse mg kg d Risk Assessment iption _ Model Figure 8 34 Toxic Waste Site spreadsheet To run the Scenario Analysis tool 1 In Excel with Crystal Ball loaded open the workbook Toxic Waste Site xls 2 Select Run gt Tools gt Scenario Analysis K Crystal Ball Note Ifthe Scenario Analysis command is not available reset the simulation and try again The Specify Target dialog appears as shown in Figure 8 35 Specify target step 1 of 2 Scenario Analysis Crystal Ball Tool Decisioneering 1998 2004 Determi
329. rs Initial seed value 999 m Sampling method Monte Carlo more random Latin Hypercube more even Sample size 500 Figure 4 3 Sampling tab Run Preferences dialog The Sampling tab sets the sampling seed value sampling method and sample size For general instructions see Setting run preferences beginning on page 80 Table 4 2 describes the settings available on this tab Table 4 2 Sampling tab Run Preferences dialog Setting Effect Use Same Sets the random number generator to generate the same set of random Sequence Of numbers for assumptions letting you repeat simulation results Random Numbers When you select this option enter an integer seed value in the Initial Seed Value field Initial Seed Value Sampling Method Determines the first number in the sequence of random numbers generated for the assumption cells integer Crystal Ball Note To reproduce the sample results shown in this manual check Use Same Sequence and use a seed value of 999 Indicates whether to use Monte Carlo or Latin hypercube simulation Latin hypercube sampling generates values more evenly and consistently across the distribution but requires more memory For more information see page 321 Crystal Ball User Manual 83 Chapter 4 Running Simulations Table 4 2 Sampling tab Run Preferences dialog Continued Setting Effect Sample Size For Latin hypercube sampling divides eac
330. rs gt Discrete Ranges before loading the data Crystal Ball Note If your data also included discrete sloping ranges you could choose Parameters gt Sloping Ranges before loading the data The data table would then have five columns and could accommodate all data types Once the Parameters setting has been made you can follow these steps to complete the data load 1 Click the More button to the right of the Name field The Custom Distribution dialog expands to include a data table as shown in Figure A 36 Crystal Ball User Manual 303 Appendix A Selecting and Using Probability Distributions 304 Q Define Assumption Cell C14 2 El mi x Edit View Parameters Preferences Help Name c14 xj No Data Available Minimum Maximum Probability Step BIDR i i OK Cancel Enter Gallery Correlate Help Figure A 36 Custom distribution with data table A column appears for each parameter in the current set selected using the Parameters menu Parameters gt Discrete Ranges was set before viewing the data table so there is a column in the data table for each discrete range parameter Because the single value and continuous ranges have subsets of the same group of parameters their parameters will also fit into the table Since the values are already on the worksheet you can click Load Data to enter them into the Custom Distribution dialog The Load Data dialog appears as shown in Figur
331. rt Type tab Chart Preferences dialog for trend charts 2 To change all certainty bands from areas to lines choose Line in the Chart Type drop down list Crystal Ball User Manual 161 Chapter 6 Analyzing Other Charts 162 To change the color of a certainty band a Select the certainty band to change from the list below the Certainty Bands button b Select a color from the Band Color list To select a different set of certainty levels or define your own click the Certainty Bands button and follow the steps in the next section Setting certainty bands Optional step You can choose Defaults at any time to restore all settings to their original default values Optional step If you want to apply the settings to more than one chart click Apply To Then specify how they should be applied see page 139 for details and click OK Otherwise go to step 7 When settings are complete click OK to accept current settings on all tabs of the Chart Preferences dialog Crystal Ball Note The Chart Series list at the top of the Chart Type tab lets you change the order of forecasts in the forecast axis For instructions see Adding removing and ordering forecasts on page 159 Setting certainty bands You can easily choose a set of certainty bands levels to display in the trend chart or create your own custom set To change or define a set of certainty bands 1 2 3 4 In the trend chart menubar choose Pre
332. rtain because you have insufficient information about a true but unknown value Examples of uncertainty include the reserve size of an oil field and the prime interest rate in 12 months You can describe an uncertainty assumption with a probability distribution Theoretically you can eliminate uncertainty by gathering more information Practically information can be missing because you haven t gathered it or because it is too costly to gather variability Assumptions that change because they describe a population with different values Examples of variability include the individual body weights in a population or the daily number of products sold over a year You can describe a variability assumption with a discrete distribution or approximate one with a continuous distribution Variability is inherent in the system and you cannot eliminate it by gathering more information For many types of risk assessments it is important to clearly distinguish between uncertainty and variability see Hoffman and Hammonds reference page 361 Separating these concepts in a simulation lets you more accurately detect the variation in a forecast due to lack of knowledge and the variation caused by natural variability in a measurement or population In the same way that a one dimensional simulation is generally better than single point estimates for showing the true probability of risk a two dimensional simulation is generally better than a
333. rtainty level of 100 since the default certainty range includes all possible values By default Crystal Ball calculates the certainty level based on the entire range of forecast values On either side of the Certainty field the certainty minimum and maximum appear The certainty level is one of Crystal Ball s key statistics because it shows the probability of achieving the values within a specific range With Crystal Ball you can determine the certainty level for specific value ranges either by moving the certainty grabbers on the forecast chart or typing the certainty minimum and maximum in the fields You can also type a certainty level in the certainty field to get a certainty range centered around the median Crystal Ball User Manual 107 Chapter 5 Analyzing Forecast Charts 108 When you move the certainty grabbers the certainty range changes and Crystal Ball recalculates the certainty level When you type minimum and maximum values Crystal Ball moves the certainty grabbers for you and recalculates the certainty level When you type the certainty level in the certainty field Crystal Ball moves the certainty grabbers to show you the value range for the certainty level you specified Moving certainty grabbers To determine the certainty level for a specific value range using certainty grabbers 1 Make sure the forecast chart you want to use is the active selected window 2 Move the certainty grabbers on the forecast c
334. s Crystal Ball Note Unlike some other preferences cell preferences must be applied to all cells of the selected type s on the chosen worksheet s or workbook s If necessary you can click the Defaults button before choosing Apply To to clear your cell preference settings and restore the original defaults Saving and restoring your models 76 The distributions you specified for each assumption cell the settings you specified for each forecast cell and the range information for each decision variable cell are saved with their spreadsheet through the Excel save process When you open the spreadsheet again Crystal Ball retains your assumption forecast and decision variable cells Crystal Ball Note When you run simulations as described in the next chapter you can also save and restore simulation results in a separate file for future display and analysis For details see page 97 Crystal Ball User Manual Running Simulations In this chapter e About Crystal Ball simulations Setting run preferences Running simulations Managing chart windows e Saving and restoring simulation results Running user defined macros Chapters 2 and 3 describe how to define a spreadsheet model in Crystal Ball This chapter provides step by step instructions for setting run preferences and running a simulation in Crystal Ball After you complete this chapter read Chapter 5 for information about analyzing simulation result
335. s Setting chart colors page 136 Show mean median mode standard deviation percentile or capability limit target marker lines Showing the mean and other marker lines page 137 Hide and show vertical and horizontal axes create and edit axis labels and change an axis scale Customizing chart axes and axis labels page 138 Format chart numbers Formatting chart numbers on page 113 Specify whether to use these preferences for more than the current chart Applying settings to the current chart and other charts page 139 128 Crystal Ball User Manual Setting chart preferences Setting preferences with hot keys Table 5 5 lists key combinations that can be used to cycle through settings available in the Chart Preferences dialog Most of these commands work on the primary distribution the probability distribution for assumptions and frequency for forecasts and overlay charts Table 5 5 Hot keys for chart preferences Hot Key Command Equivalent Description Ctrl d Ctrl b Ctrl g Ctrl l Ctrl t Ctrl 3 Ctrl m Ctrl n Ctrl p Spacebar View menu Preferences gt lt chartname gt Preferences gt View Preferences gt Chart Preferences gt General gt Density Preferences gt Chart Preferences gt General gt Gridlines Preferences gt Chart Preferences gt Chart Type gt Type Preferences gt Chart Preferences
336. s 1 In the trend chart menubar choose Preferences gt Chart gt Axis The Axis tab of the Chart Preferences dialog appears as shown in Figure 6 16 Crystal Ball User Manual 163 Chapter 6 Analyzing Other Charts Chart Preferences Si x General Chart Type Aus m Axis label T Auto Scale __ Format number Type Auto v Format Cellfomat Min Infinity Decimal pe Max infinity V7 Sand separator IV Round axis values Figure 6 16 Trend chart Axis preferences 2 By default no name displays for the value axis To add one type it in the Axis Label field 3 By default Scale is set to Auto and displays all selected bands completely To limit the display to a subset of values set Scale to Fixed and enter a minimum and maximum value By changing the minimum or maximum endpoint values you can zoom in or out on selected ranges of the trend chart 4 The Format settings are similar to those for forecast and other charts For more information see Customizing chart axes and axis labels beginning on page 138 Crystal Ball Note The number format for the axis values is taken from the first forecast that appears on the trend chart 5 Optional step You can choose Defaults at any time to restore all settings to their original default values 6 Optional step If you want to apply the settings to more than one chart c
337. s Crystal Ball User Manual 301 Appendix A Selecting and Using Probability Distributions 302 G Define Assumption Cell D11 BY lolx Edit View Parameters Preferences Help Name D11 E7 Custom Distribution 0 07 A 0 06 amp 005 004 oo w 0 02 a 0 01 0 00 6 00 8 00 1000 1200 1400 1600 18 00 20 00 Enter one or more continuous or discrete sloping ranges Minimum fie co Maximum 21 00 Min Height fo o5 Max Height o 025 Step fo 50 oK Cancel Enter _Gatery Correlate Help Figure A 34 A sloped discrete range with steps of 5 Although the bars have spaces between them their heights and the width of the range they cover are equal to the previous continuous sloped range and the total probability is the same Crystal Ball Note While a second continuous range could have extended from 15 to 20 the second range in this example starts at 16 rather than 15 to illustrate a discrete range because unlike continuous ranges discrete ranges cannot touch other ranges With Crystal Ball you can enter single values discrete ranges or continuous ranges individually You also can enter any combination of these three types in the same Custom Distribution dialog as long as you follow these guidelines ranges and single values cannot overlap one another however the ending value of one continuous range can be the starting value of another continuous range Example three This exam
338. s Crystal Ball User Manual 77 Chapter 4 Running Simulations About Crystal Ball simulations Glossary Term spreadsheet model Any spreadsheet that represents an actual or hypothetical system or set of relationships Glossary Term simulation Any analytical method that is meant to imitate a real life system especially when other analyses are too mathematically complex or too difficult to reproduce Spreadsheet risk analysis uses both a spreadsheet model and simulation to analyze the effect of varying inputs on outputs of the modeled system After you define assumption forecast and decision variable cells in your spreadsheet model you are ready to run the simulation Crystal Ball uses a technique called Monte Carlo simulation to forecast the entire range of results most likely to occur in the situation you have defined in your spreadsheet model This technique involves generating random numbers for the assumption variables For more information on Monte Carlo simulation see page 11 While the simulation is running Crystal Ball displays these assumption results in a forecast chart that shows the entire range of possible outcomes Crystal Ball provides descriptive statistics for any forecast which summarize the results numerically The Statistical Definitions chapter of the online Crystal Ball Reference Manual discusses descriptive statistics How Crystal Ball uses Monte Carlo simulation Glossary Term
339. s New York John Wiley amp Sons 1986 Morgan Byron J T Elements of Simulation Portland ME Chapman and Hall 1984 Rubinstein Reuven Y Simulation and the Monte Carlo Method New York John Wiley amp Sons 1981 Thommes M C Proper Spreadsheet Design Boston Boyd and Fraser Publishing Co 1992 Thompson K M D E Burmaster and E A C Crouch Monte Carlo Techniques for Quantitative Uncertainty Analysis in Public Health Risk Assessments Risk Analysis 12 1 53 63 1992 Yakowitz Sidney J Computational Probability and Simulation Reading MA Addison Wesley Publishing Co 1977 Probability theory and statistics Alder Henry L and Edward B Roessler Introduction to Probability and Statistics 6th Ed San Francisco W H Freeman amp Co 1977 358 Crystal Ball User Manual Biswas Suddhendu Topics in Statistical Methodology New York John Wiley amp Sons Inc 1991 Fraser D A S Probability and Statistics Theory and Applications North Scituate MA Duxbury Press 1976 Johnson Norman L and Samuel Kotz Distributions in Statistics Continuous Multivariate Distributions New York John Wiley amp Sons Inc 1972 Kapur K C and L R Lamberson Reliability in Engineering Design New York John Wiley amp Sons 1977 Larson Harold J Introduction to Probability Theory and Statistical Inference 2nd Ed New York John Wiley amp Sons 1974 Logothetis N and V Rothschild Probabil
340. s 128 chart preferences hot keys 129 charts closing 142 copying 141 customizing 128 deleting 143 displaying 140 managing 140 overlay with Two dimensional Simulation 241 pasting 141 preferences 128 printing 142 sensitivity 165 spider 213 tornado 211 trend with Two dimensional Simulation 241 chi square distribution 261 chi square test 29 122 chi squared distribution 259 Choose Assumptions dialog 35 Choose Forecasts dialog 192 closing charts 142 comparing distributions 313 comparison chart overview 30 consulting referral service 6 continuing a simulation 90 continuous distributions 249 Control Panel Crystal Ball 93 Control Panel hiding 93 Control Panel showing after reset 93 control precision defining with forecast 59 displayed percentiles 118 displayed statistics 118 increasing accuracy 318 conventions manual 4 copying charts 141 copying distributions 46 correlated assumptions practical limit 323 Correlation chart 38 Correlation Matrix example 207 tool 206 correlations accuracy 318 defined 206 specifying 33 creating distribution categories 50 creating distributions 45 Crystal Ball Monte Carlo simulation 78 system requirements 1 Crystal Ball Control Panel 93 custom distribution 294 important notes 311 customizing charts 128 D data editing 69 grouped 302 data entering custom 306 Decision Table example 227 tool 226 decision variable cells defined 18 decision variables def
341. s chapter have described how to create and customize new charts This section describes how to open copy paste print close and delete existing charts Opening charts Once you create an assumption or forecast chart it is saved with the workbook that contains it Other charts are saved with the active workbook model You can display charts again with current data anytime you rerun the model with associated workbooks open To open a chart 1 Open the model containing the chart and run a simulation or restore saved results as described in Restoring Crystal Ball simulation results on page 98 2 Open the Analyze menu and select the type of chart to display Assumption Charts Forecast Charts Overlay Charts Trend Charts or Sensitivity Charts 3 When the dialog for that chart appears check the box in front of each chart you want to display 4 Click Open Crystal Ball User Manual Crystal Ball Managing existing charts You might need to click the Crystal Ball and Microsoft Excel icons in the Windows task bar to activate charts that have disappeared behind the spreadsheet Crystal Ball Note To open a number of charts at once select the Crystal Ball data cells and choose Analyze gt Open Selected Cells All charts for the selected cells open and appear in front of any other open charts Copying and pasting charts to other applications You can copy and paste assumption forecast overlay trend and se
342. s include the breaking strengths of materials construction design and aircraft loads and tolerances This distribution is also known as the Gumbel distribution and is closely related to the minimum extreme distribution its mirror image Minimum The minimum extreme distribution is commonly used to describe the A extreme smallest value of a response over a period of time for example rainfall mfime page 267 during a drought This distribution is closely related to the maximum extreme distribution Normal The normal distribution is the most important distribution in probability A Basic theory because it describes many natural phenomena such as people s IQs page 268 OF heights Decision makers can use the normal distribution to describe uncertain variables such as the inflation rate or the future price of gasoline 7 Pareto The Pareto distribution is widely used for the investigation of distributions page 270 associated with such empirical phenomena as city population sizes the occurrence of natural resources the size of companies personal incomes stock price fluctuations and error clustering in communication circuits J Student st The Student s t distribution is used to describe small sets of empirical data A page 272 that resemble a normal curve but with thicker tails more outliers For sets SET of data larger than 30 you can use the normal distribution instead Triangular The triangular distribu
343. s obvious what effect changing a decision variable can have on the forecast results The Decision Table tool runs multiple simulations to test different values for one or two decision variables The tool tests values across the range of the decision variables and puts the results in a table that you can analyze using Crystal Ball forecast trend or overlay charts The Decision Table tool is useful for investigating how changes in the values of a few decision variable affect the forecast results For models that contain more than a handful of decision variables or where you are trying to optimize the forecast results use OptQuest for Crystal Ball OptQuest is a wizard based program that enhances Crystal Ball by automatically finding optimal solutions to simulation models This program is available with Crystal Ball Professional and Premium Editions Crystal Ball User Manual Decision Table tool Table 8 6 Comparison of Decision Table tool and OptQuest Decision Table tool OptQuest Process Runs multiple Crystal Ball simulations for different values of decision variables Results All displayed in a table Only the best solutions displayed Optimization No Yes Number of One or two Unlimited variables Variable range Small Small to large Decision Table example In the Crystal Ball Examples folder there is an Oil Field Development xls workbook you can use to experiment with the Decision Table tool This spreadshe
344. s of single values all with different probabilities use a two column format The first column contains single values the second column contains the probability of each value Ee Weight or Probability Saeass Figure A 42 Weighted values loaded in a custom distribution Crystal Ball User Manual 307 Appendix A Selecting and Using Probability Distributions 308 Mixed single values continuous ranges and discrete ranges For any mixture of single values and continuous ranges use a three column format obtained by choosing Parameters gt Continuous ranges The three column format is the same as using the first three columns shown in Figure A 36 Figure A 37 and Figure A 38 beginning on page 304 If the mix includes uniform non sloping discrete ranges use a four column format as in the first four columns of Figure A 43 and Figure A 44 To obtain four columns choose Parameters gt Discrete Ranges Mixed ranges including sloping ranges If sloping ranges are included in a mix of ranges choose Parameters gt Sloping Ranges to display a five column data table The first column contains the range Minimum value the second column contains the range Maximum value the third column contains Height of Min the relative probability height at the Minimum value the fourth column contains Height of Max the relative probability at the Maxim
345. s ranges This section uses real world examples to describe the custom distribution Crystal Ball Note For summaries of the data entry rules used in the examples plus additional rules see Entering tables of data into custom distributions beginning on page 306 and Other important custom distribution notes beginning on page 311 Crystal Ball User Manual Using the custom distribution Since it is easier to understand how the custom distribution works with a hands on example you might want to start Crystal Ball and use it to follow the examples To follow the custom distribution examples first create a new Excel workbook then select cells as specified Example one Before beginning example one open the Custom Distribution dialog as follows 1 Click cell D11 La 2 Select Define gt Define Assumption The Distribution Gallery dialog appears 3 Click the All category to select it 4 Scroll to find the custom distribution then click it 5 Click OK Crystal Ball displays the Define Assumption dialog Define Assumption Cell D11 ioii Edit View Parameters Preferences Help Name Jon x No Data Available Enter one or more values with probabilities Vale O Probabili OK Cancel Enter_ Gallery Correlate Help Figure A 27 Define Assumption dialog for custom distributions Using the custom distribution a company can describe the probable retail cost of a new product The company decides the
346. s the display range end points in terms of standard deviations defines the number of standard deviations worth of value you want to display on each side of the mean and centers values around the mean If you choose to set the display range in terms of standard deviations you could change the display range to standard deviation from the mean to look at approximately 68 of the forecast values Percentile Sets the display range end points in terms of percentiles By default the x axis value numbers are automatically adjusted to round numbers to make the forecasts easier to read The Axis chart preference settings include Round Display Range which constrains the display range to round numbers You can uncheck this setting to display actual unrounded numbers Other chart customizations are available to help you interpret the results of the simulation by viewing the data in different ways For details see Setting chart preferences beginning on page 128 Crystal Ball Note Also see Setting preferences with hot keys on page 129 for ways to change the appearance of charts without using menu commands 112 Crystal Ball User Manual Understanding and using forecast charts Showing statistics for the display range Once you have changed the display range you can display statistics for just that range To do this 1 Set the display range as described beginning on page 112 2 Note the values for the display
347. s the number of trials until the first E Geometric page 285 successful occurrence such as the number of times you need to spin a roulette wheel before you win Hypergeometric The hypergeometric distribution is similar to the binomial distribution illi page 286 both describe the number of times a particular event occurs in a fixed number of trials However binomial distribution trials are independent Hypergeometric 5 a while hypergeometric distribution trials change the success rate for each subsequent trial and are called trials without replacement Negative The negative binomial distribution is useful for modeling the if binomial distribution of the number of trials until the r th successful occurrence Neuron page 289 such as the number of sales calls you need to make to close a total of 10 orders It is essentially a super distribution of the geometric distribution Poisson The Poisson distribution describes the number of times an event occurs ll page 291 in a given interval such as the number of telephone calls per minute or Poisson the number of errors per page in a document 280 Crystal Ball User Manual Yes No Yes no Basic page 293 Using discrete distributions Table A 3 Summary of discrete distributions Continued Summary The yes no distribution is a discrete distribution that describes a set of observations that can have only one of two va
348. s to the magnitude right of the 0 line showing magnitude of sensitivity but not direction 2 Pie Shows sensitivity data as a circle divided into 2 0 proportional slices showing the magnitude of 2 sensitivity 25 5 85 2 3 While you have a series selected consider adjusting the chart color page 136 and marker line settings page 137 too Crystal Ball User Manual 133 Chapter 5 Analyzing Forecast Charts 134 4 When settings for the current series are complete follow steps 2 through 3 to customize settings for any other series in the chart 5 When all your settings are complete click OK The chart appears with your new settings Changing the chart density You can show more or less detail such as data points and columns in your chart by changing the number of bins used to group similar values The level of detail or resolution is called the density of the chart Higher densities more accurately reflect the actual distribution of data and lower densities highlight the overall shape and trend of data To change the chart density 1 Display the General tab of the Chart Preferences dialog page 130 2 Locate the Chart Bins group in the middle of the page 3 Choose a density level from the Density drop down list For example the following figure shows the lowest and highest density settings for a column chart Lowest density Highest density 4 To show a space between each
349. simulation results using 99 simulations accuracy 318 speed 321 Single Step 93 Six Sigma features 340 skewness in lognormal distribution 263 spider chart 213 static cell referencing 24 statistics view 118 stopping a simulation 90 Student s t distribution 272 Subscribe To A Category dialog 53 subscribing to distribution categories 53 support technical 5 system requirements T tables Two dimensional Simulation results 240 tasks Distribution Gallery 41 378 Crystal Ball User Manual technical support 5 Test Data example 32 tools Batch Fit 199 Bootstrap 219 Correlation Matrix 206 Decision Table 226 Scenario Analysis 231 Tornado Chart 211 Two dimensional Simulation 236 Tornado Chart description 211 example 214 tool 211 uses 211 training service 5 trend chart with Two dimensional Simulation 241 trend charts certainty levels 162 creating 155 customizing 157 overview 154 value axis 163 triangular distribution overview 274 truncating distributions 312 tutorials Test Data 32 Two dimensional Simulation example 237 tool 236 U uncertainty 236 uniform distribution defined 276 unpublishing distribution categories 55 user manual conventions 4 user defined macros 99 using simulation results 99 V value axis settings 163 variability vs uncertainty 236 variables decision defining 58 variables second order 243 viewing distribution category properties 51 example 33 virtual memory 322 who should use
350. sion variable precedent or cell independently While analyzing one variable the tool freezes the other variables at their base values This measures the effect of each variable on the forecast cell while removing the effects of the other variables This method is also known as one at a time perturbation or parametric analysis The Tornado Chart tool is useful for e Measuring the sensitivity of variables that you have defined in Crystal Ball e Quickly pre screening the variables in your model to determine which ones are good candidates to define as assumptions or decision variables You can do this by testing the precedent variables of any formula cell See page 215 for more information on precedents Tornado chart The Tornado Chart tool tests the range of each variable at percentiles you specify and then calculates the value of the forecast at each point The tornado chart illustrates the swing between the maximum and minimum forecast values for each variable The variable that causes the largest swing appears at the top and the variable that causes the smallest swing appears at the bottom The upper variables have the most effect on the forecast and the lower variables have the least effect on the forecast Crystal Ball User Manual 211 Chapter 8 Crystal Ball Tools Target forecast Material 1 Reliability Forecastaxis values 1 05 1 10 1 15 1 20 1 25 Assumptions and other variables SL cmoo o oo T on as
351. sonality you should use CB Predictor instead of the distribution fitting functionality CB Predictor is a component of Crystal Ball Professional and Premium Editions When the Batch Fit tool runs it fits each column of data to each continuous distribution For each fit of a distribution to a set of data the tool calculates the indicated goodness of fit test The distribution with the best fit is placed in the spreadsheet to create an assumption cell that you can copy to the appropriate location in your model In this example the four columns of data resulted in four assumptions that you copied to the appropriate locations The forecast is already defined for you and running a simulation produces a forecast chart of the total gross profit from the Magazine Sales workbook In the Total Gross Profit forecast chart if you replace Infinity with 5 500 you find that the certainty or probability of making this amount of profit is almost 75 Forecast Total Gross Profit i 5 x Edit View Forecast Preferences Help 500 Trials Frequency View 490 Displayed Total Gross Profit T N g n g 2 T 2 M18 15 M12 ms Fas oa 0 4 954 5 284 5 614 5 944 6 275 6 605 6935 7 265 pem Cerany 724 a a Figure 8 7 Magazine Sales profit from newsstand sales Crystal Ball User Manual 205 Chapter 8 Crystal Ball Tools Correlation Matrix tool 206 Correlations When the values of two variabl
352. ssitssecasieatensieetestssesediseedatdonscaiaateentins 95 SINGS WINGOWS cas scsresecdtecs ston tistaest cearesaeveeuesdedaeaosttadaamieatonnteavaseatttaiees 95 Multiple windows iscisisisccceedcadecasavsaveenosideasodsieectaaciaads ties cuttnanesdiessdadins 96 Saving and restoring simulation results ccceee cece cseeee cee eeeeeeseeeneenees 97 Saving Crystal Ball simulation results 00 00 eee ce eee eeeeeceeeeeeeeees 97 Restoring Crystal Ball simulation results cece eee eeeeeeceeeereeeees 98 Using PestOred Lesults ssimic aiei ei Seii Eiai 99 Running user defined Macros s sssssssessitsssissisetstisstsestisrtstnrinsinrenrisressnse 99 Interfaces insonsan o e S a a A a E ER 101 Priority rulesnsns e penai a n E E EE E ERE 101 Global MACOS ssas arasi tadan kaaneni Ea sain Kania Anene Ea aTi SASAR 102 MOO Dar Matos gasoni cote sh AR ER 102 ii Crystal Ball User Manual Contents Chapter 5 Analyzing Forecast Charts Guidelines for analyzing simulation results 0 c eee ee cece cee eeeereeel 104 Understanding and using forecast Charts 0 0 ccc eceecee cece cee ceeeeeeeeeee 106 Determining the certainty level cece cee cee eesessecneceeeeeseeneeens 107 Focusing on the display range ccccceccecesece cece cess eeecnsersenaeeseeseeeee 112 Formatting chart NUMDETS 0 0 ec ceceee eee ecceeee cee eeeeecesesseceeeeeeeesaees 113 Changing the distribution view and interpreting statistics
353. ssssssssisisensscsisivnsorecisisinusiiissnis 328 User defined functions cccccccccccccceescesseseseececeeeeceecesaeaaeaaeseeeeeeeeeeees 328 Crystal Ball User Manual v Contents Running user defined Macros o c cece eee eee e eee ee eeeeeeee cesses ceeeeseeeees 330 Special MUNCHONS sssrinin pieninkin A aSa E eSEE a EE aSa ESER iG 330 Undocumented behavior of standard functions cccccececeeeeeeeeeees 331 Incompatible range COnStruUCtS sss ssssssesssstrestetittrt sistr rtistrt ceaeeneeaaenaees 331 Labels in formulas that are not defined names ccccccccceececeeeeeeeees 331 Multiple area references 2 0 0 cece cece cece cee ceseeeeeeseeseseeseceeeseeeessecneeees 331 I D references aicsiisiisosaseiadiadcdadcadolaguansiasavecdadawadsadadvenssaboesesdeecaseediadaselss 332 Data Wa Desi EE a E EEE E AA AEA E T E 332 Other important differences nsssssiissiinnissoitesresiniinissisiissiisinenss 332 OptQuest and CB Tools sists cocieees tdeesscessessenvees steastosneastictiesnvseeaacedns 332 Precision control and cell error checking eee cece cee eeeeeeee 333 Spreadsheet updating iis icciisast esses iadaees seen tnitatieeeieeendecdsas 333 Very large models oi iicc sicetivsdiacetstaasdicccstiionres teens tainevesadviaeti in sstdasvennes 333 MEMOTY USAGE osscsisesansdetssnndesssnusedssnwanssanndasniwasddoawasddvamsvestwasnoosmasnssuaedses 334 Spreadsheets with no Crystal Ball data 0cccccccceseeeee cee ceteetee
354. ssumptions 26 the likeliest value This gives you a distribution that has 80 or four fifths of the values occurring between the two specified percentiles as in Figure 2 5 To change the parameter sets for the continuous distributions use the Parameters menu in the menubar of the Define Assumption dialog The currently selected parameter set has a check mark next to it as shown on the menu in Figure 2 5 Define Assumption Cell E8 E JE fe o x Edit View Parameters Preferences Help D Minimum Likeliest Maximum Name 10 50 90 5 50 95 15 00 bo y do y wx f350 Likeiest mm Hy soxfieso H OK Cancel Enter Galley _Conelte _ Hep Figure 2 5 10th and 90th percentiles with Likeliest parameter In addition to the standard parameter set each continuous distribution s Parameters menu has additional pre defined parameter sets that include various combinations of the standard parameters and percentiles There is also a Custom command that lets you define your own parameter set If you choose Custom in the Parameters menu you can replace any or all of the standard parameters with any percentile You will always have the same number of parameters either standard or alternate for any given distribution For example even if you choose to use custom alternate parameters for a triangular distribution you will always have three parameters either minimum likeliest and maximum o
355. st cell The sensitivity chart displays these rankings as a bar chart indicating which assumptions are the most important or least important in the model You can output print the sensitivity chart on the report or copy it to the clipboard Crist Assessment o Edit View Sensitivity Preferences Help 1 000 Trials Contribution to Variance View Sensitivity Risk Assessment 0 0 21 0 42 0 63 0 1 1 Volume of Water per Day i Ses BETE p CPF 255 Body Weight 7 2 Concentration of Contaminan 2 0 Figure 6 18 Assumptions and their effects on toxicity risk Sensitivity charts provide these key benefits You can find out which assumptions are influencing your forecasts the most reducing the amount of time needed to refine estimates e You can find out which assumptions are influencing your forecasts the least so that they can be ignored or discarded altogether e Asa result you can construct more realistic spreadsheet models and greatly increase the accuracy of your results because you know how your assumptions affect your model Crystal Ball User Manual Understanding and using sensitivity charts Creating sensitivity charts The example model used in this section Toxic Waste Site xls illustrates the risk assessment of a toxic waste site The sensitivity chart created in the steps below displays in descending order the assumptions in this model The assumption with the highest
356. stal Ball discrete values with the same probability could be entered in ranges with five columns or more Now they cannot be entered in ranges with five columns but can only be entered in single columns or ranges with six or more columns to distinguish them from sloping ranges In previous versions of Crystal Ball continuous uniform ranges with cumulative probabilities could be entered in a two column format Now a three column format is required discussed in Connected series of continuous uniform ranges cumulative on page 309 The three column sloping range format used in previous versions of Crystal Ball has been replaced by a five column format described in Mixed ranges including sloping ranges on page 308 and the section that follows it Connected series of ranges sloping Other important custom distribution notes Even if you don t load data from the spreadsheet into the Custom Distribution dialog you can still add and edit data using the data table To do this click the More button to display the data table Then you can Enter a different value in the data table and click Enter to change the data Type the minimum maximum probability and step if discrete data into a blank row and click Enter to add new data To delete a single range of data select that row of data right click and choose Delete Row To clear all data rows right click within the data table and choose Clear Distribution
357. stributed lower forecasts the ratio of the difference between the forecast mean and lower specification limit over three times the forecast long term standard deviation often used to calculate process capability indices with only a lower specification limit Cpk v One sided short term capability index for normally distributed upper forecasts the ratio of the difference between the forecast mean and upper specification limit over three times the forecast short term standard deviation often used to calculate process capability indices with only an upper specification limit Crystal Ball User Manual 353 Appendix D Using the Process Capability Features Metric Ppk upper Table D 1 Capability statistics calculated by Crystal Ball Continued Long term Short term Description One sided long term capability index for normally distributed forecasts the ratio of the difference between the forecast mean and upper specification limit over three times the forecast long term standard deviation often used to calculate process capability indices with only an upper specification limit Cpk Short term capability index minimum of calculated Cpk lower and Cpk upper that takes into account the centering of the forecast with respect to the midpoint of the specified limits a Cpk equal to or greater than 1 indicates a quality level of 3 sigmas or better Ppk Long term capability index minimum of calcu
358. stribution Q Define assumptione cei Edit View Parameters Preferences Help Namen o O Custom Distribution A Oo 2 amp L 2 amp D 7 00 Enter one or more values with probabilities Vale O Probabili oK Cancel C Ete Galley Correlate Help Figure A 28 Single values Now each value has a probability of 1 However when you run the simulation their total relative probability becomes 1 00 and the probability of each value is reset to 3333 If you want to reset their probabilities before you run the simulation follow these steps 1 Click the bar with a value of 5 00 Its value appears in the Value field 2 Type the probability as the formula 1 3 in the Probability field and click Enter You could also enter a decimal for example 0 333333 but the formula is more exact 3 Follow steps 6 and 7 for the other two bars Crystal Ball rescales each value to a relative probability of 0 33 on the left side of the screen Crystal Ball User Manual 297 Appendix A Selecting and Using Probability Distributions 298 Goetne assumption CHOSE Edit View Parameters Preferences Help Ne fT SSC Custom Distribution 2030 4 T 025 Q 020 015 4 6 00 7 00 8 00 9 00 10 00 Enter one or more values with probabilities Value Probability oK Cancel Enter_ Gallery Correlate Help Figure A 29 Single values wit
359. stribution Gallery opens Only the most common probability distributions appear in the window Crystal Ball Note The All category contains all distributions originally shipped with Crystal Ball Ifyou modify and save one of these original distributions it appears in the Favorites category unless you create and specify another category You can perform many tasks in the Distribution Gallery To see more distributions click the All folder in the category pane You can use the upper scroll bars resize the window or change the View menu settings to view all the distributions in that category e To view the description of a distribution click it the description appears in the lower pane e To have Crystal Ball select a distribution for you based on your data click the Fit button See Fitting distributions to data on page 27 for details You can also add categories add distributions to categories customize distributions share categories with other Crystal Ball users and more For details see Additional Distribution Gallery features beginning on page 41 Crystal Ball User Manual Assumption name More button Parameters Defining assumptions Select a category from the folders in the category pane and then double click the distribution you want to use A dialog appears showing the distribution type you chose for the selected cell or for the first cell in a range of cells Figure 2 3 shows
360. stribution descriptions beginning on page 253 e Discrete distribution descriptions beginning on page 280 e Custom distribution description beginning on page 294 Crystal Ball User Manual 249 Appendix A Selecting and Using Probability Distributions Crystal Ball Note Initially the precision and format of the displayed numbers in the probability and frequency distributions come from the cell itself To change the format see Customizing chart axes and axis labels on page 138 Selecting a probability distribution 250 Plotting data is one guide to selecting a probability distribution The following steps provide another process for selecting probability distributions that best describe the uncertain variables in your spreadsheets To select the correct probability distribution Look at the variable in question List everything you know about the conditions surrounding this variable You might be able to gather valuable information about the uncertain variable from historical data If historical data are not available use your own judgment based on experience to list everything you know about the uncertain variable For example look at the variable patients cured that was discussed in the Vision Research tutorial in Chapter 2 of the Crystal Ball Getting Started Guide The company must test 100 patients You know that the patients will either be cured or not cured And you know that the drug has shown a cure rate o
361. t Capability metrics appear with other forecast data similar to Figure D 12 Forecasts j Worksheet DFSS Fluid Pump xls Model i Forecast FLOW RATE FORECAST ml sec J23 Cell K23 i i Summary Certainty level is 94 9 Certainty range is from 47 2617 to 53 9283 Entire range is from 45 5614 to 53 1558 Base case is 50 6467 After 1 000 trials the std error of the mean is 0 0383 FLOW RATE FORECAST ml sec i er eas 5 i 008 Leo i pes st sascea Zoos faget 50 5950 loz 2 fl Std Dev 50 4574 2 2 00 Tos amp 3 pez 20 oot t10 i 000 D o 466452 477053 487654 498256 50 8857 519458 530059 54 0660 mi sec i Forecast FLOW RATE FORECAST ml sec J23 contd Cell K231 1 Capability Metrics Forecast values Mean 49 2453 Standard Deviation 1 2122 Cp 0 92 Cpk lower 0 55 Cpk upper 1 29 Cpk 0 55 Cpm 0 61 Z LSL 1 64 Z USL 3 86 Zsttotal 1 64 Zit 0 14 p N C below 0 05 p N C above 0 00 p N C total 0 05 1 PPM below 50 879 49 PPM above 55 92 PPM total 50 935 41 Figure D 12 Frequency report with capability metrics 352 Crystal Ball User Manual Capability metrics list Capability metrics list The following table lists and defines the statistics shown in Capability Metrics view and indicates whether each statistic appears for long term or short term data For a discussion of the equat
362. t Crystal Ball 7 gt Crystal Ball Excel opens with the Crystal Ball menus and toolbar If Excel is already running when you give this command Crystal Ball opens a new instance of Excel Crystal Ball User Manual The Crystal Ball menus and toolbar Starting Crystal Ball automatically To set Crystal Ball to start automatically each time you start Excel 1 In Windows choose Start gt Programs gt Crystal Ball 7 gt Application Manager 2 Check Automatically Launch Crystal Ball 7 When Excel Starts 3 Click OK Closing Crystal Ball To close Crystal Ball either crystal Ball Right click the Crystal Ball icon in the Windows taskbar and choose Close or e Close Excel If you want you can choose Run gt Reset Simulation to reset the model and then choose File gt Save to save it before you close Crystal Ball The Crystal Ball menus and toolbar The Crystal Ball menus When you load Crystal Ball with Microsoft Excel some new menus appear in the Excel menubar Define contains commands that let you define and select assumption decision variable and forecast cells perform Crystal Ball copy data paste data and clear data operations and set cell preferences e Run contains commands that let you start stop reset and single step through simulations freeze variables launch the Crystal Ball tools as well as CB Predictor and OptQuest if you have the Professional edition of Crystal Ball and se
363. t For an example see Figure D 6 Notice that the certainty range on the forecast chart changes to match the LSL and USL values The certainty indicates the probability of falling within those specification limits To show the certainty of different values type them Crystal Ball User Manual 347 Appendix D Using the Process Capability Features 348 into the Minimum and Maximum fields or click the certainty grabbers and drag them to a new position To add specification limit and target marker lines manually or to remove them 1 Choose Preferences gt Chart Preferences in a forecast chart window or double click the chart Click the Chart Type tab in the Chart Preferences dialog 3 Scroll down to the bottom of the Marker Lines list zi General Acs Choose series Chart type Marker lines Type Column E O Percentile O Value Color Sa z CO Base Case O Certainty Range 0 0 M LSL USL Target Figure D 7 Showing and hiding LSL USL and target marker lines 4 Be sure LSL USL Target is checked if you want to display the specification limit and target marker lines on that forecast chart To hide the markers uncheck LSL USL Target 5 When all settings are complete click OK Notice that you can display the mean and other marker lines in addition to the specification limit and target markers Crystal Ball User Manual Extracting capability metrics Extractin
364. t of wild horses in Nevada Researchers in the department travel to a particular area in Nevada to tag 100 horses in a total population of 1 000 Six months later the researchers return to the same area to find out how many horses remained in the area The researchers look for tagged horses in a sample of 200 Check the data against the conditions of the hypergeometric distribution The parameter values for the hypergeometric distribution in Crystal Ball are the population size of 1 000 sample size of 200 and an initial success rate of 100 Crystal Ball User Manual Using discrete distributions out of 1 000 or a probability of 10 0 1 of finding tagged horses The result would be a distribution showing the probability of observing x number of tagged horses Crystal Ball Note Ifyou used this distribution in a model created in Crystal Ball 2000 x you might notice slight data changes when running that model in the current version of Crystal Ball This is because some rounding might occur when converting the probability parameter used in previous releases to the success parameter used in this version of Crystal Ball Negative binomial distribution Neg Binomial Parameters Probability Shape Conditions The number of trials is not fixed The trials continue until the r th success The probability of success is the same from trial to trial Statistical Note The total number of trials needed will always be equal to o
365. t run preferences e Analyze contains commands that let you create a variety of charts and reports extract data and save or restore results The following chapters of this book explain how to use the various commands For specific information about commands see the Index at the end of this book as well as the Crystal Ball online help see page 5 for more information on help Crystal Ball User Manual 15 Chapter 1 Crystal Ball Overview The Crystal Ball toolbar To help you set up spreadsheet models and run simulations Crystal Ball comes with a customized toolbar that provides instant access to the most commonly used menu commands n S c 6 2 9 n o a E Q Pi n 5 o 7 ja Fi Q g oO of amp vo oT u pa 2 2 5 amp 8 S So PF 9 8 2 E g ggg 5 amp amp 8 S SgS Eg 5 o FTF wow Z S 3 2 c 6 g FO O 6 2 EEES ZIED LCLYssoEF amp FE BF sue eee see te FF Gee ER Rp SE Fee o OOo p W D L o E 8 amp 5 S v O S oe ee Fares eae zs EF Fee eF e Sees SF BF FS FSer gs amp F GFE FS EF FE ww Oh _ pag a A AAAA OLO FAH EH Gu OKRH CH O rA gh ni ta 41 ta tdi t APIA D kl BAZ BirPHEHD i AthMM FIBER Figure 1 1 The Crystal Ball toolbar The tools in the first three groups are from the Define menu The tools from the next two groups are from the Run menu The tools from the following two groups are from the Analyze menu and the tool in the last grou
366. t you create delete view and modify properties of and rearrange category folders in the Category pane Two additional commands let you share categories with others Publish and use categories others have shared Subscribe View Commands that change how distributions appear in the Distribution pane as Thumbnails Large Icons or Small Icons and hide or show distribution details and descriptions Help Commands that display online help for the Distribution Gallery and the selected distribution For more information on how to use these commands see the references in Table 2 2 on page 41 The Fit button at the bottom of the Distribution Gallery opens the Crystal Ball distribution fitting feature This feature can help you select an appropriate distribution for the assumption you are defining For more information see Fitting distributions to data on page 27 The Help button displays online help for the currently selected distribution Crystal Ball User Manual 43 Chapter 2 Defining Model Assumptions The Category pane Categories are groups of distributions contained in folders e Basic is the default category It contains several of the most common distributions Normal Triangular Uniform Lognormal Yes No and Discrete Uniform All contains all distributions shipped with Crystal Ball in unmodified form including those also supplied in the Basic category Favorites is the default category for
367. the whole By default the display range includes all trials within 2 6 standard deviations of the mean approximately 99 of the forecast values Crystal Ball then rounds the display range to the next even number of units For this reason outlying trials might be excluded from the display range Crystal Ball Note To display all trials change the chart axis preferences to display fixed endpoints between Infinity and Infinity for instructions see Focusing on the display range on page 112 The number of trials run for this forecast appears in the upper left corner of the forecast chart The number of trials shown the number of trials run minus the number of outlying values appears in the upper right corner The display range is the linear distance from about 15 00 to 35 00 as shown in Figure 5 1 Determining the certainty level The forecast chart also shows the certainty range for the forecast In Crystal Ball the certainty range includes all trials between the certainty grabbers In the next section you will learn how to change the certainty range Crystal Ball Note When the certainty grabbers are at Infinity and Infinity the certainty range includes every forecast value regardless of the size of the display range Crystal Ball compares the number of values in the certainty range with the number of values in the entire range to calculate the certainty level for the forecast The example in Figure 5 1 shows a ce
368. the Likeliest parameter you can estimate the Scale parameter The Scale parameter is a number greater than 0 The larger the Scale parameter the greater the variance To calculate a more exact scale you can estimate the mean and use the equation mean mode 0 57721 a where o is the Scale parameter Or estimate the variance and use the equation a 6 variance 2 T where is the Scale parameter Normal distribution Parameters Mean Standard Deviation Conditions Some value of the uncertain variable is the most likely the mean of the distribution The uncertain variable could as likely be above the mean as it could be below the mean symmetrical about the mean The uncertain variable is more likely to be in the vicinity of the mean than far away Crystal Ball User Manual LA Normal Using continuous distributions Statistical Note Of the values of a normal distribution approximately 68 are within 1 standard deviation on either side of the mean The standard deviation is the square root of the average squared distance of values from the mean Description The normal distribution is the most important distribution in probability theory because it describes many natural phenomena such as people s IQs or heights Decision makers can use the normal distribution to describe uncertain variables such as the inflation rate or the future price of gasoline The following example shows a re
369. the mouse to individually alter Crystal Ball User Manual Using probability functions the probabilities and shapes of the data points and ranges so that they more accurately reflect the uncertain variable Using probability functions For each of the Crystal Ball distributions there is an equivalent Excel function You can enter these functions in your spreadsheet directly instead of defining distributions using the Define Assumption command Be aware though that there are a number of limitations in using these functions These are listed below To view these functions and their parameters choose Insert gt Function in Excel be sure the category is set to All and then scroll down to the list of functions that starts CB i 21x Search for a function Type a brief description of what you want to do and then Go dick Go Or select a category fail Select a function CB Beta CB CustomCumul CB DiscreteUniform CB Exponential xl CB Beta2 Min Max Alpha Beta LowCutoff HighCutoff NameOf Returns a random number for the Beta distribution 4 parameter Help on this function Lox Cancel Figure A 48 Crystal Ball functions in Excel Parameters and a brief description appear below the list of functions The Cutoff parameters let you enter truncation values while NameOf is the assumption name For parameter descriptions and details on each distribution see the entry for that distribution earlier in this
370. the size of the charts by entering a percentage in the field Trend Charts Includes trend charts in the report You can scale the size of the charts by entering a percentage in the field Sensitivity Charts Includes sensitivity charts in the report You can scale the size of the charts by entering a percentage in the field Crystal Ball Note Excel charts cannot be created for assumptions defined as custom distributions Report details The Crystal Ball Custom report offers details for each section Within a section you can include any or all of them The details are described in Table 7 2 Some of the Details settings display chooser dialogs For more information about these click Help in the dialog Crystal Ball User Manual 187 Chapter 7 Creating Reports and Extracting Data Table 7 2 Custom report sections and details Section Detail Description Report Summary Report Title Name of the report default is Crystal Ball Report Date Time Date and time when the simulation included in the report was started and stopped Run Preferences Run Preferences settings for the report e Number of trials run X XXX e Monte Carlo sampling or Latin Hypercube sampling Latin Hypercube sample size e Random seed or Seed XXX e Precision Control on if true e Correlations disabled only if disabled e User macros disabled only if disabled Run Statistics First section is an exact duplicate of the
371. the x axis value where the line should appear in the Value field and click Add If you want enter an optional label and Crystal Ball User Manual 137 Chapter 5 Analyzing Forecast Charts 138 check Show Value On Marker Line to display the value on the chart You can click New to add another value 3 While you have a series selected consider adjusting the chart type page 132 and color page 136 too 4 When settings for the current series are complete follow steps 2 and 3 to customize settings for any other series in the chart 5 When all your settings are complete click OK The chart appears with your new settings Crystal Ball Note Ifthe marker lines fall outside the maximum or minimum value displayed on a chart they will not appear on the chart For example this can happen with standard deviations of plus or minus 2 or 3 for uniform distributions Customizing chart axes and axis labels You can customize the label scale and format of the main axis in Crystal Ball charts To customize chart axes 1 Display the Axis tab of the Chart Preferences dialog 2 Consider the label to use for the axis By default Auto is checked in the Axis Label group A label is automatically assigned depending on the chart type For example the value x axis of a forecast chart uses the Units entry in the Define Forecast dialog as the label To enter a custom axis label uncheck Auto and type a more descriptive label in the edit
372. this program 1 W Y Weibull distribution yes no distribution 293 Crystal Ball User Manual 379 Crystal Ball User Manual Written and edited by Barbara Gentry Eric Wainwright and David Blankinship Portions based on material written and edited by Carol Werckman Terry Hardy Cameron Harris and Roy Sargent Illustrations and screen captures by Barbara Gentry using Jasc Inc Paint Shop Pro Typeset using NewBskvll BT and Univers fonts Crystal Ball User Manual 379
373. tings are complete click OK Crystal Ball extracts the simulation data to the specified worksheet location The extracted data is arranged as columns of forecasts and assumptions and rows of data You can sort modify print or save the data in the same way as any other spreadsheet The following figures show samples of different types of data extracted with all Formatting settings checked forecasts only c Bllelejeholal e 13 14 1000 17 950 914 17 930 927 1 297 577 1 683 706 187 431 0 15 3 19 0 07 14 060 365 522 981 379 8 921 013 A B Statistics Ending Sales Year 3 Q1 Ending Sales Year 3 Q2 Trials 1000 Mean 17 081 883 Median 17 042 665 Mode Standard Deviation 1 117 986 Variance 1 249 891 924 272 Skewness 0 19 Kurtosis 3 22 Coeff of Variability 0 07 Minimum 13 695 983 Maximum 21 289 239 Range Width 7 593 255 Mean Std Error 35 354 541 033 Figure 7 9 A sample of extracted data Statistics format Crystal Ball User Manual Extracting data A Ending Sales Year 3 Q1 Ending Sales Year 3 Q2 18 13 695 983 14 060 365 15 622 926 16 312 876 16 129 311 516 837 542 16 492 819 17 288 938 16 826 501 17 603 671 17 042 665 17 930 927 17 307 813 518 215 054 17 596 651 18 592 965 17 986 610 19 018 804 18 526 765 19 659 121 21 289 239 22 981 379 B Ending Sales Year 3 Q1 Minim
374. tion s conditions Checking the binomial distribution There are only two possible outcomes the manufactured item is either good or defective e The trials 50 are independent of one another Any given manufactured item is either defective or not independent of the condition of any of the others The probability of a defective item 7 is the same each time an item is tested These conditions match those of the binomial distribution The parameters for the binomial distribution are Probability and Trials In example one the values for these parameters are 50 Trials and 0 07 7 Probability of producing defective items You would enter these values to specify the parameters of the binomial distribution in Crystal Ball G Define Assumption Cell A1 E lol o x Edit View Parameters Preferences Help Name Jan x y Binomial Distribution e ae MN E S Cat EAA Benet Probability 0 07 Sf Trials 50 OK Cancel Enter Gallery Correlate Help Figure A 19 Binomial distribution The distribution illustrated in Figure A 19 shows the probability of producing x number of defective items Example two A company s sales manager wants to describe the number of people who prefer the company s product The manager conducted a survey of 100 Crystal Ball User Manual Using discrete distributions consumers and determined that 60 prefer the company s product over the competitor s Ag
375. tion Book1 xls Sheet1 A1 Code 5539 e Workaround For best results always use standard syntax with fully valid arguments Incompatible range constructs Dynamic ranges Extreme speed does not support dynamic ranges where the OFFSET function is used on one or both sides of the range constructor For example AVERAGE Cellnamel OFFSET Cellname 2 x y e Example message Unable to interpret formula at cell location Book 1 xls Sheet1 A1 Code 5504 Invalid token e Workaround Avoid using the OFFSET function to construct dynamic ranges Labels in formulas that are not defined names Extreme speed supports defined names and their use in formulas but not the Excel Accept labels in formulas option which allows cell labels to be used in formulas without defining them as names Example message Unable to interpret formula at cell location Book1 xls Sheet 1 Al Code 5514 Undefined identifier e Workaround For best results use defined names instead of cell labels in formulas Multiple area references Extreme speed does not support multiple area references such as A1 A5 B1 C1 E1 except as used in standard functions such as SUM that accept a variable length argument list of cell ranges The only supported use of the comma is as the separator in the argument list of a standard function Crystal Ball User Manual 331 Appendix C Using the Extreme Speed Feature not as a cell range u
376. tion Continue Simulation page 92 Reset Simulation page 92 Single Step page 93 Freeze page 88 OptQuest Predictor Tools page 198 Run Preferences page 80 Analyze Assumption Charts page 18 Forecast Charts page 59 Overlay Charts page 146 Trend Charts page 154 Sensitivity Charts page 165 Cascade Table 4 7 on page 96 Close All Table 4 7 on page 96 Create Report page 182 Extract Data page 190 Save Results page 97 Restore Results page 98 Help Control Panel Help Managing chart windows You can close the forecast and other chart windows at any time and the simulation will continue Running a simulation with the windows closed decreases the time required to run the simulation Single windows To close one window click the Close icon in the upper right corner of the chart window Crystal Ball User Manual 95 Chapter 4 Running Simulations Multiple windows You can use several commands on the Analyze menu and Crystal Ball toolbar to e Close and cascade all open forecast windows and other chart windows e Open charts that have been closed e Create new charts Delete overlay trend and sensitivity charts These commands are listed in Table 4 7 following Also see Speed preferences on page 84 for ways to control whether and how often chart windows display and redraw To manage multiple windows s uty e Click a chart button in the Crystal Ball toolbar to op
377. tion describes a situation where you know the A Basic minimum maximum and most likely values to occur For example you nace page 274 could describe the number of cars sold per week when past sales show the minimum maximum and usual number of cars sold 254 Crystal Ball User Manual Using continuous distributions Table A 2 Summary of continuous distributions Continued Shape Name Summary Uniform In the uniform distribution all values between the minimum and maximum Basic occur with equal likelihood Uniform page 276 A Weibull The Weibull distribution describes data resulting from life and fatigue tests Weibull page 278 Rayleigh It is commonly used to describe failure time in reliability studies and the breaking strengths of materials in reliability and quality control tests Weibull distributions are also used to represent various physical quantities such as wind speed Crystal Ball Note As you work with the Crystal Ball probability distributions you can use the Parameters menu found in the distribution menubar to specify different combinations of parameters For more information see Alternate parameter sets on page 25 Beta distribution Beta Parameters Minimum Maximum Alpha Beta Conditions The uncertain variable is a random value between the minimum and maximum value The shape of the distribution can be specified using two positive values Alpha
378. tion or after you have stopped it you can use the Single Step command to watch the simulation process generate one set of values a trial at a time for the assumption cells and recalculate the spreadsheet This feature is useful if you are trying to track down a calculation error or verify that the values being produced for your assumption cells are valid To observe an individual trial 1 Select Run gt Reset Simulation or click the Reset button on the Crystal Ball toolbar or Control Panel IE 2 Select Run gt Single Step or click the Single Step button to run one trial of the simulation Click the button again to run another The Crystal Ball Control Panel You can use the Crystal Ball Control Panel to perform many simulation and analysis commands When you start the first run as described in Running a simulation on page 92 the Crystal Ball Control Panel appears as shown in Figure 4 10 on page 94 Once it appears you will find its buttons and menus convenient to use for stopping continuing resetting single stepping and analyzing the results Run preference settings determine whether the Control Panel appears or is hidden when a simulation runs To hide the Control Panel uncheck Show Control Panel on the Options tab of the Run Preferences dialog click the Run Preferences tool or choose Run gt Run Preferences Crystal Ball Note By default the Control Panel stays open following a reset To close it uncheck Le
379. tion to numeric values you can enter a reference to a specific cell ina parameter field Cell references must be preceded by an equals sign Cell references can be either absolute or relative You can also enter formulas and range names If necessary you can press F4 to change references from relative to absolute or back to relative This also applies to cell references in fields other than assumption parameters Note All cell references in parameters are treated like absolute references when cutting and pasting Crystal Ball data Crystal Ball always stores the cell reference in A1 format even if the Excel preference is set to RIC1 format The global RIC1 format preference is not affected by running Crystal Ball but the name ranges are in fact changed to Al format since that is the way Crystal Ball stores them To show cell references instead of current values when you enter them in parameter fields choose Parameters gt Show Cell References in the Define Assumption dialog Dynamic vs static cell references Cell references in assumption parameters are dynamic and are updated each time the workbook is recalculated Dynamic cell referencing gives you more flexibility in setting up models by letting you change an assumption s distribution during a simulation Other types of cell references are static such as the assumption name field and correlation coefficients These cell references are calculated once at the beginning of
380. tions into the two types uncertainty and variability 7 Click Next The Specify Options dialog appears as shown in Figure 8 41 238 Crystal Ball User Manual Two Dimensional Simulation tool Specify options step 3 of 3 2D Simulation Crystal Ball Tool Decisioneering 1998 2004 rSimulation Control Run outer uncertainty simulation for fio trials Run inner variability simulation for 500 trials maximum While Running Show forecasts as defined Show only target forecast Hide all forecasts lt Back Cancel __te Figure 8 41 Specify Options dialog 2D Simulation tool 8 Set the following options e Number of trials for the outer simulation 100 e Maximum trials for the inner simulation 1 000 Show Only Target Forecast is selected 9 Click Start The simulations start The tool first single steps one trial to generate a new set of values for the uncertainty assumptions Then it freezes these assumptions and runs a simulation for the variability assumptions in the inner loop The tool retrieves the Crystal Ball forecast information after each inner loop runs The tool then resets the simulation and repeats the process until the outer loop has run for the specified number of simulations CB Tools Note For more information on the 2D Simulation tool dialogs click Help in a dialog Interpreting the results The results of the simulations appear in a table containing the
381. to apply the current settings to other charts in your model you can choose what to apply and where to apply them To specify how chart settings should be applied 1 Click the Apply To button The Apply To dialog appears as shown in Figure 5 20 ce 8 s x m Apply preferences for _ m To all sensitivity charts on This sheet This workbook All open and new workbooks Cancel Hep Figure 5 20 The Apply To dialog 2 Indicate the tab or tabs of settings to apply e Choose This Tab to apply only settings on the current tab e Choose All Tabs to apply all the current settings on the entire dialog Crystal Ball User Manual 139 Chapter 5 Analyzing Forecast Charts 3 Indicate where the settings should be applied e Choose This Sheet to apply the settings to only the current sheet of the current workbook e Choose This Workbook to apply the settings to all sheets of the current workbook e Choose All Open And New Workbooks to apply the settings to all workbooks that are currently open and all workbooks to be created Crystal Ball Note The All Workbooks setting effectively changes the global Chart Preferences defaults to the settings on the current tab or all tabs depending on the setting in the previous dialog group The Apply To settings are both flexible and powerful They can be used as focused or very broad defaults Managing existing charts 140 Previous sections of thi
382. tributions to data beginning on page 27 2 Enter the assumption as described in the next section Entering an assumption Entering an assumption To enter an assumption 1 Select a cell or a range of cells Select value cells or blank cells only Assumptions cannot be defined for formula or non numeric cells Crystal Ball Note There is no absolute limit to the number of assumptions you can define per worksheet In general you should define less than 1000 assumptions decision variables and forecasts per worksheet 4 2 Choose Define gt Define Assumption Crystal Ball User Manual 19 Chapter 2 Defining Model Assumptions 20 For each selected cell or cells in the selected range Crystal Ball displays the Distribution Gallery dialog Menubar Category pane with category folders Distribution pane Distribution Gallery Cell C5 a 15 x Edit View Categories Help A ai mm Lognomal Yes No Discrete Uniform Triangular Uniform Description pane Description The normal distribution describes many natural phenomena such as IQs s heights the inflation rate or errors in measurements It is a continuous ability distribution The parameters for the nomal distribution are mean and standard deviation zl Cancel Et Help Figure 2 2 The Distribution Gallery with Basic category selected By default the Basic category appears when the Di
383. trol settings that determine when to stop a simulation based on confidence intervals for selected statistics For more information about precision control see Precision control on page 318 Crystal Ball Note The current simulation must be reset before precision control settings will take effect You can choose among the settings listed in Table 3 2 Then click OK to apply settings on the current tab to the active forecast Or you can click Apply To to apply settings on the active tab to the active worksheet the active workbook or all workbooks For more information click Apply To and then click Help in the Apply To dialog At any time you can click Defaults to restore the original default settings on the active tab in the dialog Crystal Ball User Manual Defining decision variable cells and forecast cells Table 3 2 Precision tab forecast preferences Setting Effect Specify The Activates the precision control settings for the forecast Crystal Ball uses these Desired Precision settings only if the simulation is set to stop when it reaches the specified precision For Forecast from the Run Preferences dialog For more information see Setting run Statistics preferences on page 80 The statistics available for precision control are the mean the standard deviation and an indicated percentile Check any or all Percentile If you select Percentile you can enter any percentile value greater than 0
384. ts of the simulation Buying more RAM or turning off virtual memory if possible are solutions to this problem Crystal Ball User Manual Sample size Sample size The sample size option is located on the Run Preferences gt Sampling dialog page 83 Sample size which is initially set to 500 affects Latin hypercube sampling It divides each assumption s distribution into a number of intervals of equal probability The sample size governs the number of intervals for each distribution Crystal Ball generates an assumption value for each interval according to the interval s probability distribution While any sample size greater than 100 should produce sufficiently acceptable results you can set this number higher to maximize accuracy There is no absolute limit to sample size although samples greater than 100 000 work best with at least 1 GB RAM and might take a long time to run The increased accuracy resulting from the use of larger samples however requires additional memory If memory becomes an issue reduce the sample size and consider adding more RAM Correlated assumptions There is a practical limit of about 250 fully correlated assumptions assumptions that are correlated to every other assumption and up to about 1 000 serially correlated assumptions assumptions that are correlated to one or two other assumptions for each workbook If you have a large number of correlated assumptions For coefficients that are close
385. tted statistics primarily for use in other analyses To extract formatted data see Extracting data on page 190 You can choose among the settings listed in Table 3 4 Then click OK to apply settings on the current tab to the active forecast Or you can click Apply To to apply settings on the active tab to the active worksheet the active workbook or all workbooks For more information click Apply To and then click Help in the Apply To dialog At any time you can click Defaults to restore the original default settings on the active tab in the dialog 68 Crystal Ball User Manual Working with Crystal Ball data Table 3 4 Auto Extract tab forecast preferences Setting Effect Extract Forecast Statistics Automatically Activates the auto extract feature listbox at the left The list of statistics you can extract Check the statistics you want then use the up and down arrows to rearrange their order if you want Starting Cell The first cell on the forecast s worksheet where the statistics will be copied Be sure no data entries appear to the right of this cell and below it because data could be overwritten without warning Formatting Whether to include labels on the extracted statistics and use Excel s AutoFormat for the cells Direction Whether the extraction proceeds vertically downward or horizontally to the right on the worksheet Working with Crystal Ball data Edi You
386. tton Probability Truncation grabbers i 10 00 Minimum Truncation value j iny xj q rinit Mean A Sid Dev 0 20 y ok cost C Ene _ cate Maximum Truncation value Figure 2 4 The Define Assumption dialog expanded In the expanded Define Assumption dialog you can e Enter truncation minimum and maximum values in the fields just below the distribution Use the truncation grabbers to truncate the value range e Use numeric spinners arrows to the right of the field to adjust parameter settings e Click the Less button to hide the minimum and maximum value fields and truncation grabbers Crystal Ball Note For more information about truncating distributions see Truncating distributions on page 312 You can perform these activities in both the standard and expanded Define Assumption dialog e Click the Gallery button to display the Distribution Gallery window and choose another distribution e Click the Correlate button to define correlations as described on page 33 22 Crystal Ball User Manual Additional Additional assumption features e Choose Edit gt Add in the menubar to add the currently defined assumption distribution to the Favorites category or a user defined category in the Distribution Gallery e Use other menu commands to copy the chart paste it into Excel or another application print data change the view use alternate parameters s
387. u follow one assumption along its horizontal row and the second along its vertical column the value in the cell where they meet is their correlation coefficient The matrix contains only the correlation coefficients you enter Crystal Ball User Manual Correlation Matrix tool Assumption 1 Assumption 2 Assumption 3 Assumption 1 1 000 Assumption 2 1 000 Assumption 3 1 000 Figure 8 8 Correlation matrix If you enter inconsistent correlations Crystal Ball tries to adjust the correlations so they don t conflict For more information on inconsistent correlations see Specifying correlations between assumptions on page 33 Correlation Matrix example In the Crystal Ball Examples folder there is a Portfolio Allocation xls workbook you can use to experiment with the Correlation Matrix tool This spreadsheet calculates the total expected return for an investment model In this example you will run a simulation without correlations and then add the correlations and rerun the simulation for comparison To run the Correlation Matrix tool 1 In Excel with Crystal Ball loaded open the workbook Portfolio Allocation xls Ep 2 Set the following options in the Run gt Run Preferences dialog On the Trials tab set Maximum Number Of Trials to 500 e On the Sampling tab set Random Number Generation to Use Same Sequence Of Random Numbers and set an Initial Seed Value of 999 e Also on the Sampling tab s
388. u select the Likeliest parameter you can estimate the Scale parameter The Scale parameter is a number greater than 0 The larger the Scale parameter the greater the variance To calculate a more exact scale you can estimate the mean and use the equation mean mode 0 57721 Crystal Ball User Manual Using continuous distributions where o is the Scale parameter Or estimate the variance and use the equation a 6 variance l 2 T where o is the Scale parameter Minimum extreme distribution Parameters Likeliest Scale Description The minimum extreme distribution is commonly used to describe the smallest value of a response over a period of time for example rainfall during a newem drought This distribution is closely related to the maximum extreme distribution described beginning on page 265 Define Assumption Cell Ai k lol x Edit View Parameters preferences Help Minimum Extreme Distribution Likeliest 0 00 E DE Ok Cancel Enter Gallery _Coretate Hep Figure A 12 Minimum extreme distribution Crystal Ball User Manual 267 Appendix A Selecting and Using Probability Distributions 268 Calculating parameters There are two standard parameters for the minimum extreme value distribution Likeliest and Scale The Likeliest parameter is the most likely value for the variable the highest point on the probability distribution or mode After you select
389. u want to freeze 3 When all data cells to be frozen are selected click OK Running simulations About running simulations When running a simulation in Crystal Ball you can stop continue or reset the simulation at any time You also have the option of observing each individual trial to see how the values change You can display a chart for each assumption or forecast cell or run the simulation with all chart windows closed You can select windows cascade windows open various windows at the same time or bring the spreadsheet or a chart window to the front of the other windows While the simulation is running Crystal Ball creates a forecast chart for each forecast cell using frequency distributions A frequency distribution shows the number or frequency of values occurring in a given group interval bin 90 Crystal Ball User Manual CO crystal Ball Running simulations A Forecast Net Profit MM Fi e oj x Edit Mew Eorecast Preferences Help 500 Trials Frequency View 498 Displayed Net Profit MM T w N T ho oO T T T a d X Aouenbel4 T ae N 10 0 0 0 10 0 20 0 30 0 Millions gt mim Certainty 100 q Jrintinity Figure 4 9 Net Profit Forecast dialog In Chapter 5 you can find a more detailed explanation on how to interpret the forecast chart For now remember that the chart displays the forecast results showing you how the forecast values are distributed and th
390. ual 253 Appendix A Selecting and Using Probability Distributions Table A 2 Summary of continuous distributions Continued Shape Name Summary Gamma The gamma distribution applies to a wide range of physical quantities and is page 259 related to other distributions lognormal exponential Pascal Erlang Poisson and chi squared It is used in meteorological processes to represent pollutant concentrations and precipitation quantities The gamma distribution is also used to measure the time between the occurrence of events when the event process is not completely random Other applications of the gamma distribution include inventory control economics theory and insurance risk theory Logistic The logistic distribution is commonly used to describe growth the size of a page 262 population expressed as a function of a time variable It can also be used to m describe chemical reactions and the course of growth for a population or individual Lognormal The lognormal distribution is widely used in situations where values are A Basic positively skewed for example in financial analysis for security valuation or m page 263 0 real estate for property valuation amp Maximum The maximum extreme distribution is commonly used to describe the A extreme largest value of a response over a period of time for example in flood flows Mates page 265 rainfall and earthquakes Other application
391. um Maximum 13 951 523 14 076 737 14 076 737 14 201 952 14 201 952 14 327 166 14 327 166 14 452 381 14 452 381 14 577 595 14 577 595 14 702 809 14 702 809 14 828 024 14 828 024 14 953 238 14 953 238 15 078 453 15 078 453 515 203 667 15 203 667 15 328 881 15 328 881 15 454 096 87 Growth Year 1 Q1 5 88 Growth Year 1 Q2 0 31 0 26 89 Growth Year 1 Q3 0 23 0 23 90 Growth Year 1 Q4 0 26 0 24 91 Growth Year 2 Q1 0 31 0 25 92 Growth Year 2 Q2 0 22 0 21 93 Growth Year 2 Q3 0 28 0 28 94 Growth Year 2 Q4 0 45 0 41 95 Growth Year 3 Q1 0 46 0 45 96 Growth Year 3 Q2 0 05 0 43 97 Growth Year 3 Q3 0 00 0 03 98 Growth Year 3 Q4 0 02 0 01 99 Coil Diameter in _ _ 100 Material 1 Strength _ Figure 7 12 A sample of extracted data Sensitivity Data format Crystal Ball User Manual 195 Chapter 7 Creating Reports and Extracting Data ee Sales Year 3 Q1 Ending Sales Year 3 Q2 1 18 849 027 19 620 035 2 16 454 224 16 645 784 3 16 048 233 16 565 879 4 14 838 034 14 473 412 5 14 556 109 14 399 614 6 16 234 351 16 397 570 7 16 924 035 18 552 808 8 16 344 792 16 678 518 Figure 7 13 A sample of extracted data Trial Values format 196 Crystal Ball User Manual Crystal Ball Tools In this chapter This chapter describes the Crystal Ball tools e Batch Fit tool Correlation Matrix tool e Tornado Chart tool Bootst
392. um value and the fifth column contains the Step value for discrete sloping ranges For continuous sloping ranges the fifth column Step is left blank Note that if there are uniform discrete ranges their first three columns contain the Minimum Maximum and Probability as in a four column format but the fourth column is left blank and Step is entered in the fifth column Height of Min Discrete Value 0 5 Probability 1 0 Continuous Range 1 2 Uniform Probability 1 0 Discrete Value 2 5 Probability 5 Continuous Range 3 5 Uniform Probability 4 Continuous Range 5 7 Sloping Probability 3 6 Continuous Connected Range 7 10 Sloping Probability 6 8 Discrete Range 12 14 Uniform Probability 2 Step 0 5 Discrete Range 15 17 Sloping Probability 1 4 Step 1 pil Sheet1_ Sheet2 Sheet3 Figure A 43 Mixed ranges including sloping ranges Crystal Ball User Manual Using the custom distribution CZE Assumption Cell 119 oO fe o x Edit View Parameters Preferences Help Name pooo O Custom Distribution elative Probabilit ail 10 00 12 00 14 00 16 00 Minimum Maximum Height of Min Height of Max Step FEEF 0 50 1 00 1 00 2 00 1 00 2 50 5 00 3 00 5 00 4 00 5 00 7 00 3 00 6 00 12 00 14 00 2 00 0 50 15 00 17 00 1 00 4 00 1 00 OK Cancel _ Enter Galle Comelate He
393. ummary dialog appears enter the new text in the edit field and click OK The new text immediately replaces the original Summary text To modify Description text for a distribution 1 Select the distribution in the Distribution Gallery 2 Be sure View gt Show Description is checked 3 Right click the text in the Description pane 4 In the right click menu choose Modify Description For distribution name 5 When the Modify Description dialog appears edit the description in the edit field and click OK The new text immediately replaces the original description Crystal Ball Note Alternately you can just select the description text to modify and edit it Be careful to avoid modifying descriptions accidentally Deleting distributions You can delete a distribution in the Favorites category or other user defined category However the original distribution of that type remains in Basic or To delete a distribution choose Edit gt Delete in the Distribution Gallery menubar or right click the distribution and choose Delete Crystal Ball User Manual Additional Distribution Gallery features Setting up distributions for printing To select paper print orientation and margins before printing 1 2 3 4 Double click a distribution in the Distribution Gallery to display it in the Define Assumption dialog Choose Edit gt Page Setup in the Define Assumption menubar Set paper size and source orientation
394. use short term or long term formulas depending on the time span of your model 4 Optionally specify a Z score shift value to be used in long term formulas You can specify a value from 0 to 100 Crystal Ball Note The default is 1 5 If don t want to use a Z score shift value enter 0 in the box or delete the current value and leave it empty 5 Indicate whether metrics should be calculated from a fitted distribution or directly from the forecast values For details see the next section Calculation Method 6 When settings are complete click OK Calculation Method By default Crystal Ball tries to fit a normal distribution to the forecast values You can enter a significance level to specify the threshold below which the assumption of normality is rejected The default level of 0 05 translates into a 95 confidence that a rejection of normality will be correct Other significance levels typically used are 0 01 0 025 and 0 1 which translate into 99 97 5 and 90 confidences respectively Crystal Ball User Manual Setting specification limits and targets If normality is rejected Crystal Ball will then either calculate the metrics directly from the forecast values the default or if you choose perform a best fit to select the most appropriate continuous probability distribution from which to calculate the metrics Crystal Ball Note The normality test and non normal best fit if normality is rejected
395. user defined macros In Extreme speed it is not possible to run user defined macros such as CBBeforeTrial CBAfterTrial and CBAfterRecalc during a simulation Before and after simulation macros are allowed such as CBBeforeSimulation and CBAfterSimulation If the former group of simulation macros are present they will be flagged as incompatible e Example message One or more user defined simulation macros were detected e g CBBeforeTrial CBAfterTrial etc Code 5701 e Workaround Use Normal speed to run these user defined macros during a simulation Special functions A small group of Excel functions are not supported in Extreme speed CALL CELL GETPIVOTDATA INFO HYPERLINK REGISTER ID Any forecast formulas that contain one or more of these functions will be flagged as incompatible Example message Unsupported Excel or Crystal Ball function at cell location Book1 xls Sheet 1 A1 Code 5539 e Workaround Avoid using these functions if you want to run the model in Extreme speed Crystal Ball User Manual Compatibility issues Undocumented behavior of standard functions Excel allows certain argument constructs for standard functions for example SUMPRODUCT A1 A5 B1 B5 where SUMPRODUCT normally expects two arguments These argument constructs are not supported in Extreme speed and will be flagged as incompatible Example message Unsupported Excel or Crystal Ball function at cell loca
396. usly defined for this assumption appears as in Figure 2 10 Oefine Assumption Cell D13 p is jol x Edit View Parameters Preferences Help Name Inflation Rate xi A Lognormal Distribution p gt o oo x 4 infinity xj Mean po 3 Std Dev mm Sf OK Cancel Enter Galley _Conelte Help Figure 2 10 Lognormal distribution 3 Click Correlate The Define Correlation dialog appears as in Figure 2 11 34 Crystal Ball User Manual Additional assumption features Define Correlation Cell D13 con ik Sai aie Coefficient Assumption Figure 2 11 Define Correlation dialog 4 Click Choose to select the second assumption from the Choose Assumptions dialog Crystal Ball Note You can only correlate assumptions in the same workbook The Choose Assumptions dialog provides a list of the names of all the assumptions defined in your workbook Choose Assumptions Figure 2 12 The Choose Assumptions dialog Tree view By default the dialog appears in a hierarchical Tree view If you prefer you can click the List icon to change it to List view shown in the next figure Crystal Ball User Manual 35 Chapter 2 Defining Model Assumptions gt Choose Assumptions Figure 2 13 The Choose Assumptions dialog List view 5 Choose one or more of the assumption names on the list and click OK The cell reference or name of the assumption appears in the list in
397. ution Gallery This chapter describes how to define decision variable cells and forecast cells and perform other tasks needed to complete a model definition Defining decision variable cells Decision variables are not required for simulation models but they can be useful when comparing and optimizing alternate scenarios Several of the Crystal Ball tools discussed in Chapter 8 Crystal Ball Tools use and benefit from decision variables Decision variables are the variables that you can control such as rent to charge or the amount of money to invest OptQuest Note You also use decision variables with OptQuest available with Crystal Ball Professional and Premium Editions To define a decision variable cell 1 Select a cell or range of cells Select value cells or blank cells only You cannot define a decision on a formula or non numeric cell pe 2 Select Define gt Define Decision The Define Decision Variable dialog appears Define Decision Variable Cell D10 E x Name D10 EF Bounds Type towe fT Continuous C Discrete Upper poo g Step px To yg coat tee Figure 3 1 Define Decision Variable dialog 58 Crystal Ball User Manual Defining decision variable cells and forecast cells Complete the Define Decision Variable dialog e Name is the name of the decision variable Bounds are the upper and lower limits for the decision variable range Type defines whether the
398. ution of a sample statistic approaches a normal distribution making the calculation of the statistic s standard error or confidence interval relatively easy However when a statistic s sampling distribution is not normally distributed or easily found these classical methods are difficult to use or are invalid wll Figure 8 21 Sampling distribution of a mean statistic In contrast bootstrapping analyzes sample statistics empirically by repeatedly sampling the data and creating distributions of the different statistics from each sampling The term bootstrap comes from the saying to pull oneself up by one s own bootstraps since this method uses the distribution of statistics itself to analyze the statistics accuracy There are two bootstrap methods available with this tool Table 8 5 Supported bootstrap methods One simulation method Simulates the model data once creating the original sample and then repeatedly resamples those simulation trials the original sample values Resampling creates a new sample from the original sample with replacement That is it returns the selected value to the sample before selecting another value letting the selector possibly reselect the same value It then creates a distribution of the statistics calculated from each resample This method assumes only that the original simulation data accurately portrays the true forecast distribution which is likely if the sample is large
399. utions and a mechanism for determining the combined effect of the probability distributions on the model s output Monte Carlo simulation probability Classical Theory The likelihood of an event probability distribution also distribution A set of all possible events and their associated probabilities random number A mathematically selected value which is generated by a formula or selected from a table to conform to a probability distribution random number generator A method implemented in a computer program that is capable of producing a series of independent random numbers range The difference between the largest and smallest values in a data set rank correlation also Spearman s rank correlation A method whereby assumption values are replaced with their ranking from lowest value to highest value using the integers 1 to N prior to computing the correlation coefficient This method allows the distribution types to be ignored when correlating assumptions 368 Crystal Ball User Manual relative probability also relative frequency A value not necessarily between 0 and 1 that indicates probability when used in a proportion reverse cumulative frequency distribution A chart that shows the number or proportion or percentage of values greater than or equal to a given amount risk The uncertainty or variability in the outcome of some event or decision seed value The first number in a sequence of ran
400. utions to your data series automatically Available Distributions Figure 8 2 Select Distributions dialog Batch Fit tool 3 Make sure all the distributions are in the Selected Distributions list 4 Click Next The Select Input Options dialog appears as shown in Figure 8 3 with data filled in from steps 5 through 11 Crystal Ball User Manual Batch Fit tool Select input options step 2 of 3 Batch Fit Crystal Ball Tool Decisioneering 1998 2004 Location of Data Series must be in contiguous rows and columns Sales Data lSA 1 D 361 ta Series Orientation 7 methods Data in rows Anderson Darling ji C Chi Square I First row contains headers Kolmogorov Smirnov lt Back Cancel H Figure 8 3 Select Input Options dialog Batch Fit tool 5 Click the Select Cells icon to the right of the Location Of Data Series field 6 Select the Sales Data worksheet 7 Select cells A1 through D361 f 8 Click the Return icon to return to the Batch Fit dialog 9 Select Data In Columns the default 10 Check the First Row Contains Headers option 11 Select the Ranking Method Chi Square 12 Click Next The Select Output Options dialog appears Crystal Ball User Manual 201 Chapter 8 Crystal Ball Tools 202 13 15 16 22 Select output options step 3 of 3 Batch Fit Crystal Ball Tool Decisioneering 1998 2004 Output Location Create output on a new workboo
401. ve any of the views from the window open the View menu and uncheck each view you want to close 124 Crystal Ball User Manual Understanding and using forecast charts Setting forecast preferences You can set a number of specific forecast preferences to customize how Crystal Ball calculates and displays your forecast charts These are in addition to the general chart preferences discussed starting on page 128 Table 5 3 summarizes forecast features you can control with preference settings and refers to the table or section that describes each setting Table 5 3 Available forecast preference settings How Do I Described in Change the forecast chart view Table 3 1 on page 63 Determine when the forecast window appears Table 3 1 on page 63 Fit a distribution to the forecast Table 3 1 on page 63 Set precision controls for forecast statistics Table 3 2 on page 65 Filter ranges of forecast values Table 3 3 on page 67 Automatically extract forecast data to a Table 3 4 on page 69 spreadsheet Basic instructions for setting forecast preferences Like the chart preferences forecast preferences can be set differently for each forecast chart To set forecast preferences 1 Choose Preferences gt Forecast in a forecast chart s menubar The Forecast Preferences dialog appears as shown in Figure 5 18 Crystal Ball User Manual 125 Chapter 5 Analyzing Forecast Charts 126 Forecast Preferences ma xi
402. versions of Crystal Ball However you can only choose one at a time in the Restore Results dialog Using restored results Once you have restored one or more Crystal Ball results files you can open and close restored charts create new reports using them and extract their data to spreadsheets You can create overlay and trend charts with restored results and results from the current simulation to compare data The results appear in dialogs following those for the current simulation To remove restored results from memory choose Analyze gt Close All Running user defined macros Crystal Ball 5 x supported user defined macros with the Macros tab of the Run Preferences dialog Macro names could be entered into four fields of the dialog to indicate when they should run at the start of the simulation before each trial after each trial or at the end of the simulation You can now run user defined Excel VBA macros automatically during a simulation just by naming them with predefined names Table 4 8 User defined macros Macro Name Runs CBBeforeSimulation Immediately after the Start Simulation command or Single Step command CBBeforeTrial Before random numbers have been placed in assumption cells CBAfterRecalc After Excel has recalculated the model but before a trial value has been retrieved from the forecast cells CBAfterTrial After the forecast trial values have been retrieved and entered into the forecast
403. w Parameters Preferences Help Name a2 EF Uniform Distribution 500 000 00 600 000 00 000 800 000 900 000 00 Minimum 500 000 00 Maximum fs900 000 00 Fg od ee ewer Sater _Corette _Hep Figure A 17 Uniform distribution The distribution in Figure A 17 shows that all values between 500 000 and 900 000 are equally possible Example two A manufacturer determines that he must receive 10 over production costs or a minimum of 3 per unit to make the manufacturing effort worthwhile He also wants to set the maximum price for the product at 6 per unit so that he can gain a sales advantage by offering the product for less than his nearest competitor All values between 3 and 6 per unit have the same likelihood of being the actual product price The first step in selecting a probability distribution is matching your data with a distribution s conditions Checking the uniform distribution The minimum value is 3 per unit The maximum value is 6 per unit All values between 3 and 6 are equally possible You would enter these values in Crystal Ball to produce a uniform distribution showing that all values from 3 to 6 occur with equal likelihood Crystal Ball User Manual 277 Appendix A Selecting and Using Probability Distributions Weibull distribution also Rayleigh distribution Parameters Location Scale Shape Description The Weibull distribution describes data
404. you select this option before you run a simulation Enable Correlations Activates any defined correlations between assumptions Run User Defined Macros Runs any user defined macros as part of the simulation process For details see Running user defined macros beginning on page 99 86 Crystal Ball User Manual Setting run preferences Table 4 4 Options tab Run Preferences dialog Continued Setting Effect Warn If Insufficient When checked issues a warning dialog if you don t have enough Memory memory to complete a full simulation This dialog lists several options to consider such as reducing the number of simulation trials and turning off the saving of assumption values Saving assumption values is required for calculating sensitivity charts but it can consume a significant amount of memory for large models Show Control Panel When checked activates the Crystal Ball Control Panel For more information see The Crystal Ball Control Panel on page 93 Leave Control Panel When checked continues to display the Open On Reset Control Panel after a simulation is reset Statistics preferences Run Preferences Trials Sampling Speed Options f m Calculate percentiles as Probability below a value Probability above a value m Format percentiles as 10 90 etc P10 P90 etc IV Calculate capability metrics Options Figure 4 6 Statist

"user manual"

Contents

Download Pdf Manuals

Related Search

Related Contents

&quot;user manual&quot;

Contents

Download Pdf Manuals

Related Search

Related Contents

"user manual"