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 Microsoft Excel or other analysis tools.
The following steps can help guide the simulation analysis by focusing on details as well as general trends:
Consider each forecast chart from a high-level viewpoint. Look at the shape of the distribution:
Is it distributed normally or skewed negatively or positively?
Is it "flat" (spread out on both sides of the mean) or "peaked" (with most values clustered closely around the mean)?
Does it have a single mode (most likely value) or is it bimodal with several "peaks" or "humps"?
Is it continuous or are there groups of values separated from the rest, maybe even extreme values that fall outside of the display range?
The statistical concepts in the "Statistical Definitions" chapter of the online Oracle Crystal Ball Statistical Guide can help with this part of the analysis.
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 explains these and other ways of analyzing certainty.
You can change the display range to focus on different sections of 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 discusses this analysis technique.
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.
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 describes the many ways you can format charts for analysis and presentation of results.
You can display overlay charts (Understanding and Using Overlay Charts), trend charts (Understanding and Using Trend Charts), sensitivity charts (Understanding and Using Sensitivity Charts), assumption charts (Understanding and Using Assumption Charts), and scatter charts (Understanding and Using Scatter Charts). 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 with charts and data (Creating Reports). 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 Microsoft Excel for further analysis or for further export into other analytical tools. See Extracting Data 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 Crystal Ball Tools. You can use Predictor for time-series analysis and, if you have OptQuest, you can use it to optimize simulation results.