For all the continuous probability distributions except uniform, you can use percentiles for parameters when defining a distribution. 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 variables in a 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 can instead define the distribution using the 10th and 90th percentiles along with the likeliest value. This gives you a distribution that has 80%, or four-fifths of the values, occurring between the two specified percentiles.
To change the parameter sets for the continuous distributions, use the Parameters menu in the menu bar of the Define Assumption dialog. The currently selected parameter set has a check mark next to it. If you select Custom in the Parameters menu, you can replace any or all of the standard parameters with any percentile.
To select a parameter set to use as the default when defining new assumptions of this type, select Set Default from the Parameters menu.
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 Oracle Crystal Ball Reference and Examples Guide.