DAVION

SYSTEMS

Distribution Types

 

An @Risk Monte Carlo formula defines and applies a probability distribution to your data.

The available distribution types are shown below. The choice of distribution depends on the type of uncertainty with which you are mainly concerned and the expected behavior of the costs.

 

Distribution

Expected Behavior

Typical Usage

Uniform

Closed

Uncertainty is mainly quantity uncertainty, and there is no established 'standard' price or rate.

Triangular

Closed

Reasonably well understood task in which there may be both quantity and scope uncertainty. An upper limit can be specified which is unlikely to be exceeded.

Beta

Closed

As for Triangular, but you are more certain about your estimate and less concerned about the outlying values

Normal

Open-Ended

Uncertainty is mainly quantity uncertainty, and an established 'standard' price or rate exists. Depending on the overrun probability you select, large divergences from this are possible.

Extended Triangular

Open-Ended

Scope uncertainty exists, depending on the overrun probability you select.  A hard upper limit exists, although this can be significantly greater than your specified contingency limit or high cost range.

Lognormal

Open-Ended

Significant scope uncertainty exists, depending on the overrun probability you select. No hard upper limit can reasonably be set.

 

 

Using the Lognormal Distribution

To create a lognormal distribution you need a base or mid-range cost, an upper cost limit, and an overrun probability. The upper cost limit and the overrun probability are 'built in' to the @Risk formula that Mandrel creates, and do not appear explicitly in the formula. If you change either the overrun probability or the upper cost limit (or more precisely, the size of the upper cost limit relative to the mid-range cost), you must re-create the @Risk formula.