Palisade Knowledge Base

HomeTroubleshootingPrecisionTree#N/A for Some Iterations When Simulating a Decision Tree

12.4. #N/A for Some Iterations When Simulating a Decision Tree

Applies to:
PrecisionTree 6.x/7.x when used with @RISK

I have @RISK distributions for one or more branch values or probabilities in my decision tree. I've noticed that when I do a simulation, each of these has a value in some iterations but #N/A in other iterations. When a simulation is not running, all the distributions show values.

In PrecisionTree, go into Model Settings, specifically the @RISK tab. Under "Each @RISK iteration calculate", "Values of one sampled path through the model" is probably selected. That means that @RISK randomly selects one path through the tree in each iteration; distributions in other paths aren't calculated, and therefore their vales are #N/A.

If you change that to "Expected values of the model", then during each iteration all values in the tree will be recalculated, and the value of the tree's root node (which you probably have as an @RISK output) will be calculated based on the optimal path in that iteration.

The PrecisionTree help topic "@RISK Recalculation Options" explains the two choices in a bit more detail, and "Decision Forcing During @RISK Simulation Options" explains the three choices in the second section of the dialog. However, there's a much more comprehensive explanation in the PrecisionTree example file "Using @Risk with PrecisionTree.xlsx", under Help » Example Spreadsheets.  A copy of that file is attached to this article, for your convenience.

Last edited: 2017-09-29


This page was: Helpful | Not Helpful