Different Results When Simulating Multiple Workbooks in One @RISK Session

Applies to: @RISK 5.x and Up

Issue:
You are simulating multiple copies of a workbook in a single @RISK session. There are no links between the workbooks, so you expect each workbook to produce the same results it would if simulated alone. However, the results differ. Why?

Explanation

All open workbooks participate in a single @RISK simulation. Random numbers for each iteration are drawn from a single shared stream, which you can set using a fixed seed on the Sampling tab in Simulation Settings.

• Example: A workbook has 100 distribution functions.

  • Simulating the workbook alone:

    • Iteration 1 → random numbers 1–100

    • Iteration 2 → random numbers 101–200

    • Iteration 3 → random numbers 201–300

  • Simulating two identical workbooks together:

    • Iteration 1 → random numbers 1–200 (first workbook uses 1–100, second uses 101–200)

    • Iteration 2 → random numbers 201–400

    • Iteration 3 → random numbers 401–600

Thus, results differ, though they remain statistically consistent.

Solution: Use RiskSeed( )

To ensure identical results for multiple workbooks:

• Add the RiskSeed() property function to your distributions.

• RiskSeed() gives a distribution its own independent random-number sequence.

• Example:

=RiskTriang(15, 30, 70, RiskSeed(1, 271828))

• With the same seed, identical distributions produce the same random sequence iteration by iteration, regardless of other workbooks or formulas.

Note: RiskSeed() does not work with correlated inputs.

Summary:
Different results occur because multiple workbooks share the same random-number stream. Use RiskSeed() to give inputs independent streams for repeatable, identical results.

 

Related Article:

Random Number Generation, Seed Values, and Reproducibility