Skip to main content
Simulation is the step that turns your ranges into a spread of outcomes.

​
What happens in a run

  1. The app draws one plausible set of values for every estimate, respecting your ranges and any correlations. That is one possible future.
  2. It works out the outcome of every option in that same future.
  3. It repeats this tens of thousands of times (40,000 to start).
  4. If the ranking of the options is still wobbling, it automatically runs more futures, up to 200,000, until the answer settles. You never choose how many.
Because every option faces exactly the same futures, differences between options are real and not the luck of the draw.

​
Running it

Go to Results and choose Run simulation. Reviewers and above can press it. The run can take several seconds and ends with a “Summarizing…” step that uses the AI model.
The run is blocked if any estimate is not yet known or a “must” constraint fails. The message names the cause.

​
What the simulation also allows for

  • Launch delay and a slow first-year ramp-up
  • Benefits that fade over time
  • Discounting: money later is worth less than money now
  • Year-by-year models of up to 30 years, with a full yearly cash-flow statement

​
Reproducible

Every run records how it was produced: the random seed, the number of futures, the software versions and a fingerprint of the model. Anyone can reproduce it. Runs of the same model with a different seed differ slightly, because each is a different sample of futures.
A precise simulation does not show the assumptions are right. Check the inputs before trusting the output.
Next: Set up win criteria