01Hold the total constant
Equal completed units can hide different queues and WIP.
All worksheetsIllustrative synthetic data, version 1.0
Same output. Opposite production decisions.
Look beyond one headline total and compare the decision-driving distribution.
Synthetic 478 / 478 comparison
- Scenario A
- 478 completed · inspection relief ranks first
- Scenario B
- 478 completed · release control ranks first
- Shared total
- Does not identify the action
- Required evidence
- Inputs, code version, seed, limits
- Status
- Illustrative, not a calibrated factory model
Get the simulation evidence test
How to use it.
Start here: write the production decision before choosing a model output.
02Inspect the distribution
Rank actions using the process behavior that drives the decision.
03Preserve a reproducible run
Save inputs, version, seed, results and limitations.