submove / behavioral simulation / the numbers
The numbers behind submove
Last updated: June 2026
These are the figures submove is calibrated on, with what each one means and, just as important, what it does not. We publish them in one place so they can be quoted accurately. submove is calibrated, not validated.
What the model is calibrated on
| Figure | What it counts | What it does not mean |
|---|---|---|
| 1,102 | Independent studies from the published organizational-behavior research, pooled into the evidence base the model is calibrated against. | Not 1,102 studies of submove itself. They calibrate assumptions; they do not validate any prediction. |
| 29,624 | Participants in the largest single study pool among that evidence base. | Not submove's own users or a sample of any client's employees. |
| 200+ | Calibrated coefficients spanning 32 interacting behavioral states, resolved on every tick of a run. | More coefficients is not more certainty; the model still reports ranges, not guarantees. |
| 500 | Monte Carlo runs per decision at full engagement depth, so the output is a distribution of likely outcomes. | Not 500 real-world trials. The spread is modeled uncertainty, not observed frequency. |
Calibrated, not validated
Calibrated means these figures anchor the model's assumptions to a broad evidence base. Validated would mean the model had been proven to predict specific outcomes for specific organizations, which we do not claim. submove is a decision-support and rehearsal tool that describes likelihoods, not a forecast of any individual's behavior. We would rather be trusted than impressive. See what is a behavioral simulation engine for how the figures are used in a run.
How we report results
Every run is a Monte Carlo simulation, so outputs are ranges and likelihoods rather than single points. Where a client wants it, we offer pilot retrodiction: replaying a real, recent change and comparing simulated against observed outcomes. That is a credibility and calibration path we offer as an engagement, not a claim we present as already proven at scale.
See a change before you make it
If you are weighing a people-change, you can request a simulation and tell us the situation. Engagements are invitation-only. Read more about submove.