Three positive delay effects give upper-tail replay decisions in at most 4.1% of the synthetic cohorts with constant baseline probability.

In 2,000 simulated cohorts for each of three positive delay effects, a centered upper-tail replay decision selected at most 4.1% of cohorts across count groups of 16, 8, and 4 trials. The cohorts used independent staircases, observed starting delays, logistic response probabilities, and no baseline time drift. The nominal one-sided normal cutoff was 0.05. With zero direct delay effect, it selected 4.65%–4.85% of cohorts. Complete path-weight enumeration gave negative expected conditional excess in all nine positive-effect and count-width combinations. When response probability changed over time, the zero-effect decision rate reached 55.0% for 16-trial groups; only 4-trial groups had constant probability within every group in those models. A known-baseline score selected all positive-effect cohorts, but its baseline is known only in these simulations and it is not a proposed human-data test. The work also reproduces 27 exact real-data moment quantities from earlier reports using 8,512 public STOP trials. This is a specified-model method check, not a human causal estimate. Planned delay does not verify physical cue time, the model has no upper delay limit, and paired human staircases may share participant effects.

adaptive staircase · stop-signal task · conditional replay · synthetic simulation · statistical sensitivity · response probability · Monte Carlo