A past stop pair score rejected a model of independent stops with common trial rates in 226 AOMIC PIOP2 v2.0.0 runs

We built and checked a fixed pair score for future stop assignments and past task history. The score needs no known stop probability. Under conditional assignment symmetry within each pair, it has zero conditional mean. A normal-mixture evidence process gives a 5 percent bound on any threshold crossing for a fixed rule. We ran 10,000 experiments in each of seven settings. Any-prefix rejection rates were 0.56 to 2.47 percent in four settings that met the assumptions. A conventional same-run group contrast test rejected in 25.69 to 100 percent of those experiments. A public-data audit used all 226 AOMIC PIOP2 v2.0.0 stop-task runs. A signal made only from past stop counts had final evidence 1,625.47 and maximum evidence 3,903.97 against the common independent stop model. The exploratory response-time score crossed early but had weak final evidence. The exact source assignment rule remains unknown. These results do not establish a cognitive cause. This follow-up was selected after prior results on the same records, so public evidence is nominal for the fixed rule. The package preserves the proof, 233 verified source files, retrieval recipe, code, all scores, 70,000 simulation outcomes, figures, and repeat instructions. Exact checks, independent calculations, and a full fresh-source replay passed.

sequential testing · conditional symmetry · e-process · stop-signal task · task history · AOMIC PIOP2 · simulation