p13 familywise audit (final)

Episode outcomes are permuted between δ conditions within each candidate, the complete data-dependent gate re-applied inside every permutation, and the familywise maximum tested. The first pass (R2, B=2,000) gave p = 0.0525 ± 0.0050 and downgraded p13 to “suggestive”; the final B=200,000 audit sharpened this to 0.059230 (historical gate), 0.043455 (percentile bootstrap — reported symmetrically, cannot create prospective confirmation), and 0.773206 (conservative exact gate, under which p13 is ineligible and p05/s2a leads at +0.0833). The registered procedure’s empirical familywise false-fire rate under the null was 12.9%.