v18 model prints STRESS regime for at least 3 consecutive days in Q3 2026

Resolution

VOID — the predicate is not evaluable as written. Measured from v18_regime_state (canonical_regime.py: the single source of truth for subscriber-facing v18 reads), Q3 2026 window 2026-07-01..2026-09-30, 50 rows: the label 'STRESS' NEVER APPEARS. v18 emits DEFENSIVE (ANCHOR A) - GATE OFF (19 rows) and RISK-ON - n% TOWARD ANCHOR C (31 rows across 11 variants). 'STRESS' is the regime STACK's vocabulary, not v18's. WHAT THE DATA DOES SAY: DEFENSIVE/GATE OFF ran 17 CONSECUTIVE rows, 2026-07-21..2026-08-13, across 18 weekdays (one weekday absent), so wherever that single gap falls both halves still exceed 3 consecutive trading days. THEREFORE: if DEFENSIVE/GATE OFF is stamped as satisfying 'STRESS', this resolves YES with five times the required margin. That mapping is a methodology decision and has not been made, so the call is voided rather than credited. NOT resolvable NO: 'longest STRESS run = 0' reflects a vocabulary mismatch, not an absence of stress. Reopen with a stamped mapping. Series coverage caveat: 50 rows over 66 Q3 weekdays (75.8%).