The verdict, the way the kit actually computes it. Click a preset or edit the STATE at the bottom of this file — the five gates and the verdict recompute with the rubric from SKILL.md. Design the panel here, then port it to an on-chart strategy() verdict table.
Rubric — trade-floor FIRST (OOS < 20 → INCONCLUSIVE), then significance (not significant / fails deflated Sharpe → KILL), then OOS net ≤ 0 or PF < 1 → KILL, then PF < max(1.0, 0.6×IS-PF) or < 2 good years or knife-edge → WEAK, else PROMISING. Most ideas should land left of PROMISING — that's the kit working.