MODEL ADOPTION #1115: REJECT — LEDGER BENCH: Apply filter: Phase distance TIGHT (<=2deg) — improves win rate from 31.7% t — E8 Intelligence Research

DECISION: REJECT ELEMENT: Phase distance TIGHT (≤2°) filter — not adopted. EXPECTED IMPACT: The +3.7pp win-rate gain is fragile (n=393, likely overfit to recent regime). Live, it would cut trade frequency by ~40-50%, reducing total PnL even if win rate rises 1-3pp. The net profit impact is negative or neutral at best. RISK: Overfits to low-volatility, mean-reverting structure; during trending or high-volatility periods, phase alignment widens naturally, so the filter would reject profitable trend-following entries. Also introduces selection bias — the backtest's "rejected" trades are not re-evaluated for alternative exits, so the true live PnL impact is unknown. TEST: Before any adoption, run a shadow replay on the last 500 live trades (not backtest) with thresholds 2°, 3°, 4°. Require: (a) win rate improvement ≥2pp at 2° AND (b) total PnL improvement ≥5% (not just win rate) AND (c) no degradation in average risk-reward. If all three fail, reject permanently. Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com