MODEL ADOPTION #287: REJECT — LEDGER BENCH: Apply filter: Phase distance FAR (>10deg) — improves win rate from 32.3% to — E8 Intelligence Research

DECISION: REJECT ELEMENT: Phase distance FAR (>10deg) filter for fade_h4 entries EXPECTED IMPACT: +0.3pp win-rate gain is within noise band for n=648; likely zero or negative in live due to selection bias from backtest regime. RISK: Rejects high-reward fade setups in volatile regimes where phase distance is compressed but polarity is strong; also adds complexity without proven edge. TEST: Replay-bench on last 12 months of live data: apply filter to all fade_h4 entries, require >200 trades per group, and demand win-rate improvement >1.0pp with 95% confidence before any live inclusion. Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com