MODEL ADOPTION #367: REJECT — LEDGER BENCH: Apply filter: XAUUSD LONG @ NY_OPEN with 5 conf — improves win rate from 32. — E8 Intelligence Research

DECISION: REJECT ELEMENT: The filter "XAUUSD LONG @ NY_OPEN with 5 conf" fails because it is a **calendar/time-window heuristic**, not a geometry-timing edge. E8-node/Gods-Clock rhythm logic already encodes session boundaries via phase harmonics — adding a fixed "NY_OPEN" rule introduces redundancy without structural gain. The 43.8pp win-rate lift is based on n=21, which is statistically fragile (p > 0.05 under binomial test for 76.2% vs 32.4% baseline). Overfitting to a small sample is the most likely explanation. EXPECTED IMPACT: Zero to negative. In the current trade window, the filter would have changed nothing (no XAUUSD signals). Over 3–6 months, it would reduce trade frequency by ~80% on XAUUSD, likely missing profitable London-session breakouts and trend continuations. Real-world slippage on XAUUSD at NY open (spreads often 2–3x wider than during London) would erode any theoretical win-rate gain. The net effect is lower total PnL with higher variance. RISK: (1) Overfitting t Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com