MODEL ADOPTION #410: REJECT — LEDGER BENCH: Apply filter: XAUUSD LONG — improves win rate from 32.4% to 61.2% (+28.8pp, — E8 Intelligence Research

DECISION: REJECT ELEMENT: The filter "XAUUSD LONG only" introduces a directional bias that violates E8's polarity-agnostic geometry. E8 rhythm logic is designed to exploit timing edges independent of asset direction; hardcoding a long-only rule for any instrument is a fundamental mismatch with the model's architecture. The claimed +28.8pp win rate is based on n=67 — too small to distinguish signal from regime overfit. The "same-window effect" analysis shows zero PnL impact from the only relevant trade, meaning the edge is entirely hypothetical. EXPECTED IMPACT: Negligible to negative. In a sustained bear trend, the filter would force the model to skip profitable shorts and likely take losing longs, degrading overall win rate and increasing drawdown. The backtest improvement is likely a artifact of recent gold bullishness, not a structural edge. RISK: Regime shift in XAUUSD (e.g., rate hikes, dollar strength) would cause the filter to systematically reject profitable shorts while for Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com