MODEL ADOPTION #3591: REJECT — LEDGER BENCH: Apply filter: XAUUSD LONG — improves win rate from 32.4% to 55.4% (+23.0pp, — E8 Intelligence Research

DECISION: REJECT ELEMENT: No element survives — the XAUUSD LONG-only filter is a pair-direction bias, not an E8 geometry timing edge. It contradicts the model's core rhythm logic (which is direction-agnostic) and is derived from a tiny, regime-specific sample (n=83) with no demonstrated causal mechanism tied to Gods-Clock phase or node alignment. EXPECTED IMPACT: If adopted, win rate might rise in a gold-bull regime (as backtested), but this is a curve-fit to recent trend. Over a full cycle, it would block profitable shorts in gold downtrends, likely reducing net profit and increasing drawdown variance. The +23pp is not a stable edge — it's a regime artifact. RISK: Gold regime flip to bearish → filter blocks all short profits; also, the filter may correlate with hidden time-of-day or volatility factors that degrade other pairs' timing; model becomes less adaptive and more brittle. TEST: Before any reconsideration, run a 90-day shadow test with two paper arms (LONG-only vs SHORT Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com