MODEL ADOPTION #3997: REJECT — LEDGER BENCH: Apply filter: XAUUSD LONG — improves win rate from 32.5% to 55.4% (+23.0pp, — E8 Intelligence Research

Abstract

DECISION: REJECT ELEMENT: No element survives — the XAUUSD LONG-only filter is a direction bias, not an E8 geometry timing edge. It contradicts the model's core rhythm logic, which treats direction as symmetric around the Gods-Clock phase. EXPECTED IMPACT: Win-rate lift is an artifact of a tiny sample (n=83) and a single live window (4/9 vs 4/10). Over 3–6 months, the filter would likely degrade profit in bearish gold regimes (missed SHORT wins) and add no structural edge — the 32.5% SHORT win rate is within noise for a volatile metal. Net effect: negligible to negative on PnL, with a cosmetic win-rate bump that hides real risk. RISK: Overfitting to recent gold range-bound behavior; regime shift to a downtrend would turn the filter into a systematic loss-avoider that also blocks all profitable shorts. Also, it does not fix the broader low win-rate across NZDUSD/USDJPY, so the model's core weakness remains untouched. TEST: Before any re-consideration, run a 12-month walk-forward Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com