MODEL ADOPTION #924: REJECT — LEDGER BENCH: Apply filter: USDCHF SHORT — improves win rate from 31.7% to 61.2% (+29.6pp, — E8 Intelligence Research

DECISION: REJECT ELEMENT: The filter "USDCHF SHORT only" is a pair-direction bias, not a timing or geometry rule. It does not align with E8-node/Gods-Clock rhythm logic, which is agnostic to instrument and direction. The edge is purely statistical and likely regime-dependent — no structural compatibility with the live model's core mechanics. EXPECTED IMPACT: +0% win-rate on live model because the filter would block non-USDCHF SHORT signals that E8 timing already validates, potentially discarding higher-probability entries. The reported +29.6pp is an artifact of a small, single-pair sample (n=165) and likely overfits to recent USDCHF trend behavior. RISK: Regime shift in USDCHF (e.g., trend reversal, volatility compression) would turn the filter into a net negative. Overfitting to one pair-direction also increases model fragility and reduces diversification. TEST: Replay-bench must run the filter on all E8 signals from the last 12 months across all pairs. If the filter improves Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com