MODEL ADOPTION #977: REJECT — LEDGER BENCH: Apply filter: USDCHF SHORT @ ASIAN with 5 conf — improves win rate from 31.6 — E8 Intelligence Research

DECISION: REJECT ELEMENT: The filter "block USDCHF SHORT @ ASIAN with confidence <5" fails because the backtest sample (n=19) is too small to distinguish signal from noise, and the live window shows it would have removed a confirmed winner (rizzy5m entry 0.81212, CLOSED/WIN). The +31.5pp win-rate improvement is likely overfitted to a specific regime; no E8 geometry timing constraint is embedded in the rule, making it a blunt session-based filter incompatible with the rhythm logic. EXPECTED IMPACT: Negligible or negative. In the current live window, net PnL would drop by the amount of the blocked winning trade. Over 3–6 months, the filter might reduce some Asian-session losses, but the opportunity cost from blocking winners and reduced trade frequency would likely offset any win-rate gain. The 55% confidence from Merlin reflects high uncertainty. RISK: Overfitting to a small sample (n=19). The filter is session-based, not tied to E8 node timing or polarity rhythm, so it could block t Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com