LAB #4158 NEUTRAL: LEDGER BENCH: Apply filter: Session = LONDON_OPEN — improves win rate from 32.6% to 36.4% (+3.8pp, n — E8 Intelligence Research

Abstract

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Session = LONDON_OPEN — improves win rate from 32.6% to 36.4% (+3.8pp, n=357) SAME-WINDOW EFFECT: Applying the LONDON_OPEN filter to the 10 live trades above would have excluded all 5 wkjudas entries (all outside London open) and 3 of the 5 rizzy5m entries (NZDUSD 10:20, AUDUSD 10:15, XAUUSD 09:15 are pre-London). Remaining trades: GBPUSD SHORT rizzy5m (11:50) and GBPUSD SHORT rizzy5m (11:05) — both London open. Current window outcome: 2 wins, 3 losses, 1 open, 1 rejected, 1 pending → filtered subset would be 2 trades (1 WIN, 1 WIN) → win rate jumps to 100% on n=2, but PnL delta is negligible (+0.3R at best) due to tiny sample. The apparent +3.8pp backtest gain is not visible here — it's drowned by variance. OVER-TIME PROJECTION: Over 3-6 months, the filter will reduce trade frequency by roughly 40-50% (only ~2-3 hours of London session per day). If the true edge is +3.8pp, you'd need ~1,000+ filter Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com