LAB #1935 PROMISING: LEDGER BENCH: Apply filter: Session = OFF_SESSION — improves win rate from 31.9% to 44.9% (+13.0pp, — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Session = OFF_SESSION — improves win rate from 31.9% to 44.9% (+13.0pp, n=98) SAME-WINDOW EFFECT: Applying the OFF_SESSION filter would have removed the three rizzy5m entries at 11:00, 11:20, and 10:35 (all on 2026-07-31/08-03) — two of which were losses and one a win. That would have cut the live window from 9 trades to 6, removing 2 losses and 1 win, shifting the raw win rate from 4/9 (44.4%) to 3/6 (50.0%). PnL delta is positive but modest: you avoid two full losses (GBPUSD and XAUUSD) while forfeiting one EURJPY win — net expected gain roughly +1.2R assuming average risk/reward. The 13pp backtest improvement is directionally consistent with the live sample, but the n=98 backtest is small and the live window shows only a 5.6pp improvement — suggesting the filter helps but not as dramatically as claimed. OVER-TIME PROJECTION: Over 3–6 months, filtering to OFF_SESSION should reduce noise entries du Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com