LAB #2911 NEUTRAL: LEDGER BENCH: Apply filter: Session = LONDON_OPEN — improves win rate from 31.1% to 31.7% (+0.6pp, n — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Session = LONDON_OPEN — improves win rate from 31.1% to 31.7% (+0.6pp, n=252) SAME-WINDOW EFFECT: Applying the LONDON_OPEN filter (likely 08:00–12:00 UTC+2, given entries) would have excluded all 10 live trades listed — every entry falls between 10:10 and 11:35 local time, which is inside the London session. Therefore, no trades would be removed; the filter changes nothing in this window. The backtest gain (+0.6pp) is not observable here, and PnL delta is zero — the filter is redundant for this sample, not harmful but unproven. OVER-TIME PROJECTION: Over 3–6 months, a +0.6pp win-rate improvement on n=252 is statistically weak (standard error ~2.8pp), so the true effect is likely noise. The filter may reduce trade frequency by excluding Asian/NY sessions, which could lower total PnL even if win rate rises slightly — a classic selection bias trap. Risk: overfitting to a single session's microstructure Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com