LAB #4901 NEUTRAL: LEDGER BENCH: Apply filter: Session = LONDON_OPEN — improves win rate from 31.9% to 35.7% (+3.9pp, n — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Session = LONDON_OPEN — improves win rate from 31.9% to 35.7% (+3.9pp, n=392) SAME-WINDOW EFFECT: The filter would have blocked most of the live window's trades, because nearly every entry timestamp sits in the 10:25–11:45 CET band — which is the London session, not the London open. If "LONDON_OPEN" is defined as the first 1–2 hours of London (roughly 08:00–10:00 CET), then the EURUSD, XAUUSD, AUDUSD, NZDUSD, GBPUSD, EURJPY, USDCAD and USDCHF entries at 10:25–11:45 would all have been filtered out. That removes three closed losses (EURUSD, XAUUSD, AUDUSD) and two closed wins (USDCHF, USDCAD) — a net improvement in this tiny sample, but the sample is 5 closed trades, so the PnL delta is noise, not evidence. OVER-TIME PROJECTION: A +3.9pp win-rate lift on n=392 is statistically weak — the 95% CI on that delta spans roughly -1pp to +9pp, so it is not distinguishable from zero. Over 3–6 months, if the e Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com