LAB #4796 NEUTRAL: LEDGER BENCH: Apply filter: Session = LONDON — improves win rate from 31.8% to 34.8% (+3.0pp, n=23) — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Session = LONDON — improves win rate from 31.8% to 34.8% (+3.0pp, n=23) SAME-WINDOW EFFECT: Of the 10 trades in the live window, only the EURUSD SHORT (2026-09-11 10:55, still open) falls inside the London session; the rest are timestamped 10:25–11:45 CET, which straddles the London open but the filter as stated would need a precise hour definition to be applied. If "London" means the 08:00–11:00 London window, the filter would have rejected most of these entries — including the two closed winners (USDCHF +, USDCAD +) and the two closed losers (XAUUSD, AUDUSD) — netting roughly zero PnL change on closed trades while removing the open EURUSD exposure. The +3.0pp claim on n=23 is statistically meaningless: the 95% CI on a 3pp lift at that sample size spans roughly ±19pp, so it is indistinguishable from noise. OVER-TIME PROJECTION: Over 3–6 months, a session filter with n=23 backtest support is more li Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com