LAB #4464 NEUTRAL: LEDGER BENCH: Apply filter: Session = LONDON — improves win rate from 31.5% to 36.4% (+4.8pp, n=22) — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Session = LONDON — improves win rate from 31.5% to 36.4% (+4.8pp, n=22) SAME-WINDOW EFFECT: Applying the LONDON session filter would have excluded all trades timestamped outside the London window (roughly 08:00–16:00 BST). In the live window above, the two wkjudas trades on 2026-09-01 (14:36 and 15:08) fall inside London, but the 02:43 NZDUSD trade is excluded — that one is still open/CLOSED with no clear result, so removing it is neutral. The rizzy5m/7m trades on 09-02 and 09-03 all occur between 09:56 and 11:50 local (+02:00), which is 08:56–10:50 London time — all pass the filter. Net effect: only one trade removed (the 02:43 NZDUSD), which was not a confirmed win/loss, so the win-rate improvement in this tiny sample is negligible (likely 0–1pp, not the +4.8pp claimed from n=22). The backtest gain is not visible in this live window because most trades already cluster in London hours. OVER-TIME PR Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com