LAB #4959 NEUTRAL: LEDGER BENCH: Apply filter: Session = OFF_SESSION — improves win rate from 31.7% to 41.9% (+10.1pp, — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Session = OFF_SESSION — improves win rate from 31.7% to 41.9% (+10.1pp, n=129) SAME-WINDOW EFFECT: The filter would have removed every trade in the live window, because all ten entries are timestamped between 10:25 and 11:45 CET — squarely inside the London/NY overlap, i.e. ON_SESSION by any standard definition. That means the three CLOSED/LOSS trades (EURUSD, XAUUSD, AUDUSD) would have been skipped, turning a 2W-3L window into 2W-0L on executed trades. But it also kills the two CLOSED/WINs (USDCHF 09-04, USDCAD 09-03), so the realized PnL delta is not the clean +10.1pp the ledger implies — you're deleting winners alongside losers, and the net depends entirely on whether the losers were larger than the winners. OVER-TIME PROJECTION: A +10.1pp win-rate lift on n=129 is a real signal only if the OFF_SESSION bucket isn't a thin, regime-specific slice — 129 trades over an unknown sample is small enough Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com