LAB #4891 NEUTRAL: LEDGER BENCH: Apply filter: NZDUSD SHORT — improves win rate from 31.9% to 34.5% (+2.7pp, n=139) — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: NZDUSD SHORT — improves win rate from 31.9% to 34.5% (+2.7pp, n=139) SAME-WINDOW EFFECT: Over the current window, exactly one trade matches the filter — NZDUSD SHORT via rizzy5m at 2026-09-09 11:45, which is currently REJECTED/open. Applying the filter would have promoted that trade to live, adding one more short into a window where every other executed short (EURUSD, XAUUSD, AUDUSD) closed as a loss. Net effect on the window is therefore negative-to-flat: one extra position, likely a loss, with no offsetting wins from the filter since no other NZDUSD SHORTs appear in the sample. The +2.7pp headline is invisible here because n=1 in-window. OVER-TIME PROJECTION: A 2.7pp win-rate lift on n=139 is roughly 3.7 additional wins — statistically indistinguishable from noise at that sample size (95% CI on the delta spans roughly ±8pp). More importantly, 34.5% is still a losing win rate unless average win siz Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com