LAB #4874 NEUTRAL: LEDGER BENCH: Apply filter: ETHUSD SHORT @ NY_OPEN with 6 conf — improves win rate from 31.9% to 63. — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: ETHUSD SHORT @ NY_OPEN with 6 conf — improves win rate from 31.9% to 63.6% (+31.8pp, n=11) SAME-WINDOW EFFECT: The filter as stated only touches ETHUSD SHORT entries at NY_OPEN with conf 6 — and the current live window contains zero ETHUSD trades, so nothing in the ledger above would have been filtered, added, or removed. The claimed +31.8pp lift is drawn from n=11, meaning the "improvement" rests on roughly 7 wins vs 4 losses in a single instrument/time-slot cell; that is a sample small enough that one or two flipped outcomes erase the entire edge. On this window the effect is exactly zero, which is itself the tell: the filter is untested against the actual book it would sit alongside. OVER-TIME PROJECTION: If the edge is real, it's a narrow conditional — one pair, one session, one confidence bucket — so expect it to fire rarely (a handful of times per quarter), contributing negligible aggregate Pn Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com