LAB #4156 NEUTRAL: LEDGER BENCH: Apply filter: XAUUSD SHORT — improves win rate from 32.6% to 48.6% (+15.9pp, n=35) — E8 Intelligence Research

Abstract

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: XAUUSD SHORT — improves win rate from 32.6% to 48.6% (+15.9pp, n=35) SAME-WINDOW EFFECT: Applying the XAUUSD SHORT filter would have excluded the single XAUUSD SHORT trade in the live window (entry 4598.79 on 2026-08-27, which closed as a LOSS). That trade was a loser, so removing it would have flipped the window's record from 2W/5L to 2W/4L — a raw win-rate improvement from 28.6% to 33.3%, but with only 7 total trades the statistical delta is negligible. The filter would not have touched any other pair, so the net PnL change is essentially zero (one losing trade removed, but its loss was likely small relative to the other losses). This window provides no meaningful evidence for or against the filter. OVER-TIME PROJECTION: Over 3–6 months, the claimed +15.9pp win-rate improvement (n=35) is based on a small sample and likely suffers from selection bias — XAUUSD SHORT may have been favorable during th Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com