LAB #4965 NEUTRAL: LEDGER BENCH: Apply filter: Min confirmations >= 5 — improves win rate from 31.7% to 33.2% (+1.5pp, — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Min confirmations >= 5 — improves win rate from 31.7% to 33.2% (+1.5pp, n=1935) SAME-WINDOW EFFECT: The filter would have removed the three rizzy5m/rizzy7m shorts that closed as losses (EURUSD, XAUUSD, AUDUSD) only if their confirmation counts were below 5 — but the ledger shows no confirmation counts per trade, so we cannot verify which of the 10 live entries would actually have been filtered. The three REJECTED/open trades (NZDUSD, GBPUSD, EURUSD, USDCHF, EURJPY) are ambiguous: if "rejected" means the filter already blocked them, then the filter is partially live and the +1.5pp claim is contaminated by selection. Net: on this window the filter likely would have skipped 2-4 of the 3 losses, but it also risks skipping the two wins (USDCHF, USDCAD) if those had low confirmations — no way to confirm from the data given. OVER-TIME PROJECTION: A +1.5pp win-rate lift on n=1935 is statistically thin — 95%

Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com