LAB #4160 PROMISING: LEDGER BENCH: Apply filter: Min confirmations >= 5 — improves win rate from 32.6% to 34.1% (+1.5pp, — E8 Intelligence Research

Abstract

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Min confirmations >= 5 — improves win rate from 32.6% to 34.1% (+1.5pp, n=1826) SAME-WINDOW EFFECT: Applying a minimum of 5 confirmations would have filtered out the two rizzy5m entries on 2026-08-27 (AUDUSD and NZDUSD shorts) and likely the XAUUSD short on the same day, all of which were losses. It would also have removed the rizzy5m GBPUSD short from 2026-08-26, which was a win — so the immediate net effect is ambiguous: you lose one win but drop three losses. Over the 10-trade window, that shifts the record from 3W/6L/1open to roughly 2W/3L/1open, which is a modest win-rate improvement but a smaller sample. The +1.5pp backtest gain is not clearly visible here; the filter mostly removes low-conviction noise rather than systematically cutting losers. OVER-TIME PROJECTION: Over 3–6 months, a 5-confirmation floor should reduce false signals from early-stage momentum ripples, particularly on rizzy5m w Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com