LAB #4820 PROMISING: LEDGER BENCH: Apply filter: Min confirmations >= 7 — improves win rate from 31.8% to 48.1% (+16.3pp, — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Min confirmations >= 7 — improves win rate from 31.8% to 48.1% (+16.3pp, n=27) SAME-WINDOW EFFECT: The filter would have removed the three losing rizzy5m shorts (EURUSD, XAUUSD, NZDUSD) and the losing AUDUSD rizzy7m short — all of which sit at low confirmation counts — while leaving the two USDCHF rizzy7m winners and the USDCAD rizzy5m winner intact. That converts the live window from roughly 2W/4L to 2W/0L on executed trades, a swing of about +33pp win rate on a tiny sample. The rejected/open entries (GBPUSD, EURUSD, EURJPY, NZDUSD) would also have been filtered, which is neutral-to-positive since they never resolved. Net: the filter looks like it would have dodged the exact cluster of losses in this window. OVER-TIME PROJECTION: Over 3-6 months, a +16.3pp win-rate lift on n=27 is statistically fragile — the 95% CI on that delta spans roughly -5pp to +38pp, so the true effect could be near zero or Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com