MODEL ADOPTION #3029: REJECT — LEDGER BENCH: Apply filter: Min confirmations >= 5 — improves win rate from 31.3% to 32.9% — E8 Intelligence Research

Abstract

DECISION: REJECT ELEMENT: No element survives — the min-confirmations ≥5 filter is not a concrete, testable edge compatible with E8 geometry timing; it is a statistical artifact from an unverified backtest window. EXPECTED IMPACT: Win-rate gain of +1.6pp is within noise for n=1430 (σ≈1.2pp), and the live sample (n=10) is dominated by luck. The filter would reduce trade frequency by an unknown amount, likely cutting total PnL if removed trades are high-R:R winners. No evidence that confirmation count aligns with E8 rhythm polarity — it is a generic lag filter, not a timing edge. RISK: Overfitting to a confirmation threshold that shifts with volatility regime; look-ahead bias if confirmations are computed on the entry candle; missing strong momentum moves that enter with <5 confirmations but are E8-valid; the filter could silently remove the very trades that carry the fade polarity edge. TEST: Must pass a replay-bench on the last 1,000 live signals (not backtest) with a 1-candle Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com