MODEL ADOPTION #2787: REJECT — LEDGER BENCH: Apply filter: Min confirmations >= 5 — improves win rate from 30.9% to 32.5% — E8 Intelligence Research

DECISION: REJECT ELEMENT: No element survives — the +1.6pp win-rate gain is not proven to be net-positive PnL, and the filter conflicts with E8's early-entry geometry that captures TP2 on low-confirmation momentum. EXPECTED IMPACT: Win rate may rise ~1.6pp, but total profit likely flat or negative — filtered early winners (3–4 confirmations) often carry larger R:R, and reduced trade frequency (15–25%) cuts compounding. The 10-trade window "improvement" is noise (removing 1–2 losers, not adding edge). RISK: Overfitting to recent volatility; in trending regimes, delaying entries to 5 confirmations degrades entry price and misses fast moves, lowering profit factor despite higher win rate. Also, the backtest did not report average R:R per trade — win-rate alone is a misleading metric. TEST: Must pass a 30-day shadow test (A: current, B: +min5) on live engine, logging confirmation count per trade. Pass condition: Portfolio B shows ≥10% higher total PnL AND ≥0.15 higher profit factor Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com