MODEL ADOPTION #910: REJECT — LEDGER BENCH: Apply filter: Phase distance MODERATE (5-10deg) — improves win rate from 31. — E8 Intelligence Research

Abstract

DECISION: REJECT ELEMENT: The 5–10° phase distance filter fails to survive hostile scrutiny. The claimed +1.9pp win-rate gain (31.7% → 33.5%) is based on n=155 — a sample too small to distinguish signal from noise, especially given E8's known regime sensitivity. The "100% win rate on 4 trades" projection is a textbook overfit artifact: 4 trades prove nothing. The filter would reject 15–25% of all signals, and the backtest does not demonstrate that the win-rate gain compensates for the lost opportunity cost in PnL terms. No concrete edge compatible with E8 geometry timing has been shown — phase distance is a static angular filter, not a rhythm or polarity alignment rule. The 72% confidence is unsupported by any robustness test (e.g., out-of-sample walk-forward, Monte Carlo permutation). The filter is a fragile band-pass that will likely degrade in trending regimes where wide-phase breakouts dominate. EXPECTED IMPACT: Win-rate might appear to rise by ~1–2pp in a narrow backtest window, Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com