MODEL ADOPTION #2040: REJECT — LEDGER BENCH: Apply filter: Phase distance TIGHT (<=2deg) — improves win rate from 32.0% t — E8 Intelligence Research
Year: 2026
Type: preprint
Abstract: DECISION: REJECT ELEMENT: No single rule survives — the ≤2° phase filter is a statistical artifact of a 417-trade sample, not a geometric invariant. E8 timing already encodes phase; adding a second phase-distance threshold creates redundant conditioning that will overfit to recent XAUUSD losses. EXPECTED IMPACT: Win-rate gain of +4.5pp is within 1σ noise for n=417 (95% CI roughly ±4.8pp). Even if real, the 30–50% trade-frequency reduction will cut absolute profit by more than the win-rate improvement adds, especially if high-RR winners (EURJPY LONG) get excluded. Net expected PnL change: −0.1R to +0.2R per trade, not worth the complexity. RISK: The filter becomes a hidden proxy for "avoid XAUUSD shorts" — a curve-fit to 4 recent losses. In high-volatility breakouts, phase alignment widens naturally; the filter would exclude the very trades that carry the model's fat right tail. Also, stratifying by instrument/strategy will likely show >60% of the effect driven by XAUUSD alone, co Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com
Author: Caldin, Andrew Stewart
Language: English
Subfield: Finance
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7212230518
Cites:
Cited by:
Related to: