MODEL ADOPTION #4465: REJECT — LEDGER BENCH: Apply filter: Session = OFF_SESSION — improves win rate from 31.5% to 41.9% — E8 Intelligence Research

DECISION: REJECT ELEMENT: No element survives — OFF_SESSION filter is a statistical artifact, not an E8-geometry edge. EXPECTED IMPACT: Win-rate gain is likely regression to the mean (n=129, +10.3pp driven by 2–3 favorable weeks). Filtering by session alone ignores the model's core rhythm logic — it would blind the executor to legitimate off-session FADE setups that align with Gods-Clock phase transitions, where liquidity gaps actually favor our polarity edge. Realized impact: likely −2–5% profit due to missed high-R Asian intervention moves, with no durable win-rate improvement beyond noise. RISK: (1) Overfitting to a small backtest window — the 41.9% figure collapses under walk-forward. (2) Removes news-driven off-session breakouts that occasionally produce 3–5R winners, capping upside. (3) Spread mechanism unverified — no data shows wider spreads caused the losses; could be random clustering. (4) Violates E8 principle: session is not a timing variable in the model's geometry;

Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com