MODEL ADOPTION #427: REJECT — TG AUTO: EDGE MINER: 4 claimed edges pulled from the queue to the bench — E8 Intelligence Research

DECISION: REJECT ELEMENT: No single rule survives hostile review. - **exp#420 (ICT Killzone London Open Breakout):** Timeframe mismatch (30m vs 15s-7m). Scaling down to 5m destroys the 1:3 R:R assumption — E8 sacred geometry targets are not ATR-based and do not align with breakout mechanics. Direct transferability is zero. Reject. - **exp#421 (XAUUSD LONG-only filter):** n=67 is statistically meaningless for a scalping model that executes thousands of trades. The 61.2% win rate on 1:1 R:R is irrelevant to E8's variable R-multiple exits. Over 90 days of tick data, this filter would likely block profitable SHORT scalps during XAUUSD downtrends, degrading net profit. Reject. - **exp#425/426 (Quarterly Phase filters):** n=94 and n=118 are too small. Look-ahead bias is almost certain — quarterly phases defined by future price data cannot be known at trade time. The +7.0pp and +11.7pp improvements are noise. Reject. EXPECTED IMPACT: None. All four edges are either mismatched in timeframe, Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com