LAB #4973 NEUTRAL: LEDGER BENCH: Apply filter: Quarterly phase = MANIPULATION — improves win rate from 31.7% to 32.6% ( — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Quarterly phase = MANIPULATION — improves win rate from 31.7% to 32.6% (+0.8pp, n=1563) SAME-WINDOW EFFECT: The filter would only act on trades whose entry timestamp falls in a "MANIPULATION" quarterly phase — and the live window you've shown (2026-09-03 to 2026-09-11) is a single ~9-day slice, so at most it gates a handful of the 10 listed trades, not the whole set. Given the +0.8pp lift is measured on n=1563 across quarters, the expected effect on ~10 trades is statistically indistinguishable from zero — you'd likely see 0-1 trades flipped, and the realized PnL delta would be dominated by noise, not the filter. Note the live window is already bleeding (3 closed losses vs 2 closed wins, plus 5 rejected/open), so any "improvement" here would be coincidental, not causal. OVER-TIME PROJECTION: A +0.8pp win-rate edge at n=1563 is within the range of what you'd expect from a single binary phase split by Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com