LAB #4921 PROMISING: LEDGER BENCH: Apply filter: Min confirmations >= 6 — improves win rate from 31.9% to 36.2% (+4.4pp, — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Min confirmations >= 6 — improves win rate from 31.9% to 36.2% (+4.4pp, n=944) SAME-WINDOW EFFECT: The filter would only bite on trades whose confirmation count is known and below 6 — the window above doesn't expose confirmation counts, so I can't map it trade-by-trade, but the structural read is that it removes low-conviction entries. Of the 10 live trades, 4 are already REJECTED/open and 3 CLOSED/LOSS, so if the filter had suppressed even 2 of those 3 losers, the window's realized PnL improves materially without touching the 2 winners (USDCHF, USDCAD). Net: fewer trades, same or better hit rate, small positive PnL delta on this specific window — but n=10 is far too small to confirm anything. OVER-TIME PROJECTION: A +4.4pp win-rate lift on n=944 is a real signal, not noise (roughly 42 extra wins vs. baseline expectation), and it's the kind of edge that compounds if the confirmation count is a genui Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com