LAB #5108 NEUTRAL: LEDGER BENCH: bash /opt/excalibur/fable_staging/a4e99ba48f/apply.sh # Fable build a4e99ba48f (low r — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: bash /opt/excalibur/fable_staging/a4e99ba48f/apply.sh # Fable build a4e99ba48f (low risk): Staged new read-only repair_ledger_check.py (Postgres COUNT(*) on repair_ledger, exits 0); py_compile clean; SAME-WINDOW EFFECT: Nothing in the trade window changes. `repair_ledger_check.py` is a read-only Postgres `COUNT(*)` against `repair_ledger` that exits 0 — it touches no signal path, no sizing, no entry/exit logic. The 10 listed trades (3 REJECTED/open, 4 CLOSED/LOSS, 1 CLOSED/WIN, plus the pending EURUSD short) would have executed identically, same fills, same PnL. Zero delta on win-rate and zero delta on PnL. OVER-TIME PROJECTION: Over 3–6 months this is observability, not alpha. Its value is catching ledger drift/desync early — if `repair_ledger` ever stops accumulating rows (silent write failure, migration break, connection pool exhaustion), you find out in minutes instead of after a bad reconciliation. The risk Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com