Paper execution audit: degrades honestly when the close is missing
83 paper entries were logged, all executed=False, and realized CLV is unmeasurable for every settled row -- the ledger records 'no close captured' rather than fabricating a number.
How to read this figure
The one quantity this ledger can actually speak to is placement-time divergence between the model's number and the price taken, whose median is 0.09204799999999999 probability points -- a descriptive placement-time measurement, not a realized outcome.
What it means
This is a paper-only execution ledger: 83 logged entries, none with real fills, and 37 settled by final score. Realized closing-line value is null for all 37 because no independent closing-price feed was wired for this channel, so the corpus honestly records the gap instead of backfilling a figure.
Caveats and confounds
clv_pct is null for all 37 settled rows because no independent closing feed was captured, so no better-than-close (CLV) claim is possible from this corpus.