Mechanism ledger export: the raw evidence table
Every named mechanism tested, with its recorded verdict and evidence pointer. NULL, REJECT, and NOT_TESTABLE are market-efficiency evidence. This is a descriptive measurement artifact.
How to read this figure
Each row is one mechanism run; the bucket field collapses many verdict labels (CONFIRMED_LOCAL, NULL, REJECT, NOT_TESTABLE, and so on) into confirmed / null / not_testable.
What it means
This is the raw evidence table behind the survival scoreboard: every named mechanism with its verdict and a pointer to the exact corpus and sample that produced it. NBA's back-to-back rest penalty, for instance, confirmed at an effect of -1.73 (p 0.005645490426098632) on a 4,732-game corpus.
Caveats and confounds
as_of is the recorded run_ts and is null for runs that predate run_ts stamping, so many rows carry no timestamp and a single mechanism can appear multiple times across corpora.