// POLICY AGGREGATOR · LAYER 3 · READ-ONLY
Policy Stats
Journaled outcomes bucketed by pattern_key. Feeds the learning loop: as entries reach their hindsight horizon and get labelled, each bucket's true-positive rate, mean realized capture, and suggested timing rules update. Layer 3 is read-only — Layer 4 will consume these to adjust live signal confidence + trade-plan sizing.
buckets
0
with data
0
best TP
—
best capture
—