Leak Detector — Bayesian segment scanner
What this is. Slices settled positions across book × league × market × flow × mode × sharp pool × trajectory tag × game state × stake bucket up to depth-4, applies normal-conjugate Bayesian shrinkage toward the system ROI prior
N(+0.03, 0.10²), and flags segments with persistently negative ROI under chained discipline (sample size + posterior effect + tail probability + chronological split-sample persistence). Two tiers: candidate = n≥100, posterior ROI ≤ −5%, P(<−2%) ≥ 0.85, persistence pass. Confirmed = same as candidate plus n_train≥200, posterior ROI ≤ −7%, P(<−2%) ≥ 0.95, both halves ≤ −3%, BH-FDR pass at q=0.10, replicates on 20% holdout. Pairs with the Segment Explorer (no-shrinkage bootstrap CIs) — same data, different lens.
Filters
Click Run Scan to begin.
Positions
—
awaiting scan
Slices Evaluated
—
cuboid 1-D / 2-D / 3-D / 4-D
Surviving Prune
—
n ≥ 100 floor
Candidates
—
+ persistence
Confirmed
—
+ FDR + holdout
AI EXPLANATION— plain-English read on the scan
AI
Click Explain with AI to summarize the scan.
multi-turn · context-seeded · 10-turn cap