📊 Model Accuracy — Per-Source W-L
Loading…
Why this matters:
Every pick from every module (top-25 board, whales, consensus, unders, sharp action,
locks, POD) is settled against actual outcomes. Per-source W-L tells us which modules are
actually predictive and which are noise. This is the training signal — calibration shrinkage
weights tighten on high-accuracy sources and loosen on unreliable ones over time.
--Total Picks
--Settled
--Overall Hit Rate
--Net Units (1u flat)
--ROI %
By Source
By Market
By Sport
Recent Picks
Per-Source Accuracy--