No games logged for today.
Run python "NBA Playoff Prediction.py"
to populate the log.
How these predictions are built
Every number on this page comes from a mathematical model that rates each player using EPM (Estimated Plus-Minus, from dunksandthrees.com) and DARKO (free public metric), multiplies by projected minutes tonight, and drops anyone the NBA's official injury report marks as Out. The team ratings are then fed through a simulator that adds home-court, rest, and travel to produce a fair spread and a home-team win probability.
Backtested on 897 playoff games (11 seasons, 2014-15 → 2024-25) with point-in-time EPM and per-date injury data across all seasons. After a full outlier audit, nested walk-forward threshold validation, and era split, the disciplined signals are: H2H ≥3% edge — +15.2% pooled ROI, 60.9% hit rate (n=233); ATS ≥3pt cover when H2H agrees — +12.4% pooled ROI, 57% hit rate (n=142). Nothing guarantees future results. The H2H edge badge turns green at ≥3% (bet), yellow at ≥10% (skip — model and market are reading this game very differently). Spread badge is green only when the ATS pick and H2H pick agree on the same team.