RAPM computed from playoff stint data only — separate from the regular season model. Each season is estimated independently; the multi-year leaderboard uses minutes-weighted averaging across seasons.
Playoff samples are much smaller than regular season (~15–20 games vs 82), so estimates are noisier. The luck adjustment — 3PT, free-throw, and mid-range make/miss variance muted at 75% of each shooter's deviation from his own expectation — matters most here, damping hot/cold shooting runs.
Alpha: Ridge regularization strength. α=500 (default) applies strong shrinkage, which is especially useful for small playoff samples. α=10 shrinks less and can show more separation among players with large samples.
| # | Player | Team | Min | RAPM Adj | ORAPM Adj | DRAPM Adj | RAPM Raw | ORAPM Raw | DRAPM Raw |
|---|
| # | Player | Team | Min | RAPM Adj | ORAPM Adj | DRAPM Adj | RAPM Raw | ORAPM Raw | DRAPM Raw |
|---|
Generated 2026-07-12 08:17 · Source: data/stints_playoffs.csv