wnba projections

box scores

Per-game box-score projections for every player: minutes, points, rebounds, assists, steals, blocks, turnovers, threes and shooting. Ridge regression per stat, anchored to prior-season performance by a weighted prior over twelve seasons of substrate.

how these are projected

players223 · season2026 · modelW_BOX_v1

Per-Game Projections

2026 season · sorted by PTS

team limit

Resolve WNBA per-game box-score projections. Click a column to sort.
loading…

Methodology

ridge per stat · weighted prior

Inputs: 5-feature substrate per stat per player — short-window rolling rate, long-window rolling rate, season-to-date mean, home indicator, opponent-allowed roll. Ridge alpha=1.0 fit per stat across positions; n=12 WNBA seasons, 46,784 player-games, 549 unique players.

Coverage: 223 players with sufficient 2024-25 sample. Rookies (Bueckers, Olivia Miles, Citron, Jackson) projected from W_ROOKIE_ARCH archetype centroids — see draft page for rookie archetypes once wired.

Sanity checks pass: A'ja Wilson 24.0 PTS (#1), Bueckers 18.5 (rookie #1), Olivia Miles 17.9 (rookie #2), Reese 12.1 REB (elite-big cluster ✓), Caitlin Clark 7.6 AST (PG cluster ✓).

v2 roadmap (banked): PSI/ADI advanced metrics layer (PBP fetcher landing), rookie overlay (NCAAW→WNBA transfer operator), injury filter, per-projection CI from Ridge variance.