The story of the game
Win probability after every play. Hover or drag to relive any moment.
Decided by 15 points with 9 lead changes and 23 ties. The game saw 12 major win probability swings, and the losing team's chances peaked at 84%.
Q3, 10:03, Caruso (OKC): Caruso 24' 3PT Jump Shot (11 PTS) (Dort 2 AST). It swung the win probability by +15.0 points.
Q3, 3:57, Gilgeous-Alexander (OKC): Gilgeous-Alexander 25' 3PT Jump Shot (17 PTS) (Dort 3 AST). It swung the win probability by -14.6 points.
Wembanyama (SAS) 5 pts, Williams (OKC) 5 pts
Turning points
The plays that moved winning odds the most.
Vassell 26' 3PT Jump Shot (9 PTS) (Castle 5 AST)
Joe 26' 3PT Pullup Jump Shot (3 PTS) (Hartenstein 3 AST)
Castle 26' 3PT Jump Shot (5 PTS) (Fox 2 AST)
J. Williams 24' 3PT Jump Shot (14 PTS) (Caruso 1 AST)
J. Williams 26' 3PT Jump Shot (11 PTS) (Gilgeous-Alexander 5 AST)
Bryant 3PT Jump Shot (3 PTS) (Castle 4 AST)
Who actually swung it
Players ranked by how much they moved the win probability, not by points scored.
Postgame read
OKC beat SAS 123-108 by 15 points.
The game turned at Q2, 6:25: Vassell 26' 3PT Jump Shot (9 PTS) (Castle 5 AST). That single play moved SAS's chance of winning from 49% to 64%. No other moment shifted the game more, and it went SAS's way.
No one influenced the outcome more than Gilgeous-Alexander of OKC. Across 40 plays, his actions moved the win probability by a combined 74 percentage points, the most of any player on the floor.
The tensest moment for a comeback came at Q3, 4:17, with SAS down 5 points. The model put their chance of pulling it off at 20%. In other words: difficult, but not nothing.
The strongest run that didn't fully show on the scoreboard came at Q2, 8:49, with the game tilting SAS's way. The play in the middle of it: Wallace REBOUND (Off:1 Def:2).
These numbers come from ClutchCast's champion model, the winner of a competition between six models. The winner is judged on how honest its probabilities are on games it never trained on, not just how often it picks the right winner.