What is CASE
Context Aware Scoring Expectation: every NFL play, valued by what it did to the expected final score.
The idea
At any moment in a game, the situation — the score, the clock, field position, down and distance, who has the ball — implies an expected final margin. Every play moves it. CASE is that movement, in points, for the team with the ball. Zero means the play did exactly what the situation called for.
Why it adds up
Each play is measured from the state before it to the state after, so a game's plays add up exactly to its final margin, less the small edge expected at kickoff. Offense, defense, special teams and penalties all land in the same ledger, in the same unit.
Points now, and the game after
Every play's CASE splits exactly in two. Points now is what the play did for the race to the next score — close to what expected-points models measure. The game after is what it did to everything past that: mostly the clock and the possessions left, which is where late-game decisions live.
Who wins
CASE Win Probability comes from the same projected margin that values the plays. In the final minutes it also accounts for what a margin cannot carry: timeouts, and whether the leader can run out the clock.
Plays that mattered
Contested CASE weights each play by how much it could have changed who won, so a first down in a one-score fourth quarter counts for more than one in a blowout.
What it is not
CASE describes what happened, in context. It is not a forecast, and it does not know which team is better — every game starts from the same expectation. All models are wrong; the definitions page says precisely what each number is, and the methodology says where this one falls short.
Read more
- Introducing CASE — coming soon
- How CASE works: the methodology — coming soon
- Every metric, defined
- Get the data
- The Jake Davis Analytics newsletter