Play breakdown

EPA per play asks one question of every snap: how much did it move the team's expected points? A 5-yard gain on 3rd-and-4 is a good play; the same gain on 3rd-and-12 is a punt with extra steps, and expected points added is the metric that knows the difference. Success rate is the steadier companion — the share of plays with positive EPA — which tells you whether a unit's headline number is a repeatable baseline or three long touchdowns wearing a trend coat. Our EPA-per-dropback primer covers why this framing beats yards and completion percentage for quarterbacks in particular.

Teams, players, and qualification floors

The team tab ranks all 32 offenses or defenses; team rows link to a per-team breakdown with pass-depth and run-direction splits. The player tab ranks QB/RB/WR/TE with a per-position snap floor applied server-side, so a 40-snap September doesn't sit atop a leaderboard next to full seasons. Percentile shading matches the rest of the site: color is always relative to the position cohort.

Fresh seasons start empty

Everything here is computed from play-by-play, so a new season has no rows until real games land — the page fills in from Week 1 rather than padding itself with preseason guesses. For forward-looking numbers before kickoff, the rankings board and player outlooks carry the model's projections.