Success rate

What it measures

Consistency. A play is a success when it produced positive EPA, meaning it left the offense better off than before the snap. Success rate is the fraction of plays that cleared that bar, so it treats a 6 yard gain and a 60 yard gain the same. That makes it the natural partner to EPA per play: EPA tells you the size of the wins and losses, success rate tells you how often the offense wins at all. An offense with a high EPA per play but a mediocre success rate is boom-or-bust; one with both is genuinely dominant.

League average sits in the mid 40s percent-wise. On defense, lower is better: it is the share of opponent snaps that succeeded against you.

How it is computed here

The mean of the nflverse success flag (success = 1 when the play's EPA is positive) across offensive snaps, at the same grains as EPA per play: season, week, situation bucket, and league baseline. Quarterback rows compute it over dropbacks only.

How fast it stabilizes

Faster than EPA per play. Because every play contributes a simple yes or no, one long touchdown cannot drag the number the way it drags EPA. A few games of success rate is already a decent read on an offense's floor, which is why analysts reach for it early in the season when EPA is still jumpy.

How it gets misused

• Using it alone to crown an offense. It deliberately ignores magnitude, so a dink-and-dunk offense can post a strong success rate while generating few explosive plays and fewer points. Read it next to EPA per play.
• Forgetting the defensive direction. A defense wants opponent success rate low.
• Comparing across situations without context. Success is defined relative to expectation, so a 40 percent success rate on third and long is not the same achievement as 40 percent on first down.