EPA per play is expected points added divided by the number of plays. It is the efficiency version of EPA: instead of asking how many points an offense produced, it asks how many it produced per snap, so a team that ran 50 plays can be compared to one that ran 75. It is the single team stat most analysts check first.
Updated Sept. 16, 2026.
Through Week 1, the Chicago Bears lead the league in EPA per play at +0.409, better than 98.7% of every team's start through the same week since 1999. The Denver Broncos are last at -0.344 (6.2%).
| # | Team | EPA per play | Plays |
|---|---|---|---|
| 1 | Chicago Bears | +0.409 | 70 |
| 2 | New York Giants | +0.369 | 65 |
| 3 | Jacksonville Jaguars | +0.338 | 53 |
| 4 | Baltimore Ravens | +0.321 | 60 |
| 5 | Buffalo Bills | +0.229 | 52 |
| 6 | Dallas Cowboys | +0.221 | 52 |
| 7 | Carolina Panthers | +0.182 | 62 |
| 8 | San Francisco 49ers | +0.179 | 64 |
| 9 | New York Jets | +0.168 | 60 |
| 10 | Kansas City Chiefs | +0.164 | 64 |
| 11 | Arizona Cardinals | +0.144 | 69 |
| 12 | Philadelphia Eagles | +0.071 | 51 |
| 13 | Houston Texans | +0.069 | 73 |
| 14 | Seattle Seahawks | +0.066 | 47 |
| 15 | Washington Commanders | +0.060 | 68 |
| 16 | Detroit Lions | +0.000 | 73 |
| 17 | Cincinnati Bengals | -0.002 | 62 |
| 18 | Minnesota Vikings | -0.020 | 59 |
| 19 | New Orleans Saints | -0.031 | 86 |
| 20 | Las Vegas Raiders | -0.039 | 64 |
| 21 | New England Patriots | -0.110 | 67 |
| 22 | Tennessee Titans | -0.119 | 49 |
| 23 | Green Bay Packers | -0.175 | 67 |
| 24 | Cleveland Browns | -0.184 | 49 |
| 25 | Tampa Bay Buccaneers | -0.185 | 51 |
| 26 | Miami Dolphins | -0.236 | 50 |
| 27 | Indianapolis Colts | -0.241 | 53 |
| 28 | Los Angeles Chargers | -0.244 | 51 |
| 29 | Los Angeles Rams | -0.265 | 56 |
| 30 | Pittsburgh Steelers | -0.303 | 61 |
| 31 | Atlanta Falcons | -0.321 | 57 |
| 32 | Denver Broncos | -0.344 | 46 |
Total EPA rewards volume. A team that gets a lot of possessions, or plays fast, runs more plays and has more chances to add up expected points. Dividing by plays strips that out and leaves efficiency: how much the offense gained, on average, every time it snapped the ball.
On offense, higher is better. On defense, EPA per play allowed is the same number from the other side, and lower is better. The two together explain most of a team's point differential, which is why power ratings built on EPA per play tend to track the betting market closely.
The number is small. A full-season team offense lives between roughly -0.15 and +0.15 EPA per play. That range covers the worst offense in the league and the best, so a difference of 0.05 is large, not a rounding error.
In 2024 the Ravens finished near +0.15 EPA per play on offense, the best in the league, while the Browns finished near -0.15. Over 1,000 plays that 0.30 gap is about 300 points of expected scoring, which is roughly the difference between a 12-win team and a 4-win team.
EPA per play = total EPA / number of plays, over whatever sample you choose (a game, a season, only passes, only first downs). Most published versions exclude kneel-downs and spikes. Passing EPA per dropback and rushing EPA per carry are the same idea split by play type.
Over a full season, anything above +0.10 is elite, +0.05 is clearly good, 0 is league average, and below -0.05 is a bad offense. Early in a season the spread is much wider because two games of noise have not washed out yet.
Because passing gains more per attempt and reaches the big plays that swing expected points. The average dropback is worth roughly +0.05 EPA and the average carry roughly -0.05. That gap is the main reason analysts push teams to throw more on early downs.
The raw number does not. Opponent-adjusted versions exist, and they matter most early in the season when one team has played two bad defenses and another has played two good ones.
Category: Stats & Analytics. Part of the StickToTheModel NFL Encyclopedia.