Point differential is points scored minus points allowed. It is the simplest team-strength stat and, over a season, a better predictor of next year's record than the record itself, because it does not care whether a game was won by 1 or by 30. A 10-7 team with a negative point differential got lucky in close games and tends to fall back; a 7-10 team with a positive one tends to bounce.
Updated Sept. 16, 2026.
Through Week 1, the Jacksonville Jaguars lead the league at +24, better than 95.2% of every team's start through the same week since 1999. The Cleveland Browns are last at -24.
| # | Team | Point diff | G |
|---|---|---|---|
| 1 | Jacksonville Jaguars | +24 | 1 |
| 2 | Chicago Bears | +22 | 1 |
| 3 | Kansas City Chiefs | +21 | 1 |
| 4 | San Francisco 49ers | +20 | 1 |
| 5 | Baltimore Ravens | +18 | 1 |
| 6 | Minnesota Vikings | +17 | 1 |
| 7 | Las Vegas Raiders | +14 | 1 |
| 8 | New York Jets | +13 | 1 |
| 9 | Arizona Cardinals | +12 | 1 |
| 10 | New York Giants | +8 | 1 |
| 11 | Pittsburgh Steelers | +7 | 1 |
| 12 | Cincinnati Bengals | +6 | 1 |
| 13 | Buffalo Bills | +5 | 1 |
| 14 | Seattle Seahawks | +3 | 1 |
| 15 | Philadelphia Eagles | +2 | 1 |
| 16 | Detroit Lions | +1 | 1 |
| 17 | New Orleans Saints | -1 | 1 |
| 18 | Washington Commanders | -2 | 1 |
| 19 | New England Patriots | -3 | 1 |
| 20 | Houston Texans | -5 | 1 |
| 21 | Tampa Bay Buccaneers | -6 | 1 |
| 22 | Atlanta Falcons | -7 | 1 |
| 23 | Dallas Cowboys | -8 | 1 |
| 24 | Los Angeles Chargers | -12 | 1 |
| 25 | Tennessee Titans | -13 | 1 |
| 26 | Miami Dolphins | -14 | 1 |
| 27 | Green Bay Packers | -17 | 1 |
| 28 | Indianapolis Colts | -18 | 1 |
| 29 | Los Angeles Rams | -20 | 1 |
| 30 | Denver Broncos | -21 | 1 |
| 31 | Carolina Panthers | -22 | 1 |
| 32 | Cleveland Browns | -24 | 1 |
Wins are a coarse measure: a one-point win and a 30-point win count the same. Point differential keeps the margin, and margins carry information that records throw away. The Pythagorean expectation, borrowed from baseball, converts a point differential into an expected win total; teams that win more games than their points say they should are running hot in close games, which does not persist.
A rule of thumb: over a 17-game season, roughly 37 points of differential equals one win. So a team at +150 projects to about 12.5 wins, and a team that went 11-6 at +40 was about two wins better in the standings than on the field. The betting market knows this, which is why win totals are set closer to point differential than to last year's record.
The stat's weakness is game script and garbage time: a 34-3 lead that becomes 34-24 records +10 rather than +31. Efficiency stats like EPA per play address that, and opponent-adjusted point differential addresses schedule. But as a first look at how good a team is, nothing beats the simple margin.
The 2022 Vikings went 13-4 with a point differential of -3, an NFL record for a team with that many wins. They were 11-0 in one-score games. The Pythagorean expectation had them at 8.5 wins, and in 2023 they went 7-10.
Point differential = points for minus points against. Pythagorean win expectation = games x PF^2.37 / (PF^2.37 + PA^2.37), with the exponent fit to NFL history. Approximate conversion: 37 points of differential per win over a 17-game season.
Because close games are mostly luck, and records are made of close games. Point differential measures how much a team outplayed opponents rather than how often the bounces went its way, and it predicts the following season's record better than the record itself does.
About 37 over a 17-game season, or roughly 2.2 points per game. A team at +74 for the season projects to be about two wins above .500.
A formula that turns points for and against into an expected record: points for to the 2.37 power, divided by points for plus points against each to that power, times games. Teams that beat their Pythagorean record tend to regress the next season, and teams that fall short tend to improve.
Category: Stats & Analytics. Part of the StickToTheModel NFL Encyclopedia.