When the Denver Broncos fell to the Indianapolis Colts on Sunday, just three teams remained undefeated in the 2015 NFL season: the New England Patriots, Cincinnati Bengals and Carolina Panthers.
For the Patriots, starting a season 8-0 is old hat; they also did it in 2007, the year they navigated an entire (regular!) season without a loss. And although the Bengals hadn’t started a season 7-1 or better since 1988, they also had the league’s sixth-best record in the four seasons leading up to this year — they weren’t exactly strangers to the top of the standings.
Only Carolina, who was essentially a .500 club over the preceding four years (and worse than that last year) seems to have taken the NFL completely by surprise. Based on our preseason Elo ratings — FiveThirtyEight’s favorite system for estimating an NFL team’s skill level at a given moment — there was a 1.8 percent probability that the Panthers would be 8-0 through Week 9. By contrast, New England had a 6.6 percent probability of being unbeaten at this stage.
The Panthers have improved their rating by 116 Elo points (the equivalent of 4.6 points of per-game scoring margin) since the start of the season, and, according to our simulations, they’re on track to win 4.6 more games than was projected before the year. That’s an unusually big jump. But the manner in which Carolina has improved is also unconventional. While the typical big first-half Elo gainer does it with a significantly improved passing game, the Panthers’ aerial attack is virtually the same as it was last season — league average, basically, in the eyes of expected points added (EPA).
Although quarterback Cam Newton gets the headlines for the Panthers’ offense, which has improved its overall EPA by about half a standard deviation since last year, almost all of the change is owed to a more efficient rushing game. (Newton himself is on pace for about the same production in the rushing game as last season.) Of all the teams who improved their Elo as much as the Panthers have, only about 14 percent did so with a passing attack that, like Carolina’s, didn’t improve relative to the league. Generally speaking, to win more games in the NFL, you need to throw the ball more effectively.
That is, unless you improve your pass defense — which is exactly what the Panthers have done this year. Using the EPA grades I introduced last week, which rate teams on a scale in which 100 is average and one standard deviation is 15 points, only four defenses have improved more against the pass than Carolina’s has this season:
| PASS DEFENSE | |||
|---|---|---|---|
| TEAM | 2014 | 2015 | Δ GRADE |
| Denver | 111 | 138 | +26 |
| Washington | 65 | 91 | +26 |
| N.Y. Jets | 89 | 113 | +24 |
| Philadelphia | 97 | 120 | +23 |
| Carolina | 102 | 125 | +23 |
| Atlanta | 82 | 103 | +21 |
| St. Louis | 105 | 123 | +18 |
| Tennessee | 84 | 101 | +17 |
| Oakland | 79 | 90 | +12 |
| Pittsburgh | 87 | 98 | +11 |
But perhaps the most overlooked factor that explains Carolina’s rise is that they weren’t in a terrible place to begin with. Although they went 7-8-1 a year ago, giving them one of the worst records of any playoff team ever, they still finished the season with a solid 1,551 Elo rating and were projected by Elo to win 9.2 games before the season started. (Yes, playing in the NFC South helps, too.) Based strictly on preseason numbers, there was a 2.9 percent probability that the Panthers would improve their Elo rating as much as they have — making their surge more likely than the Elo improvements of either the Patriots (2.6 percent) or Bengals (1.1 percent).
Carolina still represents an unlikely success story, particularly because of how the team has achieved its undefeated record. But while its 8-0 record is surprising, the fact that Carolina is good isn’t. Elo is optimistic about the Panthers’ chances — according to our simulations, only New England has a better chance of winning the Super Bowl.
