Skip to main content
ABC News
So … Are The Buffalo Bills Actually Good?

For just the third time since 1992, the Buffalo Bills have started a season 3-0, after they beat the Cincinnati Bengals on Sunday in Orchard Park. The team is currently thriving under second-year quarterback Josh Allen, who has led game-winning drives in two of the Bills’ three games this season. Don’t look now, but the Bills are tied for the 11th-highest Super Bowl probability in the league, according to the FiveThirtyEight model.

Circle the wagons, Buffalo!

But naturally, the Bills are staring at a potential reality check this week, in the form of the also-undefeated New England Patriots. So, in this week’s column, we must ask: Are Allen and the Bills actually good? And will it even matter against the Pats? Let’s dive into the early numbers.

Allen has built on a polarizing rookie season with his play so far in 2019. On the one hand, Allen has won games with a combination of timely passing (he leads all quarterbacks this season in yards compiled during the final five minutes of one-score games) and impressive running (only Baltimore’s Lamar Jackson has more rushing yards among QBs). On the other hand, he ranks just 26th out of 35 qualified QBs in passer rating, with below-average rates of completions, yards, touchdowns and interceptions per attempt. In that regard — winning despite across-the-board mediocre passing numbers — Allen is well on his way to joining a bizarre group that includes Jake Plummer, Trent Dilfer, Kordell Stewart and Mark Sanchez.

Our own quarterback ratings basically split the difference between the good and the bad. With an Elo QB rating of 161, Allen isn’t great, but he’s not terrible either. In fact, he’s almost precisely average — just 4 rating points below the typical NFL starter — and by Buffalo standards, that’s not too shabby. The last time the Bills’ starter had an Elo rating that good relative to average, it was Tyrod Taylor in the middle of the 2017 season. (That happened to be the only year this century that Buffalo made the playoffs.) Compared with QBs from earlier 3-0 Bills starts, Allen is no Jim Kelly or even Ryan Fitzpatrick, but he is better than Trent Edwards was back in early 2008:

Comparing Josh Allen’s Bills to other 3-0 Buffalo starts

Quarterback and team Elo ratings for Buffalo Bills teams that started a season 3-0, 1985-2019

QB Elo vs. Avg, First 3 Gms Before Game 4
Year Starter Gm 1 Gm 2 Gm 3 Qb Elo Rating Team Elo w/ QB Won Gm 4?
1988 Jim Kelly +8 -48 -36 +6 1498
1991 Jim Kelly 263 326 62 108 1701
1992 Jim Kelly -3 201 150 50 1681
2008 Trent Edwards 53 67 11 -21 1535
2011 Ryan Fitzpatrick 170 176 105 31 1510
2019 Josh Allen 19 32 21 -4 1520

QB Elo ratings are the number of Elo points a team would gain or lose if the main quarterback was replaced by a league-average starter from that season.

Source: ESPN,

Those 2008 Bills would beat the Rams in Week 4 to run their record to 4-0 … before closing the year on a 3-9 skid to finish out of the playoffs. The 2011 Bills also fizzled after a promising first two months — which included a rare win over New England — losing seven straight and eight of their final nine games to waste a 5-2 start. So Buffalo has been tantalized by fast starts before, only to fall flat later on. But all the Kelly-led Bills on the list above went to the AFC championship game, and two of them went to the Super Bowl.

Fine — Allen and Kelly, his Hall of Fame predecessor, still have little in common aside from the charging bison on their helmets. But Allen has at least shown some hints of improvement as a passer so far in his second pro season. Despite his still-mediocre overall numbers, Allen has been more accurate this year, cutting his rate of “poor throws” (according to from a league-worst 24 percent in 2018 to just 6 percent — which ranks fourth-best this year. Some of that can be attributed to a dose of shorter throws,1 but the NFL’s Next Gen Stats estimate that Allen is completing passes at a rate 0.9 percentage points higher than expected this season, even after adjusting for the depth of the pass and other particulars, while he checked in at 7.7 percentage points below expected last season (second-worst in the league).

Allen has gotten more accurate

Rates of good and bad passing actions (per attempt) for Buffalo Bills quarterback Josh Allen by season, 2018-19

Year Passes Actual vs. Expected Interception% Poor Throw%
2018 320 52.8% -7.7% 3.8% 24.4%
2019 103 64.1 0.9 2.9 5.8

Sources: Pro-Football-Reference, NFL Next Gen Stats

So Allen may be getting better with his arm while remaining a threat with his legs. And as a team, Buffalo ranks ninth in total expected points added (EPA) per game so far this season, including 10th in special teams EPA, eighth in defensive EPA and sixth in defense against the pass specifically. ProFootballFocus grades safeties Micah Hyde and Jordan Poyer both as among the top 10 in the league at the position this season, two of 10 total Bills defenders who rank among the top 50 at their respective positions per PFF (and four who rank among the top 25). Last year’s Bills had one of the best EPA defenses in football as well, but they were undone by poor passing offense and special teams; if this year’s team has improved in those regards, Buffalo could have surprising staying power.

Of course, the Bills have also played one of the easiest schedules in the NFL so far. The Jets, Giants and Bengals collectively had an average pregame Elo rating of 1415 going into their contests against Buffalo; only the Cowboys, whose opponents have averaged a pregame rating of 1392, have played a softer group of creampuffs this season. And now, here come the Patriots — the second-highest-ranked team in our Elo ratings. New England’s current rating of 1686 is the 13th-highest the Pats have ever had coming into a game against the Bills in the history of the rivalry. (If you can call it that — New England has won five straight and 33 of the past 37 meetings dating back to 2000.) The Pats’ defense hasn’t allowed a touchdown since last year’s AFC title game, and some folks are already talking about the possibility of another undefeated season in Foxboro.

So let’s be honest: A lot of signs point to the Bills’ magical season-opening run coming to a screeching halt against New England this week. (Our model gives the Pats a 65 percent chance of winning, despite being on the road, and that’s conservative compared with the probability implied by the Vegas line.) But we also give Buffalo a 70 percent chance to make the playoffs — and for a franchise with only one playoff game under its belt since 1999, that’s as good an excuse for a raucous parking-lot celebration as anything else.

Looking Ahead: Week 4

Best matchup: No. 4 Dallas (-1) at No. 14 New Orleans, 8:20 p.m. ET Sunday

Matchup quality: 85th percentile2
Matchup evenness: 89th percentile

As we wrote about in an earlier edition of the column, Dak Prescott of the Cowboys has been one of 2019’s breakout quarterbacks: According to our Elo QB value metric, only Kansas City’s Patrick Mahomes has been better on a per-game basis than Prescott this season. Now he’ll take Dallas and its 3-0 record into New Orleans to face Teddy Bridgewater and the Saints in our top game of the week. All things considered, Bridgewater was fine in his first game as Drew Brees’s stand-in, grading 73 Elo points better than average against the Seahawks last week — though he still ranks 25th among current starters in our QB rankings. New Orleans will probably rely heavily on Alvin Kamara again after the running back racked up 161 scrimmage yards versus Seattle, but our model gives the Cowboys a slight edge (53 percent chance of winning) on the road.

See also: Minnesota at Chicago (80th/70th); Philadelphia at Green Bay (85th/62nd).

Biggest playoff implications: No. 11 Minnesota at No. 7 Chicago (-2.5), 4:25 p.m. ET Sunday

Potential shift in playoff odds: 29.4 total percentage points

The NFC North has turned into a surprisingly strong division, with every team featuring an above-.500 record at the moment. That’s a good recipe for crucial early-season divisional matchups, and no Week 4 game is more important than Sunday’s showdown between the Vikings and the Bears. Minnesota rebounded from a tough loss to the Packers with a convincing win over the Raiders last week; Chicago finally got a good game out of Mitchell Trubisky en route to a Monday night romp over Washington. The winner of Bears-Vikings will see its playoff odds rise to over 50 percent, but the loser will drop into the 20-to-25 percent range. We give Chicago a 59 percent chance of coming out on top.

See also: Cleveland at Baltimore (26.3); Philadelphia at Green Bay (25.3).

Best QB duel: No. 1 Patrick Mahomes (KC) vs. No. 13 Matthew Stafford (DET)

See also: No. 3 Tom Brady (NE) vs. No. 12 Josh Allen (BUF); No. 7 Carson Wentz (PHI) vs. No. 8 Aaron Rodgers (GB)

FiveThirtyEight vs. the Readers

As a weekly tradition here at FiveThirtyEight, we look at how our Elo model did against everybody who made picks in our forecasting game. (If you entered, you can find yourself on our leaderboard here. I am currently in 984th place!) These are the games in which Elo made its best — and worst — predictions against the field last week:

Elo’s dumbest (and smartest) picks of Week 3

Average difference between points won by readers and by Elo in Week 3 matchups in FiveThirtyEight’s NFL prediction game

ARI 65% ARI 55% CAR 38, ARI 20 +9.8
LAC 66 LAC 56 HOU 27, LAC 20 +9.6
TEN 67 TEN 61 JAX 20, TEN 7 +5.7
TB 69 TB 64 NYG 32, TB 31 +5.2
LAR 61 LAR 66 LAR 20, CLE 13 +1.1
CHI 59 CHI 63 CHI 31, WSH 15 +0.7
NE 91 NE 95 NE 30, NYJ 14 -0.2
DAL 91 DAL 95 DAL 31, MIA 6 -0.3
PHI 70 PHI 69 DET 27, PHI 24 -0.5
GB 74 GB 76 GB 27, DEN 16 -0.9
KC 69 KC 68 KC 33, BAL 28 -2.7
MIN 74 MIN 72 MIN 34, OAK 14 -3.2
SEA 70 SEA 72 NO 33, SEA 27 -4.3
BUF 74 BUF 69 BUF 21, CIN 17 -5.0
SF 79 SF 72 SF 24, PIT 20 -5.8
IND 54 ATL 52 IND 27, ATL 24 -8.2

Home teams are in bold.

The scoring system is nonlinear, so readers’ average points don’t necessarily match the number of points that would be given to the average reader prediction.

Week 3 was a victorious one for the readers! The average FiveThirtyEight picker beat our Elo model by 1 point this time around, giving the field its first win of the season. While Elo shrewdly took the Colts over the Falcons and had its faith in the 49ers and Bills rewarded, the readers picked up big points shading the Cardinals, Chargers, Titans and Bucs, all of whom eventually lost. Interestingly, the readers and the computer agreed on the favorite in all but one game last week (that Colts-Falcons game), but the human prognosticators won with smart probabilistic thinking. Nice job, all!

Among individual readers, congrats to Bruce C, who led all (identified) readers in Week 3 with 170.9 points, and to Matt Melchior, whose total of 407.0 points leads the full-season contest. Thanks to everyone who played — and if you haven’t, be sure to get in on the action! You can make picks now and try your luck against Elo, even if you missed Week 3.

Check out our latest NFL predictions.


  1. Allen’s average air yards per pass has dipped from 10.5 to 8.2 this season, according to ESPN’s Stats & Information Group.

  2. In terms of the harmonic mean of both teams’ QB-adjusted Elo ratings, relative to that figure for all regular-season NFL games this year.

Neil Paine was the acting sports editor at FiveThirtyEight.