We are now halfway through the 2021 NFL season, and things are not as they once seemed. Which teams have changed our minds over the first half -- for better or worse?
Entering the year, ESPN's Football Power Index (FPI) doles out ratings for every team. With each game, it updates its opinion on teams based on their performance in all three phases of the game, while considering strength of opponent and factors such as home-field and rest advantage. It always keeps its prior rating in mind, but stubborn, FPI is not.
Today, we're looking to see which teams FPI has upgraded -- and downgraded -- the most so far this season. Here are two important things to keep in mind:
FPI is predictive. It is not a reflection of the standings, or even exactly how teams have played. It is predicting how strong teams will be going forward for the rest of the season.
The rating indicates how many points better or worse the model believes a team is relative to the average NFL team. For example, the FPI-leading Buccaneers have a rating of plus-7.8. That means it would make Tampa Bay a bit better than a touchdown favorite over an average team (think: Steelers) on a neutral field.
Let's look at the biggest FPI risers and fallers since the preseason, starting with one of the NFC's top teams.
BIGGEST UPGRADES
![](https://a.espncdn.com/combiner/i?img=/i/teamlogos/nfl/500/ari.png&h=110&w=110)
Arizona Cardinals
Preseason FPI Rating: minus-0.7
Current FPI Rating: plus-7.3 (up 8.0 points)
FPI definitely did not see this coming. The model was fairly down on Arizona going into the season. And frankly, Kyler Murray's hype had extended beyond what the former No. 1 overall pick had shown on the field -- he was 14th in QBR in 2020. But that hype was also about promise, the potential that Murray realized this season. His accuracy has improved a great deal: His off-target rate jumped from 12th-best in 2020 to second-best this season, and his completion percentage over expectation went from 10th to second, per NFL Next Gen Stats.
So the No. 1 reason why Arizona (8-1) has obliterated expectations, moving from 21st in FPI to third: a QB who is firmly in the MVP race halfway through the season.
The Cardinals' offense has become less DeAndre Hopkins-focused, in a good way. Hopkins was targeted on 28% of his routes last season, and that number has dropped to 20% this season. But in lieu of that volume has come efficiency. Murray's QBR when targeting Hopkins has jumped from 86 to 97, and his yards per attempt increased from 9.1 to 10.3. (Murray's numbers to targets other than Hopkins have also improved.)
The last thing -- and it's a big one -- is the defense. FPI went into the season thinking the Arizona defense was average, and maybe even a shade below. Instead, it ranks second in expected points added (EPA) per play on defense and is particularly strong against the pass. Edge rushers Markus Golden and Chandler Jones have delivered 9.0 and 6.0 sacks, respectively, and the Cardinals have even gotten surprisingly strong play from cornerback Robert Alford, who has allowed the third-lowest yards per coverage snap (0.6) as the nearest defender to the target, per NFL Next Gen Stats.
![](https://a.espncdn.com/combiner/i?img=/i/teamlogos/nfl/500/buf.png&h=110&w=110)
Buffalo Bills
Preseason FPI Rating: plus-4.4
Current FPI Rating: plus-7.7 (up 3.3 points)
Defense, defense, defense. It's hard to imagine the Bills being the second-most improved squad in football despite having a pretty disappointing offensive performance thus far -- Buffalo ranks just 12th in EPA per play on offense -- and entering the season so high on the board. But that defense, whew.
The Bills (5-3) have recorded two shutouts and lead the league by far in EPA per play on that side of the ball. They're not just turnover-reliant, either. Even if we ignore interception and fumble plays, they remain at No. 1.
![](https://a.espncdn.com/combiner/i?img=/i/teamlogos/nfl/500/dal.png&h=110&w=110)
Dallas Cowboys
Preseason FPI Rating: plus-0.7
Current FPI Rating: plus-3.8 (up 3.1 points)
The 6-2 Cowboys spun the regression roulette wheel and hit it just right. Dallas has long looked dangerous if it could get its defense together. Well, look what we've got here.
Cornerback Trevon Diggs' impact cannot be overstated, as his interceptions are a big reason why the Cowboys rank ninth in defensive EPA per play, though FPI is less bullish going forward. The Cowboys defensive FPI ranks 15th.
But an average defense just might be enough. Why? Because the Dallas offense is awfully productive, and that's true via both the air and the ground. The Cowboys rank fourth in EPA per play on pass plays and seventh on run plays.
![](https://a.espncdn.com/combiner/i?img=/i/teamlogos/nfl/500/lar.png&h=110&w=110)
Los Angeles Rams
Preseason FPI Rating: plus-2.9
Current FPI Rating: plus-5.7 (up 2.8 points)
The 7-2 Rams have demonstrated a dramatic increase in passing efficiency from last season, from 0.05 EPA per play to 0.27, the best in the league. Now some of that was expected with the switch from Jared Goff to Matthew Stafford at quarterback, which is why the Rams had a fairly strong FPI in the preseason. But they've still exceeded those expectations.
Stafford has indeed been a big part of that, and he leads the league in QBR. The Rams' efficiency on deep passes has dramatically increased, from a QBR of 38 last season to 95 this season.
But I also want to highlight another unit playing well for the Rams: the offensive line. It ranks third in pass block win rate, and that has surely made Stafford's life easier in Los Angeles.
![](https://a.espncdn.com/combiner/i?img=/i/teamlogos/nfl/500/phi.png&h=110&w=110)
Philadelphia Eagles
Preseason FPI Rating: minus-3.8
Current FPI Rating: minus-1.2 (up 2.6 points)
Here's an interesting one. FPI went into the season believing the Eagles (3-6) were going to be one of the five worst teams in football. To be simply below-average is quite the expectation-beater.
Jalen Hurts has been just that -- a little below the average, but not awful. He ranks 20th in QBR, and the Eagles have recorded the second-most EPA on designed QB rushes of any team, along with the most on scrambles. In fact, their entire rushing attack is efficient, ranking third in EPA per play and sixth in EPA per play on designed running back runs.
While the defense is below-average, the pass rush has certainly flashed, even with defensive end Brandon Graham out for the season. Javon Hargrave is third in pass rush win rate at defensive tackle behind two future Hall of Famers in Aaron Donald and J.J. Watt, and Fletcher Cox, Josh Sweat and Derek Barnett are all top-10 at their respective positions, too.
![](https://a.espncdn.com/i/infographics/20170710_nfl_fpr/charts/_end_rule.png)
BIGGEST DOWNGRADES
![](https://a.espncdn.com/combiner/i?img=/i/teamlogos/nfl/500/kc.png&h=110&w=110)
Kansas City Chiefs
Preseason FPI Rating: plus-6.6
Current FPI Rating: plus-2.1 (down 4.5 points)
This truly is the stunner of the season. We're talking about Patrick Mahomes, Tyreek Hill, Travis Kelce and Andy Reid. That group can't get it done? Remember that FPI still incorporates a prior on all these teams, so the initial bumps in the road didn't immediately dissuade the model on the Chiefs (5-4). But eventually, the struggling play made FPI move off its mark.
Opponents are blitzing Mahomes significantly less after he recorded a 96 QBR against the blitz last season. And year over year, his numbers don't look great.
Interception rate increased from 1.0% to 2.8%
Fumble rate increased from 0.7% to 1.1%
Completion percentage over expectation (NFL Next Gen Stats) dropped from minus-0.9% to minus-2.7%
QBR fell from 78 to 54
But this isn't all on the offense. The defense has been miserable, ranking 27th against the pass and 29th against the run in EPA per play. Peak Mahomes can almost certainly withstand a below-average defense. But down-year Mahomes certainly can't withstand a horrific one.
![](https://a.espncdn.com/combiner/i?img=/i/teamlogos/nfl/500/mia.png&h=110&w=110)
Miami Dolphins
Preseason FPI Rating: plus-1.2
Current FPI Rating: minus-2.7 (down 3.9 points)
How naïve we all were just a few short months ago, thinking the Dolphins were a pesky challenger to the Bills in the AFC East. It turns out they're not even close, even after upsetting the Ravens on Thursday night. This is a team that cannot run the ball, as it ranks 31st in EPA per play on the ground. And the passing game? That's grounded by a porous offensive line that makes life difficult for Tua Tagovailoa. Miami ranks dead last in pass block win rate, a virtually impossible situation for success.
As a result, Tagovailoa has an average time to throw of 2.56 seconds, the third-least in the league. His QBR is 48.4 when opponents have at least one pass rush win, but it's 65.6 when they do not. That is hardly unique to Tagovailoa but just demonstrates the impact of a poor offensive line.
The defense, too, has been a disappointment. While cornerback Xavien Howard was a Defensive Player of the Year candidate a season ago, his numbers have regressed in 2021. He's not alone out there, of course, and Miami's defense has fallen back to 21st in EPA per pass play (down from fourth a year ago) -- though it is trending up after the past couple games.
![](https://a.espncdn.com/combiner/i?img=/i/teamlogos/nfl/500/wsh.png&h=110&w=110)
Washington Football Team
Preseason FPI Rating: minus-0.2
Current FPI Rating: minus-4.0 (down 3.8 points)
Let's be fair here and note right up top: This is a team that lost its quarterback. And while Ryan Fitzpatrick may not be an elite QB, we're still talking about someone who finished eighth and fifth in QBR in 2019 and 2020, respectively, behind bad offensive lines. I think it's safe to assume Fitzpatrick would likely have outperformed Taylor Heinicke (currently 26th in QBR).
But the story here is all about defense, which was supposed to be Washington's strength with edge rushers Chase Young and Montez Sweat leading the way. That has not panned out, to say the least. Washington (2-6) ranks 29th in EPA per play on defense, with most of the issues coming against the pass. It's a good reminder of the volatility of defense from year to year, and that it's less reliable than the other side of the ball.
![](https://a.espncdn.com/combiner/i?img=/i/teamlogos/nfl/500/det.png&h=110&w=110)
Detroit Lions
Preseason FPI Rating: minus-5.4
Current FPI Rating: minus-8.6 (down 3.2 points)
The winless Lions are in the exact same FPI rank as they were in the preseason: 31st. But their rating has dropped by over a field goal. In other words, the model is telling us it always thought the Lions were a bad football team, but it now feels confident they're particularly bad.
What has gone wrong? Well, a few things. Jared Goff is keeping company with Davis Mills and Zach Wilson in the QBR rankings, and the team still has no wide receivers that scare you. D'Andre Swift has been a nice fantasy player, but his rush yards over expectation are the lowest in the league (per NFL Next Gen Stats). And that's the just the offense. Did I mention that's the Lions' better side of the ball?
![](https://a.espncdn.com/combiner/i?img=/i/teamlogos/nfl/500/nyj.png&h=110&w=110)
New York Jets
Preseason FPI Rating: minus-4.9
Current FPI Rating: minus-7.7 (down 2.8 points)
Similar to the Lions, the model thought the Jets were bad in the preseason. Then they stepped on the field and proved the model hadn't gone far enough.
The biggest issue here is Zach Wilson. Here are his season ranks:
Last in completion percentage over expectation (NFL Next Gen Stats)
Second-to-last in QBR
Last in interception rate
Second-worst sack rate
Does it make it better or worse that the offensive line has actually been OK in protection? A bad rookie season doesn't spell doom and gloom for the rest of Wilson's career (he has to get his job back from Mike White first), but the 2-6 Jets' outlook for the rest of 2021 is pretty miserable. And to be clear, both sides of the ball are a problem here too.