AGL Injury Numbers Point to Ravens Rebound in 2022

Baltimore Ravens QB Lamar Jackson
Baltimore Ravens QB Lamar Jackson
Photo: USA Today Sports Images

NFL Offseason - For more than two decades now, Football Outsiders has collected data from the NFL's weekly injury reports and transformed it into adjusted games lost. Adjusted games lost doesn't just add up total injuries. It accounts for both absent players and those playing at less than 100%, and it gives more weight to injuries to expected starters and situational players than to expected back-ups. As such, AGL estimates the impact of injuries on teams and provides a comparable total that often succinctly explains why teams improved or declined from one year to the next. Tom Brady deserves much of the credit, but the Bucs' rise from 5-11 and last place in the NFC South in 2018 to 11-5 and Super Bowl champions in 2020 followed a two-year ascension from 30th to first in AGL. The 49ers' fall from 13-3 and NFC champions in 2019 to 6-10 and fourth place in the NFC West in 2020 followed a one-year jump of 70.8 AGL, the most in football from 2019 to 2020. And the Ravens … well, I'll get to them shortly.

We haven't changed our methodology, but AGL has become harder to compare across seasons the last two years. In 2020, COVID-19 cost teams 401.0 AGL, which was 15% of the overall total. And while COVID AGL fell by nearly half at 219.3 last season, the NFL also expanded its regular-season schedule to 17 games. Those factors combined to push AGL to a new record of 2,806.7, or 87.7 per team in 2021. For many, that total will serve as confirmation of the NFL's hypocrisy for expanding its schedule while championing its investment in player safety initiatives. But it is worth noting that 2021 even with the extra week had fewer non-COVID AGL than 2016 (80.8 versus 82.2). And prorated to 16 games, 2021 continued the plateau of AGL that started in 2014.

Change in AGL Over Time

If one assumes that AGL would have continued the upward trajectory it followed prior to 2014 if left unchecked—as the average player continued to become stronger and faster—then the NFL likely has made the game safer.

Team Results

Coming off back-to-back seasons in the top seven in total DVOA, the Ravens were poised to test the true value of running backs when they lost J.K. Dobbins, Gus Edwards, and Justice Hill—the three members of their expected backfield committee—to preseason ACL and Achilles injuries. But instead, those losses proved to be the tip of an injury iceberg that sank the Ravens' 2021 season.

2021 AGL Results
Team 2021
Rk 2020
KC 38.6 1 65.2 11 -26.6 17.5%
BUF 42.8 2 63.1 9 -20.3 27.7%
ATL 48.1 3 48.0 3 +0.1 -29.6%
MIA 50.9 4 65.7 12 -14.8 -11.1%
LAR 54.3 5 45.6 2 +8.7 21.6%
SEA 63.2 6 88.3 20 -25.1 11.8%
CIN 63.2 7 88.8 21 -25.6 0.0%
DAL 69.9 8 118.5 28 -48.6 30.9%
LAC 70.2 9 109.1 27 -38.9 8.5%
TEN 73.4 10 57.5 6 +15.9 -3.0%
NE 75.6 11 134.8 31 -59.2 23.3%
PHI 76.0 12 128.1 30 -52.1 3.8%
CHI 79.6 13 87.1 19 -7.5 -10.0%
GB 80.1 14 73.1 15 +7.0 11.5%
CAR 80.8 15 73.1 14 +7.7 -24.1%
TB 81.3 16 30.6 1 +50.7 27.6%
HOU 82.3 17 58.2 7 +24.1 -24.2%
CLE 85.9 18 89.7 22 -3.8 4.7%
JAX 87.2 19 63.4 10 +23.8 -31.3%
PIT 89.8 20 55.2 5 +34.6 -10.4%
MIN 89.8 21 83.5 16 +6.3 2.7%
ARI 92.3 22 95.7 23 -3.4 10.9%
IND 96.5 23 72.9 13 +23.6 9.8%
LV 101.6 24 86.9 18 +14.7 -5.2%
NO 107.0 25 49.2 4 +57.8 4.4%
WAS 107.4 26 97.0 24 +10.4 -10.0%
NYG 109.2 27 104.4 25 +4.8 -29.6%
DEN 118.1 28 108.4 26 +9.7 -1.7%
SF 119.5 29 166.6 32 -47.1 19.5%
DET 122.4 30 83.7 17 +38.7 -27.1%
NYJ 158.5 31 123.9 29 +34.6 -18.3%
BAL 191.2 32 59.6 8 +131.6 -2.1%

September signees Devonta Freeman and Latavius Murray staved off the decline most backs see as they enter their thirties with above-average rushing DVOA rates. But by November, running backs were the least of the Ravens' injury worries. All-Pro players Ronnie Stanley and Marlon Humphrey missed time with ankle and pectoral injuries. First-round rookie Rashod Bateman had a groin injury that delayed his debut until mid-October. Twelve expected starters, replacement starters, and important situational players landed on the non-COVID injured reserve at some point in the regular season. And in Weeks 15 and 16, the team was missing nine defensive backs, including each of their four expected starters. They lost both games as part of seven straight losses to end their season. All told, the Ravens racked up 191.2 AGL. It was the most by a team in our database, and it remains the most when prorated to 16 games.

Most AGL Per 16 Games, 2001-21
Team Season AGL Per 16 DVOA
BAL 2021 191.2 180.0 -2.1%
CHI 2016 171.6 171.6 -8.8%
SF 2020 166.6 166.6 5.4%
NYJ 2019 160.1 160.1 -17.7%
NYG 2015 152.0 152.0 -7.4%
NYJ 2021 158.5 149.2 -18.3%
NYG 2013 140.5 140.5 -16.5%
WAS 2017 138.8 138.8 -2.6%
NYG 2014 136.1 136.1 -7.8%
NE 2020 134.8 134.8 -7.1%

The Ravens did get 12 games from former MVP quarterback Lamar Jackson, and that no doubt contributed to their relative success on offense (2.0% DVOA, 17th) versus defense (9.3%, 28th) despite a slightly bigger total of 103.2 versus 88.0 AGL on his side of the ball. Still, the Ravens made an impressive effort to even threaten a playoff berth before their collapse to an 8-9 record. The team's -2.1% overall DVOA was second best of the 10 teams with the biggest pro-rated AGL totals this century. And AGL and DVOA correlated in 2021 and correlate in general with a coefficient of about -.30. Aided by the easiest projected schedule by DVOA, the Ravens are a good bet to bounce back to contention in 2022 with better injury fortune even after Cincinnati's maturation and Cleveland's offseason additions in their division.

Speaking of the Super Bowl runners-up, the Bengals were one of the three teams in the final four of the playoffs that finished in the top eight in fewest AGL. And the top two teams by DVOA, the Cowboys and Bills, finished there as well. One could argue that the Chiefs and Bills needed Patrick Mahomes and Josh Allen more than they needed the smallest two AGL totals for their success, but it clearly helps to have fewer injuries. The Super Bowl champion Rams make the best case for that. With an extreme preference for stars over prospects and quality depth, the Rams rode healthy seasons from the bulk of their blue-chip players such as Matthew Stafford, Cooper Kupp, Aaron Donald, and Jalen Ramsey to the title. Robert Woods, Cam Akers, Sebastian Joseph-Day, and Justin Hollins were the only expected Rams starters to miss extended time this season, and the team made aggressive moves to replace those losses with their Odell Beckham signing and Sony Michel and Von Miller trades.

The 49ers deserve the same recognition I gave the Ravens for their NFC title run despite the fourth-most AGL. In fact, the 49ers authored the lone above-average DVOA season with a top-10 all-time AGL total just a few years ago in 2019. But those overachievements may obscure a problem. The 49ers have suffered from 93.6 AGL or more in all five of Kyle Shanahan's seasons with the team. And all five of those totals landed the team in the bottom third of teams those seasons. Maybe that's poor luck, but AGL correlates with itself (0.25 from 2020 to 2021) about as well as it does with DVOA. And signing a player with a checkered injury history such as Jason Verrett without a strong backup plan suggests a cavalier attitude toward injury risk. The Rams have suffered less than half as many AGL as the 49ers in Sean McVay and Shanahan's tenures (229.1 versus 572.8), and the former situation makes life so much easier for a head coach.

The Bucs were poised for injury regression after finishing first with just 30.6 AGL in their Super Bowl season. But they jumped from 30.6 AGL in 2020 to 81.3 in 2021 and were one just three teams that increased by 50.0 AGL, or even 40.0. Their increase was still less than half of the 131.6-AGL spike the Ravens suffered, but like the Ravens, the Bucs saw some a concentration of injuries at the same positions and at the same times of the year. In September and October, those injuries clustered at defensive back where Carlton Davis, Antoine Winfield, Sean Murphy-Bunting, and Jamel Dean all missed time. In December and January, they clustered at wide receiver, where Chris Godwin, Breshad Perriman, and Cyril Grayson either missed time or were forced to play injured, which left Tom Brady with Mike Evans, Rob Gronkowski, and little else to target after Antonio Brown was dismissed from the team.

Two of the most surprising playoff teams, the Patriots and Eagles, owe at least some of their resurgence to their more than 50.0-AGL declines from 2020 to 2021. In 2020, the Patriots had almost three times as many COVID losses as any other team. Linebacker Dont'a Hightower and running Branden Bolden were the only two players from among those opt outs who returned to help the 2021 team, but avoiding those unexpected absences meant that Bill Belichick's uncharacteristic free-agency spending successfully filled the team's biggest holes. And while the Eagles still fared worse on offense (46.4 AGL) than defense (29.5) in 2021, they avoided the cluster of offensive line injuries that cost them Brandon Brooks, Lane Johnson, Isaac Seumalo, and Andre Dillard in 2020. Quarterback Jalen Hurts invites some pressure with a style that leans on scrambling and downfield passing, but the Eagles improved from a 60% pass block win rate in 2020 to 67% in 2021, third best in football.

Offense vs. Defense

The Saints faced a daunting task in replacing Drew Brees, their quarterback starter for the previous 15 seasons. And that effort was made dramatically more difficult by 76.5 offensive AGL, the second most on that side of the ball.

Offensive and Defensive AGL, 2021
Team Offensive
Rk Defensive
Rk Dif
ATL 17.4 1 30.7 11 +13.4
KC 17.7 2 20.9 3 +3.3
BUF 19.3 3 23.4 5 +4.1
DAL 19.4 4 50.5 25 +31.1
TEN 29.7 5 43.7 20 +14.0
LAR 30.3 6 24.0 6 -6.4
CHI 30.5 7 49.1 24 +18.5
CIN 31.9 8 31.3 12 -0.6
TB 34.4 9 46.9 22 +12.5
LAC 36.0 10 34.2 13 -1.8
NE 39.2 11 36.5 17 -2.7
ARI 40.0 12 52.4 26 +12.4
SEA 41.1 13 22.1 4 -19.0
DEN 42.0 14 76.1 30 +34.1
MIA 42.2 15 8.7 1 -33.5
PIT 42.3 16 47.5 23 +5.2
LV 42.7 17 58.9 27 +16.2
CAR 43.5 18 37.2 18 -6.3
SF 45.0 19 74.5 29 +29.5
PHI 46.4 20 29.5 9 -16.9
MIN 48.5 21 41.4 19 -7.1
CLE 50.8 22 35.1 14 -15.7
IND 51.7 23 44.8 21 -6.9
HOU 53.4 24 28.9 8 -24.5
GB 54.4 25 25.6 7 -28.8
NYJ 57.1 26 101.3 32 +44.2
DET 61.6 27 60.8 28 -0.8
JAX 67.9 28 19.2 2 -48.7
WAS 72.1 29 35.4 15 -36.7
NYG 73.5 30 35.6 16 -37.9
NO 76.5 31 30.5 10 -46.0
BAL 103.2 32 88.0 31 -15.2

Jameis Winston surprised with a 13.6% passing DVOA that was the second best of his career and would have made him 10th among regular starters if he had hit the 200 pass attempts to qualify. But his knee injury forced the Saints into quarterback plans C, D, and E in Trevor Siemian, Taysom Hill, and Ian Book. And that deep into their depth chart, the Saints would likely have fallen from contention even if they weren't also missing important offensive pieces such as linemen Terron Armstead, Ryan Ramcyzk, and Andrus Peat, and receivers Michael Thomas, Tre'Quan Smith, and Adam Trautman.

New Giants head coach Brian Daboll was complimentary of incumbent quarterback Daniel Jones in his introductory press conference and has since made just a minor addition at the position in Tyrod Taylor. That may surprise some outsiders after Jones failed to crack even -10.0% passing DVOA for this third consecutive season, but Daboll can add injury issues to his list of Giants instabilities in Jones' tenure. Last season, the Giants suffered 73.5 offensive AGL, third-most in football. And that at times left Jones with career backups and practice squad players such as Collin Johnson, Dante Pettis, C.J. Board, and David Sills as primary targets. Three years is more rope than many quarterback prospects see, but Jones can fairly claim he has yet to enjoy a fair chance to evaluate his capacity to become a franchise quarterback.

Robert Saleh spearheaded the 49ers' transformation from a bottom-10 defense by DVOA in 2017 and 2018 to a top-six one in 2019 and 2020. But his first year as Jets head coach saw his new defense trend the other way from 21st to 32nd in DVOA, and injuries were a big reason why. With 101.3 defensive AGL, the Jets were the only team to surpass the Ravens on either side of the ball. And those injuries struck seemingly every blue-chip defender the team had, from edge rusher Carl Lawson to safeties Marcus Maye and Lamarcus Joyner. Saleh will need better defensive health as much as a sophomore Zach Wilson breakout to turn his new team around.

The theory that Russell Wilson forced a trade to Denver to enjoy a better supporting cast cannot count on DVOA for support—the Seahawks and Broncos had nearly identical 3.3% and 3.0% defensive DVOA rates in 2021, and the former team was better on both offense and special teams. But Broncos fans can take some solace in their team's 76.1 defensive AGL, third most in football. Pass-rusher Bradley Chubb and linebacker Josey Jewell would likely have been difference-makers on their own in their returns from ankle and pectoral injuries. But free-agent additions Randy Gregory and D.J. Jones should balance the team's defensive line and, with improved health, give Wilson a shot at his first top-10 DVOA defense since 2016.

Later this week in Part II: a look at AGL by position group.


17 comments, Last at 28 Mar 2022, 10:12am

1 as the average player…

as the average player continued to become stronger and faster

  1. Do you have metrics quantifying stronger and faster across full-rosters?
  2. What's your assumed mechanism for stronger+faster=increased injury count? (Or rate? Not sure which you are assuming)

This may be asking a lot for a free article, but that first table would benefit greatly from a 2020 DVOA and a 2021-2020 DVOA diff, too, to see the effect of +/- on AGL. (You strongly suggest a correlation with the 2019-2020-2021 49ers, for instance)

Also, poor Atlanta. Healthy and terrible.

2 Maybe that's poor luck, but…

Maybe that's poor luck, but AGL correlates with itself (0.25 from 2020 to 2021) about as well as it does with DVOA. And signing a player with a checkered injury history such as Jason Verrett without a strong backup plan suggests a cavalier attitude toward injury risk. 

Do you think the correlation is based on roster selection, playing style, or both? Does AGL travel with a coaching staff? Or (glances at the NY teams) is it franchise-rooted?

5 I would argue it's a cursed…

I would argue it's a cursed indian graveyard, but they built it on an uninhabited swamp.

Maybe it's a cursed mobster graveyard and they are taking the under.

On the other other hand, Saleh did come from SF, so play-style is still on the table.

7 I wondered tongue in…

I wondered tongue in cheekily about this last year as McCarthy seemed to take his injury record with him to Dallas while GB improved. The effect got weaker this year.

In the 14 seasons since 2008 when McCarthy joined the league (13 of which he coached in since he didn't coach in 2019) Dallas had their 9th and 14th worst AGL under McCarthy. I used the AGL without COVID that they provided last year for that too to keep as apples to apples as I could.

GBs 3 non McCarthy seasons out of those 14 are 7th, 4th, and 9th. So GB does appear to be healthier without McCarthy and Dallas still appears to be more injured with him.

Of course like I said last year this was mostly tongue and cheek since there is no where near the data in the sample I used to draw any real conclusions. However I think the data might be out there to determine if there is an effect even with how noisy the data is going to be. Of course even if one is found it might be measuring how a coaching staff uses the injury report and not if a team actually suffers more injuries since AGL is based on that injury report data.

8 Easiest projected schedule

Do the Ravens really have the easiest projected schedule anymore?  You did your projections before they had to play Russell Wilson, Tom Brady, and Watson twice.

9 Titans injuries

Surprised the Titans ended up in the top 10 healthiest teams considering the stat about them using so many players (yes, I know a lot of that was intentional, but still).

10 Historically, the Ravens…

Historically, the Ravens have been a pretty good agl team. This year's faceplant is rather unusual.

A way too early and unfounded conjecture would say that DeCosta isn't so averse to drafting injury prone players vs Ozzie. I have no sense of that's true as I don't follow the team, but if the Ravens continue to be a mash unit in the future seasons, I'll be more convinced of this.

12 I don't think that's really…

I don't think that's really accurate, the only one who fits the bill is Marquise Brown and he was one of the few survivors last season. Bateman got ravaged by Covid in college but didn't get injured until after being drafted. The bigger problem is EDC whiffing on mid-round picks - 4 picks at guard and a fullback, and none have won the starting job.

That, and opposing linemen falling on the back of the legs of 2 of the Ravens' best (and most expensive) blockers. It's still unclear if Nick Boyle or Ronnie Stanley can ever play again.

11 Totally agree that it is too…

Totally agree that it is too early to blame DeCosta for anything.  However, when you are replacing Ozzie, it is always a step down, it would be like replacing Brady or Manning.

I think Ozzie was the best GM in any sport in his era, I feel that the man must have the most incredible ability of looking at film of college athletes and making judgments.  Then he created his board and ranked the players from 1-300.  He drafted hall of famers, and over-performers in later rounds.  The man was a genius at player assessment.

16 "Last season, the Giants…

"Last season, the Giants suffered 73.5 offensive AGL, third-most in football. And that at times left Jones with career backups and practice squad players such as Collin Johnson, Dante Pettis, C.J. Board, and David Sills as primary targets. Three years is more rope than many quarterback prospects see, but Jones can fairly claim he has yet to enjoy a fair chance to evaluate his capacity to become a franchise quarterback."

I have a hunch when the position article comes out that the Giants WRs won't be that bad, actually. I'd argue as far as the targets go, their receiver depth held up pretty well and Jones didn't really have the excuses you mentioned as primary targets.

Their top 5 WRs last year were Golladay, Toney, Shepard, Slayton, and Ross. These were the top 5 in WR yards, Ross was one target behind Johnson for 5th in that category, and - excepting CJ Board's 6 targets - they were the top 5 in WR yds/tgt. Combine that with Engram and Rudolph and you've got your 7 main targets.

Only looking at box scores for these 7 players (so a player could've played but not have been targeted), of the 11 games Jones started:

  • 4 had 4 of the receivers, at least 1 TE
  • 4 had 3 of the receivers, both TEs
  • 2 had 3 of the receivers, 1 TE
  • 1 had 2 of the receivers, both TEs

So while some combination of Johnson, Board, Pettis, Sills, and Pharoh Cooper were targeted in 9 of the 11 games, they had at least 3 of the receivers, including at least 2 of Golladay/Slayton/Toney, in 10 of them.

17 Packers d agl

Maybe I'm missing something but Z smith missed 16 games.  J Alexander missed 13.  2 starters alone missed 29 games, yet the packers are listed at 25.6.

How do you count this?