2014 Adjusted Games Lost
by Scott Kacsmar
Once upon a time a series of hamstring injuries made us wonder what exactly the Giants were going to get out of Odell Beckham Jr. He was the last 2014 first-round rookie to make his NFL debut, doing so in Week 5. Quickly we found out he could play as he put in one of the best rookie seasons ever in just 12 games.
But that quarter of a season he missed makes us wonder what more he could have achieved in 2014. The Giants had a lot of those "if only" dreams last year as they became the first team to lead the league in Adjusted Games Lost in back-to-back seasons. This is the fifth year in a row the Giants ranked 22nd or worse, and they have missed the playoffs in four of those seasons.
With Football Outsiders' Adjusted Games Lost (AGL) metric, we are able to quantify how much teams were affected by injuries based on two principles: (1) Injuries to starters, injury replacements, and important situational reserves matter more than injuries to benchwarmers; and (2) Injured players who do take the field are usually playing with reduced ability, which is why AGL is not based strictly on whether or not the player is active for the game, but instead is based on the player's listed status that week (IR/PUP, out, doubtful, questionable, or probable).
As long as NFL teams are solely responsible for producing weekly injury reports, we cannot say that every single injury has been accounted for, but secrecy is an unavoidable aspect of this part of the game.
Obviously every team would prefer a low AGL, but some teams will inevitably have a tough year of injuries. The following table lists the AGL totals and rankings for 2014 along with the results from 2013 for comparison. Teams are sorted from healthiest in 2014 (Denver) to most injured (New York Giants). This data is only for the regular season.
|Team||2013 AGL||2013 Rank||2014 AGL||2014 Rank|
|Team||2013 AGL||2013 Rank||2014 AGL||2014 Rank|
The league-average AGL went from 67.6 in 2013 to is 74.3, making it four years in a row that injury totals around the league went up in our AGL database (complete for 2002-2014). There are a couple of issues here. AGL numbers count starters (including players who take over starting roles after other injuries, such as five different Chargers centers) and important situational players. If we look only at actual starters, we see that the average AGL has increased from the mid-40s in 2007-2010 to the lower 50s the last couple seasons, and now 55.8 in 2014. So teams are clearly reporting more injuries now than they did in the past. In addition, we have better information to identify reserves who play a significant amount of snaps on a consistent basis and therefore should count in AGL. This is an area where we know we can improve our past years of data, and that is certainly on the to-do list. We have already made some improvements to 2013, which is why last year's numbers may look different in the table above.
A team's third wideout and nickel cornerback are practically starters today, and the rotations used in the defensive front seven are eye-opening for some teams. Starting defensive linemen in 2014 played an average of 64.2 percent of the weekly snaps. There were only 33 instances of a defensive lineman playing 100 percent of his team's snaps in any given game. New England's Rob Ninkovich accounted for 12 of those ironman performances. Chandler Jones had four complete games of his own, meaning 16 of the 33 belonged to the Patriots.
The correlation for AGL between 2013 and 2014 was 0.40, the second highest year-to-year correlation on record. This is more in line with results from 2009-2012, as opposed to the smaller 0.12 correlation between 2012 and 2013 AGL.
The correlation between 2014 AGL and 2014 team DVOA was -0.35, and the correlation between 2014 AGL and 2014 regular-season wins was -0.33. Both correlations are a little stronger than what we have observed over the years.
Denver, Green Bay, and Pittsburgh all made the playoffs, but exemplify the misfortune of bad timing in regards to injuries. Maybe things would have ended differently if Peyton Manning (quad), Aaron Rodgers (calf), and Le'Veon Bell (knee) had been healthy in January. Denver had the 10th-biggest decline in total AGL since 2002, but Green Bay made this season's biggest improvement in the rankings from 31st to third. That is the third-biggest decline in total AGL since 2002. Mike McCarthy's teams have usually been poor at AGL, but this was a fortunate year, especially on offense. We are probably over-crediting the short-term IR for center J.C. Tretter, because fifth-round rookie Corey Linsley was able to step in Week 1 and earn that starting job even after Tretter was healthy.
Philadelphia, noted for Chip Kelly's foray into sports science last year, had the best AGL in 2013 and ranked fifth this season. They are only the third team since 2002 to lead the league in AGL and finish in the top five the following season. Maybe this team is onto something with preventing soft tissue injuries. The big problems for the Eagles were focused along the offensive line, plus a broken collarbone for Nick Foles and a torn Achilles for DeMeco Ryans. Sometimes bones are going to break in tackles regardless of how much prep work goes into each week.
After they ranked second in AGL in 2013, we expected the Chiefs to fall back to the pack this year. They didn't just fall back; they plunged to 26th, the seventh-largest year-to-year increase in AGL since 2002. The Chiefs were hit hard and fast with Derrick Johnson, Mike DeVito, and Jeff Allen (three starters) headed to injured reserve after Week 1. The worst news was Eric Berry's diagnosis of Hodgkin's lymphoma, and like everyone around the league, we wish him the best in his recovery.
The two Super Bowl teams, New England and Seattle, ranked in the teens in AGL. The Patriots lost Jerod Mayo for the second straight season, but overall had a pretty healthy year with red-flag players like Rob Gronkowski and Darrelle Revis able to finish the season. Seattle's injury issues, of course, were much stronger in the playoffs, which are not included in the AGL totals above. The Super Bowl loss ended up with practically the entire "Legion of Boom" secondary gutting out tough injuries.
The Colts ranked 30th in AGL for the third year in a row and have ranked 24th or worse in nine straight seasons. Jon Torine was the strength and conditioning coach from 1998 to 2011, but he was replaced in 2012 by Roger Marandino. Despite three head coaches since 2002, Ryan Grigson's annual roster purge, and new philosophies on both sides of the ball, the Colts continue to be one of the NFL's most injured teams each season. Yes, I just copied most of last year's paragraph, because nothing changed in Indianapolis on the injury front. Among the bottom 10 teams in 2014 AGL, the Colts were the only team to make the playoffs. The 2012 Packers, 2013 Colts and 2014 Colts (two games) have the only four playoff wins by teams with more than 100 AGL.
Then we have the Giants, looking to put together an injury dynasty. After setting the benchmark with 141.3 AGL in 2013, the 2014 club has the second-worst AGL on record at 137.1. Running back and defensive back remained two major problem areas despite new roster additions, but wide receiver was also hit hard, as were the linebackers.
Here are the 2014 AGL splits for offense and defense:
The 2013 Giants' record for worst offensive AGL (80.9) only stood one year. San Diego edged them out this year with the use of roughly nine thousand centers, though the 2014 Giants also rank among the ten most-injured offenses in our database.
Pittsburgh had the 10th-lowest offensive AGL since 2003. Ben Roethlisberger only sat out seven regular-season snaps (all in a blowout win) after missing none in 2013. Starting right tackle Marcus Gilbert missing four games proved to be the only real significant injury to the offense until Le'Veon Bell went down late in the year.
[ad placeholder 3]
The 2014 Raiders now have the worst defensive AGL in our database. Just ahead of them are the 2014 49ers, who would probably rank as the worst if we included suspensions (the Aldon Smith factor). This was a really tough season for that unit, with Glenn Dorsey, Patrick Willis, Chris Borland, Ian Williams, Jimmie Ward, and NaVorro Bowman all on injured reserve by season's end.
Carolina had the third-biggest ranking drop in defensive DVOA from 2013, falling from 15th to third. That sure did not feel like a healthy unit, but injuries were not a big factor in the performance decline. That was more about the major roster turnover in the secondary and Greg Hardy's off-field situation, which limited him to one game but does not factor into AGL. The Panthers also finished the season much better on defense after that miserable start. The Giants fell 19th spots in DVOA from sixth to 25th. That is a unit that can blame injuries, with the third-worst defensive AGL.
Injury Reporting Tactics
Everyone knows the injury reports are not always on the level in the NFL. Seattle safety Kam Chancellor was probable for Super Bowl XLIX, but of course we found out later that he played that game with a torn MCL and a deep bone bruise on the outside of his knee. Chancellor didn't even know if he could play two days before the big game. Call it gamesmanship or outright deception, but some teams like to cover up as much truth as possible.
We can use the given data to determine which teams may be trying to deceive the opponent with the likelihood of a player playing. Generally, probable is a very high likelihood the player will play, especially if the player is a starter or key reserve. Just fewer than 95 percent of players listed as probable played in 2014. Only Carolina (89.5 percent) and Tennessee (86.2 percent) were under 90 percent. Questionable is supposed to be a 50/50 proposition, and doubtful might as well mean out these days. Out of 216 doubtful players in 2014, only Seattle's Alvin Bailey was active that week, and he did not play.
The grayest area comes with listing a player as questionable. Only 55.7 percent of questionable players played in 2014, which is actually the lowest we have ever tracked and a lot closer to the purpose of that designation. In 2013, 61.3 percent of questionable players played and 69.3 percent in 2012. So a downward trend that pushes closer to 50 percent is a good thing.
The range of active questionable players was wider this season. The Steelers only played one of their league-low 12 questionable players (8.3 percent). Some of the other teams on the low end were Denver (third at 35.3 percent) and Seattle (fifth at 40.9 percent) after having the two lowest rates in 2013. The Falcons, after 84.2 percent in 2013, were still in the top six at 65.4 percent in Mike Smith's final year on the job.
The 2014 Redskins with Jay Gruden in his first year as head coach led the way with 79.0 percent of his questionable players playing. You can find Gruden's staff doing a suspect job as early as Week 1. Defensive lineman Barry Cofield sprained his ankle in Week 1 and was placed on short-term IR. In November, Cofield revealed to CSNWashington.com that he had a groin injury that was bothering him a lot. He had groin surgery while out for the ankle, but the Redskins never disclosed any groin injury for Cofield. The team contends Cofield was healthy for Week 1, but it's easy to be skeptical.
Gruden has a long way to go to catch up to the master of injury report shenanigans: Bill Belichick and the Patriots. While New England only had the fifth-highest rate of active questionable players (66.3 percent), the Patriots blew the league away again with 104 questionable players -- 30 more than runner-up Tampa Bay. The other 31 teams averaged 31.2 questionable players. New England only used probable 36 times, the third-lowest total in 2014.
[ad placeholder 4]
Rex Ryan's Jets finished second in most probable players for the second year in a row. Interestingly enough, the Houston Texans had the most probable players again (170) despite the switch to Bill O'Brien, a member of Belichick's coaching tree. Houston was one of three teams (Atlanta and Tennessee the others) to never use doubtful in 2013. The Falcons and Texans repeated their actions in 2014, but the Titans used it in Ken Whisenhunt's first season on the job. Sean Payton and the Saints were the third team to never use doubtful in 2014.
The Saints were involved in one of the most puzzling injury transactions of the season. Rookie linebacker Khairi Fortt was drafted in the fourth round and made the final 53-man roster. He was placed on short-term IR with an undisclosed leg injury suffered in the preseason, yet Fortt insists his hamstring was healthy since Week 1. The Saints cut Fortt on October 6, which is downright odd given he was a fourth-round pick and was deemed valuable enough to warrant the use of the season's lone short-term IR tag. The Bengals quickly added him to their practice squad, but he was cut in November and picked up by the Jaguars. He appeared in three games for his third team last season.
After three years of its existence, I have to crown the 2014 Saints with the worst use of the short-term IR designation. Fortt's story is just another example of the cynicism in believing teams when it comes to injuries.
Tomorrow we will look at the AGL results by unit.
17 comments, Last at 10 Mar 2015, 11:34pm
#17 by DisplacedPackerFan // Mar 10, 2015 - 11:34pm
Green Bay adopted a lot of the "Sports Science" stuff this year as well, they were doing a lot of the same stuff as Kelley was doing in Philly. So that is another data point to watch to see if some of those methods actually work. They also completely reworked their practice schedule in an effort to try an cut down injuries as well. I still feel it's just coincidence and luck, but I'm hopeful that they found something and did turn things around. It was nice to not wonder who had to be next up this year like I've had to pretty much every year for the last decade.
#13 by chemical burn // Mar 05, 2015 - 2:51pm
Man, Philly's defense was so healthy this year - basically Kendricks and Ryans were the only guys to miss time. Depth is going to continue to be a big issue for this team...
#9 by bravehoptoad // Mar 05, 2015 - 1:41pm
Wow. San Francisco's defensive AGL was the worst ever (if you count the Aldon Smith suspension), but the team finished #5 in defensive DVOA. No matter how good your depth, that's got to take great coaching.
So naturally most of that defensive coaching staff was fired.
#11 by coboney // Mar 05, 2015 - 2:39pm
No one ever said, sadly, power struggles make for rational decisions. Lord knows it didn't in San Fran when it came to the coaching situation.
#12 by chemical burn // Mar 05, 2015 - 2:49pm
Yeah, I don't know what I'd do if I were a SF fan - I'd be in total despair. Especially since the Tomsula hire looks like such a joke.
#14 by Karl Cuba // Mar 06, 2015 - 10:58am
I think most of the coaches on staff are good, experienced guys. The problem is with the OC, DC and head coach, which is less than ideal.
#15 by chemical burn // Mar 06, 2015 - 4:27pm
Yeah, but doesn't losing Harbaugh kill you? Coaches that good don't come along with any regularity.
#6 by ChristopherS // Mar 05, 2015 - 12:22pm
What's remarkable isn't that the Giants top the list both times - it's how far beyond any other team they are. This is especially true in 2013, when second-worst GB was closer to average than to NYG.
Or: between 2013 and 2014 the r^2 is 0.160 with NYG and 0.035 without them.
#7 by chemical burn // Mar 05, 2015 - 1:00pm
What's explanation? I know they have a few key players who just seem injury prone, but that's a pretty outlandish thing you pointed out.
#10 by ChristopherS // Mar 05, 2015 - 2:34pm
Oh, personally I have no idea. My guess: just an outlier. (but quite an outlier indeed)
#5 by awsalick // Mar 05, 2015 - 11:31am
Thanks, Scott. I've always liked this article.
One thought: In order to determine the gravity of each injury, could you try running the numbers with 'Player salary lost' as a measurement instead of 'Games Lost'? It would be nice to weight an injury to a star player instead of a replacement-level rookie, and salary might be a way to account for that somewhat. (It obviously would underweight 2nd+3rd year stars still on their first rookie contract, but it still might be better than nothing.)
#3 by herewegobrowni… // Mar 04, 2015 - 7:12pm
CLE 41.1 3 (for 2013)
Wow, this is surprising. If Brian Hoyer had played 14 games starting from the first he had we could've realistically seen a better year than '14 (even though his '14 completion percentage was actually lower than Jason Campbell's 2013 percentage.)
#1 by theslothook // Mar 04, 2015 - 5:59pm
As a colts fan, I'm tired of this. I don't know what else to say.
#2 by Dr. Mooch // Mar 04, 2015 - 7:06pm
Correlation between 2014 opening day active roster and 2014 AGL is 0.44. Only 3 teams were older than the Colts at that point, so they're not making it any easier on themselves.
#4 by lvsilbs // Mar 05, 2015 - 9:54am
As a Giants fan I couldn't agree with you more on the tired of being consistently injured thing.
#8 by jklps // Mar 05, 2015 - 1:40pm
Colts fans have had the enjoyment of rooting for a team led by Peyton Manning or Andrew Luck every season since 1998 except for one season. It is hard to be sympathetic to their fans at all.
#16 by Paul R // Mar 09, 2015 - 11:36am
There's a difference between seeking sympathy and expressing frustration. If you've got one of the fastest cars in the field, but you don't win the race, you get frustrated. That's natural.