October 30, 2014

Outside Foxborough – The Injury Effect

logoby Bill Barnwell
[email protected]

The Miami Dolphins came into this season with high hopes following the acquisitions of quarterback Trent Green, wide receiver Ted Ginn, Jr., and outside linebacker Joey Porter over the long NFL offseason. Replacing the departed Nick Saban was offensive guru Cam Cameron, who’d successfully crafted the Chargers offense into one of the best in football. With Ginn and Chris Chambers stretching defenses deep, Porter and defensive player of the year Jason Taylor coming off the edges, and Green bringing a stability and authority that the quarterback position hadn’t seen since the Marino days, things were looking up for Dolphins fans.

14 games later, the Dolphins are the laughingstock of football, a team that induces tears from their owner upon winning their first game of the season, with a pass thrown from one Chargers castoff, Cleo Lemon, to another, Greg Caramillo. Cameron is likely one and done as a head coach, Chambers has ironically been shipped to San Diego, Ginn’s failed to make an impact at wideout, and while Taylor’s put up 10 sacks, Porter’s only mustered 3.5.

More than all that, though, it’s injury which has struck the Dolphins at their core. Green suffered a severe concussion in Week 5 that’s called his career into jeopardy. Ronnie Brown, who had been among the NFL’s best backs, went down two weeks later. That was also middle linebacker Zach Thomas’ last game of the year. Other starters have missed time: Vonnie Holliday, Travares Tillman, Channing Crowder, Matt Roth, and David Martin have all missed time with injuries this year.

On the other hand, the Patriots have enjoyed some remarkable health this season. Losing Richard Seymour for the first six weeks of the year would have been a problem for most teams, but the combination of Jarvis Green and Mike Wright (himself now gone for the remainder of the season) did an admirable job in his stead. Meanwhile, the Patriots’ long-term injuries consist of Sammy Morris, Wright, and Rosevelt Colvin, who managed to play for most of the season. Ben Watson is the only other injured player of note on the roster.

Obviously, the differences between the Patriots and the Dolphins are not a series of injuries. That being said, would this be the same Patriots team if their starting quarterback, best skill position player, and core middle linebacker had all gone down for the season by Week 7? Obviously not. Injuries are often discussed when it comes to why a team’s had a disappointing season, but when they have a great one, staying healthy is almost never one of the reasons you’ll hear bandied about as a reason why.

And, well, they should be. The effect of injuries on a team can be huge, and while everyone knows they’re bad, there’s been little in the way of research on how much they hurt, which hurt more, and what causes them.

Last year for this very website, I analyzed the injury rates of Patriots defensive backs and attempted to find reasons or trends in the data to explain this injury rash that had broken out. I later updated and expanded my findings inĀ Pro Football Prospectus 2007, the annual published by me and my colleagues at Football Outsiders. With that as a starting point, I’ve done much more research into the trends of injuries and begun to quantify the likelihood and effects of injuries. The results, if confirmed with more data, could be staggeringly important: They could guide teams on how to construct their rosters, while offering gamblers and fans likely picks for teams that will rebound or decline in a given season.

We’ll be doing a two-part feature on the injury effect here on Patriots Daily. This week, I’ll be looking at the year-to-year effects of injuries, why they seem to happen, and what the results are on teams. Next week, we’ll get more specific, looking at offenses and defenses and what injuries are more important than others.

First, the data: The NFL injury reports, from 2001 to 2006, as compiled by Football Outsiders intern Chris Povirk. Injuries were weighted with a simple metric that measured the effect of an injury: A player who was listed as “Out” or placed on IR was scored with four points for each week he was in that role; a player who was listed as “Doubtful” three, “Questionable” two, and “Probable” one.

Following that, we compared the injuries to what we determined were the team’s 22 offensive and defensive starters heading into the season, its “projected” starters. For example, in 2006, although Tony Romo ended up being a huge part of the Cowboys’ success, Drew Bledsoe was the expected starter going into the season. We focused specifically on their injury rates, since the majority of a team’s salary cap, training camp, and focus is placed into those starters. This differs from last year’s research, which looked at a team’s 53-man roster equally.

We then split up the data by position and squad (which will be discussed next week) and compared it to wins and DVOA, our metric which analyzes teams based upon their play-by-play performance after correcting for down, distance, situation, and opponent. (For 2006 teams, in order to analyze their success this year, we used DVOA through 15 weeks, and did not include this year’s wins on their research.) This analysis was not only confined to one year, but we examined a team’s health from year-to-year and compared it to injuries from year-to-year. We used correlation coefficients, which gauge the relationship between two variables, to analyze whether there were significant comparisons between the injury data recorded and a team’s performance. When looking at the correlation coefficients, remember that football relationships tend to be less significant than those of most other comparisons because of the myriad variables involved, so what would be a small relationship in other fields is actually a decent relationship when it comes to football.

The answer to that question is, unequivocally, yes. Over the six seasons that we looked at, injury rates had a dramatic effect on a team’s amount of wins, DVOA, and the change of the two from season to season.

Comparing a team’s injury rate to the number of wins they had in that season, we found a moderate relationship of -.22, indicating that there’s a negative relationship between injuries and wins. That shows some importance to injury rate, but that injuries aren’t a death knell to a team. Where the relationship becomes more intense is when we look at the difference in a team’s wins from a season ago. The correlation between injury rate in a given season and the difference in that team’s wins from a season before is -.38, or nearly twice as strong. When we correlate the difference in a team’s injury rate from year-to-year and the difference in that team’s wins, we get a whopping -.50 as the correlation, a huge indicator.

That’s borne out when we look at the injury outliers in our data. The most injured team in our study were the 2004 Tennessee Titans, who collapsed at the end of their multi-year run by going from 12 wins down to 5. Number two were the 2006 Cleveland Browns, who, of course, have been one of the surprise packages of this season.

On the flip side, the healthiest team we saw across the data were the 2003 Dallas Cowboys. They won ten games that year after winning five the year before. The year after, with a more normal injury rate, they went 6-10. Second were the 2002 Bills, who went 8-8 after going 3-13 in 2001; in 2003, with an average amount of injuries, they went 6-10.

Another issue that comes up is, as you can see, that teams tend to revert back to an average amount of injuries: Simply put, there’s no team that stays snakebitten with injuries. The correlation between a team’s injury rate and the same metric from the previous season is 0.02, or virtually nil. A team that’s extremely banged-up might get hurt a lot the next year, too, but they’re just as likely to stay healthy.

On the other hand, a team that does see a dramatic change is likely to see their fortune change with it. The biggest injury shift involves the Chicago Bears. A healthy Bears team in 2001 went 13-3; the next year, as the most injured team in football, they went 4-12. The year after, they got healthy and got back up to 7-9, but then suffered a serious injury bug in 2004 and stayed at 7-9. Finally healthy in 2005, they went 11-5.. The 2006 Browns were the second-largest shift, and they lost two more games that year than in 2005, while the third-placed 2004 Titans lost seven games more, as we mentioned.

That 2004 Titans team was a veteran team that had been through several years together before collapsing, which of course, brings up some comparisons to this year’s Dolphins, one of the older teams in football. It obviously brings up age as a possible indicator of likely injury. Surprisingly, though, age has little to do with the equation.

Taking the average age of the 22 starters and comparing it to injury rates reveal a correlation of 0.04: Again, virtually no correlation whatsoever. It should be noted that the Dolphins defense is significantly older than the Dolphins offense, and the correlation between defensive injury rates and age (+.18) is much higher than that of offensive injury rates and age (-.02).

Comparing injury rate to DVOA shows that the changes aren’t just superficial and related to lucky or unlucky wins. The correlations are similar to those we saw when looking at the relationship between wins and injury rate, but slightly weaker. For injury rate in a year and a team’s DVOA, there’s a -.16 correlation. If we compare injury rate to the difference in DVOA from the year prior, though, the correlation is -.4, and if we then compare the difference in injury rate to the difference in DVOA, it’s -.46. In short, there’s a very real and dramatic decrease in both the success and performance of a team when they get hurt more.

Looking at the 2006 season, this data would’ve given us several teams to highlight as potential flukes or sleepers because of their injury rates. One of them was the aforementioned Browns, who have stayed healthier and done much better than expected. The Jaguars were the second-most injured team in football, and have also rebounded with a nice year. Third was Tampa Bay, another team who’ve done well somewhat surprisingly. Our DVOA projections pegged all three of these teams as likely to improve.

On the other hand, there are several teams we’d expect to struggle with injury. First and foremost would be the Rams. Although we didn’t expect them to lose their entire offense, there was a strong likelihood that they would lose one or more of their key players to injury, and at times this year, they were without their starting quarterback, running back, and left tackle. Number two, surprisingly, would be the Cowboys, who illuminate the fact that injury rates aren’t infallible. They’ve stayed very healthy and had a superb year. Number three would be the Houston Texans, who will be at least one game better than they were last year, and number four would be San Diego, who have seen their performance drop some, although that likely has more to do with Norv Turner than injury.

Next week, we’ll analyze offense and defense and find out who’s more likely to get hurt and what the results are.

Comments

  1. Interesting stuff.
    I don’t know that you’ll be able to single out the effect of injuries. Those good to bad and bad to good teams may have suffered a different amount of injuries in the 2 years, but they also were facing a schedule made harder (or easier) in the second year as a result of their record in the first year. You’d also hope a bad team has the benefit of a good draft in that second year, although with the way some teams draft thats far from certain.

  2. Interesting stuff, Bill. Where did the 2003 Patriots fall in the list? They’ve been talked about often as a team that won despite an inordinate number of man hours lost to injury.

    Also, I considered Chad Scott a significant injury. It happened in training camp, so it’s easy to forget, but he was having a good camp and looked to be challenging as a starting corner.

  3. CoreyCroom says:

    This information is not suprising to me, and seems crucial to NFL GMs. I feel like Piloi and Belichick might know this already, accounting for some of their management edge over other NFL teams. Injuries are probably the strongest random element determining success or failure in the NFL. One of the most under-rated aspects of the 2003 Pats squad that won the SB was overcoming a rash of key injuries to LB and OL early in the year and hanging in long enough to get healthy down the stretch. Without Brady, my speculation is this 14-0 team would likely be in the 7-7 to 9-5 range. This is a perfect hypothetical to imagine the huge impact of a random injury. Unfortunately, only one reality exists, so we can’t play out the hypothetical 2007 Pats with Cassel at QB for a control to the thought experiment. Bottom line, be prepared to minimize the impact of injuries as an NFL GM in the future. Two Jabar Gaffneys might truly be better than one Chad Johnson when random injuries are taken into account.

  4. Bill Barnwell says:

    The 2003 Patriots are in the top ten for most injured teams in the timeframe, actually.

Leave a Reply