Are NFL Teams Faking Injuries?

Sorry for the recent absence, I returned from Beirut and went right into recruiting season.  Once I’ve accepted an offer I’ll start posting again. In the meantime, I’ve got a guest post (unedited) from Jared Cohen (previous posts include the 4th down chart and the kick return strategy post).  You can find the original here. and follow the author on Twitter @jaredscohen.

——–

Given all the animated discussion over the Patriots tactics against the Ravens in their divisional round playoff game, I thought it would be as good a time as any to post some gamesmanship research.

If you read about the game – you know the Ravens were a bit upset with the Patriots usage of receiver eligibility to disguise their offense. The response from the Patriots was, well, Patriots-like. If it’s not against the letter of the law, it’s all good (unless it’s videotaping other teams, in which case even the law doesn’t matter).

Clearly, the NFL is a league where teams will look for any edge, even if it means pushing the bounds of fair competition.

So it’s with that issue in mind that I started digging into the possibility that players are faking injuries.

As a Philadelphia sports fan, I’m generally inclined to assume that my teams will ultimately lose, and so once the Eagles started running Chip Kelly’s offense, I was quick to accuse every injured defender a liar and a cheat (not to their faces of course).

The Eagles run a very high-tempo offense, one that doesn’t allow opposing defenses to leisurely make substitutions or get a full play clock to catch their breath. It’s a major feature of their strategy, and one that opposing teams would love to minimize, particularly if they aren’t well prepared for it.

One way to slow down the pace of the Eagles offense would be for an opponent to use their timeouts while the Eagles offense is in full-swing. But since a team only has three timeouts per half, they’re a little too valuable to burn. An injury however, is an official’s timeout – these are unlimited – and there’s no cost to the injured team outside of the last two minutes of a half, except that the injured player must sit out for the next play.

So in the current NFL world where fake injuries don’t have a cost (apart from having the ‘injured’ defender miss a play) and can help defenses maintain an easier pace – you could see why an Eagles fan might look at an opposing defender’s injury with suspicion.

Could the Eagles opponents be faking injuries to slow them down? The idea is one that makes the rounds in Eagles bars, but one that’s hard to actually evaluate. So this is my attempt to try.

Others have analyzed NFL injuries via metrics like games lost (i.e., players who aren’t active on game day because they’re injured), but to my knowledge, this is the first attempt to use play-by-play data to look at in-game injuries for trends and whether teams might be faking against the Eagles or other high-tempo teams.

The analysis is a bit long, so below are some quick takeaways:

– The Eagles suffered (or inflicted depending on your point of view) the most defensive injuries against the in league in 2014, and are 2nd in the league when adjusted for a per-play basis
– Across the league, there is a significant positive correlation between running more offensive plays and a higher per-play rate of defensive injury
– Such a correlation could be attributed to fatigue, but this correlation does not hold for the three other possible game situations (own offense, own defense, offense against) – these show no strong relationship between running more plays and a higher per-play rate of injury
– Taken together, these last two points support my hypothesis that players fake injuries against higher tempo offenses

Data Collection and Methodology:

I gathered play-by-play data from all the regular season games this year, and identified all the in-game injuries noted in the descriptions. In case you haven’t read play-by-play before, each play has its own line and explanation, and any play that resulted in an injury timeout is noted. Below is an example:

2-10-DET 40 (14:05) (Shotgun) 10-E.Manning pass incomplete deep middle to 80-V.Cruz (27-G.Quin). DET-27-G.Quin was injured during the play.

If an injury was noted as a stoppage, it was recorded. In an ideal world, we’d eliminate injuries that are serious and clearly not fakes, but there’s no detail on the injuries in the game data, so we have to take the major with the minor.

The play-by-play injuries were then coded as to whether they occurred to the offense, defense, or on special teams (e.g., kick coverage). There were approximately 700 total observations, and while it’s possible that not all injuries were noted in the play-by-play data, this is the only comprehensive source for such information. Given that there are ~700 injury stoppages in our set, that works out to 2-3 injury timeouts per game, which sounds possible but could also be low. It’s possible that whoever officially creates the play-by-play gets lazy and misses some, my assumption here is that if any injuries are somehow missed, they aren’t biased towards one particular side of the ball.

After gathering the data, one additional adjustment is for play frequency. Simply put, the more snaps a player gets, the more likely they are to sustain an injury. Therefore, any team that runs more plays is more likely to see a higher absolute number of injuries. To account for this, I also looked up the total number of plays for each team’s offense and defense during the course of the year – to understand the rate of injury rather than the total number.

Output:

Let’s start with the absolutes. I found 692 injuries in the play by play data, 66 of which were special teams plays. I took these out, because they aren’t central to the question of are teams faking injuries to slow down offenses. Of the remaining injuries, I looked at whether they happened to an offensive player or a defensive player and which team they occurred against, below is the data from this season:

Not a shocker to see the Eagles at the very top of that list, and indeed they led the league in defensive injuries against this season.

However, as I already noted, this metric can be misleading. The Eagles offense runs more plays per game than any other team, so we would expect them to be near the top of this list. We need to adjust our data for the number of offensive plays – and we can examine the rate at which opposing defensive players get injured against the Eagles and whether they are still an outlier.

So as we see when we look at it on a rate basis (number of injuries/number of total offensive plays), the Eagles are still close to the top of the league, and roughly 50% above the league average. Houston is just above them, and while no one would consider their offense up-tempo, the fact that the Eagles are so high would be consistent with the theory that opposing teams might be faking injuries to slow them down.

Now, before we get any further down the faking rabbit hole, what if there’s a simpler explanation that doesn’t involve fake injuries? There’s another obvious possibility to explain why the Eagles are so high in defensive injuries against. What about the idea that as you run more plays, players get more physically exhausted, and therefore are naturally more susceptible to injury?

That seems possible, right? So let’s examine that idea a bit.

The first thing we can do is very simple, does injury frequency vary by quarter? If teams get physically tired during the course of the game and that leads to more fatigue and more injury, there should be more injuries as the game goes on:

Interesting. This sort of muddies our waters a bit.

In absolute terms, the number of injures rises dramatically as the game goes on. Injury stoppages in the fourth quarter occur at 2x the rate they do in the first quarter. Part of that can be explained by the fact that the clock stops more frequently in the fourth quarter than the others (and thus more plays), but that wouldn’t explain a 2x difference. I would want to check against the sheer number of plays run by quarter, but I don’t have that data without a bunch of more work.

Still – it looks like that thinking may be reasonable, injuries increase as the game goes on. But it’s also interesting to note that the increase is much more pronounced on the defensive side of the ball. We’ll come back to that later.

For the time being, let’s move on to looking for evidence of fake injuries.

As a general framework for this analysis, I’ve split the types of injury stoppages into four buckets:

1. While on defense, your own team suffers an injury (Own-Defense)
2. While on defense, your opponent suffers an injury (Opponent-Offense)
3. While on offense, your own team suffers an injury (Own-Offense)
4. While on offense, your opponent suffers an injury (Opponent-Defense)

We’ve been focused on bucket #4 thus far, and saw that on a per-play basis the Eagles are close to the top of the league in terms of defensive injuries against on a per-play basis. We also saw that overall injuries increase as the game goes on – but it seems much more prevalent on the defense, which is the side that would be interested in faking injuries.

So can we look a bit deeper to see if play frequency increases injury risk across each type of injury stoppage? The idea that running more plays increases the rate of injury should not be exclusive to offense or defense – although it appears that way at first glance – it’s hard for me to believe that defensive players are in any worse shape or take any harder hits than offensive players.

To take a look at the issue, I ran some basic correlations across each of those four injury types, looking at the number of plays run and the rate of injury. Just to clarify, I summarized the four below:

1. Your defense runs more plays and gets injured more often (this would be a bad defense)
2. Your defense runs more plays and your opponent gets injured more often
3. Your offense runs more plays and gets injured more often (this would be a good offense)
4. Your offense runs more plays and your opponent gets injured more often

Again, if the rate of injury increases with more plays, we should see relationships in each of these situations. So what do we see?

#1 – So earlier we saw defenses suffering more injuries as the game goes on…and yet, when we look at number of defensive plays per game and the rate of defensive injury, there really doesn’t seem to be any relationship. Teams with defenses that are on the field a lot don’t seem to get injured at a higher rate than those who execute fewer plays.

#2 – Our next picture shows a similar lack of correlation, this time between defensive plays per game and the rate of opponent offensive injury. This idea would be that if an opposing defense is really bad, your offense gets more plays, and might get hurt more frequently. But the data shows nothing that looks like a relationship.

#3 – Now we’re on the offensive side of the ball, looking at whether an offense that runs a lot of plays suffers a higher rate of injury. There’s actually a relatively weak negative correlation between running lots of offensive plays and suffering offensive injuries. If you want to believe in things like Chip Kelly’s Sport Science program, you would expect a negative relationship as teams that employ high tempo offenses are more adequately prepared to stay healthy while running it. While a very slight relationship exists, it doesn’t look to be that large, if it even exists at all.

#4 – Hmmm…now it’s officially interesting. When we look at the rate of defensive injury against offensive plays per game, there is our most significant positive relationship. A correlation of 0.39 is significantly more than we’ve seen in the other three instances, and it’s also the only one where there is a clear incentive to fake injuries.

Taken alone, this relationship might be explained by the fatigue theory, but I think it’s tougher to make that argument when you don’t see anywhere close to the same relationship in all other situations. When a defense is bad and on the field a lot, they don’t get hurt more often, when an offense is good and runs lots of plays, they don’t get hurt more often, and when a defense is bad and their opponent runs a lot of plays, they don’t get hurt more often. The only ones who show a substantial increase in injury stoppages as plays increase are opposing defenses.

To me, that’s pretty freaking suspicious. Either opposing defenses are the only ones who suffer from fatigue-related injuries…or maybe some of the injuries aren’t injuries at all.

Now, this is far from 100% conclusive. It may be that defensive players naturally get more fatigued than offensive players due to their roles (i.e., offensive players can take more plays off because they know the play calls)…but I don’t really buy that. I think there’s at least a little bit of shenanigans.

It’s also an entirely different question as to how much this even matters. Any fake injury will happen on the margins, as you see the number of total injury stoppages remain relatively small (2-3 total per game). But for an Eagles team that narrowly missed the playoffs, the marginal differences matter.

Solutions

So is there a way to address teams that fake injuries? There are certainly options, but some of them are just impractical. The NHL has a penalty for diving, but you really can’t ask the officials to diagnose injuries and try to penalize fakers. You could charge a team a timeout, which the NFL already does if an injury occurs in the last two minutes. That’s much easier than trying to penalize teams, but also provides incentive for coaches and players to hide injuries (also, what do you do in the case of a ‘Body Bag Game’?)

One idea I think might actually be workable, is to tweak the NFL’s current rule for injured players. As it stands today, an injured player who causes a stoppage has to miss at least one play. Well, if you want to eliminate fake injuries, you should raise the cost to those players for faking, and you can do that simply by making them sit out longer. What if, when a player is injured and causes an official stoppage, they must sit out not for just one play, but for the remainder of that series or until a change of possession?

Missing the rest of a series is a bit more significant than missing just one play, and is something that could balance the equation on faking injuries. It also dovetails nicely with the NFL’s stated emphasis on player safety (interpret my use of the term ‘stated’ as you will, based on your own level of cynicism)

If there are fake injuries happening, such an increase in missed time might be enough to keep anyone from acting hurt. Requiring a player to miss the remainder of a series also isn’t as significant as forcing them out for the rest of a quarter or a game.

Some would argue that this isn’t even a problem worth focusing on. But if fast-paced offenses gain greater acceptance in the NFL (which will happen if more of them succeed), the issue will only become more prominent (beyond the realm of the paranoid Eagles fan) and could materially impact the game.

Summary Data

Below is a table of all the raw data I used here, as a reference:

Bonus – Jevon Kearse All-Stars

One last thing I did with this data, after pulling it together, was dig through and sum up all the specific players who sustained injuries in a game this season.

I wanted to look into it because I was really interested in what I’ve termed the ‘Jevon Kearse All-Stars.’ It may just be a bad memory on my part, but one of the things I really remember about Jevon Kearse’s tenure with the Eagles was his tendency to hurt himself and fall to the ground like he got shot. I feel like his injuries always looked more serious than they actually were. It’s possible I’m misremembering, and if so I apologize to the Freak. But with that said, here were the league leaders in injury stoppages in the NFL this year:

Now I’m not accusing these guys of faking injuries, these just happened to be the guys with the most injury stoppages in the play-by-play data (excluding special teams, which most of these guys don’t play anyway).

Enjoy your spot on the Kearse All-Stars guys – the trophy (it’s an ace bandage) is in the mail!

Advertisements

Week 17: Eagles vs. Cowboys Pre-Game Notes

Win or go home; the playoffs start today for the Eagles.  As I wrote in my breakdown, the Eagles are, deservedly, significant favorites and are likely to win.  However, that doesn’t mean they will. Here’s where thinking probabilistically becomes difficult for many fans and commentators.

According to Vegas, if this game was played a very large number of times, we would expect the Eagles to win about 70% of the time.  That’s a LOT.  Moreover, I think the line is too low.  At my line, I’ve got the Eagles between 75-83%.  Very encouraging, but it raises an interesting question:

If the Eagles lose, did they “choke”?

Let’s assume for a moment that the Eagles do lose tonight.  The low-level analysis will involve “choking” and talk about things like cracking under the pressure or Chip Kelly not being ready for such a big game.  Mostly, it will revolve around one supposed fact, that the Eagles SHOULD win this game.   Read that again:

The Eagles SHOULD win this game.

It sounds right, but it’s complete bullshit.  That’s not how the sport (or the probabilities the sport is based on) works.  The correct way to say it is that, ex-ante, we know that the Eagles are LIKELY to win.  The problem with “should” is that it implies a level of control that the team simply doesn’t have, regardless of what generic sportswriters would have you believe.

Let’s use an analogy.  Imagine you’re playing poker (Texas Hold ‘Em).  You’re heads up, looking at the river.  There are no more strategic moves to be made, the outcome of the hand depends entirely on which card comes out.  Of all the cards remaining in the deck, 75% of them will result in a Win, 25% a loss.

If a Loss card comes out, did you “choke”?  Similarly, before the card is drawn, “should” you win?

Of course not, that’s ridiculous.  The fact is, the entire strategy of Poker (and many games/sports in general) is to shift the odds in your favor as much as possible (and maximize the expected payout when you do).  Unfortunately, that’s all you’re doing, shifting the odds.  As anyone who truly understands what that means knows, even after you’ve shifted the odds, there is still a chance for a loss.  In Poker, it’s sometimes called a Bad Beat.

Circling back to the game tonight, if the Eagles lose, it may just be a Bad Beat.  The Eagles are the better team, but if they played tonight’s game a large number of times, they’d STILL LOSE a decent percentage of the time (30% according to the spread).  Just as the Eagles MUST win a large majority of the games if it played a large series, they also MUST lose as well.  Unfortunately (when you’re a favorite), you don’t get a series, you get one game.

The ball takes lucky/unlucky bounces.  The refs don’t see everything correctly.  Players will have bad games, coaches will have bad games.  They are not robots, their performance varies.  If those bad games coincidentally happen to occur when the stakes are high, the knee-jerk reaction is to say the player/coach “choked”.  Or to somehow imply that the stakes themselves forced them into a sub-optimal performance.  I’m not ruling this out.  It’s possible, and it may even be probable (at the extreme margins).  However, it’s much more likely that they simply caught a Bad Beat.

As anyone who plays poker knows, when that happens, the “right” thing to do is to shake your head and check your calculations.  If they were correct, you don’t do anything different the next time (unless you really did screw up), even though the results this time were bad.  Process, not outcome.

Now, a few actual notes:

– The Eagles are 7 point favorites. E = R ((60 – T) / 60) + C.  That means, at least to start, they should be playing a relatively low-variance game.  Over any stretch of time, we expect the Eagles to outplay the Cowboys.  That puts the onus on Dallas to change the conditions.  Note though, that at 7 points, all that would take is an early TD from Dallas.

– Don’t let Dez Bryant beat you.  He’s the only elite weapon the Cowboys have.  Demarco Murray is having a good year, but if the Eagles offense performs anywhere close to expectations, Murray isn’t going to be able to keep the Cowboys in the game.

– Don’t get blitz-happy.  A frequent reaction to a back-up QB is to blitz him.  Yes, getting pressure on Orton is important, but when you blitz, you actually make it EASIER for receivers to get open, you just (hopefully) diminish the time the QB has to make that decision.  If I were calling the game, I’d want to see if Orton can move the ball against the base defense before dialing up an blitzes.  Remember, blitzing is a high-variance move.  If you’re a big favorite, you should blitz sparingly.

– Don’t get sloppy.  Penalties and turnovers can equalize a skill-advantage pretty quickly.  There’s no reason to believe either will be a problem, but that doesn’t mean the players shouldn’t be reminded to play under control (Cary Williams especially).  The Eagles have averaged 53.4 penalty yards per game this season, Dallas averages 57.5.    Similarly, both teams have averaged +0.7 turnovers per game, which is 4th best in the league (7-way tie).

– Watch out for “David” strategies.  We’ve looked at these from the underdog’s perspective, but since Dallas is in that position tonight, it’s a good chance to view things from the other perspective.  So, watch out for surprise-onside kicks or fake punts.  Expect a few 4th down plays.  Prepare to be blitzed.  I don’t think Garrett has the balls (or brains) to fully deploy these strategies, but that doesn’t mean we won’t see one or two instances. They’re high-risk/high-reward; the Eagles really want to focus on that high-risk part.

– Enjoy.  This has been a lot of fun, and regardless of what happens, Eagles fans should be pretty excited about the Chip Kelly era.  The team has already met my expectations (surpassed them in some respects), so it’d be silly to consider the season anything less than a great success, no matter what the outcome of tonight’s game is.

That said, it’s Dallas, the Eagles are the better team, and it’s for a playoff spot.  I think the Eagles roll.

Notes from Yesterday

Just a few notes from yesterday’s game:

– I don’t understand Chip’s decision to dial down the offense in the second half.  It makes complete sense to become more conservative and to take fewer risks when you have a lead (think equation).  HOWEVER, when the opposing team is basically begging for you to take a shot, you should take it.

Early in the game, it was clear the Eagles were going to take shots downfield when they had Cooper matched up one-on-one with a DB with no safety over top.  It almost led to an early TD (Cooper lost sight of the ball).  Anyway, late in the game the Redskins were packing 8 in the box and playing a single deep safety.  That means you’ve got both D-Jax and Cooper against a CB, and the safety can only help on one of them.

Somehow, a situation Chip was hoping for and targeting early in the game lost its appeal.  Keep in mind that this is not a high-risk play.  Throwing it deep to Cooper when he’s in single coverage is very unlikely to produce an outcome worse than an incomplete pass.

Given that we saw this exact same situation play out last time the team played the Redskins, after which Chip claimed he learned his lesson, I’m worried this will be a recurring issue.  Obviously, that would necessitate having big leads, which would be awesome, but it’s still a bad habit.  My only guess as to the reasoning is that Chip still doesn’t fully trust Foles.

– Overall a good win, but let’s remember that the Redskins aren’t a good team.  We’ll learn a LOT more about the team when it faces Arizona and Detroit after the bye week.  To date, the Eagles “best” win came against a Green Bay team playing with its 3rd string QB.  It remains to be seen whether the Eagles rank within the “mediocre” division of the NFL.  They’ve lost against Dallas and San Diego…which would suggest they’re at the bottom of that subset of teams.  If so, they’ll have trouble against the Cardinals.

– Still researching the topic, but safe to say that Nick Foles is at least close to doing something unprecedented.

He now has a career rating of 97.6, with 22 TDs and just 5 interceptions.  He’s also rushed for 3 TDs.

His rating this season is currently 128.  The single-season record is 122.5 (Aaron Rodgers).

Under Chip Kelly, he’s seen significant playing time in 6 games…he’s won 5 of them.

His career interception rate is now 1.2%.  The NFL Record for a career rate is 1.7% (Aaron Rodgers…yeah, he’s really good).

As I showed at the end of last week, few QBs have, at ANY point in their careers, had a career rating that as high as Foles does now.  The fact that Foles has it 15 appearances and 11 starts into his career is a very good sign.

Naturally, it’s a safe bet that Foles won’t maintain this level of play.  The next question, though, is:

What are the odds a “bad” QB could have a stretch of games like this?

How about a “mediocre” QB?

We could probably turn to Bayesian analysis to help out, but for now, it’s enough to know that the odds of either situation aren’t very good.  When you then consider that fact that he’s doing it to start his career, I think it’s safe to say Foles’ odds are now pointing heavily in favor of at least “solid NFL starter” and potentially much higher.

– Last point.  Chip was correct in going for it on 4th and 1.  There’s just not much to gain from punting the ball there, especially in comparison to the relatively high likelihood of maintaining possession.  I was much more concerned about the play-call.  It looked like a delayed handoff, which would be an inexplicable call (especially to Bryce Brown).  However, it may also have just been a miscommunication.  Unfortunately, the announcers had already stopped calling the game and were too busy to bother talking about it.  I don’t think we even got a replay.  I’ll have to review the film, but my first impression was: right strategy, wrong play.

 

P.S. It’s week 12 (practically) and the Eagles are entering their bye week in first place.