Writers of Pro Football Prospectus 2008

10 Nov 2017

Do Thursday Games Cause More Injuries?

by Zach Binney

It's not hard to find players going on the record about how much they hate Thursday Night Football (TNF). Their concerns center around player health, safety, and fitness to play after just a four-day rest period.

These concerns were on display in last night's TNF matchup, where Doug Baldwin attributed his groin injury to short rest and Richard Sherman reportedly ruptured his Achilles after nursing soreness in the tendon for at least two weeks (when his Achilles first popped up on the injury report). Would Baldwin have avoided injuring his groin and Sherman avoided exacerbating his Achilles if the game had been played on Sunday instead? It's impossible to know in any individual case, but I was curious if these concerns would show up in NFL-wide injury reports. Are there higher rates of reported injuries for weeks with short rest, as opposed to normal or even extended rest?

The injury rates below are given per 1,000 athlete-exposures (AEs). One AE is defined as one player participating in one practice or game. AEs are calculated based on the 53-man active roster, with reductions made for players held out of practices or games for ongoing injuries (not including IR, which is separate from the 53-man roster). An injury is defined as any new issue appearing on the NFL's weekly injury reports, and the week the injury occurred is taken as the week before it was first reported.

As an example, in Week 8's reports, Karl Joseph of the Raiders popped up with a new groin injury. This injury was assumed to have occurred in Week 7's TNF game. (It could also have occurred in Week 8's practices, but that's less common.) Because the Raiders played on Sunday in Week 6, this injury is said to have occurred on four days' rest.

We excluded injuries occurring in Week 1 because there is no well-defined rest period, and Week 17 because data for that week is unavailable for non-playoff teams. Additionally, the data below is for 2007 to 2015; 2016 and 2017 are excluded because the removal of "Probable" from game status reports makes comparisons with earlier years difficult.

All Injury Rates

There were no substantial differences in injury rates across different rest periods (Figure 1, blue bars). TNF and other short-rest rates were not elevated; nor were the rates after bye weeks lower, as one might hypothesize if additional rest protected against injuries.

Surprisingly, injury rates for TNF games (14.5 per 1,000 AEs, 95 percent CI 13.5-15.6) were lower than those for the typical seven-day rest period (16.1 per 1,000 AEs, 95 percent CI 15.8-16.4). It's possible this difference could be due to players only playing at 900,000 rather than a million miles an hour on short rest, leading to less violent collisions, but there's a more likely alternative explanation for the difference: it's a quirk of public NFL injury reporting.

Teams typically issue their first injury reports on Tuesday of each week; that's two days after a Sunday game but five days after a Thursday game. Any injury that resolves in that two- to five-day period (to the point the player can fully practice) then wouldn't be reported after a TNF game, but would after a Sunday game. If just 10 percent of injuries (16.1-14.5/16.1) resolve in that period, that would fully explain the lower TNF rate we saw among all injuries. That seems plausible, especially since there's no clinical reason I can imagine for shorter rest to be associated with fewer injuries. If it's substantially more than 10 percent that resolve in this two- to five-day period, then injury rates might actually be elevated with short rest and our analysis missed it, but without better quality (non-public) data that's just speculation.

It's possible that only certain types of injuries may be more common on less rest, though. A recent paper in the other football found around 20 percent higher rates of "muscle injuries" during shorter rest periods (three or fewer days' rest, versus six or more). We decided to look at leg muscle injuries to see if we could replicate this study for the NFL.

Leg Muscle Injury Rates

We defined leg muscle injuries as any calf, hamstring, groin, quadriceps, or thigh injury. Once again, there were no substantial differences in injury rates across different rest periods (Figure 1, red bars). TNF and other short-rest rates once again were not elevated.

Figure 1: Injury Rates with 95 percent Confidence Intervals (CIs) by Rest Period for All Injuries (Blue) and Leg Muscle Injuries (Red)

Limitations

This analysis only includes injuries that were severe enough to merit an injury report listing, which is a threshold with some arbitrariness and variation across teams. It does not account for "normal wear and tear" or how ground down and beat up a player feels after a Thursday versus a Sunday game. I'll trust players who tell me it's rougher on their bodies, even if it doesn't show up in injury report counts.

It's also important to note that the injuries we're attributing to a given rest period could have happened either during the game following that rest period or the subsequent week of practice. This is a time during which players could have been in a weakened state from shorter rest, so our analysis makes sense. We were unfortunately not able to account for any effect shorter rests have on injuries in the practices during the shorter week itself, but previous studies have shown only around 15 percent of NFL regular season injuries occur in practices, limiting these issues.

Conclusions

There's little evidence from injury reports to suggest TNF games and other short rest periods are associated with different injury rates from the typical Sunday-to-Sunday schedule. However, this study is limited by the use of public injury report data, which only accounts for discrete reported injuries that limit players in practices or games rather than wear and tear. Everybody's hurt, but only a few of those are reported as injured. I am in no position to tell an NFL player how they feel after four versus seven days of rest, and analyses that better measure the players' actual health before and after these games are needed.

Posted by: Zach Binney on 10 Nov 2017

13 comments, Last at 14 Nov 2017, 10:20am by jtr

Comments

1
by Bisquick :: Fri, 11/10/2017 - 7:09pm

Is it really that freakin' hard to schedule every team to play its Thursday game... the week after their bye? You can still have football on most Thursdays, with players who are well-rested and not complaining. Sure, it'd make it a little harder to work out the schedule, but that's what we have all these computery gizmos for. I'm sure they can figure it out, if they want to.

3
by renangms :: Sat, 11/11/2017 - 1:28pm

It's not hard. It's impossible.

Bye weeks go from week 5 to week 11. There are more than 2 teams on bye every week.

Unless:
- you start bye weeks on week 2 and keep through the end of the season or
- reduce the numbers of games on TNF and a lot of teams will not play which is a competitive advantage or
- you play multiple games on TNF which is a horrible idea.

12
by leviramsey :: Mon, 11/13/2017 - 1:20pm
  • Add a second bye week for every team for the week before their TNF appearance
  • Add a 17th game for each team as an interconference game at a neutral site (international or Hawaii or college stadia). Define a neutral site as a site more than 200 miles from any stadium that hosts multiple games for any NFL team, or a site agreed upon by both teams (if Dallas and Houston agree to play a game in Austin or San Antonio, power to them). Interconference games would be selected via a schedule stack (which would ensure a Super Bowl rematch every year). For this season, the matchups might be: NE/GB, PIT/ATL, KC/SEA, HOU/DAL, OAK/DET, MIA/NYG, DEN/TB, TEN/WSH, BAL/ARZ, IND/MIN, BUF/PHI, CIN/NO, LAC/CAR, NYJ/LAR, JAC/CHI, CLE/SF.
  • Drop two preseason games
  • Create a spring league for players who haven't accumulated 1000 plays from scrimmage or didn't play in the previous season. This replaces the dropped preseason games for the on-the-edge players, and also provides a training ground for officials, TV, etc.
2
by erniecohen :: Fri, 11/10/2017 - 7:52pm

The league released their own study a couple of years ago, and I think they found like 30% fewer injuries in Thursday night games.

The other question is, how do teams do the following week? You would expect the extra 3 days of rest to give Thursday night teams an advantage, but I've also seen it claimed that there is a recovery problem.

4
by Tim R :: Sat, 11/11/2017 - 4:06pm

Are ratings not way down on the Thursday night games?

Would they not just be better off having 2 games on a monday?

5
by Tim R :: Sat, 11/11/2017 - 4:06pm

Are ratings not way down on the Thursday night games?

Would they not just be better off having 2 games on a monday?

6
by scraps :: Sat, 11/11/2017 - 9:35pm

Almost everyone, when discussing Thursday games, says it's a four-day rest (as you do too). But it looks to me it's three days rest: Sunday football, Monday rest day one, Tuesday rest day two, Wednesday rest day three, Thursday football. What am I missing?

7
by Hextall_27 :: Sat, 11/11/2017 - 11:33pm

Most games are 1pm or 4pm Sunday and finish by 8pm Sunday.

The Thursday game is at 8:20pm on Thursday.

This is 96 hours (or 4 days) after the end of the 4pm slot games.

I think this is where the 4 days rest term comes from.
The players also sleep 4 times between games.

In baseball terms it would be considered 3 days rest.

8
by jtr :: Mon, 11/13/2017 - 9:31am

Another factor affecting the lower injury rates on Thursdays could be lower participation for banged-up players. The rest period is shorter, so a player nursing a relatively minor injury from Sunday sometimes has to sit out Thursday when he may have been able to play the following Sunday on full rest. So we have a somewhat-injured player who would be at some risk of re-aggravating that injury being replaced by a healthy backup, which would reduce the risk of injury.

10
by CaffeineMan :: Mon, 11/13/2017 - 11:24am

I think this may be the case as well. I think teams should have expanded rosters to handle this. Of course, I think there should be some kind of developmental league as well, to go with this, but both cost money and thus it won't happen.

11
by Zach Binney :: Mon, 11/13/2017 - 1:08pm

That's a great point, jtr. Finding what epidemiologists would call a proper "counterfactual population" for Thursday Night Football is difficult. Basically this would entail forcing Sunday teams to issue their first injury report on Tuesday and, importantly, make decisions about who will play on Sunday based on their ability to play on Thursday if they had to.

The TNF rates here are DEFINITELY underestimated. By how much? It's an open debate, but I think the 10% or so that explains the lower rates is totally plausible. Is the underestimate so large that we're actually missing a truly higher rate for TNF vs. 7-day rests? It's possible, but I'm somewhat less confident in that given the NFLPA's trumpeting of an analysis of the NFL's own internal data that shows no higher rates (e.g. https://www.washingtonpost.com/news/sports/wp/2017/01/25/thursday-night-...). Now there are lots of conflicts of interest here (interests of former vs. current players, etc.), but without the NFL or multiple teams opening their books to independent researchers like me this sort of injury report analysis is probably the best we can hope for.

13
by jtr :: Tue, 11/14/2017 - 10:20am

If you have data for DNP's by the day of the week, you could look at how many players are too hurt to participate on the average Sunday vs how many are too hurt to play on the average Thursday to at least see how many extra games missed there are. Alternatively, an approximation might shine through if you look at how often a player doesn't participate in Thursday practice but still plays on Sunday; this player would almost certainly not play if his team had TNF that week.

9
by big10freak :: Mon, 11/13/2017 - 9:58am

I know the Packers loved and hated playing the Lions on Thanksgiving Day.

Loved it because if GB won they had a road division win and then had the longer layoff to the next game which helped for the December push to the playoffs (typically)

Hated it because the Lions were always jacked up and more importantly guys were tired and banged up playing on short rest.

When asked about 'regular' Thursday games before he left the team TJ Lang said that Thanksgiving was different and ordinary Thursday games just sucked.