Writers of Pro Football Prospectus 2008

13 Sep 2009

FO in the NYTimes: Some NFL Injuries Hurt More Than Others

Football Outsiders returns to the Keeping Score column of the Sunday New York Times with this roundup of data from our huge injury database. Included: offensive injuries are more important than defensive ones; injuries to running backs generally don't hurt a team's running game at all; older players don't actually get injured more often; and more. (Free NY Times registration required.)

Posted by: Bill Barnwell on 13 Sep 2009

9 comments, Last at 16 Sep 2009, 1:01pm by tuluse


by Dales :: Mon, 09/14/2009 - 6:56am

"Offensive line inconsistency holds particular significance for fans of the Jets and the Giants. They became the first teams in five years to have two consecutive seasons (2007-8) in which their five starting offensive linemen started all 16 regular-season games. That is extremely unlikely to occur for a third consecutive season, and replacing a starter can have a drastic effect on a team."

When heads have come up twice in a row, it is really unlikely heads will come up that third time.

by dmb :: Mon, 09/14/2009 - 10:16am

I see what you're getting at, but having a coin come up heads isn't an improbable event by itself; maintaining a perfectly continuous offensive line is. I don't think Bill meant it's "extremely unlikely to occur for a third consecutive season" because it happened the past two years; it's simply unlikely to happen in any context.

If you're playing a board game and your opponent does well because he happens to roll two consecutive sixes (with a 6-sided die), you're not falling for the gambler's fallacy to suggest it's unlikely that he'll roll another six. It's just equally unlikely as it was before.

by tuluse :: Wed, 09/16/2009 - 1:01pm

However, injuries and health aren't discreet events, so the coin flip analogy isn't very good.

by billsfan :: Mon, 09/14/2009 - 10:20am

A somewhat related post from Advanced NFL Stats:


(I also like the Eagles)

by LinebackerU :: Mon, 09/14/2009 - 7:35am

"9. Hamstring injuries linger — and N.F.L. teams don’t know."

I'm not sure I understand the implication here with the Mangini part. Does it mean that Mangini reported Dwight with a thigh injury so teams would think there was a better chance for him to play than there really was? But if that's the case isn't it an indication that NFL teams DO know?

by A Dallas Fan (not verified) :: Mon, 09/14/2009 - 1:04pm

"But players participated in 77 more games than expected when listed with a shoulder concern on the injury report."

Is this including or not including Brady's famous shoulder that has been probable for years?

by Bill Barnwell :: Mon, 09/14/2009 - 1:47pm

Missing the point on the Offensive Line analysis (the AdvNFL piece doesn't apply here in the slightest).

You're suggesting -- at least, from what I can tell, that I'm engaging in the gambler's fallacy, that because the team had no injuries a year ago, that they're due to have injuries this year. It's the same criticism that came up when we were talking about the Cowboys last year.

That's not what I'm saying at all. What I'm saying the data shows is that there's no correlation, whatsoever, for the injury rates of the five starting offensive linemen on a given team from year-to-year. So what happened last year has nothing to do with what happened this year.

DMB has the right idea -- It's like rolling a die. Say, if you roll a 1, your offensive line stays healthy, but if you roll a 6, your offensive line is in shambles. The Giants and Jets each rolled a 1 each of the last two years. I'm not saying they're likely to roll a 6 this year; I'm saying that they're no more likely to roll a 1 than they are any other number, and there's a 5/6ths chance they're NOT going to roll a 1.

Eric Mangini knows about hamstring injuries. One other team knows, too, but I don't have the data in front of me (Carolina?). No one else really notices it.

The player participation data does include Brady's trick shoulder, but remember -- we're comparing it to all players who are listed as "Probable", not just general injuries. So we might have "expected" Brady to actually miss, oh, two of those games with the shoulder injury. He doesn't have a huge impact.

by Bill Barnwell :: Mon, 09/14/2009 - 1:55pm

Oh, and this article was featured in the Freakonomics blog today, which was certainly a personal thrill.

by weinsteinium (not verified) :: Wed, 09/16/2009 - 12:46pm

What does H.G.L. mean for the curse of 370. What you seem to be saying is that injuries appear to be random, unaffected by team, player or age. Does that mean that 370 carries is a statistically anomaly?