Does the likelihood of injury increase with the number of QB rushes?

Submitted by MGoLiteral on September 27th, 2010 at 5:53 PM

Does the likelihood of injury increase with the number of QB rushes?



September 27th, 2010 at 6:11 PM ^

Wow, what an interesting question and concept. Someone asked a similar question (read: the exact same question) earlier today. He was pointed to pretty much anywhere else on this blog, since that has been one of the major topics of discussion.

For the best research, look here.

In other news, when is the points threshold being increased?


September 27th, 2010 at 6:23 PM ^

That diary doesn't answer my question. I'm not interested in the relationship between rush/pass ratio and injury. Rush/pass ratio is not a very good metric in my opinion. I want to know whether the total number of QB rushes is correlated with the probability of injury.

EDIT: Also, I'm not asking whether "spread" QBs are more likely to get injured than non-"spread" QBs. I'm only interested in the relationship between number of rushes and likelihood of injury.

Zone Left

September 27th, 2010 at 6:32 PM ^

It really is answering your question.  Think about it, the QB is going to hold the ball on QB runs and passes, not on handoffs, which is basically the other option. 

You asked:

Does the likelihood of injury increase with the number of QB rushes?

Presumably, given a large sample size (like the five years of data in the diary), the higher or lower likelihood of a QB getting injured who runs more would be answered.  Regardless, the forum six threads below this is about the same topic.


September 27th, 2010 at 7:40 PM ^

I'd say that's a difficult question to answer from a QB perspective.  I believe the reason a non-spread QB (i.e., pocket passer) was used compared to a spread (i.e., mobile QB) was because those are two distinct styles.  If you compare the ratio of injuries/rush to the whole QB set, then my guess is it would be skewed given the varied running ability of QBs.  In other words, my guess (with no evidence) is that a pocket passer is more likely to get hurt running the ball than a mobile QB running the ball.

I think the best analysis would be done within the mobile QB group.  However, it will be difficult to determine when a mobile QB got injured - was he running or was he hit back in the pocket. 


September 27th, 2010 at 6:13 PM ^

executive summary - not really.  but that topic has been discussed at length on this board as MGoTarHeel pointed out.  i know it wouldn't be easy to search for a diary like that, but look around a little because the contributors to this blog have done some really fantastic research and probably get annoyed seeing questions they've dedicated a lot of time to asked again.


September 27th, 2010 at 6:13 PM ^

How many different ways can this question be asked?

I kind of wish we were back to the four daily threads on the weather. At least my head wouldn't hurt as bad.


September 27th, 2010 at 6:15 PM ^

There are a few excellent diary entries on this very subject.  I've been looking for them for several minutes but am unable to find them.  I'll keep looking but the essential point was that if you divide QBs into 3 categories where category A QBs have a relatively low rushing%, category C QBs have a relatively high rushing% and category B QBs are between A and C then category C QBs get hurt the most, then category A QBs, then category B QBs.

Essential, the data suggested that if a QB runs occasionally, they are less likely to be hurt.  The reasoning was that QBs like that are more likely to evade big hits from lineman which is what will generally hurt category A QBs (if you're going to be hit, better that it's a corner than a DE coming from the left side).

I don't remember if a statistical analysis was performed to determine if the difference was statistically significant.  It was also discussed in the comments that it's possible all injuries are understated since players can be hurt in practice and subsequently never make the field- if only the best category B and C QBs were able to survive practice hits while category A QBs could almost always survive practice hits, then B and C guys would actually be getting hurt more- we just wouldn't know about it.

I'll see if I can find the diary since it's much clearer, more complete, and more useful than that admittedly garbled summation.

Edit: Ah the hazards of a long comment, the link I was referring to is already posted above.


September 27th, 2010 at 6:31 PM ^

but I have been wanting to complain about this, using Denard as a case study, ever since Saturday. Denard takes a bad angle to the ground and looks like he might be out for a game or more, and Dr. Saturday is posting "Ok, Denard Robinson's durability is officially questionable."

Then he writes about the 5 carries Denard had in the game without realizing how ridiculous it is to talk about the overall number of carries DR gets... since his injury came after only 5!  To make the numbers even more ridiculous, only 2 or 3 of those carries actually ended with a tackle.

Thing is, every run has the same probability of ending badly, so the more there are, the more they add up.  But it's an arithmetic progression, and carry #5 is just as risky as carry #25.  Urgh.


September 27th, 2010 at 6:44 PM ^

that's not the appropriate question. The question is in 50 snaps, is Denard tackled more than Chad Henne (e.g. of a standard drop back passer) was and are the hits that either take "harder" than the other (Denard is more likely to get hit by DBs not 300lb linemen).

I don't know the answer. The answer to the OP's question is obvious.


September 27th, 2010 at 6:58 PM ^

It appears that the board has conflicting answers to my question. Some think it's obvious that the likelihood of injury is a positive function of the number of QB rushes (I am referring to the "does GEICO sell auto insurance" crowd), while others claim that there is no relationship and cite the diary post that finds no significant relationship between rush/pass ratio and injury.

I'll settle with the fact that no one knows the answer to my original question: Does the likelihood of injury increase with the number of QB rushes?

Perhaps someone with enough statistical skills, data, and time will be able to answer my question in the future. 


September 27th, 2010 at 7:25 PM ^

Perhaps someone with enough statistical skills, data, and time will be able to answer my question in the future.


Actually, people with statistical skills, data, and time have answered your question in the past. Emphatically. Maybe next time you should use some Googling skills, research, and time to educate yourself rather than treating this board like Yahoo! Answers. Start by reading 'The Mathlete'. He's done some killer stuff, but maybe the "fact" that nobody knows the answer will deter you from looking further.


September 27th, 2010 at 9:42 PM ^

people don't have answers because it's OBVIOUS. mobile qbs don't have a higher likelihood to get hurt than other qbs. however, the odds of being hurt increase every time you go on the field. a player being tackled is more likely to get hurt than one not being tackled. your question is just blatantly obvious, hence the conflict. you are more likely to get hurt having people trying to get the ball from you. 


September 27th, 2010 at 7:36 PM ^

Through basic statistics, if we had a large enough sample size and assuming every QB was equal, we could determine the ratio and it would ALWAYS be that ratio if a particular QB ran enough to remove statistical ambiguity (law of large numbers).  Let's say it turned out that 10% of the time a QB rushes he gets injured.  That ratio implies that if the QB rushed 10 times he should get injured once and if he rushed 1,000 he should get injured 100. 

So to answer your question, no, the likelihood of getting injured shouldn't change based on the number of rushes, all else equal (it's that all else equal bit that makes your question impossible to answer).  What this ratio does say, however, is that the more a QB rushes the more likely he is to get hurt at least once.  But this is obvious and shouldn't need to be investigated/proved. 

What makes our situation unique/awesome is that Denard is an excellent rusher and is less likely to get hurt (because he's so quick) than somone like Navarre trying to rush it 29 times a game.

Dark Blue

September 27th, 2010 at 7:03 PM ^

Denard Robinson could walk out in front of a bus and get hit tomorrow. Injuries can happen at any time, until he gets seriously injured lets not worry about it.


September 27th, 2010 at 7:15 PM ^

The answer is it depends on your assumptions about frequency of hits, “hardness” of hits, and other falls (like Denard hitting his knee hard on the ground).

Does a running QB gets hit more than a drop back QB? It depends on if the running QB slides/ goes out of bounds. It also depends on the protection of a drop back QB (obviously an awful line = more hits and a great line = less hits).

The research done by MGoBloggers has centered around running QBs vs. drop back QB injuries, which has shown that there really is no difference. However, if you are asking if Denard (a running QB) increases his likelyhood of injury by running 15 times to 16 times in 1 game, the answer likely is no (assuming you believe the MGoBlogger research).

El Jeffe

September 27th, 2010 at 8:43 PM ^

You asked

Does the likelihood of injury increase with the number of QB rushes?

and then complained that no one gave you an answer. Here is my answer: of course it does. As jwschultz and a few others hinted at above, there is some nonzero probability that any qb rush will end in injury. If you add all those probabilities up, you get an increased cumulative probability with every additional rush. That part is simple.

What I think you are asking is more complicated, and that is, is there empirically a correlation between qb injuries and number of rushes? Here you should start with MCalibur's diaries, but while you're there, consider this: there may be no empirical correlation, but that would only be because of confounding variables like speed, shiftiness, quality of blocking, etc. If you controlled for the right confounds, you would have to find that there is a correlation.

To see this, imagine two versions of Denard (impossible, I know, but this is what it means to "hold constant" additional variables). One version runs 25 times a game and one runs 5 times a game. This is similar to bluetell's quiz above. Given there is some probablility of injury on every rush, it has to be true that 25-rush Denard has a higher likelihood of injury over the course of a game or season than 5-rush Denard. Actually, can't you kids and your new-fangled video games test this?


September 27th, 2010 at 9:21 PM ^

had a long, fruitless back and forth with someone who insisted this was a version of the Gambler's Dillema. Except that Denard isn't a deck of cards, but an object banging violently into other objects, all intent on tackling him. The real analogy might be to a car. Assume that on the bulk of those plays he gets hit, even if he scores. Now consider a car--is a car that gets in 50 crashes likely to be more banged up than one that's in five? 

Lots of confounding variables, obviously. He's getting hit at the second level, not in the backfield--cool, and we are delighted that that lessens his chances of major injury. But you don't have to think RichRod's some monster intent on hurting Denard to acknowledge the simple fact that the more he's exposed the more chances for him to sustain injury.