Writers of Pro Football Prospectus 2008

28 Jan 2010

Stat of the Day: Quarterbacks Getting Hit

As part of our ongoing Stat of the Day series, we're digging deep into our spreadsheets to run a new stat every weekday until Super Bowl XLIV. Today, we're going to look at quarterback hits from the perspective of the offense. A quarterback hit is any play where a defender knocks the quarterback to the ground after a pass attempt. I'm not counting sacks here, but our count of hits does include plays canceled by penalties. (Often, that penalty is the hit itself.) Official scorers are really hit-and-miss on this stat, so a lot of them get added to the official PBP after the fact when we send the league a list of "PBP mistakes" compiled by FO game charters.

By the way, I've noticed that the discussion threads on these Stat of the Day posts seem to be roughly one-third to one-half complaints that each stat doesn't take every play into account or is otherwise imperfect. I have to chuckle at these inevitable complaints, but I do hope one thing is clear: At Football Outsiders, we try to be honest about our limitations. We don't claim that our stats measure things they don't measure. Stats that come straight out of play-by-play only include things that are listed in play-by-play. Stats from game charting are going to be subject to charter bias and are limited by lame TV camera angles. That's just how it is. Some of us still find these things interesting.

Anyway, on to today's lists.

First, here are the quarterbacks with the most hits, led by the same guy who led the league in this category in 2008. The total for "passes" is there as a guide -- I didn't do percentages because hit totals include plays canceled by penalty but pass totals do not.


Player

Team

QB Hits

Passes
9-D.Garrard JAC 85 565
8-M.Schaub HOU 65 607
13-K.Warner ARI 58 537
12-T.Brady NE 54 587
9-T.Romo DAL 53 591
17-J.Campbell WAS 52 555
4-B.Favre MIN 51 569
6-J.Cutler CHI 49 593
8-M.Hasselbeck SEA 47 529
12-A.Rodgers GB 46 600

Other quarterbacks with a lot of hits, but fewer passes so they don't show up on this list: Trent Edwards (31 hits, 207 passes); Ryan Fitzpatrick (29 hits; 248 passes); and Matthew Stafford (42 hits, 398 passes).

Here are the quarterbacks with the fewest hits. The order is determined by hits divided by passes -- but again, since one of those categories includes plays canceled by penalty and the other doesn't, this isn't really a true percentage. It just seemed like a good way to get a fair list.


Player

Team

QB Hits

Passes
8-K.Orton DEN 27 577
18-P.Manning IND 29 591
5-D.McNabb PHI 24 485
7-B.Roethlisberger PIT 31 563
10-V.Young TEN 15 271
9-D.Brees NO 30 538
7-C.Henne MIA 27 479
6-M.Sanchez NYJ 23 391
17-P.Rivers SD 33 519
9-C.Palmer CIN 32 496

Posted by: Aaron Schatz on 28 Jan 2010

60 comments, Last at 31 Jan 2010, 11:24am by Eddo

Comments

1
by Arkaein :: Thu, 01/28/2010 - 7:23pm

I'd be curious to see a chart that showed how many times QBs took hits that resulted in penalties.

It seemed to me that Aaron Rodgers took a lot of hits that should have been flagged as late hits (or at least would have been flagged if the same hits had occurred against some other QBs), but I don't have any hard numbers to back this up. A ratio hit penalties to total hits might be indicative of some QBs getting treated differently by the refs.

Or it might not, but it would still be interesting.

4
by Yuri (not verified) :: Thu, 01/28/2010 - 7:46pm

I see your Aaron Rodgers and I raise you Donovan McNabb. I'd think he takes at least one hard "should have been penalized" hit per game. Though when dude got ribs broken by a helmet (on a run) the league apparently reviewed and said it was OK.

I am also not sure how I feel about the little rant in the beginning about overly critical readers... Sure on every internet message board there's a "this sucks" category, but then again, there are people who are suggesting improvements, and FO staff should be perhaps more forthcoming, on, for example, why they think the stats are interesting.

Incidentally, why not have a "total beatdown index" (hits + sacks + run tackles)?

38
by ABW (not verified) :: Fri, 01/29/2010 - 11:37am

For as long as I've been reading FO, they have been real good about taking suggestions from readers.

I read a lot of sports statistics sites, and frankly I'm sick of people complaining that the stats are incomplete or don't include something or implying that the entire stat is invalid because it's producing an obviously incorrect result for a particular player or team. If there are reasons why the stat is producing misleading results for someone, by all means post it and let those of us who are not as familiar with that player/team know. But hopefully, we all know that statistics do not paint a complete picture and every statistic is just shining some light and rigor on one particular aspect of the game, and we don't need to have this debate in EVERY FUCKING THREAD here and on BP and PFR and Advanced NFL stats, etc. etc. etc.

7
by billsfan :: Thu, 01/28/2010 - 7:48pm

The NFC Championship Game (and a season full of Patriots games) got me wondering the exact same thing.

(I also like the Eagles)

22
by jmaron :: Thu, 01/28/2010 - 9:54pm

The NFC playoffs had several key no calls for what looked to me like blatant roughing the passer penalties:

1) OT Arz sacks Rodgers - forget the facemask - the Arz player hit Rodgers in the head before the face mask - that should have been 1st down GB - instead Arz wins the game.

2) New Orleans got away with a hit to head on Warner on a key sack early in the game against Arz as well. The score was 21-7 Saints, but instead of 1st and 10 at the Arz 42 it was 2nd and 18 at the Arz 27. That is a huge play.

3) Favre hit low on an int - late 3rd quarter 21-21 - Vikings would have had a 1st down at the NO 19.

The penalty has become a huge play in many instances because it often changes or should have changed a dramatic play (sack, sack-fumble, ints) because bad things tend to happen to the offence when the QB is under duress.

21
by jmaron :: Thu, 01/28/2010 - 9:40pm

I suspect that just about every fan thinks their QB didn't have nearly as many flags thrown for roughing the passer as should have been and conversely are convinced other teams famous QB's get preferential treatment.

I for one was incensed in the 2nd GB game when R. Edwards got a roughing the passer penalty at a key time in the game in GB. It looked like a perfectly clean hit - not late, no throwing the QB to the ground. But I bet you watched the same play and thought it's about time they called one.

52
by Arkaein :: Fri, 01/29/2010 - 6:05pm

You could well be right (in the general case, can't remember the specific play you mention), which is why I'd like to see a deeper analysis.

On some further consideration, if I am right, I wonder if mobile QBs draw flags less often than pocket QBs. It did seem like a lot of Rodgers rough but no flag hits were times when he escaped the pocket. Maybe refs subconsciously are less likely to flag QB hits on QBs who act more like RBs?

60
by Eddo :: Sun, 01/31/2010 - 11:24am

I was under the impression that the roughing rules do change slightly when the passer is outside the pocket. If not it would make sense to allow a bit more leeway to defenders; after all, the intentional grounding rules are eased-up in such situations, it would seem fair to give the defense a break, too.

2
by burbman (not verified) :: Thu, 01/28/2010 - 7:42pm

I would have thought this stat would show "why team X sucks" or "why team Y is great" but I see five playoff quarterbacks on each list. Go figure.

3
by Led :: Thu, 01/28/2010 - 7:44pm

Interesting that Big Ben is in the second chart. Also interseting in that there doesn't seem to be a compelling relationship between number of hits taken and passing efficiency. High and low DVOA guys sprinkled throughout both charts.

6
by Temo :: Thu, 01/28/2010 - 7:47pm

Actually, it seems like if you made a group-based DVOA rating, the "gets hit a lot" group would be better than the "doesn't get hit a lot" group.

8
by tally :: Thu, 01/28/2010 - 7:59pm

No, they actually seem pretty equal.

Gets hit a lot:

Player DVOA rank
9-D.Garrard 23th
8-M.Schaub 6th
13-K.Warner 13th
12-T.Brady 2nd
9-T.Romo 7th
17-J.Campbell 24th
4-B.Favre 4th
6-J.Cutler 30th
8-M.Hasselbeck 34th
12-A.Rodgers 9th
Average rank 15.2

Doesn't Get Hit:

Player DVOA Rank
8-K.Orton 17th
18-P.Manning 5th
5-D.McNabb 20th
7-B.Roethlisberger 8th
10-V.Young 14th
9-D.Brees 3rd
7-C.Henne 21st
6-M.Sanchez 36th
17-P.Rivers 1st
9-C.Palmer 19th
Average rank 14.4

Just as many good and bad/mediocre QBs sprinkled into each group. There's basically no obvious correlation between DVOA and QB hit rate.

29
by Temo :: Fri, 01/29/2010 - 1:10am

You're right on the data you presented, but it's not close to being a comprehensive enough analysis to say "There's basically no obvious correlation between DVOA and QB hit rate."

There's a bunch of other stats that we would need to see first.

49
by tally :: Fri, 01/29/2010 - 5:24pm

I qualified it as "no obvious correlation," with obvious being the operative word.

Ideally, you'd have a list of QB hit rates and DVOA for every QB with a minimum # of drop backs and look at r-squared. However, the two lists already provide hit rates on 20 QBs, which is a decent sample size. I'd feel pretty confident in saying that there would be little correlation.

50
by countertorque :: Fri, 01/29/2010 - 5:24pm

There's a correlation between DVOA and QB hit rate or there isn't. This data for this year shows there isn't. There are many other questions that could be asked and answered that involve other statistics. But, they do not affect the truthfulness of the statement "There's basically no obvious correlation between DVOA and QB hit rate," which is pretty clearly true.

54
by tally :: Fri, 01/29/2010 - 6:31pm

The data from this article isn't complete, which is why I felt it necessary to qualify my statement. I'm just saying that I couldn't see any correlation from the limited data set, not whether there is any.

12
by Mr Shush :: Thu, 01/28/2010 - 8:08pm

Totally hit reply to the wrong comment. Ignore this post.

48
by Ben Johnson (not verified) :: Fri, 01/29/2010 - 5:11pm

I think these charts need to include sacks, or sack rate. I think comparing the ratio of hits:sacks would be interesting.

5
by Temo :: Thu, 01/28/2010 - 7:46pm

Please get better pass blocking O-Lineman for Romo. Thanks.

13
by Mr Shush :: Thu, 01/28/2010 - 8:08pm

Funny, I kind of feel like a large part of Romo's value comes from his ability to survive and be effective behind a line that's not very good at pass protection - much like Roethlisberger. He's not that accurate, he doesn't always make the best decisions, but he's incredibly good at moving around and buying himself time for a guy to get open enough that his deficiencies as a passer don't matter so much. I almost think getting him better (and therefore presumably more expensive) pass protectors would be a waste of cap space. Oh, wait . . .

19
by Brendan Scolari :: Thu, 01/28/2010 - 9:10pm

Except the Steelers line actually was fine this year, whereas the Cowboys was terrible at pass blocking (albeit superb at run blocking). The guys the Cowboys are paying a lot (Flozell, Davis, Gurode) are all great run blockers so I'd say even if your premise were true (and I don't think it is personally) they are still getting fairly good value out of them.

27
by Staubach12 :: Thu, 01/28/2010 - 11:56pm

Except that Romo has developed persistent back problems that will end his career early if he doesn't start getting better pass blocking soon. Also, I think you overestimate his deficiencies as a passer; he had a rating of 97.1 and only threw 9 interceptions this year. Any quarterback improves with better protection.

9
by Dales :: Thu, 01/28/2010 - 8:01pm

B-Roth on the least-hit list?

O.o

Not saying that is wrong. Just saying it is... o.O

11
by Mr Shush :: Thu, 01/28/2010 - 8:07pm

Yeah, that really jumped out at me, too. Did the Steelers offensive line take a huge step forward, or has Roethlisberger got better at getting the ball away quickly? Or both?

14
by alexbond :: Thu, 01/28/2010 - 8:32pm

I think Roethlisberger is good enough at breaking tackles and shaking guys off that he doesn't go down much. If you had a stat for "touched by a defensive player in any way at some point in the play", I imagine Ben would be very high on the list. It seems like every third pass play, a guy has a clean shot at Ben and he pump-fakes then just kind of spins away from the guy or ducks under him or stiff-arms him or something then throws it.

20
by Brendan Scolari :: Thu, 01/28/2010 - 9:23pm

Are you sure that the QB has to go down for it to go down as a hit? I'm not sure that's the case.

It's interesting that PFF also has the Steelers line rated collectively as much higher than the Cowboys, the two lines are rated at 2.4 and -23.2 respectively. I'd guess that a large portion of the perception has to do with confirmation bias (the Steelers have had a poor line for years) and the fact that Big Ben's avoided sacks (literally shaking guys off) look much flashier than someone like Romo's (whose footwork allows him to do so in a much less noticeable manner).

I'd dispute that Roethlisberger had much worse protection this year than someone like Peyton Manning, or certainly Romo, and I don't think there's much objective evidence to say otherwise.

57
by dsouten :: Sat, 01/30/2010 - 12:03pm

Yes, the definition of Hit is that the QB must hit the ground. So all the guys that bounced off Roethlisberger don't count.

17
by Siren :: Thu, 01/28/2010 - 8:53pm

I suspect that it is a perception effect. You need to have a sack column and a total column to expose a few things. What we are measuring here is the number of sacks that the QB has averted by reading the situation and getting rid of the ball, either to target - Which is part QB, part reciever - or to ground, which is all the QB. If you really want to consider a QB's 'ditch the ball' skill, think of each hit and sack as a 'break in protection', and each hit as a 'sack averted'

Consider:


...........Hits..Sacks.%Averted
T.Brady.....54.....16.....77
P.Manning...29.....10.....74
T.Romo......53.....34.....61
HWSNBN......51.....34.....60
A.Rodgers...46.....50.....48
B.Roeth.....31.....50.....38

So, while the Pittsburgh offensive line is giving Ben the same level of protection as Romo and Farve are getting, when he does get touched, he's going down still holding the ball. But, he's got that habit of throwing the ball to a receiver with a half-ton plus of D line and linebackers hanging off him. So, without the success rate for the situation, we can't really say if it's a worth while strategy for him.

18
by Scott C :: Thu, 01/28/2010 - 9:10pm

Eh.

Or you could say that guys that don't get hit that much but have low %averted (rivers) are good at getting the ball out before getting hit.

Its a lot more complicated, and the above analysis assumes all "hits + sacks" are not something that could be avoided by throwing the ball sooner or sidestepping a rusher.

30
by The Guy You Don't Want to Hear (not verified) :: Fri, 01/29/2010 - 1:11am

I like that in the chart he shall not be named but in your description he is named.

31
by Bobman :: Fri, 01/29/2010 - 2:42am

Hey, I didn't get it in the chart, so thanks for explaining. That's awesome.

So he's Voldemort now? Kinda makes sense since Voldy sought a way to live forever, every time you think he's dead he keeps coming back, and he could never decide which telly to buy at Sears.

36
by ammek :: Fri, 01/29/2010 - 9:53am

And I like that after 19 not-always-discreet years, he still has the most misspelled surname among active QBs (although a lot of people cop out with 'Ben').

10
by Danish Denver-Fan :: Thu, 01/28/2010 - 8:07pm

Surprised to see Roethlisberger, Cutler and Rodgers so far down the list.

Maybe my surprise is a testament to my slight Cutler-crush and the awesomness og R and R. Thats the only reasonable explanation anyway..

28
by tuluse :: Fri, 01/29/2010 - 12:26am

You have to add sacks and hits together to get a true list of "hits"

42
by J. Morse (not verified) :: Fri, 01/29/2010 - 12:40pm

I'd be interested to see Cutler's trendline. He was getting creamed regularly until they benched Orlando Pace. The hits seemed to dip dramatically after that.

15
by Packerpalooza (not verified) :: Thu, 01/28/2010 - 8:32pm

I think the distinction is just that more of Rodgers hits resulted in sacks. Certainly in the first half of the season.

It is also true that Rodgers took some BRUTAL hits that were never flagged. Mike McCarthy had to be wearing out his UPS number sending tapes to the league office asking "WTF?"

16
by Larry Hagman's Liver (not verified) :: Thu, 01/28/2010 - 8:51pm

I totally expected to see Matt Cassel at the top of the hit list. Surprising that at least 10 other QBs got hit more.

23
by starzero :: Thu, 01/28/2010 - 10:03pm

and here i was thinking the colts' liine was terrible.

or maybe it is, and manning is just that good.

24
by Brendan Scolari :: Thu, 01/28/2010 - 11:05pm

They're alright, it's mostly Manning though.

32
by Bobman :: Fri, 01/29/2010 - 2:51am

Last year these charts agreed with you since he was among the most hit and least sacked. This year he was least sacked, but also not hit much. Part of that was replacing 2nd rounder Tony Ugoh at LT with 6th rounder Charlie Johnson, 2nd rounder Mike Pollack with Arena League OG Kyle DeVan, and the return of Ryan Lilja who was out all last year. A lot of it IS Manning, but they tweaked the line just enough to give him the extra half second or so he needs. And for them, that's loads more important than run blocking.

Based on these charts for 2008 and 2009, and my own faulty eyes, it seems that the Colts OL went from below average last year to above average this year. That's not to say they're actually good, but with Manning, above-average seems to be good enough. Of course, Joe "Hit In The Backfield Again" Addai might have an issue with that statement. (New nickname suggestion: HIBA) His fumble last week was basically every run for the last 2.5 seasons distilled to a five-second sequence.

One wonders what happens next year when Howard Mudd ACTUALLY retires as OL coach, instead of just faux retiring to lock in benefits, and then coming back for 2009 as a consultant.... I guess they burn that bridge when they come to it.

25
by Botswana Meat Commission FC (not verified) :: Thu, 01/28/2010 - 11:18pm

Is it possible that the most-hit quarterbacks tend to have slightly slower releases? So the league's fastest releases would be least hit (Manning, McNabb, etc.).

26
by whodat :: Thu, 01/28/2010 - 11:49pm

This is okay, but can we add stats for concussions, with weightings according to lobe as well as grade?

33
by Paul R :: Fri, 01/29/2010 - 9:16am

David Garrard leads the first list by a significantly uncomfortable margin. If I was him, I'd print out five copies of that list and tape them to the OL's lockers.

34
by dbostedo :: Fri, 01/29/2010 - 9:37am

"...I'm not counting sacks here..."

Aaron - Is there any way this could in bold? Or a larger font? I think several of the commenters have missed it...

35
by huston720 :: Fri, 01/29/2010 - 9:46am

I think most readers realize that these stats aren't counting sacks, but couning sacks as well as hits can help add context. For example if as pointed out above, hits are basically averted sacks where the QB got rid of the ball just in time. Also if both hits and sacks are low then that gives us a good idea that the QB and/or line are playing extremely well.

37
by djanyreason :: Fri, 01/29/2010 - 11:27am

Very quick and dirty and incomplete analysis here, but I compared rank in QB hit rate (assuming the worst 10 in total hits as the worst 10 in hit rate, which is probably incorrect) with rank in ASR for the team, and found almost no correlation between the two (.187). Some QBs (e.g., P. Manning, V. Young, Brees, Campbell) had very close ranks between hit rate and ASR. Some had vast differences (e.g, Roethlisberger 4th least hit, 4th highest ASR; Brady 5th most hit, 2nd lowest ASR).

I found that rather odd.

39
by djanyreason :: Fri, 01/29/2010 - 11:38am

Taking this one step further, and grouping QBs by similar hit/sack rates:

Low hit, high sack: Roethlisberger (4th least hit, 4th most sacked), McNabb (3rd least hit, 13th most sacked), Sanchez (8th least hit, 10th most sacked)

High hit, low sack: Brady (5th most hit, 2nd least sacked), Warner (2nd most hit, 7th least sacked), Schaub (3rd most hit, 8th least sacked)

Low hit, low sack: P. Manning (2nd least hit, 1st least sacked), Brees (6th least hit, 4th least sacked)

High hit, high sack: Garrard (most hit, 5th most sacked), Campbell (3rd most hit, 6th most sacked)

There's obviously, a 12 team gap in the middle for QB hit rate, so its hard to find guys in the middle of both, but Carson Palmer is 10th least hit and 10th least sacked.

41
by huston720 :: Fri, 01/29/2010 - 12:23pm

Thanks for the information, this is exactly why it is important to consider sacks and hits together. It shows that Brees and Manning are good at getting rid of the ball and also setting up their protects (and have good offensive lines) since rushers just aren't getting to them much either for hits or sacks. Meanwhile it shows that brady is helping his o-line out by getting rid of the ball just in time, and Ben and Mcnabb hold onto the ball a bit too long, so if you get there you can probably get the sack. Interesting data that mostly agrees with conventional wisdom.

58
by Joseph :: Sat, 01/30/2010 - 6:27pm

Cue some writer saying O-line play (esp. pass blocking) is SUPREMELY important to be able to get to the SB.
I will say this--this post (#39) shows how well the Saints' O-line is--the Saints running games was first in (rushing) DVOA as well.

40
by Johnny (not verified) :: Fri, 01/29/2010 - 11:43am

As a Pats fan, I knew Brady was going to be in the Top 10, but No. 4 is a tad higher than I thought. Not surprised, though. The whole, "Pats have a great o-line and Brady gets tons of time" myth pushed by the media and Pats fans was a fascade this year. The o-line, especially the right side with Kazcur (puke) and Neal, were awful almost every game except against the "great" pash rushes of the Jags and Titans.

43
by t-locster (not verified) :: Fri, 01/29/2010 - 1:57pm

This stat needs to involve how much time lapsed before the QB got hit. There is probably going to be people out there thinking Hass should have done better compared to other QB's getting hit the same amount of time, but if your getting hit with an average 2 secs to throw compared to someone getting 3-4 secs. to throw, then you can't compare the hits. Sorry if someone already brought this up, I didn't read all the comments.

44
by huston720 :: Fri, 01/29/2010 - 2:26pm

Well time elapsed before the hit isn't going to be in the play by play obviously. But also if a QB is consistently only getting a couple seconds to throw then we would expect that to show up on these stats with lots of hits and also lots of sacks. Then film study could be used to try and determine how much of the blame is o-line and how much is qb, though both will share some of the blame.

45
by R O (not verified) :: Fri, 01/29/2010 - 2:47pm

I'm more than surprised at no mention of Rothlisberger being on the least hit list.

I found this astounding to be honest.

46
by Kal :: Fri, 01/29/2010 - 3:04pm

It's not that surprising from an anecdotal scenario; think about all the times that Ben had sitting back for seconds or scrambling for his life to throw something. His mobility tends to mean he doesn't get hit as often. Also, I suspect that he doesn't get hit often but when he does he gets hit hard; people don't tend to just randomly run into him.

It also might be that a 'hit' doesn't count unless he's knocked down after the ball is released, in which case he takes a lot of hits but isn't actually tackled as much. Hard to say.

59
by TTLG :: Sun, 01/31/2010 - 2:11am

It also might be that a 'hit' doesn't count unless he's knocked down after the ball is released

Yeah, that's it. When he's pressured, and gets hit, but stays on his feet, that's a hurry. I would bet that he was one of the most hurried QBs this year from the few games I watched.

Why is that a chimp does not straddle a goat and ride into the sunset?
-Werner Herzog

47
by huston720 :: Fri, 01/29/2010 - 4:19pm

Note that while he isn't hit very much he is sacked quite a bit. Probably means that when tacklers get to him they either bring him down, or he escapes and is able to get the ball away without being knocked to the ground.

51
by R. Carney (not verified) :: Fri, 01/29/2010 - 6:01pm

The only exception to the rule of the best qbs getting sacked less than the others is Aaron Rodgers, who is unfortunately stuck behind a group (although they showed signs of life towards the end of the year) who needs to wear helmets on and off the field. Favre would have gotten KILLED behind that line this year. You can'tcontrol how often you get hit, but you can control how often you get sacked. Just another reason why Manning, Brees, Rivers, and Brady are the best in the business.

53
by R O (not verified) :: Fri, 01/29/2010 - 6:26pm

Favre would have gotten rid of the ball much quicker.

So he would have been sacked about half as often as Rodgers...and thrown 3 times as many interceptions.

55
by utvikefan (not verified) :: Sat, 01/30/2010 - 1:05am

Thank you guys for more great stats. The thing that stood out at me...enough to think perhaps its wrong is how many passes McNabb had. How can he have that few of passes when they don't run the ball? Crazy. I am almost thinking decimal error ; P.

56
by RickD :: Sat, 01/30/2010 - 2:31am

McNabb was injured early in the season. He missed two games and ended up with fewer than 40 attempts through the Week 5 (including the bye week).

But he did have 55 passes against the Chargers, FWIW.