Writers of Pro Football Prospectus 2008

29 Sep 2005

QBs Under Pressure

Over at the PROTRADE site, Ben Alamar takes a look at the best and worst quarterbacks of 2004 in a variety of situations: 3rd downs, blitzes, and red zone. (You may remember Ben as the guy who helped us create the DVOA-based win projection system.) The moral of the story: blitz Marc Bulger but not Peyton Manning, and try to trap Tom Brady in third-and-long.

However, there are a couple of places where I would take issue with Ben's assumptions. He's using STATS Inc. numbers here, and -- as I have mentioned numerous times on this site -- I have no idea what STATS Inc. means when they refer to a play as a "blitz." If a 3-4 defense sends a linebacker, does STATS Inc. count that as a blitz? STATS Inc. also gives a YAC (yards after catch) number for each quarterback, and Ben takes that out, saying that quarterbacks who throw short passes should not be rewarded for the efforts of their receivers. But often YAC is about hitting an open man in stride, not a receiver who is able to avoid multiple defenders as he runs down the field. Even a screen pass shows quarterback talent; one of Tom Brady's greatest skills is his ability to allow blocks to develop, look off defenders, and then, when they've all committed, throw the screen pass at just the right time.

Posted by: Aaron Schatz on 29 Sep 2005

24 comments, Last at 02 Oct 2005, 3:30pm by David Keller

Comments

1
by Al (not verified) :: Thu, 09/29/2005 - 12:11pm

For more evidence of a QB's impact on YAC, you don't have to look much further than alien-era Kurt Warner. I'm not sure I've ever seen a quarterback do a better job of hitting his receivers in stride than the pre-injury Warner in St. Louis. It seemed like every throw was perfectly placed a foot in front of Holt, Bruce or Hakim, allowing them to catch the ball in perfect stride, never having to slow down. Oh, the memories of having both Holt and Bruce as my fantasy wide receivers and watching each of them have 80+ yards and at least a TD every week.

2
by Carl (not verified) :: Thu, 09/29/2005 - 12:12pm

Aaron, doesn't your new deal with Fox allow you to use the Stats, Inc. package? They're owned by the same parent corporation.

3
by RowdyRoddyPiper (not verified) :: Thu, 09/29/2005 - 12:19pm

It's official, ProTrade is on the map, they've just been blocked by the proxy on my work network. Good job.

4
by princeton73 (not verified) :: Thu, 09/29/2005 - 12:45pm

alien-era Kurt Warner

was Warner the alien, or Vermeil?

(I mean, we KNOW Martz is a Venusian)

5
by Aaron (not verified) :: Thu, 09/29/2005 - 12:48pm

Well, so are the New York Post and FOXNews, but I still write for the New York Sun and I won't be appearing on Hannity and Colmes anytime soon. It's something I'm looking into, but don't hold your breath.

6
by Vern (not verified) :: Thu, 09/29/2005 - 1:26pm

A classic steaming pile of nonsense. So, YAC is all the receiver's credit, but on a long pass it's all the QB? So I guess the receiver just appears magically in the right spot in the right time, and is open too? How many INTs are a result of a bad route being run, misreading the blitz etc.? While we're at it, isn't beating the blitz mostly a function of the O-Line?

I applaud the desire to pick apart situations, but doing it partially is worse than not at all.

If you're going to pick apart plays, you would need to know how everyone on the team did their role. Then you would have things like, QB X does the best when the O-Line sucks, QB Y makes the most out of poor route runners, QB Z does the least with wide open receivers, and so on.

7
by Jim A (not verified) :: Thu, 09/29/2005 - 1:51pm

Arrgh, with access to those highly-coveted blitz stats, couldn't he have compiled some much simpler stats, such as completions, attempts, yards, INTs, and TDs on blitz plays? Or even just NFL QB rating? That would have been far more useful.

8
by Ben (not verified) :: Thu, 09/29/2005 - 4:23pm

I agree with you all that taking out the YAC is controversial. The arguement goes that a more accurate passer will lead to higher YAC. This is logical, but the stats just do not support it. After controlling for the distance of the throw, there is no statistically signifcant relationship (and it is not close on significance) between passing accuracy and YAC. Even without controlling for the pass distance, the simple correlation between accuracy and YAC is 0.03.

And Vern, I wish I had the data to do what you suggest.

9
by princeton73 (not verified) :: Thu, 09/29/2005 - 4:47pm

The arguement goes that a more accurate passer will lead to higher YAC. This is logical, but the stats just do not support it.

no, the argumant goes that a certain type of QB playing in a certain system (e.g. Montana in his prime & the "alien era Warner") have an ability to increase YAC in a manner COMPLETELY INDEPENDENT of classic "accuracy" measurements

however, I suspect if you included YAC in the calculations to come up with the various lists in the article, it wouldn't change the rankings very much

10
by Vern (not verified) :: Thu, 09/29/2005 - 5:26pm

The point of following and ranking player stats is to guage how that player helps a team win more. Trying to answer the question - if we move this player from team A to team B, will the team(s) improve/decline?

Arguably, the most important skill that a team depends on from their QB to help win games is decision making, or "managing the game", particularly when under pressure situations. (Even if you argue that it is only one aspect of winning, it is certainly the least understood, and thus most deserving of more analysis. So much so that QB types like Roethlisberger, Brady, even Montana go way later in drafts than they should.)

Yet, attempts to better rate QBs seem always to focus on the QB as a passer (or even define QB as little more than passer). You don't need more stats when you can already spot nice spiral thrown accurately during a game or even at a combine.

If at the end we're still throwing up our hands and falling into the mantra of "they played on good teams" for some and "we'll never know if they'd have done better over here" for others, then how useful is the stat?

11
by JG (not verified) :: Thu, 09/29/2005 - 5:35pm

RE #4:
princeton73, given that 'Men are from Mars and Women are from Venus,' are you questioning Mike Martz's manhood?

12
by JG (not verified) :: Thu, 09/29/2005 - 5:36pm

RE #4:
princeton73, given that 'Men are from Mars and Women are from Venus,' are you questioning Mike Martz's manhood?

13
by Purds (not verified) :: Thu, 09/29/2005 - 5:39pm

Ben:

I can see why you chose to take YAC out (even though I might disagree with that choice), but then don't you also need to take screens out of the Yards per attempt equation, as it's not the QB's fault that his coordinator called a short pass play by design?

14
by Pat (not verified) :: Thu, 09/29/2005 - 5:46pm

Ben: (#8)

Wait, why are you controlling for distance? There's only a slim window of distance where accuracy should improve YAC - probably between 0-10 yards, where the chance of completing a pass is probably fairly flat anyway.

Past that (10+ yards) I'd guess that YAC has a lot more to do with the passing tree (and the defense's coverage) than the QB's accuracy.

15
by steelers homer (not verified) :: Thu, 09/29/2005 - 8:21pm

I notice Ben Roethlisberger's rating is low compared to other quarterbacks, because Ben's recievers don't drop many of his passes. And when you throw out all the dropped passes the other quarterbacks suddenly seem much better.

Well, yeah, but ...

16
by Kachunk (not verified) :: Thu, 09/29/2005 - 8:23pm

RE #14: I don't think that's true. If a receiver has a defender beat deep, then accuracy has a lot to do with YAC. If the receiver has to jump for the ball then the DB has a chance to catch up to him and hit him in the air, thus no YAC. If the ball drops perfectly over the guy's shoulder and into his hands then he just keeps running, and possibly beats the DB completely. It depends on how the play is designed--some routes, like comebacks are really not designed for YAC, but others are, it just depends on the QB's ability to place the ball accurately.

17
by Pat (not verified) :: Thu, 09/29/2005 - 9:48pm

#16:

I don't agree. For one thing, a defender rarely gets wide open deep for long enough for a QB to realize he's wide open, and throw accurately to him. The QB generally launches the ball much earlier than that, and it's usually thrown in such a way that only the WR can get it.

QB accuracy has a ton to do with YAC for success of very short yardage passes (screens, etc.) but for longer passes, the QB is often placing the ball not for maximum YAC, but to simply not throw an interception.

18
by mactbone (not verified) :: Thu, 09/29/2005 - 9:53pm

You're all missing the obvious point - the Bears don't have a QB. I think most of us have known this but to see it so starkly in print? It makes you want to give Angelo a wedgie. High floor - what a crock.

19
by Jim A (not verified) :: Fri, 09/30/2005 - 2:08am

I never really understood this obsession with coming up with the perfect QB rating. It seems like everyone and his brother has his own system that is touted as better than the official NFL rating. Like most of them, I think this one glosses over some very critical assumptions, many of which are discussed here.

Instead of leaping toward the end result, I believe we should be re-examining the steps along the way. The data from Stats Inc. provides a unique opportunity to study some new and interesting things, such as the blitz and YAC. I hope the author will consider revisiting these ideas in greater detail with future research.

20
by Joey (not verified) :: Fri, 09/30/2005 - 2:35am

Re: #1
Amen. He was amazing. I read a quote from him saying arm strength is overrated because he hardly ever throws as hard as he can. I guess you don't have to when the ball always ends up exactly where you want it, like he was throwing back then.

21
by Theo (not verified) :: Fri, 09/30/2005 - 4:28am

you could extract all the YAC after 15 yards

22
by Dennis (not verified) :: Fri, 09/30/2005 - 2:56pm

Re #8 - What is your definition of "accuracy"? Are you just talking about completion percentage? If a QB does make a difference in YAC, then it should be somewhat consistent for QBs from year to year. Have there been any studies done on that?

Re: #17 - The accuracy makes a huge difference on YAC for longer passes. Of course the QB is trying to throw it so only the receiver can catch it. But there's a huge difference in YAC between hitting the receiver in stride and the receiving having to dive or be off balance to make the catch.

23
by Mr Shush (not verified) :: Sat, 10/01/2005 - 8:02am

Intriguingly (for me at least) David Carr comes out absolutely horrible - worse than I imagine even his most virulent detractors would have him - in Ben's table of overall performance, but right around the middle of the pack in not just one but all three "pressure situations".

On a side-note, I know it's obvious, but I imagine Randy Moss is quite a help to a quarterback who wants to complete third and longs, and I don't suppose anyone would be too surprised to discover that the already considerable value of TO and Westbrook in the passing game is magnified in the red-zone.

Actually, come to that, what happens with touchdowns after the catch? Suppose McNabb throws a shovel-pass to Westbrook on the four, and he takes it in for six: we know that Donovan, Madden Curse and all, gets no credit for the YAC, but surely to be consistent that ought to mean no credit for the score, either? Is this so?

24
by David Keller (not verified) :: Sun, 10/02/2005 - 3:30pm

What about Eric Dickerson? No one threw better passes than Eric. Not one interception.