Writers of Pro Football Prospectus 2008

09 Jan 2006

2005 Quick Reads: Week 18

You thought the Quick Reads were over, but no!

Posted by: Aaron Schatz on 09 Jan 2006

46 comments, Last at 10 Jan 2006, 3:33pm by ernie cohen


by Stiller Fan in Cle (not verified) :: Mon, 01/09/2006 - 7:23pm

I just liked the Bettis comment...I'm not wure what annoyed me more, the network repeatedly cutting to Bettis' parents, or the whole AJ Hawk's girlfriend is the other team's QB's brother in the Fiesta Bowl...

Network coverage is getting to me more and more...I blame this site...

by Craig B (not verified) :: Mon, 01/09/2006 - 7:39pm

I know it was only one play but what was Carson Palmer's DPAR?

by LnGrrrR (not verified) :: Mon, 01/09/2006 - 7:42pm

Hehe I thought it was funny...because it looked like they were jumping every time the camera was on them. Did they know the camera was on them and start jumping?
Or were they just jumping constantly?

by Queequeg (not verified) :: Mon, 01/09/2006 - 7:57pm

The comment on Kitna from Tim Gerheim seems contradictory to me. If Kitna "doesn't really care how open a guy is" surely he wouldn't, as Tim rightly identifies, “have held on to the ball way too long� throughout the second half. I would suggest that it was Kitna's indecisiveness in the pocket and his reluctance to throw the ball, rather than his ‘gunslinger’ tendencies, that cost the Bengals.

by admin :: Mon, 01/09/2006 - 8:03pm

Sorry, meant to stick that in. 2.3 DPAR for Palmer.

by Will Allen (not verified) :: Mon, 01/09/2006 - 8:09pm

I've never been too enamored with Eli Manning, but I suspect that Roethlisberger's DPAR will drop dramatically next Sunday if the Colts rush for well over two hundred yards, while holding the Pittsburgh rushing attack to an anemic performance. I'd even wager ol' Ben would toss three interceptions, although none of them might be as awful as Eli's Excellent Across the Field Adventure.

by DJAnyReason (not verified) :: Mon, 01/09/2006 - 8:20pm

Re: #6

Welcome to non-sequitor theater!

Its nice to know that somebody's going to carry the water for the "Ben's just a product of his run game and defense" troops

by Dman (not verified) :: Mon, 01/09/2006 - 8:32pm

Yeah, I remember when we used to say the same thing about brady.

by Harry (not verified) :: Mon, 01/09/2006 - 8:35pm

If Dillon has become as mediocre as DPAR says, his return in the Bronco's game may not be the difference maker Pats fans hope. Maybe they'd be better off with Pass.

by Reinhard (not verified) :: Mon, 01/09/2006 - 8:51pm

The Asante Samuel route where he turned around out of man coverage and the ball was coming right at him and he made a great grab? Or am I forgetting one

by Dave (not verified) :: Mon, 01/09/2006 - 8:53pm

RE #9: For the entire regular season, Dillon is 13th in DPAR and 15th in DVOA among qualified rushers. So, it's possible that his undistinguished work this weekend is an anomaly, or it's possible he's not going to be very productive this postseason.

The one thing that jumps at me from his season-long stats is his 49% success rate, which would concern me if I were a Pats fan.

by Will Allen (not verified) :: Mon, 01/09/2006 - 9:17pm

#7, welcome to needlessly abrasive posting theater! Congratulations, you're the winner.

I'll simply note that I'm not "carrying water" (hey, you win the cliche contest as well!) for anyone. However, judging qb performance on a game in which the opposition runs for over 200 yards, while one's own running game is almost completely non-productive, can be a little tricky. Now, if we can get past the silliness, do you care to dispute my assertion, that if the Colts were to rush for over 200 yards next Sunday, while the Steelers get nothing from the running game, that Roethlisberger's DPAR would likely suffer significantly?

by Matt Weiner (not verified) :: Mon, 01/09/2006 - 9:20pm

#3: I thought the camera cut to them whenever Bettis did something good. So they were already jumping up and down.

by Starshatterer (not verified) :: Mon, 01/09/2006 - 9:34pm

Reinhard (#10 )--

My understanding was, that Samuel was supposed to mimic Cover 2 on that route, then release the deep guy to the saftey and jump the short outside route if the Jaguars were passing from that formation.

In other words, Samuel's jumping that route for the pick was a defensive playcall, based on Jacksonville's tendencies.

by BlueStarDude (not verified) :: Mon, 01/09/2006 - 10:28pm

re: Cadillac Williams (-1.9 DPAR)

Doesn't this pretty much put him over the top as the overrated rookie of the year? Kyle Orton was overrated by the Skip Bayless and Mike Greenberg types who count QB "wins", but Cadillac's production seems to have been overstated by almost everybody.

But then, how much of his poor DPAR/DVOA was due to the Tampa O-line?

by karl (not verified) :: Mon, 01/09/2006 - 11:10pm

And if the opposing team scores 500 points and holds the steelers to -67 points, Roethlisberger's DPAR is sure to be horrendous. Look, its a silly argument. The guy is efficent, it helps that his team is balanced (good running, efficient passing), but that shouldn't be a knock against him - just look at what Eli, Orton, Grossman, Brunell, Frerotte, Simms, et al have done over the entire season in similar settings. No his team doesnt pass much and they don't depend absolutely on his passing, but when he does throw the ball he does it very well.

by karl (not verified) :: Mon, 01/09/2006 - 11:13pm

re #15:

I wish you had a vote for the ROY awards, actually I wish you had all of them, my boy Ronnie should've walked away with that one, leaving matt jones choking on dust.

by CaffeineMan (not verified) :: Mon, 01/09/2006 - 11:30pm

The one thing that jumps at me from his season-long stats is his 49% success rate, which would concern me if I were a Pats fan.

Which puts him 11th (actually same as Tiki Barber) higher than his DPAR rank, which is 13th, so I'm not sure what you're getting at, unless you're trying to make the point about relative value of success rate versus DPAR. Dillon has dropped off since last year for sure, but I'd still WAY rather have him than Pass.

Now if you want to see something really worrying, look at the Pats adjusted line yards: ranked 26th, stuffed 23rd, although they're 3rd in power situations. But look especially at Left Tackle behind the two rookie O-linemen: 2.47, ranked 31st. Ugh.

by DGL (not verified) :: Mon, 01/09/2006 - 11:31pm

#6, #12: Eli's picks didn't happen because the Panthers had 223 yards rushing and the Giants had 41; they came as a result of poor judgement, poor location of the ball, and poor decisions. In other words, he looked a lot like Ben in last year's AFC Championship game. If he throws a couple of balls away instead of forcing, he has maybe one pick -- and a DPAR of, what, maybe -10.

In week 13, if you recall, the Bengals had 100 yards of rushing and effectively shut down the Steelers running game (95 yards, 3.4 YPC). Ben's DPAR was a season-high 13.8. And that was with three picks.

by Dave (not verified) :: Mon, 01/09/2006 - 11:41pm

Re #18:

Oops. Yeah, should have looked harder at that one.

by Will Allen (not verified) :: Mon, 01/09/2006 - 11:52pm

DGL, if you hold the opinion that the quality of a quarterback's decison-making has, on average, zero correlation to how the opposition is performing, and how his own running game is performing, or you are unable to discern the qualitative difference between rushing for 100 yards and 200 yards, or 95 and 41, well, you and I aren't watching the same games.

Karl, please re-post where I was knocking Roethlisberger. I was not. I made an assertion that when the opponent rushes for well over the league average, while the the quarterback's team rushes for well below the league average, said quarterback's DPAR is very likely to suffer greatly. I believe this is likely to be particularly trues when the quarterback's team is like Eli Manning's and normally rushes the ball very well. Of course, I could be very wrong. Perhaps I'll go back and check what Ben's DPAR was the last time the Steelers gave up over 200 yards rushing while gaining less than 50 on the ground.

by RCH (not verified) :: Tue, 01/10/2006 - 12:22am

#18 - Excellent point. I've noticed that Dillon seems to be particularly futile running left, a problem as he seems to prefer that direction (always tucking the ball under his left arm and never switching).

by David (not verified) :: Tue, 01/10/2006 - 12:52am

re: #18

I wonder how much of their success at power runs is attributable to their preternatural ability to get the first down with QB sneaks and how much of it results from line blocking for the RBs. For the former, only the center and guards really matter in their battle against DTs; for the latter, it's the line and the backs.

by Nick (not verified) :: Tue, 01/10/2006 - 12:57am

Will Allen -

This event, Steelers held to less than 50 rushing while the opponent runs for over 200 is so rare as to be ridiculous. The last time it happened - I checked - was Pittsburgh @ Chicago on 12/13/1992 Cowher’s first season as coach, since that date they've held 9 opponents to less than 50 while running for over 200 themselves.

Phil in 2004
NE in 2004
Min in 2001
Jax in 2000
Cle in 1999
Indy in 1996 - Playoffs
Buf in 1993
NO in 1993
Cin in 1993

This is clearly a thinly veiled effort to suggest that Big Ben is a system QB, or God-forbid, a game manager a-la Brad Johnson, rather than the effective QB he's shown to be. Remember, his DPAR reflects that he plays extremely well in given situations . . . no credit for 8 yarders on 3rd and 15. Once in 13 years for that type of rushing disparity, if Vegas will give me odds on it happening versus the Colts I'll take that. Heck, Eli was 11 years old when this last happened.

by DGL (not verified) :: Tue, 01/10/2006 - 1:28am

#21: Good luck on that. I checked NFL gamebooks back through 2002 and couldn't find a game where the Steelers gave up more than 169 yards rushing (11/17/03 against the 49ers). And I don't think this site has Tommy Gun's DPAR from that game available.

I'm not going to disagree that an opponent's ability to stop your running game is likely to have an impact on a QB's DPAR (although I don't have the week-by-week DPAR to research it -- given that it's a variation on the original question that led Aaron to develop DVOA, maybe one of The Outsiders knows or could look into the correlation between rushing yardage and QB DPAR. On the one hand, if the rush is getting stuffed, a good QB will have more pass attempts and, if successful, a higher DPAR. On the other hand, if the rush is getting stuffed, a bad QB will have more pass attempts and more sacks, interceptions, and incompletions and a lower DPAR...). At least, it's the conventional wisdom of how to beat the Giants -- stop Tiki and make Eli beat you.

On the other hand, the conventional wisdom of how to beat the Steelers has been stop the running game and make Ben beat you. The Colts stopped the run (86 yards/3.4 YPC) and Ben played poorly (I don't know his DPAR, but except for one less pick, his game wasn't unlike Eli's: 17-26, 111 yards, 2 picks); the Bengals stopped the run (95 yards, 3.4 YPC) and Ben had the best game of his career to date (29/41, 386 yards, 13.8 DPAR). So from this admittedly small sample of two, I wouldn't say it's a foregone conclusion.

As far as opponents rushing yardage affecting QB DPAR, I don't see it. Dominating time of posession is important to minimize the number of posessions the opponents get; if the opponents are averaging 16 yards and one first down per posession, you could let 'em have the ball 20 times and you'll still win.

by Purds (not verified) :: Tue, 01/10/2006 - 1:33am

Wow, Will, these guys don't get it. Let me try with a different QB's name.

I’ve never been too enamored with Eli Manning, but I suspect that Brady's DPAR will drop dramatically next Sunday if the Broncos rush for well over two hundred yards, while holding the NE rushing attack to an anemic performance. I’d even wager ol’ Tom would toss three interceptions, although none of them might be as awful as Eli’s Excellent Across the Field Adventure.

Get the point? That Eli's terrible performance was certainly negatively influenced by his team's terrible running game, and Carolina's great running game. He's saying you cannot just isolate the DPAR of two guys playing the same weekend when so many other factors were so dramatically different.

(And, I think Will used Ben as the model because the FO guys compared the two.)

by Purds (not verified) :: Tue, 01/10/2006 - 1:36am


Good arguments in opposition to Will's. Your post wasn't up when I posted mine.

But, let's remove the argument away from our favorite players, and talk in terms of positions and influences on the game of various offensive success rates, as DGL has.

by MJK (not verified) :: Tue, 01/10/2006 - 1:39am

We Pats fans aren't counting on Dillon's "return in the Bronco’s game ... [to] be the difference maker Pats fans hope." We counting on:

Tedy Bruschi instead of Monty Beisel,

Ellis Hobbs instead of Duane Starks,

Dillon instead of Pass,

Kevin Faulk instead of Amos Zereoue

Willie McGinest with two hands instead of Willie McGinest with only one

A strong safety that has played the same position for more than one consecutive game

to be the difference makers.

by MJK (not verified) :: Tue, 01/10/2006 - 1:40am

Oops... I meant a strong safety that has played the same position for multiple consecutive games. I guess every player has always played their position for one consecutive game. :-)

by RowdyRoddyPiper (not verified) :: Tue, 01/10/2006 - 1:41am

Purds, beat me to the punch. Though I think Will used Roethlisberger because he was the highest rated (by DPAR) QB of the week. Can we waste 5-10 posts discussing that?

I still am glad Manning had a horrible game. I have a deep dislike for him going back to the 2004 draft...see I really wanted the Steelers to take Rivers, but the whole Eli thing (not wanting to play for SD) precipitated a situation where NY would end up with Eli and Rivers would end up in SD. I was very much gloom and doom about Ben (though I will never admit this using my real name). I just didn't buy into the mid-major hype. In fact I'm not big on QBs from outside of the Big Ten.

Needless to say, I'm glad Eli got his way, and Big Ben is in the 'burgh. Hey, I never claimed to be a good GM.

by Starshatterer (not verified) :: Tue, 01/10/2006 - 1:46am

All of you bashing Will Allen: here is what he was trying to say:

"People should be less harsh on Eli Manning. Yes he had a bad day, but so would [INSERT NAME OF A KNOWN GOOD QUARTERBACK] if his running game produced under 50 yards, while the [INSERT NAME OF THE TEAM OF THE KNOWN GOOD QUARTERBACK] defense gave up 200 yards rushing."

Nick (#24 ), let's try one of your counter-examples:

"People should be less harsh on Eli Manning. Yes he had a bad day, but so would [TOM BRADY] if his running game produced under 50 yards, while the [PATRIOTS] defense gave up 200 yards rushing."

Now let's check Brady's statline from last year's Halloween massacre: 25/43 for 271 yards, 2 TDs, 2 INTs. Passer rating of 72.9 or so. That's bad.

I can't calculate DPAR on the fly, but since Brady was sacked a four times and fumbled once, it probably wasn't good. Not as bad as Eli Manning's day this past Sunday, but I'm guessing Brady last year had just a teensy bit more expected of him than did Eli this past weekend.

So yes, Eli sucked this past Sunday. Who wouldn't? That, I think, was Will Allen's point.

by DGL (not verified) :: Tue, 01/10/2006 - 2:00am

Purds, Will:

No, no, I get it, I just don't necessarily agree with all of it.

Specifically, I don't agree with the implication that Carolina running for 220 yards is a significant contributing factor to Eli having a DPAR of -11.6.

Yes, if the opponents run for 220 yards, they will likely dominate TOP (like the 3:1 that the Panthers had). And yes, if the opponents dominate TOP 3:1, the QB will likely run fewer plays and have fewer pass opportunities, and hence a lower DPAR (although I recall some games from the Rams "Greatest Show On Turf" days where the Rams got beat in TOP by 2:1 or worse and won by scores like 45-17 -- I don't think anyone has Kurt Warner's DPAR from those days, but I'd bet it's up there).

But I think that the TOP (and indirectly the opponents' running success) influence on QB DPAR is a multiplier. So if a QB is playing well, but his team loses TOP by, say, 2:1 or 3:1, his DPAR will be lower -- maybe only 6 instead of 12. But ISTM that to have a DPAR of -11.6, a QB has to be playing pretty poorly.

Yes, if a team is well behind and time is short (both of which can result from the opponents having run the ball well all game), a QB will be more likely to make poor decisions and poor throws. But again, in the Giants game, Eli's first pick came with 7 minutes left in the third quarter down 10-0: neither far behind nor time short. His second pick came, down 17-0 true, but still with 4:53 left in the third quarter.

In the first half, with the game either 0-0 or the Giants down 7-0, so before TOP or the Panthers' running game could really have any impact on what Eli was doing, the Giants had 9 passing plays. By my estimation, 3 were successful (10 yards on 2-8, 15 yards on 2-9, 8 yards on 1-10), one was partially successful (3 yards on 1-10), and 5 were unsuccessful (incompletions on 3-2, 3-18, 2-9, 3-9; sack on 2-11). That's just not good. If the Panthers hadn't been dominating TOP, maybe the Giants would have had one or two more drives in the half -- but there's nothing to indicate they would have been any more successful if they'd had them, nor that Eli's DPAR would have been any better.

by CaffeineMan (not verified) :: Tue, 01/10/2006 - 2:04am

Yeah the flip side of that stat is interesting as well: the Pats line is ranked 8th on runs at Right Tackle.

by DGL (not verified) :: Tue, 01/10/2006 - 2:12am

Purds, Will, Star,

Sorry if it's coming across as bashing; not intended that way. I'm trying to debate the following points either made or implied by Will (and Purds, my post #32 was being written before I saw your post #26...):

1. Opponent rushing success was a significant factor in Eli's poor showing.

2. Lack of Giants' rushing success was a significant factor in Eli's poor showing.

I'll tentatively concede #2 -- it fits with the conventional wisdom, but I'd be really interested in seeing FO do a statistical analysis, because if we all beleived everything CW said, we wouldn't be here.

I'll also concede #2 up to a point: namely, I believe opponent success in running the football will reduce the magnitude of the DPAR, but not the direction. In other words, I'd expect to see an inverse correlation between opponent rushing yardage and DPAR magnitude. The more the opponents have the ball, the fewer opportunities a QB will have to perform -- either well, resulting in a lower magnitude positive DPAR, or poorly, resulting in a lower magnitude (closer to 0) negative DPAR.

My anecdotal evidence is in #32 regarding Sunday's Giants game - Eli had a bad game, or if you prefer, the Panthers played really good defense, regardless of what the Panthers offense did.

I'd really like to see those hypotheses tested with data. (Hello crack FO statisticians?)

by DGL (not verified) :: Tue, 01/10/2006 - 2:14am

Duh. My second "concede" should have been "#1". Can I blame lack of sleep?

by CaffeineMan (not verified) :: Tue, 01/10/2006 - 2:14am

Oops, wow, teach me not to refresh before I post. :D

Dave, an interesting question about the sneak thing, I have no feel for it. But don't forget that power situations include 1st and 2nd downs and goal-to-go from the two or one yard line. Not many sneaks there and I recall Dillon being pretty effective. As a matter of fact, I'm quite happy to have him back in the Pats lineup even if he's only used as a "Bus Lite." :D

by LnGrrrR (not verified) :: Tue, 01/10/2006 - 2:18am

I'd say "sucking" and "posting a big goose egg" are two wholly different levels of suckitude. :)

by Justus (not verified) :: Tue, 01/10/2006 - 2:19am

Will Allen - I, along with a few others apparently, understood the point you were making even if the rabid Stillers fans seem determined to misconstrue it as "clearly a thinly veiled effort to suggest that Big Ben is a system QB". (Disclaimer: I'm a Steelers fan.)

by CaffeineMan (not verified) :: Tue, 01/10/2006 - 2:21am

Ellis Hobbs instead of Duane Starks

Can I get a Hallelujah, brother? :D

by CreamyG (not verified) :: Tue, 01/10/2006 - 2:45am

Mark Brunell is clearly ranked too low because he is super smart and tough and was fighting his guts out. . Winning play-off games is way better than this. ALL YOUR DPAR ARE BELONG TO GIBBS!

by Justus (not verified) :: Tue, 01/10/2006 - 2:46am

(Hopefully someone will correct if the following is wrong.) I believe that DPAR is a counting statistic. The entire point of DPAR is to account for number of plays that the player is involved in. If you are involved in few plays you cannot have a high DPAR, even if those few plays are all excellent. This is why David Garrad is #12 in DVOA but with only 176 pass attempts comes in at #23 by DPAR. Meanwhile Favre with 632 passes is #20 in DVOA but #15 in DPAR.

If you throw 3 picks and your team only has 8 possessions then it is going to be awfully tough to climb back to a positive DPAR, especially when the running game is dead and the defense is able to limit the impact of each of your very few passes. That means each one of your passes is at or below the average pass of a "replacement level player", because the average replacement player doesn't face that same type of coverage. You also have fewer passes than a "replacement level player" who will have the opportunity to throw more passes.

The result is a negative DPAR.

If you throw 3 picks but your team has 15 possessions then you have more opportunities to work yourself out of that hole. If your running game is working then you have more of an opportunity for those passes to be longer ones.

If opponent running success leads to a TOP difference then it will limit how high your DPAR can rise. If the TOP difference increases the pressure on the QB to "make plays" (because you've only got 8 minutes left to score three times, for instance) then it could increase the number of below-replacement level plays. It doesn't seem a stretch to suggest that this behaviour is more likely to be seen in young quarterbacks experiencing their first playoff game. Evidence suggesting that would be that Manning didn't start throwing INTs until half-way through the third quarter at which point they were down 10-0, Barber had 29 yards rushing (8 on them on a meaningless run to end the half), and it seemed the only way they were going to win was for Manning to take the offense on his shoulders.

I agree it isn't conclusive but it is highly suggestive.

by Will Allen (not verified) :: Tue, 01/10/2006 - 3:17am

Thanks to all for more clearly explaining what I was attempting to say. I'll pat myself on the back by claiming (and I swear, it's true!) that I said to the person I was watching Giants game with that Manning was going to be tossing interceptions in the second half by the bushels.

Now, this wasn't a stretch, of course, to anybody who had seen Manning this year, but I was even more dead certain than the normal anti-Eli type, because with the Panthers dominating the time of possession, and with a quarterback, Manning, who had gotten very good run support for most of the year, the fact that the Giants weren't running the ball at all meant that Manning was going to feel extreme pressure to make throws he would otherwise have a better chance of avoiding, because the Giants had so few possessions to work with. There are damned few QBs who would have a good DPAR on a day when their team yielded well over 200 yards rushing, while their own rushing attack had less than 50, and those that did would not be on teams which were normally rush dependent.

My crack about finding such a game for Rothliesberger was perhaps a little too abstract; the point was that such games are pretty hard to find, and thus are perhaps not the best ones to use when trying to compare quarterback performances.

by ernie cohen (not verified) :: Tue, 01/10/2006 - 7:03am

Assuming that DPAR is correctly designed, so that at least the sign of a QB's performance is independent of game situation, it doesn't matter how many yards your D has given up. It might matter whether you (and more importantly, the opposing D) think that you won't be able to stop them in the future, because it effectively puts you in a more desparate game situation, but I don't know whether this makes it harder to pass (because the opponents know you are passing) or easier to pass (because the opponents are playing prevent).

The main effect of a good running game on QB evaluation is that the opposing D is more likely to play the run on your passing plays. Moreover, PIT has a somewhat greater propensity to run compared to other teams with equally good running and passing games, further inflating QB value per play.

by Will Allen (not verified) :: Tue, 01/10/2006 - 12:30pm

Ernie, I think it matters how those yards are yielded. When large stretches of time go by between possessions, because the other team is having long, successful running drives, and one's own team isn't running at all, the pressure on a QB really builds, especially when the QB normally receives good run support. I wasn't surprised to read of the example above, when Bulger had a good DPAR when his team yielded a lot of yards running while running very little themsleves, because the Rams have always been a pass-centered team in the Martz era.

DPAR is as good a statistical evaluation that exists for QB play, but it cannot completely seperate the team effect on individual performance, especially in outlier games which have unusual attributes.

by Oswlek (not verified) :: Tue, 01/10/2006 - 2:28pm

MJK #28,

You forgot a few:

Seymour instead of an injured Jarvis Green.

Colvin instead of Chad Brown.

Good Wilfork instead of sucky Wilfork.

And the last one I can think of is Troy Brown. He was inactive in the earlier game and is an improvement over the last dime back and #3 WR.

There is way more going on here than a casual observer of the Pats can really see.

Oh and this statement

A strong safety that has played the same position for more than one consecutive game.

is accurate. To play in more than one consecutive game the player would have had to had played in the previous one. This was the ridiculous stretch of 4 or 5 games that the Pats started a totally different SS, often a player that wasn't even on the team the week before.

by ernie cohen (not verified) :: Tue, 01/10/2006 - 3:33pm

re: #44

One of the assumptions of DPAR (and all of the DP analyses I've seen) is that the "game situation" ignores the past. It doesn't take into account how tired the defense is, how many times the QB has been sacked, etc.

Keep in mind that, much like win shares, DPAR isn't supposed to measure how good someone is; it is supposed to measure more accurately what they did.