Writers of Pro Football Prospectus 2008

01 Dec 2008

2008 Quick Reads: Week 13

This week in Quick Reads: How Brian Westbrook's 110 rushing yards make him the best back of the week, and how Michael Turner's 120 rushing yards make him the worst.

Posted by: Bill Barnwell on 01 Dec 2008

70 comments, Last at 04 Dec 2008, 4:45pm by black president


by The Hypno-Toad :: Mon, 12/01/2008 - 8:28pm

Interesting stuff about Wesbrook and Turner. Out of curiousity, why doesn't there seem to be anything from the Broncos-Jets game? It's entirely possible that it's there and I'm just not seeing it.

by Dales :: Mon, 12/01/2008 - 8:39pm

Yeah, Favre and Cutler are AWOL.

by kevinNYC (not verified) :: Mon, 12/01/2008 - 8:43pm

The deep pass hasn't been a part of the Giants offense at any point this season after an offseason in which they sought to make it more of a factor. The loss of Burress puts more pressure on Eli to be excellent in reading coverage and decision making. The strength of Burress (and WRs like Randy Moss even for a great QB like Tom Brady) is he doesn't have to be open and he can even be double-teamed and still make plays. One thing that will help Eli in the transition is that Kevin Gilbride appears to have turned into a very competent OC and playcaller.

I was surprised Jason Campbell finished so high despite the Skins getting a 3rd of their yards on the final 2 meaningless drives. Was Michael Turner really the least productive RB or was that so something could be written about him?

by Kurt :: Mon, 12/01/2008 - 11:03pm

I don't think FO metrics are able to recognize when a drive is meaningless. Which doesn't make them useless, but you have to combine them with some common sense.

by Aloysius Mephistopheles (not verified) :: Mon, 12/01/2008 - 8:45pm

I didn't realize that DYAR includes defensive pass interference penalties. Does it penalize the receiver for offensive pass interference?

by countertorque :: Mon, 12/01/2008 - 10:45pm

I believe this stems from a study that FO did that showed that DPI is the only penalty that has a significant effect on the outcome of the game. I'm too lazy to look for a link or a reference. In any case, since it was signifcant, it had to be included in DVOA to improve the correlation with wins and losses. I guess they then extended this to include it in individual player stats. Players would not get penalized by OPI, because that penalty doesn't significantly affect the outcome of the game.

I'm not sure I agree that a QB or a WR should get a benefit for pulling DPI. I think the correct conclusion to draw is that DPI is way too severe and is distorting the outcome of the game in a way that other penalties are not. But, perhaps drawing DPI is a repeatable skill that certain players are better at.

by countertorque :: Mon, 12/01/2008 - 10:56pm

Actually this statement from the FO Basics page seems to disagree with me completely.

"Teams with more offensive penalties generally lose more games, but there is no correlation between defensive penalties and losses. The penalty that correlates highest with losses is the False Start, and the penalty that teams will have called most consistently from year to year is the False Start."

So, I must be smoking too much crack.

by jimbohead :: Tue, 12/02/2008 - 4:24am

here's the article that deals with pass interference:


To sum up, they include it because they found that, for some players (Favre to Driver for instance), it was relatively consistent from year to year, implying skill rather than chance.

by Bobman :: Mon, 12/01/2008 - 8:47pm

This is such utter BS. What, NOBODY from the Colts/Browns game shows up here? I'm shocked... SHOCKED! Not even mister Prevailing Swirly Wind? He had a stellar game!

by shake n bake :: Tue, 12/02/2008 - 12:21am

Mr. Prevailing Swirly Wind was dominant. One of the best defensive performance I've seen in a long time.

by Dales :: Mon, 12/01/2008 - 9:00pm

The Campbell numbers surprise me, too. First, let's look at 3rd downs.

3-6-WAS 36 (13:44) (Shotgun) 17-J.Campbell pass short middle to 83-J.Thrash to WAS 40 for 4 yards (21-K.Phillips).
3-9-WAS 40 (9:09) (Shotgun) 17-J.Campbell pass short left to 47-C.Cooley to WAS 46 for 6 yards (21-K.Phillips).

3-12-NYG 45 (15:00) (Shotgun) 17-J.Campbell pass incomplete short right to 82-A.Randle El.
3-12-WAS 41 (10:33) 17-J.Campbell pass short right to 26-C.Portis to NYG 44 for 15 yards (21-K.Phillips, 93-J.Alford). WAS-26-C.Portis was injured during the play. His return is Questionable. (Note, I believe this play was mostly Portis doing what Portis does)
3-10-NYG 49 (2:50) (Shotgun) 17-J.Campbell pass incomplete short middle to 83-J.Thrash.

3-8-WAS 22 (12:36) (Shotgun) 17-J.Campbell pass incomplete deep right to 12-M.Kelly (23-C.Webster) [24-T.Thomas].
3-6-WAS 32 (7:29) (Shotgun) 17-J.Campbell pass short left to 82-A.Randle El to WAS 40 for 8 yards (24-T.Thomas).

3-7-WAS 23 (14:07) (Shotgun) 17-J.Campbell pass short left to 47-C.Cooley to WAS 29 for 6 yards (35-K.Dockery).
3-9-WAS 31 (11:01) 17-J.Campbell pass short middle to 82-A.Randle El to WAS 39 for 8 yards (35-K.Dockery).
3-13-NYG 41 (3:58) (Shotgun) 17-J.Campbell pass short left to 89-S.Moss to NYG 34 for 7 yards (97-M.Kiwanuka)
3-6-WAS 34 (2:18) (Shotgun) 17-J.Campbell pass short right to 87-T.Yoder to WAS 43 for 9 yards (58-A.Pierce).
3-13-NYG 46 (:49) (No Huddle, Shotgun) 17-J.Campbell right end pushed ob at NYG 30 for 16 yards (21-K.Phillips).
3-10-NYG 17 (:03) (Shotgun) 17-J.Campbell pass short middle to 89-S.Moss to NYG 6 for 11 yards (24-T.Thomas). FUMBLES (24-T.Thomas), RECOVERED by NYG-37-J.Butler at NYG 6. 37-J.Butler to NYG 6 for no gain (89-S.Moss).

DVOA and DYAR may not get more predictive if it punished those 3rd-and-8 7 yard completions, but it would be more just! ;-)

by Tim Kirk (not verified) :: Mon, 12/01/2008 - 9:29pm

I'm actually more interested in the RB figures for the Denver-Jets game, any chance of having something posted here? I was rather expecting Hillis to make the top 5 after a very solid seeming performance against a highly ranked rush defence.

by George A. Brownfield (not verified) :: Mon, 12/01/2008 - 9:31pm

OK (Steelers fan here) trying to understand how Big Ben ends up 21st this week (and possibly 23rd if they post Favre and Cutler). Compared to say, Peyton Manning - Ben has more TD's, less picks, higher YPC, against a team with a winning record instead of the Browns (OK but maybe a just as bad defense) and grades out 10 places lower? And if we are recognizing drops it doesn't matter how his opponent grades out. Kurt Warner throws three picks and grades out higher?

by Matt W (not verified) :: Mon, 12/01/2008 - 10:26pm

As a fellow Steelers fan, I'm guessing it's probably all the incompletions Ben threw -- he had 16, Manning had 6.

by Temo :: Mon, 12/01/2008 - 11:00pm

Firstly, YPC doesn't really tell you much. Most people only care about YPA.

Also, notice that Roethlisberger's EYDS (effective yards) is actually higher than Manning's, even if his DYAR is lower. What this tells you is that Roethlisberger did more, but he did more with more attempts than Manning.

by BlueStarDude :: Mon, 12/01/2008 - 9:39pm

RE: TO and how his "Ninety-eight yards and a 21-yard pass interference call are nice, but it was [against] the Seahawks."

But isn't that what the "D" in DYAR is for?

by Key19 :: Tue, 12/02/2008 - 6:35pm

I don't think that the comment was made in regards to his DYAR total this week, but for why they're still skeptical about him.

by widderslainte :: Mon, 12/01/2008 - 9:44pm

Of course, Ryan does benefit from having a good running game -- and more importantly, a running game that attracts defenses' attention.

Wait a second, didn't you just tell us how crappy Turner was?

by Derek (not verified) :: Mon, 12/01/2008 - 11:40pm

A. That was just for this game, not for the entire season. For the season, Turner is 10th in DYAR and the Falcons are 4th in Rushing DVOA.

by Anonymous111 (not verified) :: Mon, 12/01/2008 - 10:33pm

I would assume that Westbrook's directional splits have a lot to do with running behind the 2nd (and then 3rd) string RG.

by HC (not verified) :: Tue, 12/02/2008 - 12:14am

No Cutler....hummmmmmm.......either your system for evaluating Quarterbacks is screwed or somebody complete forgot about the Jets-Broncos game. Somebody forgot right?

by Bill Barnwell :: Tue, 12/02/2008 - 1:54am

Hey -- there were some issues with the PBP from Denver-Jets that apparently prevented it from being processed in time for Quick Reads.

by Hummingbird Cyborg :: Tue, 12/02/2008 - 2:25am

Any chance it will be there later.

I don't mind if it isn't up for a bit, but I'd definitely be interested to see it.

by hector :: Tue, 12/02/2008 - 3:52am

It's a shame drops can't be factored into the equation. Boldin had a couple, one a sure thing in the end zone. Obviously, all incomplete passes are not created equal, nor should they be counted as such, but it's a tricky scale to evaluate and quantify.

by morganja :: Tue, 12/02/2008 - 4:19am

I have to strongly, strongly disagree with the assertion that Rodger's INT was 'basically a Hail Mary'.

If that was what it was judged by the system then the system is deeply flawed. He threw the INT with a minute and a half left in the game, on a second down, with two timeouts left and down by only four.

How is that a Hail Mary situation? For every other quarterback in the league that is called a two minute drill situation. If your ratings discount an interception thrown at that critical point in a game, with that down and distance and situation, than your system essentially sucks.

I would think that you simply misspoke, but with Rodgers ranked so high, it seems apparent that either your data is simply flawed, or that your system needs a serious overhaul.

by ammek :: Tue, 12/02/2008 - 4:48am

I agree, it wasn't (or shouldn't have been) a Hail Mary. Rather, it was underthrown into quadruple coverage. I guess the intended receiver, Greg Jennings, will have his DVOA suffer for it, but apparently not Rodgers. Hmmm.

Rodgers had a nice outing against a very good pass defense, but got a tad lucky with his DPAR/DVOA with respect to a high snap which the Packers lost. FO stats consider that fumble recoveries are essentially random, but from watching the game it was clear that Rodgers ought to have recovered that one. I think in some way, in an ideal system, he should be penalized for it.

Also, Romo's intentional grounding penalty cost him 20+ DPAR???

by resident jenius :: Tue, 12/02/2008 - 8:28am

I think calling it a Hail Mary is really more of a symptom of having no (i think I’m right here) NFC South fans writing for FO. Was it a Hail Mary? Nope. But really who cares? Rodgers did have one heck of a game.

by morganja :: Tue, 12/02/2008 - 4:36am

I don't mean to sound too harsh. But the fact that an Int in that situation is discounted pretty much throws the whole system into doubt. We're not talking about a last second heave in a desperate attempt to win a game, I can understand how that is discounted, but a pass in that situation should, if anything, be magnified by its obvious importance.

by MJK :: Tue, 12/02/2008 - 12:27pm

I don't believe the system discounts the INT--my impression is just that Bill was discounting the INT in his description. I don't know for certain, but I suspect that Rodgers had the best day even including the INT. The reason is that I think I remember Aaron stating at one point that the only plays that are actually ignored is in obvious "garbage time" when one team is so far behind that winning is pretty much impossible, especially when one or both teams put their 2nd stringers in, and that those situations were made as a judgement call by the person running the analysis, not picked by the algorithm. If that's right, I would be shocked that any play at the end of a close game would be ignored--if Rodgers had thrown for a TD and not an INT, the play certainly would have been counted, so I have a hard time thinking the INT was discounted. If I'm wrong, then I would agree with you--it's daft for being ignored. But I doubt that.

Remember, the play can be counted, and yet not "count" that much. The "AR" in DYAR means "above replacement", which means, like DVOA (where the "OA" is "over average", the play is normlized relative to some average metric (for DYAR it is the average play of a "replacement" level player, for DVOA it is the league average) in the specific situation. I'm guessing converting a long pass when down by more than a FG under two minutes is something that not too many replacement QB's can do and that INTs are more common in that situation than in many others, so Rodgers throwing an INT in that situation probably hurts DYAR (and DVOA) a lot less than, say, throwing an INT in the first quarter. Just like throwing an incomplete on 3rd and 30 probably hurts a lot less than on 3rd and 3.

by taxistan :: Tue, 12/02/2008 - 7:06am

You are nuts for thinking Michael Turner was a bad back in the Falcons victory over the Chargers.

by ammek :: Tue, 12/02/2008 - 7:28am

Did you read the explanation? All the evidence you need is there. Feel free to query it intelligently. "You are nuts" doesn't qualify. Also, DPAR is calculated by a computer program; reprimanding a computer is the very definition of "nuts".

It should be added that DPAR is a cumulative statistic: that is, the more negative plays a running back compiles, the lower his DPAR will be. Turner's failures weren't necessarily catastrophic on an individual basis; but there were a lot of them, each one chiselling away at his numbers. So by anointing him "worst back of the week", Quick Reads is actually lambasting Mike Smith and the Falcons' sideline for continuing to call Turner's name, even though he was repeatedly losing value. This appears particularly idiotic in view of the heavy load Turner has already carried this season, especially with a postseason berth (400+ carries?) very much in reach.

I admit I expected Tomlinson's numbers to be worse; any chance we could see them, please? Especially as Atlanta's run defense has been horrible.

I guess Reggie Bush didn't have enough carries for a dreadful DPAR, but his DVOA can't have improved.

by crack (not verified) :: Tue, 12/02/2008 - 10:57am

That a back can account for over a third of a teams yards in a win and still be deemed the worst back is a legitimate reason to question the value of a stat.

The fact that the D in D(voa|yar) is bootstrapped from itself may be part of the issue here.

by Wanker79 :: Tue, 12/02/2008 - 1:53pm

I guess I'll just have to reiterate ammek's initial question. Did you even bother reading the explanation in Bill's opening paragraphs?

by dbostedo :: Tue, 12/02/2008 - 3:40pm

Another problem with your comment is that you think it matters whether or not the Falcons won (and the implication that you'd be more OK with saying Turner had a mediocre day if they Falcons had lost).

Maybe I'm reading too much into this...

by crack (not verified) :: Wed, 12/03/2008 - 1:31am

Yes winning matters! What is people's problems with this premise? The point of D(YAR|VOA) is supposed to be that they are more strongly correlated with winning than box score stats.

And this isn't saying he had a mediocre day, it's saying he had the most negative day of anyone at his position. I'm not saying Turner should be in the top 5, i'm not even saying he should be positive. He's -41!

If a player has the most negative impact of any player at his position while accounting for half his team's offensive plays and the team loses then it makes sense. High utilization of a poorly performing player should lead to a loss.

by Eddo :: Wed, 12/03/2008 - 12:36pm

Of course winning matters, and DYAR and DVOA correlate to winning, but what Turner's negative DYAR is saying is that the Falcons would have won by even more if they had stopped using him so much. Why is that so hard to grasp?

For example, when Rex Grossman had a 1.3 QB rating, but the Bears still won, did that mean he wasn't the worst-performing QB that week? Of course not; it means that the rest of the team was good enough to overcome his poor performance.

Now, you can legitimately argue Turner was not the worst RB this week. Personally, I think there's value in getting 31 carries when your team has the lead; however, that's nearly impossible to model in something like DYAR. The argument that the Falcons won, so Turner's day couldn't have been that bad is just silly, though.

by crack (not verified) :: Wed, 12/03/2008 - 1:41pm

I grasp the claim that they would have won by more if they had used less Turner, I just believe it's false.

What was Rex's DYAR for that game?

I'm not saying it couldn't have been bad. I haven't once stated that. I even conceded it was likely bad. I consider replacement level bad. I guess if you consider replacement level good then we could argue that. Being below it is bad.

My biggest question is with the D adjustment, it is the complete black box in the whole thing. If SD played at 7% D-DVOA yesterday, instead of their season long 14%, because some linebacker's ingrown toenail healed how much would that have affected Turner's DYAR? It wouldn't have. Now, I understand FOs can't base a system on the toenail health of an LB, but there are other issues that plague self-referential higher order stats and sometimes performances like this can help point them out. Does anyone know how to get the YAR for the game from the premium DB? I'd like to know it.

by black president (not verified) :: Thu, 12/04/2008 - 4:45pm

I think you make an abundance of excellent points, and furthermore I would like to add that numbers are gay.

*doffs cap, does a soft shoe*

by chubbypuppy (not verified) :: Tue, 12/02/2008 - 9:54am

Late to the party but Rodgers interception was a TERRIBLE decision.

There was legitimate time left in the game. The Packers had 2 timeouts. There were not one but TWO guys open 10-15 yards underneath.

While Rodgers is down the list of guilty parties from Sunday's debacle (hey coverage units, are you awake?) that interception was, well, Favrenunian in nature.

It was grotesque. And incredibly stupid.

So I agree with above and VEHEMENTLY disagree with the author's note in the Rodgers entry.

by Jon Coit (not verified) :: Tue, 12/02/2008 - 11:29am

At this point, I'd have thought the point about Turner is obvious: while he piled up standard stats that look impressive, he performed poorly compared to the league average against an atrocious defense. The fact that the team won is utterly beside the point, on par with "When X back rushes more that 25 times a game, the team is 15-1." At the outside, Turner's performance might suggest alterations to the yards required for a successful play (which have been raised in the last two years IIRC).

The beef with Rodgers is mainly with the tendentious text in the explanation. He was obviously penalized for the INT as an unsuccessful play, the play wasn't "discounted" as folks have implied in the rating. It was balanced out by the other successes. Sheesh.

by crack (not verified) :: Tue, 12/02/2008 - 12:26pm

I guess I disagree. A team putting together a winning effort seems important to me. Yards seem different than attempts to me. A back accounting for a third of the offense in a loss may indicate indifference by the opposition. In a close win it doesn't. The point of NFL games is to win. It's possible ATL could have won without Turner, but then they'd need to do something on the half the offensive plays he used. If he was truly the worst back it would be easy, but I question that ranking and I question that the Turnerless ATL off would have even effective enough to win. If removing a back ranked worst in the league from your offense takes you from a win to a loss what's the value of the ranking?

by Eddo :: Tue, 12/02/2008 - 12:41pm

"It's possible ATL could have won without Turner, but then they'd need to do something on the half the offensive plays he used."
Remember, the "R" in DYAR stands for "replacement". Turner's negative DYAR is saying that you could have used a readily available back instead of Turner on all his carries and gotten even better production.

Turner's negative DYAR is not saying that he was an absolute negative for the Falcons. In fact, look at Effective Yards; Turner has 68. What the -41 DYAR is saying, however, is that another RB would be expected to put up 109 EYds, given the same number of carries Turner had.

by crack (not verified) :: Tue, 12/02/2008 - 2:15pm

I don't think that's right. Eyds isn't figured that way. The Eyds would be more like 127. I think it's (dvoa+1)*(league avg/attempt)*(attempts) (http://www.footballoutsiders.com/dvoa-ratings/2008/dyar-mailbag).

Which is weird because it brings in (league avg/attempt) and that is based on many games right in line with Turner's.

But aside from the enigma that is eyds, who is the R reference back right now? Ryan Grant appears to have a -1 DYAR for the season so he's about as close as it gets. Do you think Ryan Grant would have had more success? I guess he might have, but I don't think so, I think he would have been as good at best. So what is this stat telling us? What can we take from the list?

I suspect the D in D(voa|yar) is a problem (at least in the player level stats), I have no way of knowing since I don't know how it's computed or even what the Dless YAR and VOA are. I also don't know how to test them, with the team level DVOA you can regress it against W-L. It apparently does pretty well. I don't know what I could compare DYAR to since it's counting (thus not pace adjusted) and related to individual performance, which is apparently subjective since some people even think wins don't matter. I could keep following it blindly and claiming anyone who thinks it has problems is a troll, but I'll probably slowly stop looking at it.

by Bowl Game Anomaly :: Tue, 12/02/2008 - 3:46pm

You're confusing average performance with specific performance.

What these numbers say (ignoring for the moment the impact of the O-line on rushing performance) is that Michael Turner had a bad game. They do not say that Turner is a bad back. They say that Ryan Grant's typical performance would have been better than what Turner did. Of course, Turner is usually even better than Grant and if Turner had played up to his usual standard, he would have had an even more effective game than the theoretical Grant game. So yes, the system is claiming that a typical Grant game would have been more effective than Turner's performance, but also a typical Turner game would have been much more effective than his actual performance. Turner was atypically bad, to the point that he hurt his team rather than helping. Usually, though, he helps a lot.

(Formerly "The McNabb Bowl Game Anomaly")

by crack (not verified) :: Wed, 12/03/2008 - 12:07am

But it's not just saying he had a bad day, it's saying his day had a larger cumulative negative impact on his team than any other RB. Yet despite him having the largest cumulatively negative impact of any RB and accounting for half of their offensive plays, his team won. I question the value of a stat that produces that result.

And secondarily, he doesn't usually help a lot according to FO stats, his dvoa is like -.7%.

by Bowl Game Anomaly :: Wed, 12/03/2008 - 6:07pm

Two points:

1. Yes, Turner hurt his team a lot. They still won. Why is that a problem? Is it impossible for a team to win when it gets a very negative contribution from a key player? Note that San Diego is not a good team, and also note that the Falcons got only 3.3 yards per rush (less than San Diego's 3.7) but made up for it by achieving 9.0 yards per pass (San Diego had a pathetic 4.0). So Atlanta won in part because their outstanding efficiency in passing made up for their serious inefficiency in rushing, and also because they forced San Diego to play very inefficiently in the passing game.

2. Turner's -.7% DVOA may seem poor, but it really isn't. The more carries you have, the more difficult it is to achieve high value on most of them. (This is due to several factors: larger sample size, a greater physical toll, more targeted gameplanning by the opponent, and so on.) FO measures cumulative value (DYAR) compared to a replacement player, not an average player, because you don't have to be above average to be valuable. You just have to be better than the alternative (i.e. a replacement). Turner is average, and average is valuable. (Replacement level is something like -12% or -13% DVOA, which is well below average.)

(Formerly "The McNabb Bowl Game Anomaly")

by Eddo :: Tue, 12/02/2008 - 3:53pm

I won't argue with your EYds point, as I'm not entirely sure I used them correctly either.

But as for replacement level, let's use Ryan Grant. On Sunday, Michael Turner averaged 3.9 yards per carry over 31 carries. Grant is averaging 3.8 yards per carry on the year. So, on the surface, Grant would have put up just about the same per-carry production as Turner. Now, factor in that San Diego's defense is awful, and you get an adjusted figure well below zero.

The whole point of DYAR is to go beyond basic numbers. You point out that Turner had over 1/3 of Atlanta's yards, which is true, but he also was involved in almost 1/2 the plays! What DYAR is saying is that the Falcons would have been even better if they had not kept running Turner into the line and thrown some more or used Norwood instead.

by crack (not verified) :: Wed, 12/03/2008 - 1:15am

Ugh the site ate a comment I had spent about an hour writing (granted much of it was trying to track down something tangotiger wrote and getting sidetracked).

Suffice it to say that I'm skeptical of black box analytical processes, and that higher order stats aren't of any intrinsic value. Many attempts to enlighten have ended up obfuscating instead.

by E :: Tue, 12/02/2008 - 12:55pm

About Rodgers, I'm not sure that your comment is correct. I believe that while he may have been slightly penalized for an unsucessful play, DYAR/DVOA do "discount" the (negative) value of an interception that is a hail mary, such that it is not nearly as bad as other interceptions. As several commenters have noted, that INT was not a hail mary and if anything was the kind of negative play that DYAR/DVOA should have magnified, rather than discounted. I don't think there's anything wrong with "the system" - more likely it was a game-charting error that then got inputted into the system as a hail mary, rather than what it was. I wonder whether Bill and the team could fix that input and whether/to what extent that would drop Rodgers this week.

by Bill Barnwell :: Tue, 12/02/2008 - 12:21pm

The Rodgers' interception wasn't counted in the system as a Hail Mary -- it was counted as a legitimate interception. I was just saying that it basically amounted to a Hail Mary in reality.

by Eddo :: Tue, 12/02/2008 - 12:45pm

I respectfully diagree, Bill. On a Hail Mary, an interception give the same result as an incomplete pass: the game is over. However, Rodgers's interception came on second down, trailing by less than a TD, with enough time and timeouts to move the Packers down the field. An incomplete pass there would not have ended the game, but the interception did.

by E :: Tue, 12/02/2008 - 1:20pm

In that case my comment above was obviously wrong (and like others I disagree that the pass amounted to a Hail Mary).

by chubbypuppy (not verified) :: Tue, 12/02/2008 - 12:47pm


That assessment is incorrect. A "Hail Mary" is a last gasp effort with minimal time left in the game, say less than 10 seconds. The Packers had two timeouts and a 1:30 to play. Rodgers threw the ball up for grabs on a 2nd and 10. While not a situation likely to result in a TD it was most definitely not a "Hail Mary" type scenario.

And even setting aside the context Rodgers made the WORST possible decision. There were no less than FOUR Panthers in the immediate vicinity. Two Packer receivers were just sitting at around the Packer 35 with nobody within 10 yards of them. It is almost a certainty that a completion to either one would have had GB at their 40 with 1:10 on the clock assuming the receiver got out of bounds. (Very likely).

So now a team has 60 yards to go with 70 seconds against a Carolina secondary that had stunk it up most of the game.

I repeat, this decision was NOT the primary or even the secondary reason GB lost. But it was a REALLY stupid throw. I think Rodgers has a fine future but boy when he goofs he GOOFS.

by MCS :: Tue, 12/02/2008 - 1:35pm

He's just having fun out there!

by Jon Snow (not verified) :: Tue, 12/02/2008 - 1:02pm

What kind of play-by-play issues? Isn't the validity of the analysis dependent on a complete sample? Was there something with the numbers of Cutler, Favre, Hillis, or Jones that didn't fit well with the Quick Reads supposition?

Would love to see this (again) once those guys are figured in.

by The Hypno-Toad :: Tue, 12/02/2008 - 1:56pm

"Was there something with the numbers of Cutler, Favre, Hillis, or Jones that didn't fit well with the Quick Reads supposition?"

I must be misinterpreting what you're saying there... Because it sounds like you are accusing FO of intentionally leaving those guys off the list because they didn't fit what FO wanted Quick Reads to mean. If they were going to intentionally omit information that makes their model look weird, the Eagles wouldn't have shown up in DVOA for the past 4 years.
I mean, I would like to see the numbers from that game too, but I'm not ready to go all conspiracy theory about it.

by Shattenjager :: Tue, 12/02/2008 - 6:32pm

Perhaps this is nit-picking, but as I recall, Aaron commented in an Audibles a few weeks ago that DVOA has not been over-rating the Eagles for years. It's the pre-season projections for the Eagles that have been so high. Obviously, this season, the DVOA seems to be sharing the projection system's crush on the Eagles, but I don't believe it's been a 4-year-long issue.

That said, your point is still valid that the Eagles would be hidden in this year's DVOA if they were going to hide data in such a manner.

by Bowl Game Anomaly :: Tue, 12/02/2008 - 3:53pm

Someone with such a good nick should make more intelligent comments. (On the other hand, if your namesake is a bastard...)

Snarkiness aside, I assume that what he meant was that there were errors in the copying and pasting of the PbP numbers. I can't imagine that the Outsiders are trying to hide the results of that game because they are troubling.

FO is not afraid of controversy. How many weeks this season have the Eagles been ranked higher than the Titans? But please, by all means keep posting until you draw the dreaded Football Outsiders Message Board Curse down on yourself and your team.

(Formerly "The McNabb Bowl Game Anomaly")

by Eddo :: Tue, 12/02/2008 - 3:59pm

"Isn't the validity of the analysis dependent on a complete sample?"
Replacement level isn't calculated on a week-by-week basis, but rather over the past several seasons, so no, the validity of the published DYAR figures for a given week is not dependent on a complete sample for that week.

by Rich Conley (not verified) :: Wed, 12/03/2008 - 2:48pm


The "R" part is based over multiple seasons, but the D is not. You can't have a proper DYAR without a complete D (Defense). Without a complete sample, the D is not correct.

by Eddo :: Wed, 12/03/2008 - 11:59pm

Isn't the "D" part in Quick Reads is determined prior to the week in question? If you're truly RIGHT, Rich, then Quick Reads is always incorrect, as the Monday night game is never included.

by Jon Snow (not verified) :: Tue, 12/02/2008 - 5:00pm

Please, no more curses!!!

I knew Hillis' DYAR was probably near the top this week (ended up he was 2nd behind Westbrook) and I just thought that it was odd that an analysis comparing the best and worst backs of the week would do so without knowing Hillis' position in the weekly DYAR. Evidently, that information was available - it just wasn't posted.

Hillis being 2nd doesn't impact the storyline at all, so nope, no conspiracies here; but yes, I am a bastard. Additionally, I know nothing. ;)

by morganja :: Tue, 12/02/2008 - 1:04pm

Thanks for the explanation. I'm glad the system doesn't discount the INT. It was a horrible decision at a critical moment, and frankly was the moment that the Packers lost the game. But if it was just your commentary that considers it a Hail Mary and not the system, we can vehemently disagree without worrying about the system.

Rodgers had a very good game, but certainly not atypical of the West Coast offense which regularly shreds the Panther secondary for yards, but somehow often not for wins.

If there is a subjective element which lets a person gauge whether this throw came in a critical situation or just 'garbage time', I would recommend that they reconsider.

by ammek :: Tue, 12/02/2008 - 3:20pm

Having followed the 2008 Packers, I can assure you that in each game there have been multiple moments that lead you to say to yourself, "That's it. They've buggered it up again."

For me, Sunday, they were, in order:
- Rodgers' delay of game penalty on 1st and goal;
- seeing the offense come out in a 'heavy' formation on 3rd-and-goal from the 1;
- having to kick off after the field goal;
- watching Delhomme basically heave the ball into space, and the long wait for it to come down with crushing inevitability in a Panther's hands;
- Rodgers' game-ending pass into quadruple coverage.

by BroncosGuy (not verified) :: Tue, 12/02/2008 - 1:49pm

OPI is obviously less damaging than DPI due to the yards involved. I am offering another reason OPI might not correlate to losing. Many believe that defensive penalties don't correlate because, while obviously damaging, they may also correlate to something positive, namely aggressiveness. Might OPI have this same hidden benefit? Given how often receivers still push off and get away with it, it be better to be aggressive in this area and suffer the occasional OPI.

As for the DPI being skill- or chance-related: to me, this penalty is officiated so poorly that it is nearly random, rendering the underlying data meaningless.

by Joseph :: Tue, 12/02/2008 - 3:36pm

I think the reason that DPI is counted in DYAR/DVOA is because it is assumed (at least by the FO stat machine) that the receiver would have caught the ball. Thus DPI deprives the WR/QB the chance to accumulate those yds, and thus his DPAR/DVOA. As the penalty is an automatic 1st down & spot foul, the offense is given the result of the completion without the stats that would come from it. My guess is that the offensive DPAR/DVOA wouldn't "add up" with the individual DPAR/DVOA, although I doubt that it actually does.
(Or does it? Does the sum of all receivers & runners DPAR add up to the total offense's? [i.e., minus the QB passing DPAR, but including his rushing DPAR] Or is it possible with the different baselines?)

by morganja :: Tue, 12/02/2008 - 2:35pm

The flaw in that otherwise good argument is in the phrase "get away with it". Getting caught should show a negative effect, even if the upteen times they get away with it they show a positive effect. I'm looking at you, Irvin.

by BroncosGuy (not verified) :: Tue, 12/02/2008 - 6:15pm

morganja: I agree the OPI is still a negative play; that wasn't my point. My thesis is that OPI empirically might actually correlate positively to overall passing success because of the umpteen extra completions you allude to. This thesis has nothing to do with the mechanics of DYAR etc. In fact, it has little to do with anything at all, but it seemed mildly interesting at the time.

by Kevin A (not verified) :: Tue, 12/02/2008 - 5:33pm

Glad to see where Peyton Manning was on the last. I have seen a lot of OMG Peyton Manning only had 125 yards and 2 picks he stinks on message boards with media shows going right along. Of the 6 incompletions, the 1st was dropped in the end zone (nice play) 2nd was a heave ho pick with :03 left in the 1st half. He had a left end zone fade to Wayne that wasn't close, a dropped screen by Addai, and a misfire on 3rd & 3 across the middle to Wayne when had finally had it with the failure to get 3rd & short on the ground. It was a bad pick, but other than that, it was an okay game and 11th is about right.

I think about 2 games that the Cleveland game reminded me of. The 2004 divisional game - right down to the poor conditions and the 6-3 halftime score. Running game couldn't get going and the Pats weren't going to give up anyting deep, D wore down in the 2nd half 20-3 final score, and the Peyton chokes stories go at full throttle.

The second game that it reminded me of was the Philly game in 2006. 2 deep zone where you couldn't even see receivers on the screen. Manning was 14-20 183, 1/1. However, Addai had 170 yards rushing and 4 TD's. This was the evil twin brother game/beta experiment. Cleveland was daring and inviting the Colts to run, gave up 156 yards a game during the season when they were acutally trying to take it away and allowed 90 yards on 29 carries including stopping 3rd & short three times and 4th & 1 once. The Colts could get 5 on 1st and 4 on second and then couldn't get 3rd & 1 and Manning wouldn't throw it on 3rd & 1 with a 2 deep zone. It is like they are better getting the 2nd &9s and 3rd & 8s - he is a better hitting at 0-2 than the running backs are at 3-1.

When you take his deep ball challenges and couple them with running challenges, he is getting everything done this year on short and mid passes when everyone knows that it is coming. It seems like a tough formula to be able to do anything in January.

by The Ninjalectual :: Wed, 12/03/2008 - 11:46am

Touchdowns do not equal yardage "bonuses." Dyar is retarded. Bring back DPAR!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!