Writers of Pro Football Prospectus 2008

printable version

Football Outsiders

Innovative Statistics, Intelligent Analysis



23 Jun 2009

FOA 2009 Preview: New York Giants

The release of Football Outsiders Almanac 2009 is just around the corner, and to get people excited, we've decided to release some sample material from the book. Over the next two weeks, we'll post a bit from each of the three main sections: NFL team chapters, college conference chapters, and "skill player" comments.

To start off, here's a free download of the New York Giants chapter from FOA 2009, written by Bill Barnwell. Let your friends know all about it and spread the word for us.

Download the Giants chapter here.

Pricing for FOA 2009 will be as follows:

  • $12 for the PDF download
  • $20 for the print-to-order book
  • $30 for a package that includes the KUBIAK fantasy projections workbook along with the PDF download
  • $60 for the PDF download, KUBIAK workbook, and a year subscription to Football Outsiders Premium -- since Premium is normally $40 and KUBIAK is normally $20, this essentially means you get Football Outsiders Almanac 2009 for free.

We want to also offer packages that combine the print-to-order book with KUBIAK and/or Premium, but we're not exactly sure how we will handle that yet. We're also exploring a deal that will allow FO Premium members to get a discount on ESPN Insider membership, since so much FO material is now appearing behind that subscription wall.

Hope everybody enjoys our Giants writeup!

Posted by: Aaron Schatz on 23 Jun 2009

102 comments, Last at 12 Aug 2009, 8:48am by TerraByte

Comments

1
by young curmudgeon (not verified) :: Tue, 06/23/2009 - 12:42pm

"skill player comments"--I apologize to the poor, dead horse that I keep beating, but here I go again:

To call ends and backs "skill players" is to imply that playing the other positions requires no or less or only a modicum of skill. This, in turn, (to me at least) suggests that the other guys can just be plugged in to their positions and start performing, i.e., they don't need to learn skills or develop more skills. However, year after year, we see rookie running backs and receivers stepping right off the campus and doing well (not all, of course, but enough). We also see many offensive lineman having to play for two or three years before they develop proper NFL technique. This doesn't fit with the notion that some players are designated "skill players" and others not. It seems to me that if you are playing NFL football, you're pretty darned skillful. Why insult the linemen by suggesting that they are less so than the other guys?

BTW, this isn't personal--I'm a skinny guy who didn't even play football, but was a distance runner.

3
by dbostedo :: Tue, 06/23/2009 - 12:57pm

I would suggest that you start thinking of the term "skill player" as technical football term meaning WR, RB, QB. Don't try to apply the usual definition of each word. Think about it like "Guard". The other offensive lineman guard the QB too - that term doesn't imply that they don't. And then there's the amount of tackling offensive tackles actually do...

"Skill player" used to mean "QB, RB, or WR" has become way too much a part of the vernacular, I think, to bother with trying to change it.

8
by billsfan :: Tue, 06/23/2009 - 1:23pm

There is no law save usage, as they say. Dr. Z often used "skill position" to refer to Left Tackle.

The Quarterback isn't always 1/4 of the way back, nor the halfback halfway, nor the fullback all the way; the tailback doesn't always have a tail. Wide receivers run skinny posts. Tight ends line up loose. The gunner isn't usually armed. The dime-back isn't the 10th guy in the secondary.

Football is full of archaic and sometimes contradictory terminology. Get used to it.

54
by Jimmy Oz (not verified) :: Tue, 06/23/2009 - 7:45pm

The tackle isn't allowed to tackle
The tight end isn't always that tight.
The split end isn't as painful as it sounds.
The nickelback sometimes has musical talent.

61
by Goodbye Ladies :: Tue, 06/23/2009 - 10:51pm

Sir, you win at life!

The "Goodbye, Ladies" Draft Report

71
by Temo :: Wed, 06/24/2009 - 10:13am

Split ends aren't painful if you're talking about hair.

59
by vague (not verified) :: Tue, 06/23/2009 - 9:23pm

bravo on the "Gunner isnt USUALLY armed"

102
by TerraByte (not verified) :: Wed, 08/12/2009 - 8:48am

You're right about offensive tackles. We should rename them False Starters.

19
by Concrete Jungle :: Tue, 06/23/2009 - 2:30pm

I know of a coach from NFLe level who used to refer to Big Skill Guys and Little Skill Guys, which really prevents anybody getting too upset about it.

68
by Levente from Hungary :: Wed, 06/24/2009 - 6:20am

As I remember FO did discuss in an article how improper this term is. I guess that is why it is between quotation marks.

BTW I also have never played football and I'm also a skinny distance runner :-)

2
by James-London :: Tue, 06/23/2009 - 12:44pm

Any news on the cost of the print to order version for us "Limeys"? How much will delivery be and will it be available through Amazon?

Phil Simms is a Cretin.

6
by bubqr :: Tue, 06/23/2009 - 1:14pm

Same question here. This year, Won't have the chance to get it directly in the US, I hope shipping won't be too expensive...

17
by fyo :: Tue, 06/23/2009 - 2:22pm

I've been able to buy it LOCALLY here in Denmark every year (at a very good price, too), despite it being a small market, non-English speaking and NFL not being all that popular.

Based on that, I have a hard time imagining that you won't be able to find a UK store that carries it.

24
by Vince Verhei :: Tue, 06/23/2009 - 2:54pm

Actually, you won't be finding it in any stores, anywhere.

51
by Theo :: Tue, 06/23/2009 - 6:48pm

I've bought it at a dutch online store. You bought it IN store?

67
by Levente from Hungary :: Wed, 06/24/2009 - 6:17am

I did buy it in bookstores (once in Taiwan, once in Hungary). In the first case they had it on stock, the second time I ordered through them (with Amazon.com shipping would have cost me more than the book itself!!).

So I was lobbying for the electronic version for a while so I am glad to hear the developments. Plus I can read it while on the move (impossible with the printed version :-)

32
by HoleInZone :: Tue, 06/23/2009 - 4:25pm

Seconded. Hopefully the annual effort to get my hands on a hard copy of PFP or equivalent in good pre fantasy draft time wont be even more difficult this year.

Great job though, obviously.

4
by td (not verified) :: Tue, 06/23/2009 - 1:01pm

I've avoided certain things that I would have otherwise purchased because of not wanted to deal with Paypal. Is that an option here?

5
by qed :: Tue, 06/23/2009 - 1:12pm

I'd like to second the request for a non-PayPal option.

23
by sethburn :: Tue, 06/23/2009 - 2:49pm

I also would appreciate a non-PayPal option. Also, do we know when the dead tree option will become available?

7
by JasonK :: Tue, 06/23/2009 - 1:15pm

Hooray! Real FO content! (Twitpolls are a weak substitute.) And on my favorite team, to boot!

So, there is a new QB "accuracy index." And the Giants are your NFCE favorites, so perhaps the annual "DVOA always overrates the Eagles" arguments will die down a bit. Overall, it's a great read, and spot-on on most of the team's issues. My minor nits with the Giants excerpt are that 1) I suspect that you're a little too confident in Sintim getting the starting SAM job, 2) Rocky Bernard isn't going to be starting at DT absent injury, and 3) this is probably just a semantic thing, but a lot of the bit about the DL getting "tired" down the stretch has been attributed to injuries that weren't made public during the season.

100
by Kevin from Philly :: Mon, 06/29/2009 - 2:26pm

The second leg of the Eagles jinx has fallen! The first was obviously the decision of the Mayors office to never, ever again put the stupid team gear on top of Billy Penn's hat. Today, FO has declined to name them pre-season favorites. Now, if we could just get Ed Rendell to stop the friendly bets with the birds opponents, we'll finally get a parade in February!

9
by bubqr :: Tue, 06/23/2009 - 1:27pm

And ho, yeah : I'm a big FO fan, and I've been for years, but I don't care what stats say when the Giants play the Eagles without Burress : They are clearly not the same team.

28
by bravehoptoad :: Tue, 06/23/2009 - 3:34pm

Clearly.

30
by JasonK :: Tue, 06/23/2009 - 4:05pm

This has been a core FO caveat from the beginning: Matchups matter. An advantage in big-picture stats doesn't always transfer over to individual matchups in individual games. Hixon comes out better than Plaxico does on season-wide effectiveness statistics, but when each of them finds himself lined up across from Sheldon Brown, Plax gets open while Hixon struggles to get off the line cleanly.

10
by Mark S. (not verified) :: Tue, 06/23/2009 - 1:30pm

Really good read. Obviously as a Giants' fan the lack of depth for the O-line is the biggest concern, with the inexperience of the receiving corps being a close second. I don't suspect there will be much margin for error in the NCFE this year. Can't wait for the season, though!

11
by Art (not verified) :: Tue, 06/23/2009 - 1:31pm

On the second page of this article in the second column and second paragraph you state that the Giants held the Patriots to 13 points in the SB XLII win. Of course we know that the Patriots actually scored 14 points.

I would not have bothered bringing this up, but in all the years of reading your materials this is the first time you have given the Giants more credit than they deserve and I thought you would want to quickly address this error.

12
by Anonymous2 (not verified) :: Tue, 06/23/2009 - 1:35pm

Regarding the Giant's passing offense being better without Burress, I wonder how sensitive that conclusion is to that Seahawks game. The Seahawks had an outlier of a bad D last year. Had Burress been active for that game, how do you know the score and passing stats wouldn't have been just as high, or even higher? If you remove that game from the analysis, does it appear that the Giants are better off with Burress than without?

The article calls their 3-3 streak a "demise." The Giants finished 12-4, basically a .750 team at heart. Why would we think their losses would need to be be evenly paced throughout the year?

Also, the mean prediction for the Giants is 10.0 wins, but the forecast probabilities are heavily skewed in favor of the 11+ wins category (42%). The 8-and-below categories total only 27%. I don't see how that would be possible without the distribution being highly irregular.

I should stop here. I know this is just a draft.

14
by Conor :: Tue, 06/23/2009 - 1:39pm

Never mind, got it fixed.

13
by JasonK :: Tue, 06/23/2009 - 1:37pm

Typo patrol:

-- The Patriots scored 14 points in SB42, not 13. (Not 17, either.)

-- Something is off in a few of the Catch% stats. Both Steve Smith and Sinorice Moss seem to have been attributed the "C%" figures for the other, more famous NFL receivers referred in your tables as "S. Smith" and "S. Moss."

25
by Vince Verhei :: Tue, 06/23/2009 - 2:57pm

We'll fix that SB score in the final book.

15
by chemical burn :: Tue, 06/23/2009 - 1:48pm

I'm curious why the author of that chapter ignores the fact that FO's own numbers confirm the conventional wisdom about the Giants' offense when Plaxico is out: their rushing offense is significantly worse because teams don't need to commit safety help to stopping the pass. It's right there in table #1: 21.1% w/Plax and 11.3% w/o.

That is a huge change! The Giants' strength last year was their punishing running game that teams literally could not stop. They dominated and controlled games in the first half of the season. After Plax was gone, the running game fell off a cliff, they had to rely more on their passing game, they could control the clock less, their probably defenses played more actual minutes (and, hey, got more tired!) because as we all know a good running game keeps your defense off their field.

The analysis of this article is just wrong - it ignores the glaring fact that losing Plaxico really did matter! Either it mattered only SOME (the rushing attack declines slightly and the defense is worse even if you include the Seattle game) or it mattered A LOT (if you don't include the Seattle blowout). And it's dubious to include the Seattle blowout because a defense (especially one as hapless as Seattle) wouldn't have as much of chance to change their strategy based on one game... Late in the season, defense knew they could in on stopping the run because Plax was gone... and it worked. I can't believe anyone would argue otherwise...

18
by foos (not verified) :: Tue, 06/23/2009 - 2:28pm

They did rush for 301 yards without Plax against the Panthers in week 15.

20
by Mark S. (not verified) :: Tue, 06/23/2009 - 2:37pm

Brandon Jacobs also missed the Week 12, 15, and 17 games, along with the second half of the Week 14 game (after having a very good first half vs. PHI). I'd imagine that put a ding in the rushing performance, as well.

Which isn't to say "chemical burn" doesn't have a point. I do think the numbers are underselling how much the G-Men missed Plax. I just don't think it was game-breaking as some people think.

21
by Independent George :: Tue, 06/23/2009 - 2:49pm

The Running game declined some without the threat of Burress stretching the field to concern opposing safeties, but it wasn't enough to drag the offense down; the team's offensive DVOA was 20.6% with Burress in the lineup and 27.8% over the eight games he wasn't around.

That doesn't sound like they ignored anything.

Incidentally, I agree with you and the conventional wisdom - my own lying eyes told me the offense wasn't the same without Plax, but as was also cited in the article, they faced a tougher slate of defenses. The point is, my own lying eyes also told me their defense went from suffocating to slightly above average, which is the main point of the article.

26
by chemical burn :: Tue, 06/23/2009 - 3:03pm

Maybe ignored is too strong, but certainly the analysis of the numbers is all wrong: the rushing game was far worse and the defense was worse after Plax was gone for the season. The weakness in the running attack has frequently been attributed to Plax being gone. I also think having a weaker running attack plays into the author's own point that the thin defense seemed to start wearing down as the season went on - a great rushing attack is a defense's best friend.

To argue that the offense "didn't miss a beat" with Plax gone is just wrong: in that timeframe, their rushing attack fell off a cliff. With Plax gone teams could start keying in on the run at the expense of playing the pass, opposing defenses stopped the run - does anyone actually think the numbers (and our lying eyes) don't say that's exactly what happened?

The Giants passing attack improved, so I'm not arguing the Giants are helpless - they're just a different offense. But it makes no sense to think that their running game will be as dominant (or even "really good") this year without Plax in the line-up - the numbers clearly say otherwise... Plax had a measurable positive effect on the Giants' offense...

27
by Pat (filler) (not verified) :: Tue, 06/23/2009 - 3:20pm

because as we all know a good running game keeps your defense off their field.

God, please let this stupid myth die. Please. The only thing a good running game does is allow you to close out games quickly in the second half once you've built a lead from actually passing the ball.

It does not "keep your defense off their field." You know what does that? Keeping the ball. You know what else helps keep the defense off the field? Scoring. The amount of actual time that passes between scoring a TD and when the defense takes the field is longer than between when a punt occurs and when the defense takes the field.

As to the rest of your conclusions, you're not actually reading the article. The Giants offense was better during the weeks when Plaxico was out. The decline in the running game was more than compensated by the improvement in the passing game. If opposing safeties were playing in the box more without Plaxico... well, the Giants should hope that happens, considering you win more games with a strong passing game, which the Giants had.

31
by tuluse :: Tue, 06/23/2009 - 4:19pm

That's what the paragraph says, but the table disagrees.

It has Off Pass w/ Plaxico 29.5% Run Off w/ Plaxico 21.1%

Off Pass w/o Plaxico SEA excluded 29.9%, Run Off 11.3%.

34
by chemical burn :: Tue, 06/23/2009 - 4:34pm

In the first half of the season, the Giants controlled time of possession because they ran the ball more effectively. If they have the ball more, their defense is on the field less. How is that stupid myth?

Now the conventional extrapolations from that starting point might be mythical, but if the author is making the point that the defense seemed to mysteriously start wearing down in the second half of the season, doesn't it matter that the Giants weren't controlling the time of possession (because they weren't running as effectively?)

36
by chemical burn :: Tue, 06/23/2009 - 4:41pm

Now, obviously a strong passing game is better (and leads to more wins in the modern game), but the stronger passing attack coincided with a weaker rushing attack and a weaker defense.

If you don't accept my premise that the weaker rushing attack and defense AREN'T compensated for by the improved passing attack (and the main argument in favor of that idea is that they, you know, won fewer games), that's fine. But my point is that losing Plaxico affected the Giants in their running game and defense.

Please explain why you think those units suddenly got much, much weaker once Plaxico was out. Somebody mentioned that Jacobs was hurt and playing at less than 100% in that timeframe - that seems very possible to me. Or maybe defense depth really was the problem and replacing sub-par backups with mediocre backups will right the defense decline.

But the stats don't support the analysis that Barnwell is putting out there - in this case, they support the conventional wisdom.

47
by Pat (filler) (not verified) :: Tue, 06/23/2009 - 6:17pm

"If you don't accept my premise that the weaker rushing attack and defense"

I don't accept your premise that Burress could have anything significant to do with the defense. He didn't take the offense from "good -> bad." He didn't even take the rushing offense from "good->bad." The total offense didn't change. The rushing offense went from "very good->good."

There's no way that has a significant effect on the defense. The idea that a strong rushing offense leads to a fresher defense is just a total and complete myth.

It gets perpetuated because several of the strong defenses recently have been run-heavy offenses: Chicago, Pittsburgh, and Baltimore come to mind. So you get this dopey myth that a great running game leads to a fresh and furious defense.

But that's just recent coincidence. It's not cause-and-effect; there have been plenty of counterexamples. Tampa Bay in 2002 is a good one - their running game was awful, averaging under 4 yards/carry. They passed way more than they ran (58% pass/42% run) and still had a very good defense.

90
by DeltaWhiskey :: Fri, 06/26/2009 - 7:27am

So we're to accept your n=1 study refuting the claim, rather than your n=3 example?

How about instead the correlation between Defensive DVOA and Rushing DVOA since 2002

r=.07 (ns).

92
by Jimmy :: Fri, 06/26/2009 - 8:32am

OK I am fairly statistically illiterate, isn't a correlation of .7 pretty high?

Have you run the numbers to get that figure?

Of course correlation doesn't indicate causation but that would be a pretty strong relationship of some description.

93
by DeltaWhiskey :: Fri, 06/26/2009 - 9:18am

.7 very high, .07, not so high (Sorry, should have had a leading zero - like below).

I did run the correlation (r=0.07), and the results suggested there was no relationship between Rushing DVOA and Defensive DVOA.

94
by Eddo :: Fri, 06/26/2009 - 10:40am

"...the results suggested there was no relationship between Rushing DVOA and Defensive DVOA."

Um, isn't that exactly what Pat was saying?

96
by DeltaWhiskey :: Sat, 06/27/2009 - 2:13am

Yes, but he provides three arguments that fail to support his argument and only one that does.

98
by Pat (filler) (not verified) :: Sat, 06/27/2009 - 6:43am

Correction: I only provided one counterexample to three examples. I provided quite a few arguments. But I left a correlation test as an exercise for the reader.

It should also be noted that with only 3 data points in support of a correlation, even 1 against the correlation significantly weakens it.

99
by DeltaWhiskey :: Mon, 06/29/2009 - 2:25am

But even those were poor examples, as they don't fit the myth either. Since 2002,
BAL's avg DVOA rank for rushing was 19, for defense 4. For CHI, rushing was 22, defense was 10. For PIT, 17 and 6 respectively, and TB 22 and 8.

Across this time period, these four teams have had, on average, respectable defenses and average to below average running games.

101
by DeltaWhiskey :: Thu, 07/02/2009 - 6:17am

One more observation.

While overall, there is no relationship between the running game and defensive performance from the period of the 2002 to 2008 season, it occurred to me that perhaps at the edges there might be a relationship, so I ran 8 more sets of correlations on a sub-sample of the original sample. Each sample had an n=35. Thirty-five was the magic number due to the following – first, I looked at the relationship between the running game and defensive performance for each year’s top and bottom five teams in each category (rushing DVOA and defensive DVOA). I then decided that perhaps taking the bottom and top 35 teams in each category made more sense. After completing the analysis, only one correlation reached the p<0.05 level of significance – for the top 35 teams, rushing effectiveness as measured by rush DVOA correlates with defensive DVOA (r = .34, p < 0.05). Note that the correlation is positive, which indicates that higher rushing efficiency is associated with poorer defensive efficiency. That is, it appears that the teams that run the ball better tend to play poorer defense.

NOTE: I fully admit and accept that these findings may represent a Type I error.

95
by Jimmy :: Fri, 06/26/2009 - 1:30pm

OK so I fail reading comprehension, my bad.

38
by Mark S. (not verified) :: Tue, 06/23/2009 - 4:48pm

I think what he's saying is that when you look at stats that correlate most closely to a successful team and actual wins (like something found here: http://www.twominutewarning.com/correlations.htm) that points per drive and success per play run are more predictive of a winning effort than time of possession. So while it's great to do both, it's better to have an efficient offense than to control the clock.

43
by Pat (filler) (not verified) :: Tue, 06/23/2009 - 5:54pm

"If they have the ball more, their defense is on the field less. How is that stupid myth?"

Because it's not true?

Suppose one team has a time-of-possession of 35 minutes, and the other has a time-of-possession of 25 minutes. Does that mean that the defense was on the field for 35 minutes for one team, and 25 minutes for the other team?

My God, no! You have no idea how much of that time was between downs, and you have no idea how much time went on when the clock stopped. Time-of-possession means nothing. If one team runs 10 plays in 3 minutes of game clock, and one team runs 5 plays in 5 minutes, which defense is going to be more tired?

It's a stupid myth. Please, let it die. Defenses do not get tired from "being on the field more." They get tired by playing more plays.

46
by chemical burn :: Tue, 06/23/2009 - 6:13pm

Ok - but didn't the Giants also have a dominant relative plays per drive in the first half of the season last year? Didn't FO have a whole thing about how they would conduct long, rushing attack-based drives with many, many plays per drive?

And also doesn't it stand to reason that a good average run is going to be shorter than a good average pass, so running-based drives are normally going to have more plays than passing-centric drives? That conforms with your ideas that passing attacks are more efficient methods of scoring and also that time of possession on the game clock has no real relationship to actual, real-time of possession?

But really, forget the defense - you said "the stats say Giants offense was better in the second half of last year, end of story" and I said "the overall offense might have been better, but the rushing attack was markedly worse and the defense was markedly worse - what's the reason?" What DO YOU think the reasons for the change are?

49
by Pat (filler) (not verified) :: Tue, 06/23/2009 - 6:33pm

And also doesn't it stand to reason that a good average run is going to be shorter than a good average pass, so running-based drives are normally going to have more plays than passing-centric drives?

Yes, but those plays are also going to lead to a shorter game, which means fewer overall plays. The effect on "number of plays/drive" is small compared to the effect on "game clock used/play," which is large.

What DO YOU think the reasons for the change are?

Rushing offense:
1) Jacobs's health.
2) Reduced favorable opportunities.
3) Small-sample size. 18% vs 22%, 11% vs 22%, they're all basically identical.

Defense:
1) No Osi Umenyiora.
2) No Michael Strahan.
3) No Osi Umenyiora or Michael Strahan.

Did I mention no Osi or Strahan? Tuck and Kiwanuka always had Pro Bowl and Hall of Fame players to help reduce the number of double-teams they saw. It's not exactly surprising that they started out pretty well, but faded as teams adjusted or they wore down as the season went on.

Kiwanuka admitted that the injury he suffered the first week of the season affected him for most of the season. Tuck struggled with bone spursin his feet late in the season, as well.

Neither Kiwi nor Tuck started an entire season before. To attribute their decline (along with the rest of a defense whose defensive line is the best component) to Burress is, well, bizarre. It's not exactly surprising that they struggled late in the season as they had more plays than they've ever had before.

53
by Pat (filler) (not verified) :: Tue, 06/23/2009 - 7:01pm

Sorry, to clarify the whole "rushing offense leads to better defense" debunking:

The number of plays/drive thing doesn't matter that much. It's a small effect, especially since rushing plays tend to be shorter than passing plays (so passing plays are more tiring, which any defensive player will tell you is true).

The number of plays/game matters more, but you don't reduce the number of plays/game unless you're actually winning. Which means that a rushing offense doesn't lead to a better, fresher, overall defense: it just leads to being able to close out a game better. No sane offensive coordinator is going to say "I've got to run a bunch on this drive, even though we're down 10 points."

But for closing out games, you don't need a "great" rushing offense. You just need a good one, and the Giants had one.

35
by Mark S. (not verified) :: Tue, 06/23/2009 - 4:37pm

But really why should we exclude one game (SEA) from an analysis? Is that game not count? Did nobody else get a boost in statistical performance from playing poor defenses in 2008? If you think it was an aberration, fine, but then you should also exclude the game where their offense performed its aberrant worst.

37
by chemical burn :: Tue, 06/23/2009 - 4:46pm

The reason I think that game should be discounted is that the "Plaxico matters" argument hinges on teams changing their game planning based on observing proven weaknesses in the offense scheme. If Seattle had a even competent defense last year, I wouldn't want to discount it but it's not unreasonable to think they weren't great at game-planning last year... It's the same reason I don't read too much into the early Browns loss - one game isn't a trend. Half a season of consistent results is...

39
by Mark S. (not verified) :: Tue, 06/23/2009 - 4:59pm

But surely you understand the argument that if you want to discount aberrant games in the search fro "true level", you should discount an equal number of aberrant games on both sides?

"If Seattle had a even competent defense last year, I wouldn't want to discount it but it's not unreasonable to think they weren't great at game-planning last year"

Maybe they were...or maybe they were great gameplanners and just weren't that talented, or were a particularly bad matchup for a power-running team like the Giants. I am sure there were plenty of teams that had great gameplanning last year, and plenty of teams that had poor ones. There are also good matchups and bad ones. There's nothing to say that wouldn't show up and balance out over time in the numbers though. But it doesn't ultimately make sense to cherry pick which performances you do and don't want to include though.

40
by tuluse :: Tue, 06/23/2009 - 5:10pm

The talking heads on ESPN say it takes 3-4 weeks of game footage to gameplan properly, so we should probably only count the games after Plaxico had missed 3 weeks. We're heading towards very small sample sizes though.

41
by chemical burn :: Tue, 06/23/2009 - 5:10pm

No, you're absolutely right and maybe I am cherry-picking. But I do think there's a difference between Plaxico being out for one game and the defensive coaches of Seattle being able at best to say "ok, don't worry about the pass and just concentrate on the run - maybe that will be a good idea, maybe not" and after Plax was totally gone defenses being able to REALLY say, "forget about the pass, if we just concentrate on the run, we can definitely stop it. don't worry about their wide-outs and matching them up with safety help."

42
by chemical burn :: Tue, 06/23/2009 - 5:12pm

Honestly, though, everybody - why was the rushing offense so much worse in the second half of the season if Plaxico isn't a big part of the reason?

45
by Pat (filler) (not verified) :: Tue, 06/23/2009 - 6:06pm

1) It's possible it was just opportunity cost.

You have more favorable opportunities to run (especially with the Giants, where Jacobs is an efficient back) when you've got the lead and dominating your opponent, which the Giants weren't.

Just as an example, if a team consistently gets to 3rd and short, they can run on that down, and if they've got a strong offensive line, they'll do well on that down. If a team consistently gets to 3rd and long, they can't run on that down, and so they never will get a chance to actually exercise a strength of the team. (This is just one example, which is a good offense vs. bad offense, but you can make the same argument with a team that's playing from behind late in the game vs. a team with the lead late in the game).

DVOA controls for this somewhat (by equalizing opportunities that you have), but you simply can't magically correct for opportunities a team did not have. That's why it's safer to look at total offensive DVOA, since a team always has an opportunity to make a play, even if you don't always have a viable opportunity to run.

2) "11.8%" vs "22.1%" is not "so much worse." Both are "good to very good."

48
by chemical burn :: Tue, 06/23/2009 - 6:21pm

I'm not saying they became bad. I'm say they became noticeably worse in an area of strength. And their rushing attack changed more than their defense. Again, none of this works in favor of Barnwell's analysis...

I like the idea that opportunity was a factor - is their any evidence of this being the reason? It seems unlikely that being really good at rushing in second and long situations and just not getting enough of them could send suppress their DVOA that much, especially since DVOA corrects for it...

50
by Pat (filler) (not verified) :: Tue, 06/23/2009 - 6:43pm

"And their rushing attack changed more than their defense."

No way.

If you include Seattle, it was a change of -2.4% in rushing DVOA, and a change in defensive DVOA of 6.9%. If you don't include Seattle (which, again, you have no justification for doing, as it's completely arbitrary - why not toss the St. Louis game, then, too? Same blowout level win), the change is baaaaarely more: -9.8% rushing DVOA change vs 9.1% defensive DVOA change.

Except: defensive DVOA is all plays on defense. Rushing DVOA is less than half the plays on offense. In terms of "total impact to the team," the defensive change was easily twice as important as the offensive rushing decline.

56
by chemical burn :: Tue, 06/23/2009 - 8:24pm

I'm not going to argue this too much more because I'm fighting a losing battle - even if losing Plaxico did change the Giants for the worse, it did so in a very indirect manner. I'm arguing that it affected the strategies of the opposing defenses and that's purely speculative (or rather, based only on what my eyes told me when I watched those games). The Giants running attack got weaker, but you say it doesn't matter, overall. Fair enough, I guess.

My secondary argument is even more speculative (and truthfully, my heart's not in it), that their defense was hurt by their running game weakening (you know, like a cascading injury hurts more than just the initial point of injury - come on, it makes a little bit of sense!)

For the record, I don't buy any of the reasons put forth for the defensive decline (your reasons of losing Osi and Strahan are weak since the decline happened mid-season and those injuries happened before the season). And the secondary reason of "they had to play their back-ups more and more as the season went on" or "their main guys got tuckered out" are as purely speculative and flimsy as anything I've said about Plax.

There's an decline in the Giants running game without Plaxico, even including the Seattle blow-out. And again, the only reason not to include the Seattle blow-out is based on speculation (or what everybody claims to have seen) that the opposing defenses changed their strategies.

Anyway, It's funny, too, because I've been reading this site for years and I've never really posted, mainly because of you, Pat - I mean that in a nice way. I always felt like I didn't need to post because you were saying what I was thinking, especially because you're a non-hysterical/mean-spirited Eagles fan... oh, how the tables have turned...

78
by Dales :: Wed, 06/24/2009 - 10:08pm

I think that teams that are constructed well and led well tend to have, at GM and coach, rational actors.

I believe the Giants have won enough lately and have demonstrated a talent level that suggests they are such a team.

As such, I think it makes sense to look at what the rational actors who are running the Giants did this offseason-- first and foremost, brought in a ton of defensive line depth.

To me, this strongly suggests that the people with the most direct knowledge of what was happening, and who have the most vested interest in success for the team, and who have proven themselves to be capable at correctly analyzing things, felt that the defensive line was a primary reason for the Giants' performance towards the end of the year.

Also obviously, by spending two more draft picks on wideouts, they see that as an area of concern as well. But with them passing on free agents, it seems they think it is more of a long-term matter than a pressing need. Meanwhile, they loaded up on the DL right away.

57
by chemical burn :: Tue, 06/23/2009 - 8:37pm

I also feel like you're arguing it two ways at once: defenses either get tired out and it affectes their performance, or they don't. You're arguing that the defense got tired out while at the same time arguing nothing really changed for the team late in the season - they got just better on offense!

The whole Kiwi/Tuck never played a full season is in line with the idea that the Giants defense wore down as the rushing game weakened. In-game exhaustion seems like as plausible a factor as season-long exhaustion (to my mind, more plausible - Tuck and Kiwi weren't rookies and had been conditioning for pro season for a while. also, there's 9 other guys on that defense that were perfectly used to playing for a full season). If depth was factor, it should have been showing itself for the whole season and the obvious decline should have been apparent (and a similar factor to DVOA) whenever the back-ups were in.

Maybe the good running game helps a defense is largley a myth, but surely you're not so obstinate as to insist a good running game could never conceivably help a defense? Either these guys were tired or the somehow experienced an unexpected decline in skill.

If you're going to argue injury, again, that should've manifested itself sometime earlier in the season. I guess I can buy it's just a coincidence that this all happened with the Plax drama...

74
by Pat (filler) (not verified) :: Wed, 06/24/2009 - 3:25pm

I'm arguing that lack of a running game does not lead to a defense tiring during a game.

I am not arguing that a defense with poor depth cannot "wear down" (due to injuries) during a season.

The whole Kiwi/Tuck never played a full season is in line with the idea that the Giants defense wore down as the rushing game weakened

No, it isn't! The idea that Tuck developed bone spurs in his feet due to the Giants not running the ball is, well, bizarre.

If depth was factor, it should have been showing itself for the whole season

No, it wouldn't. It'd show primarily at the end of the season. Injuries are progressive - teams start the season mostly healthy, end the season mostly injured. Backups aren't that important early on, and are very important (especially since some of them start) by the end of the season. This is not really controversial.

but surely you're not so obstinate as to insist a good running game could never conceivably help a defense?

The only way a good running game helps a defense is because it's good offense, and even then, good offenses don't make defenses better. They make them less necessary. When an offense with the lead keeps the ball for the last 5 minutes of a game, you don't say "wow, great defense" and it certainly doesn't help the defense's stats (in a pace-free stat).

If you're going to argue injury, again, that should've manifested itself sometime earlier in the season

I have no idea where this logic comes from, honestly. The loss of Osi and Strahan seriously cut the Giants defensive line depth. As the season went on, Kiwi and Tuck both picked up injuries. This meant that the Giants had to start two injured DL, whereas if they had Osi and Strahan, they would've just leaned on those guys a bit more.

Shallow depth doesn't show up early in the season. It shows up late in the season, as any Redskin fan can attest to.

44
by Pat (filler) (not verified) :: Tue, 06/23/2009 - 5:59pm

The article include the Seattle game in the "better without than with." Even if you exclude the Seattle game (for which you have zero justification to do so) the offense isn't going to be that much worse; "total" DVOA isn't "run+pass/2". It's weighted by play, and they passed more than they ran.

The "excluding Seattle" statistics are mainly there to show that the Seattle game isn't biasing the statistics a ton. It shouldn't actually be in the table, because there is no justification for arbitrarily excluding the best game in one sample and not the other, and it leads to posts like this, where people say "OMG, look, if you exclude the games where NYG's offense did well without Plaxico, they did crap without Plaxico!"

62
by Anonymous2 (not verified) :: Tue, 06/23/2009 - 10:56pm

Pat-The suggestion was to exclude the Seattle game merely to test the sensitivity of the statistical inference to a single game. This is a common method in statistics of gauging the robustness of the conclusion you're trying to make.

If a single game completely reverses your inference, then it's likely your data isn't strong enough to make any inference at all. There are only 16 games in a season. Who's to say that if they played 1 more game with Plaxico, he wouldn't explode for 150 yds and 2 TDs? Or if they played 1 more game without him they wouldn't stumble miserably. There's probably too much sample error involved to make any conclusion.

The suggestion was not to pick a game to exclude in order to bolster one point of view or the other.

75
by Pat (filler) (not verified) :: Wed, 06/24/2009 - 3:28pm

Pat-The suggestion was to exclude the Seattle game merely to test the sensitivity of the statistical inference to a single game. This is a common method in statistics of gauging the robustness of the conclusion you're trying to make.

Yes, you're right. It's called jackknifing, and is a resampling method akin to bootstrapping.

But you do not use the jackknifed set other than to test the robustness. You have no justification to. You certainly do not use it as the baseline measurement, which, of course, is the unaltered set.

16
by drobviousso :: Tue, 06/23/2009 - 1:51pm

That's a great price for the print to order. If there's an option for a dead tree book and kubiak, I'll be taking it.

22
by Jon :: Tue, 06/23/2009 - 2:49pm

Nitpicking: Chase Blackburn started over Gerris Wilkinson for most of last year. There's been talk about Robbins and Cofield missing time with injuries, so while Bernard and Canty will help, the Giants may not be able to just sub in fresh linemen at will.

29
by moses74 (not verified) :: Tue, 06/23/2009 - 3:48pm

love it love it love it

Now I have that out of the way, I wanted to say that I was interested in the comment about the change in Tom Coughlin's aggressiveness index value, and wondered if there was any data about how that figure changed during the average career, if Coughlin was atypical, or if it is standard trend. I'd love a graph. I'd even write the article if you gave me the numbers!

33
by tuluse :: Tue, 06/23/2009 - 4:27pm

I have a couple of minor problems with the names for stats.

1) Success Rate for defensive players. This should be stop rate, or you should take the inverse and call it success allowed rate. Every other defensive stat lower is better, but not this one.

2) Aggressiveness Index. My problem isn't with the name on this one, but the abbreviation, AI is stupid. Can't it be AGG or AGI or something like that?

52
by Theo :: Tue, 06/23/2009 - 6:55pm

Darn good comments.

Oh. Looking forward on buying the dead tree version over here on the other side of the pond.
The others made my train rides to work a lot better.

72
by Temo :: Wed, 06/24/2009 - 10:18am

Sorry, Madden has conditioned me to equate AGI with Agility.

Though it would be interesting to see what kind of time Tom Coughlin could put up in the shuttle.

55
by GsK (not verified) :: Tue, 06/23/2009 - 8:03pm

Maybe I missed it, but do you have an exact release date? When can I a) download the PDF and b) have the book shipped? How about a size (page count) and how big a file?

Can't wait and need more info!

58
by dsouten :: Tue, 06/23/2009 - 9:19pm

One of the few complaints I've had about PFP, and again about this Almanac sample, is that the blurbs for the skill positions players are consolidated and listed alphabetically rather than included in the team chapters. I always end up reading to the end of a team chapter, and then having to flip through the QB, RB, WR, and TE sections to get the "complete story" on that team. The only reason to organize them by position would be if this were a fantasy football guide, but this is a REALITY football guide and the teams should be the focus. I am an avid fantasy player and all that - but if people want fantasy guidance from FO they can download the sortable KUBIAK excels, or look up the KUBIAK projections in each team's chapter.

60
by blorrrbgpglpgl (not verified) :: Tue, 06/23/2009 - 9:58pm

Seriously? You guys really think Barnwell's content is what you want to use to lure people in? This is NOT something I would use to 'spread the word.' Put up one of Tanier's chapters - that's something I wouldn't be actively embarassed to link a friend to.

63
by Big Johnson (not verified) :: Wed, 06/24/2009 - 1:14am

Barnwell I think you did a great job on the Giants page. The analysis is fresh and doesn't just concede to the idea that well known players HAVE to make a team better. I am sick of every article saying how every player is a good player. Some players stink but have gotten more than their fair share of opportunities in great situations. AND IF 1 MORE PERSON SAYS MATT FORTE IS A GREAT PLAYER I MIGHT TAKE A SHIT ON THEIR PORCH

66
by tuluse :: Wed, 06/24/2009 - 3:05am

OK, Matt Forte isn't great.

He is very good though.

69
by Pouringlizards (not verified) :: Wed, 06/24/2009 - 6:57am

Hope you've got some rubber gloves, a good scrubbing brush and some bleach...

73
by Big Johnson (not verified) :: Wed, 06/24/2009 - 3:01pm

lol i just added them to the grocery list

64
by Key19 :: Wed, 06/24/2009 - 1:27am

Really good stuff. Looking forward to the Almanac!

65
by Dan :: Wed, 06/24/2009 - 2:15am

I'm surprised not to see any mention of fourth round RB Andre Brown - holder of this draft's highest speed score - in the discussion of who will replace Derrick Ward. Didn't you write the article on speed score, Bill?

70
by Jimmy :: Wed, 06/24/2009 - 9:37am

As far as the whole with Burress / without Burress thing goes I have a thought.

When PB plays the safeties and linebackers can't crowd the line as much as Manning will simply throw the ball to PB, this makes the running game more effective.

When PB doesn't play the safeties and LBers can crowd the line to stop the run. This reduces the effectiveness of the running game but leaves too few defenders in pass coverage to prevent Manning completing passes to his WRs and causes the passing game to pick up the slack for the running game.

If it is the case that Manning is a very good QB, the extra space created for the passing game by removing Burress (and thus having defenders creep up on the line) could actually improve the passing game enough to compensate for any loss in the running game so the overall offensive DVOA could be unaffected (or even improve) despite Hixon not being as good a player as Burress.

So the Giants could be able to produce at a similar level on offense without Burress. The problem could be in decreased flexibilty on offense (ie fewer ways to beat the defense) which could allow strong defenses to create a gameplan that can stop them.

76
by Dan :: Wed, 06/24/2009 - 5:03pm

It sounds plausible, but the Giants offense did about as well against top defenses without Burress as they did with him. They played 6 games against top 4 defenses (DVOA < -18%), 3 with Plax and 3 without him. With Plax, against Baltimore, Philly, and Pittsburgh, they had offensive DVOAs of 37%, 29%, and -1% (avg = 22%). Without Plax, against Minnesota, Philly, and Philly again in the playoffs, they had offensive DVOAs of 43%, 15%, and -2% (avg = 19%).

84
by Jimmy :: Thu, 06/25/2009 - 10:01am

I see your point but there probably aren't enough games to say (definitively) one way or the other - 3% doesn't seem much.

I would say that if you just look at the games against Philly then going from 29% with Burress to 15% without Plaxico and then -2% in the playoffs it does start to look like a sharp decline. Part of my theory is that whilst the Giants still have a good offense without Burress they have lost a very good weapon and it could well make them more predictable. The fewer ways a defense has to worry about getting beaten the more emphasis they can place on neutralising the weapons the offense has. It seems that Jim Johnson might have found that he could effectively defend the Giants without Burress by the playoff game. Of course that could just be one game and maybe should be dismissed until further evidence presents itself (damn you season, hurry up).

77
by Bill Barnwell :: Wed, 06/24/2009 - 6:20pm

The conventional wisdom was that the Giants' offense suffered without Burress. The running offense DID suffer without Burress, but the passing attack more than made up for it. The small sample arguments I've read on boards discussing the chapter is nonsense; if the numbers confirmed the conventional wisdom, no one would be saying "Oh, well, maybe the Giants would have had a better offense if they'd played three more games without Burress."

79
by D :: Wed, 06/24/2009 - 11:16pm

Quick question. Have you changed the DVOA formula? The numbers in the book don't match those on the site.

91
by Vince Verhei :: Fri, 06/26/2009 - 7:37am

Quick answer: Yes.

80
by Jay Paradise (not verified) :: Thu, 06/25/2009 - 12:01am

Looks like you have an incorrect calculation:

In the defensive secondary section, you say "Webster led the league by breaking up an impressive 32.1 percent of the passes thrown at him".

Looking at the chart, it shows 25 pass deflections on 69 targets. That is 36.2%. Am I missing something?

The only other suggestion I would make is to chose your top 5 best stats for each position and just stick to those for you book charts. There is way to much clutter that is hard to follow on a printout. Using lots of categories online is fine because you can link easily to their descriptions, but in printable form it is difficult to comprehend.

81
by Anon-FO (not verified) :: Thu, 06/25/2009 - 2:20am

Since i moved out the US, can i pay you in sheeps, goat and cheese?

82
by KevinM (not verified) :: Thu, 06/25/2009 - 3:44am

While I love statistical analysis of FO, one major thing is overlooked. Burress routinely murdered the Philadelphia Eagles and to a lesser extent, the Dallas Cowboys. In Burress' last 6 games versus the Eagles, he scored at least 1 TD in 5 of them (2 in the playoff meeting). The Giants had quite a bit of success running the ball on the Eagles during that stretch. Once Burress was gone, the Eagles did a much better job slowing the Giants running game... correlation or causation?

I still feel that the Giants would've made the Super Bowl without Plax if they didn't have to play the Eagles. Those 2 teams were so even, that to take away such an important player was too much to overcome.

85
by Pat (filler) (not verified) :: Thu, 06/25/2009 - 3:25pm

Actually, you're overlooking one important thing: Burress wasn't all that good last year.

The last time Burress "murdered" the Eagles was in 2007, not 2008. In 2007, in the one game he played in, he had a grand total of 1 catch for 17 yards and a TD. That play came when the Giants accurately read a corner blitz and had Burress matched up on Dawkins, a serious mismatch. Other than that, he had exactly zero catches.

But look more generally at Burress last year. He had exactly 1 100-yard receiving game: in week 1. He only came within 20 yards of 100 yards in one other game: week 2. Burress was a non-factor last year even when he was on the team. Everyone seems to constantly be saying "Burress is too much of a weapon not to have an effect!" but, well, he wasn't.

Of course, you can say "well, that's because teams were double teaming him!" and that might be true - but when you can double team a guy and hold him to 1 catch all game? That's one heck of an effective double team.

Once Burress was gone, the Eagles did a much better job slowing the Giants running game... correlation or causation?

Coincidence. Brandon Jacobs was injured mid-game in the second game. Until that point, he was rushing for 5.3 yards/carry, so the Eagles weren't really "slowing" the Giants running game at all. When Jacobs came back in the playoffs, he rushed for nearly 5 yards a carry again against the Eagles, and the Giants put up almost 150 yards rushing, so again: not really "slowing."

I have no idea why people are ascribing so much to Burress last year. Look at what he actually did. He wasn't that good a receiver last year.

Those 2 teams were so even, that to take away such an important player was too much to overcome.

Plax played in 10 games last year and put up 454 yards. That's 45 yards/game average. If he's so important... why did he do pretty much nothing for them?

83
by Art (not verified) :: Thu, 06/25/2009 - 8:04am

In the second paragraph you state that the Giants went 3-3 after the 10-1 start including the playoff loss. They actually went 2-4.

86
by KevinM (not verified) :: Thu, 06/25/2009 - 4:51pm

"When he's not there, a huge, huge part of their offense is taken away, the ability for him and (quarterback) Eli (Manning) to be on the same page and sometimes playing playground ball to the weak side," Dawkins said. "As a defense, you didn't have to be that concerned about rolling to one guy or another." (NY Daily News 1-11-09)

"Plaxico always seemed to make the big plays in the red zone," said Eagles defensive coordinator Jim Johnson on a day when the Giants failed to score a touchdown. "That makes a big difference." (NY Daily News 1-11-09)

Do you know better than those guys Pat?

The Giants ran for 219 yards in their first meeting against the Eagles in 2008. They got 226 combined in the next 2 games. Some Giants fans sloughagree with you that Plaxico simply wasn't that good last year. I think he had an effect on the game even without big numbers. That wouldn't have mattered if the Giants faced the Panthers, Vikings, Cardinals, or Falcons. They could've beaten those teams. Beating the Eagles without Plax was too much to ask.

87
by Pat (filler) (not verified) :: Thu, 06/25/2009 - 6:05pm

I know that they're talking about Burress's career, and not Burress in 2008. If you take away "Burress the average player over his career" and don't have to plan for him, things change a bunch. But they weren't going to face that guy, period. He wasn't on the team last year.

I think he had an effect on the game even without big numbers.

It's not that he didn't put up "big numbers." He barely put up any numbers. I'm sure he had an effect on the game, but I'm also sure that other players had much bigger effects.

The Giants ran for 219 yards in their first meeting against the Eagles in 2008. They got 226 combined in the next 2 games.

And losing their starting running back in the middle of the game in the second meeting had nothing to do with that, huh? They ran for nearly 150 yards in the playoff meeting. The problem was the second game, and that was due to Jacobs being injured during the freaking game.

I'm sorry, but when people claim that magic Plaxico being missing explains a decline in a running game when the starting running back gets injured makes absolutely no sense whatsoever. Jacobs got injured, running game declines. This is not rocket science.

88
by speedegg :: Thu, 06/25/2009 - 7:39pm

Whoa! Great stuff. Is FO taking pre-orders for the deluxe package? Worth it's weight in gold.

As for the Plaxico impact on running backs, people pull out the TiVo of the State Farm NFL Match-up. They broke down how the Giants' offense worked with and without Plaxico and how defenses played them. Plaxico didn't need to score, his presence forced the safeties to play off him instead of in the box. This opened up lanes for the running back(s) to gouge the defense for yards. But, maybe the game chartering would show something different...

89
by KevinM (not verified) :: Thu, 06/25/2009 - 11:04pm

What's Derrick Ward, a practice squad player or the second ranked DVOA RB in 2008? He didn't do anything against the Eagles or Cowboys after Plax shot himself, yet he carved up the Panthers and played very well against Minnesota. Once again, I'm back to my point about Burress being more important against certain teams.

Explain to me how Eli had such wonderful performances against excellent defenses with Burress not doing much (Pittsburgh for example), then was incapable of doing anything once Burress was gone? It seems clear to me that Burress' loss changed how teams defended the Giants.

97
by Pat (filler) (not verified) :: Sat, 06/27/2009 - 6:36am

He didn't do anything against the Eagles or Cowboys after Plax shot himself,

Why is everyone paying so much freaking attention to Plax? Jacobs missed both of those games. I'm waiting for someone to start insisting that Burress was the reason that Jacobs got injured.

Brandon Jacobs must be the most unappreciated guy in the league suddenly. He had the #2 DYAR in the league last year and rushed for over 1000 yards despite being injured later in the year. And instead of his loss being the reason that the running game struggled, instead, it's the guy who put up a DVOA of less than 5% and had less than 50 yards receiving/game.

Explain to me how Eli had such wonderful performances against excellent defenses with Burress not doing much (Pittsburgh for example)

WTF? Manning put up a "wonderful" performance against the Steelers? 19/32, 199 yards, 1 TD, no ints is "wonderful"? He put up an only slightly-worse performance with the Eagles in the second game - slightly lower yards/completion, slightly lower completion percentage, but not tremendously worse. And his game against the Redskins, Panthers, etc. were all much better.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.

More information about formatting options

CAPTCHA
To skip this, please log in.