Writers of Pro Football Prospectus 2008

Most Recent FO Features

FitzgeraldLar04.jpg

» Scramble for the Ball: The DVOA Schism

Mike and Tom try to figure out what kind of secret sauce Arizona is feeding the media to sit at the top of the power rankings and in the middle of our DVOA rankings.

07 Sep 2006

2006 DVOA Projections

by Aaron Schatz

OK, folks, here they are: the 2006 DVOA projections. You'll find more discussion of these projections in the FOXSports.com preseason power rankings, which combine the DVOA projections and the summertime subjective power rankings that we did in June (50% strength for each).

(Here's the requisite link to an explanation of DVOA, which stands for Defense-adjusted Value Over Average and measures a team's performance on every play of the season compared to league average in the same situation, adjusted for opponent. I know a lot of people are coming here from various message boards and this is just going to look like a jumble of pointless numbers. Trust me, there is a method to the madness, and over the past six seasons DVOA has been a far more accurate predictor of future performance than wins or points.)

Offense, defense, and special teams DVOA are all projected separately using a system based on 2000-2005 numbers. The equations include a number of variables based on performance over the past two seasons in different splits (by down, passing vs. rushing, red zone vs. whole field) plus variables based on recent draft history, coaching experience, quarterback experience, and even weather.

Strength of schedule was then figured based on the average projected total DVOA of all 16 opponents for 2006 (yes, projected performance, not 2005 performance), and then wins were projected based on offense, defense, special teams, and strength of schedule.

There are no manual adjustments -- the numbers we are presenting here are exactly what the projection system spit out. A few of them will look strange to you. A few of them look strange to us. Two years ago, we thought the projection system was broken when it said San Diego would have a huge offensive season. Last year, we didn't understand why it said Tampa Bay would win the NFC South. There's a good chance that something that looks strange now will turn out to be correct by the end of the season. And some of them will end up wrong, because nobody's perfect. (Some readers have a problem with this concept, but what are you gonna do?)

The projections here are updated from Pro Football Prospectus 2006 based on some variables related to player experience, injury history (improved for Cincinnati and Miami), and roster movement (such as downgrading the Kansas City offense with Roaf retiring, upgrading the defense with Ty Law signing, and downgrading the Patriots for the Deion Branch situation).

Notice that the defensive projections are grouped together much closer than the offensive projections, so it's much better to be projected as the third highest offense than it is to be projected as the third highest defense.


TEAM
TOTAL
DVOA
TOTAL
RANK
OFFENSE
DVOA
OFF.
RANK
DEFENSE
DVOA
DEF.
RANK
SPECIAL
DVOA
S.T.
RANK
SCHED SCHED
RANK
SEA 39.8% 1 32.0% 1 -8.4% 2 -0.7% 18 -6.4% 29
IND 27.6% 2 30.6% 2 -0.2% 19 -3.2% 30 -2.2% 21
SD 21.2% 3 16.3% 5 -5.3% 8 -0.4% 17 2.9% 10
KC 19.4% 4 14.8% 6 -5.6% 4 -1.0% 22 5.2% 3
CIN 19.1% 5 22.2% 3 2.8% 24 -0.3% 16 5.8% 2
DEN 18.3% 6 18.9% 4 -0.7% 17 -1.3% 24 7.3% 1
PIT 15.6% 7 14.0% 8 -0.9% 16 0.6% 12 4.4% 5
PHI 12.0% 8 4.4% 10 -5.2% 9 2.4% 4 -2.3% 22
CHI 11.2% 9 -2.7% 17 -13.7% 1 0.2% 14 -9.5% 32
CAR 11.1% 10 8.9% 9 -3.2% 12 -1.0% 20 -3.2% 24
ATL 8.4% 11 1.6% 12 -5.5% 5 1.3% 8 -1.7% 19
WAS 7.2% 12 -0.5% 15 -5.4% 7 2.3% 5 -2.0% 20
NE 6.8% 13 14.1% 7 6.0% 27 -1.3% 25 -5.6% 28
JAC 4.2% 14 2.5% 11 -1.3% 15 0.4% 13 0.5% 16
NYG 2.0% 15 -0.2% 14 -2.3% 13 -0.2% 15 4.6% 4
OAK 1.8% 16 -5.4% 21 -5.7% 3 1.4% 7 4.2% 6
TEAM
TOTAL
DVOA
TOTAL
RANK
OFFENSE
DVOA
OFF.
RANK
DEFENSE
DVOA
DEF.
RANK
SPECIAL
DVOA
S.T.
RANK
SCHED SCHED
RANK
TB 0.0% 17 -1.9% 16 -3.3% 11 -1.4% 27 4.1% 9
BAL -0.3% 18 -4.9% 20 -5.5% 6 -1.0% 21 4.2% 7
MIA -0.4% 19 -8.2% 22 -3.5% 10 4.3% 1 -4.9% 27
DAL -2.4% 20 -4.0% 18 -0.7% 18 0.9% 10 0.3% 17
TEN -7.4% 21 0.9% 13 9.8% 29 1.5% 6 1.1% 13
DET -8.7% 22 -4.8% 19 3.2% 25 -0.7% 19 -6.8% 30
CLE -8.8% 23 -12.2% 26 -2.2% 14 1.2% 9 4.2% 8
GB -13.4% 24 -9.9% 23 2.2% 22 -1.4% 26 -8.0% 31
ARI -17.5% 25 -10.5% 24 1.5% 20 -5.6% 32 1.0% 15
MIN -19.4% 26 -16.4% 28 1.9% 21 -1.0% 23 -4.7% 26
HOU -21.0% 27 -13.4% 27 8.4% 28 0.8% 11 1.1% 14
BUF -23.3% 28 -20.5% 32 5.7% 26 2.9% 2 -2.9% 23
STL -24.6% 29 -11.8% 25 10.6% 31 -2.3% 28 2.6% 12
SF -25.6% 30 -18.0% 30 2.2% 23 -5.3% 31 0.2% 18
NYJ -26.7% 31 -19.2% 31 9.9% 30 2.4% 3 -4.2% 25
NO -34.1% 32 -16.9% 29 14.6% 32 -2.6% 29 2.8% 11

Next, we've projected the entire season 5,000 times. We give the mean projected wins for those simulations, as well as the standard deviation, the percentage of time each team won the division, and the percentage of time each team made the playoffs. Thank you to Dr. Ben Alamar, who wrote the code to create the simulations.


NFC
East Division Playoffs Win SD
PHI 43.2% 61.4% 9.9 2.49
WAS 33.3% 53.9% 9.2 1.98
NYG 16.9% 35.6% 7.8 2.09
DAL 16.3% 33.9% 7.5 2.46
North Division Playoffs Win SD
CHI 59.4% 69.0% 10.6 2.94
DET 19.6% 36.1% 7.7 1.93
GB 15.4% 30.4% 7.2 2.82
MIN 7.1% 15.5% 5.9 3.54
South Division Playoffs Win SD
CAR 45.8% 62.0% 9.9 2.14
ATL 38.0% 56.4% 9.0 3.38
TB 18.1% 34.5% 7.5 2.17
NO 1.3% 3.5% 3.6 4.51
West Division Playoffs Win SD
SEA 79.5% 79.8% 13.6 2.66
ARI 2.4% 12.8% 5.6 2.31
SF 1.5% 7.8% 4.8 3.10
STL 1.7% 7.0% 4.6 3.88


AFC
East Divison Playoffs Win SD
NE 49.4% 56.5% 9.6 3.03
MIA 30.5% 40.3% 8.4 2.68
BUF 6.8% 10.3% 5.4 2.87
NYJ 5.7% 8.8% 5.2 2.93
North Divison Playoffs Win SD
CIN 41.6% 55.0% 9.8 3.41
PIT 37.3% 52.0% 9.6 3.27
BAL 14.1% 25.8% 7.4 2.26
CLE 7.3% 14.5% 6.3 2.19
South Divison Playoffs Win SD
IND 66.0% 72.6% 12.0 3.32
JAC 18.0% 38.6% 8.5 2.42
TEN 8.4% 21.4% 7.1 3.35
HOU 3.0% 8.3% 5.4 3.02
West Divison Playoffs Win SD
SD 37.7% 59.2% 10.4 2.24
KC 31.7% 54.4% 10.0 2.83
DEN 29.3% 52.2% 9.7 3.03
OAK 13.3% 30.2% 7.9 2.91

Posted by: Aaron Schatz on 07 Sep 2006

138 comments, Last at 19 Sep 2006, 5:13pm by John Morgan

Comments

1
by Dr. Evil (not verified) :: Tue, 09/05/2006 - 11:11pm

Why is Seattle's secondary considered mediocre? Trufant is a top talent who granted hasn't played up to his potential but has been injured. Herndon is servicable. Hamlin is a lock Pro-Bowler when healthy. Boulware does nothing but make plays and get picks that win games.

How is that mediocre?

1
by Dr. Evil (not verified) :: Tue, 09/05/2006 - 11:11pm

Why is Seattle's secondary considered mediocre? Trufant is a top talent who granted hasn't played up to his potential but has been injured. Herndon is servicable. Hamlin is a lock Pro-Bowler when healthy. Boulware does nothing but make plays and get picks that win games.

How is that mediocre?

3
by Scott de B. (not verified) :: Tue, 09/05/2006 - 11:12pm

So you don't predict New England's defense to improve at all from last year?

4
by Luz (not verified) :: Tue, 09/05/2006 - 11:15pm

almost time for the real football to start........

4
by Luz (not verified) :: Tue, 09/05/2006 - 11:15pm

almost time for the real football to start........

6
by Zac (not verified) :: Tue, 09/05/2006 - 11:38pm

Is there an echo here?

I see the system has finally stopped expecting big things out of the Buccaneers.

Your pre-season subjective rankings don't interest me much, so I'm just going to talk about what the projection system spit out.

Playoff Teams (based on projected wins)
NFC
West-Seattle-13.6 wins
North-Chicago 10.6 wins
East-Philadelphia-9.9 wins
South-Carolina 9.9 wins
WC-Washington (9.2 wins) & Atlanta (9.0 wins)

AFC
South-Indianapolis-12.0 wins
West-San Diego-10.4 wins
North-Cincinnati-9.8 wins
East-New England-9.6 wins
WC-Kansas City (10 wins) & Denver (9.7 wins)

And it's picking Green Bay to finish ahead of Minnesota! Gosh, I hope so!

7
by paytonrules (not verified) :: Tue, 09/05/2006 - 11:58pm

Is Minnesota really that bad? I've read the chapter in Football Prospectus and I'm still not seeing the 5.7 wins. The offensive line should be better, FP refers to their front 4 as possibly being a monster, and they get 6 games against the Bears, Green Bay and Detroit. This was a fashionable Super Bowl pick a year ago - I could see 9 wins, which could win the division.

8
by jonnyblazin (not verified) :: Wed, 09/06/2006 - 12:22am

I happen to agree that Minny has a pretty good shot at the North division title. Didn't their defense play quite well during the 2nd half of the season or am I hallucinating?

"Steve McNair is not going to be healthy for 16 games." Well its a good thing his backup is Kyle 'Breakout' Boller (at least according to KUBIAK).

9
by Ryan Mc (not verified) :: Wed, 09/06/2006 - 12:25am

"Does anyone remember the last player to star in NFL Europe and use that as a springboard to play an important role on an NFL team?"

Sure, my memory isn't so short that I've already forgotten Kurt Warner, Dante Hall and Jake Delhomme. Heck, Warner and Brad Johnson's careers alone justify the existence of NFLE. And as Pats fans, I'm sure you'll remember that NFLE was a big part of Adam Vinatieri's development (according to the team website, the fact that he doubled as a punter particularly impressed the team)

10
by Jake (not verified) :: Wed, 09/06/2006 - 1:06am

The projection system is down on Minnesota because of the passing game and mediocre D, I think. Something about Brad Johnson going down.

11
by Jake (not verified) :: Wed, 09/06/2006 - 1:14am

The Chiefs projection higher than I expected. I know their run defense has been decent, but I want to see what effect Herm Edwards on Gunther's play calling. Far too often the Chiefs gave up big plays, although last year in particular it was atrocious tackling as much as the 8 man blitz.

12
by Joe (not verified) :: Wed, 09/06/2006 - 1:46am

What they had to say about the Rams...

"Steven Jackson has the most commonly misspelled name in the NFL, so he can't be happy about Stephen Davis coming aboard to confuse things even further."

Such insight.... How much do these guys get paid to do this?

13
by Crushinator (not verified) :: Wed, 09/06/2006 - 1:51am

What they had to say about the Rams…

“Steven Jackson has the most commonly misspelled name in the NFL, so he can’t be happy about Stephen Davis coming aboard to confuse things even further.”

Such insight…. How much do these guys get paid to do this?

Pot, meet kettle.

Well, not completely. After all, you're not being paid for your worthless comment. and the actual interesting stuff are the numbers, not the comments.

Unless knowledge like "Dwight Freeney is a good pass rusher" like most sports writers will give you is somehow more useful.

14
by Bobman (not verified) :: Wed, 09/06/2006 - 1:54am

The Ron Dayne comment is priceless, and even better when compared to Peter King's blather about Ron Dayne gaining 1,200 yards for Kubiak in Houston. Which Ron Dayne is he talking about?

The stat in the Indy blurb, about teams losing 1,000 yd rushers with 4+ YPC averages is, frankly, amazing. Even more amazing is that the Rams lost Greg Bell and got worse!

Since numbers never lie, I can't wait for my well-deserved 15-1 season fom the horseys. That'll offset a season I lived through with them about 15 years ago when they had only 1 win. And after that, they still owe me about a dozen more like it.

And considering how many "experts" are picking the Panthers (and Cowboys) for the SB, the projections here are not quite as rosy.

Last note: I wholeheartedly agree with Aaron's Jags comment re Offense vs Defense: come on, ask a dozen schoolkids, or football fans, or Klingons for that matter: which side of the ball is better in Jax? D, as in Del Rio. Not O as in Omigod, Fred Taylor got hurt at Jimmie Smith's retirement party!

15
by Bobman (not verified) :: Wed, 09/06/2006 - 1:57am

Duh! That's Jimmy. (damn, now I'm sure to be left off his Christmas list. Again)

16
by Matthew Furtek (not verified) :: Wed, 09/06/2006 - 2:14am

The forces will collide in Week 1 on MNF.

Vikings vs. Redskins.

Two teams whose preseason performance is the opposite of what is expected.

'Twill be very interesting.

17
by Becephalus (not verified) :: Wed, 09/06/2006 - 2:29am

Here are my 2 cents:

Well, I hate to just be a complete homer, but the Minnesota projection seems like the most off base one to me. 6 wins? They should get 3 or 4 wins in the division, and they play BUF and the NYJ. Not to mention they still have one of the better HFA in the NFL and are more or less a lock for 5 home wins each year (In past 5 years they are 26-14 at home and 11-29 on the road). Those years also include two pretty bad teams, which are worse than this years model IMHO.

Are they are really going to get only 1 or 2 wins in their other 8 games (4 of these at home)? I highly doubt it. I see them as a 8 or 9 win team. I think you are underestimating the value of the improved coaching staff.

Anyway, I also tend to think ATL will not get 9 wins this year, more like 7, and that the projection for OAK seems a bit high given their difficult schedule.

Like I said just my thoughts. Keep up the informativework.

18
by Jake (not verified) :: Wed, 09/06/2006 - 3:36am

The projections seem to be pretty high on Oakland. They like Aaron Brooks (I think).

19
by Jake (not verified) :: Wed, 09/06/2006 - 3:44am

I'd like to see projected VOA ranks for o/d/ST as well.

20
by jonnyblazin (not verified) :: Wed, 09/06/2006 - 4:18am

I must say a lot of the defensive rankings seem counterintuitive to me. Seahawks 2nd, Raiders 3rd (!!!), Chiefs 4th, Falcons 5th, Chargers 8th- these teams don't strike me as defensive powerhouses.
Panthers 12th, Giants 13th, Steelers 16th, Broncos 17th, Vikings 21st, Patriots 27th- I think these are probably going to be some of the strongest defenses in the league. I guess you can't argue with the numbers, but my gut tells me Steelers and Panthers aren't average defenses.

21
by James C (not verified) :: Wed, 09/06/2006 - 6:20am

Re 12

"Such insight…. How much do these guys get paid to do this?"

That is just plain rude. The Outsiders got off their arses and built a publication, a website and a system of statistical analysis which far exceeds anything that was previously available and all you can do is turn up on their website and bitch because you didn't understand their joke. They didn't just schmooze their way into a job with an existing media entity or try to get attention by spouting obnoxious opinions. They created something unique and I for one am grateful that they bothered. If they are getting fairly compensated then all power to them, because they took a risk with their time and effort and it paid off.

22
by someone (not verified) :: Wed, 09/06/2006 - 8:13am

So, from another thread, Zimmerman and Banks are disgraces to punditry for not predicting more 12+ win / 12+ loss teams. There's never been less than 6 in any given season and the average is over 8 according to the FO guys over there. But your statistically modelled predicted records contain two 12+ win teams and one 12+ loss team? How confusing.

23
by kleph (not verified) :: Wed, 09/06/2006 - 8:17am

can i get a witness!

24
by Scott de B. (not verified) :: Wed, 09/06/2006 - 8:39am

But your statistically modelled predicted records contain two 12+ win teams and one 12+ loss team? How confusing.

That's because these are median averages, not actual predictions. Thus the Patriots are listed as having 9.6 wins.

25
by Alex (not verified) :: Wed, 09/06/2006 - 8:40am

RE: 22
The wins here are averages, which are based mostly on numbers. These are the projections to get you general placement, pundits' picks are the ones with their magical system to determine wins and losses.
Statistics hedge their bets. Pundits aren't supposed to.

Color me surprised as well at the defensive rankings. It's great to see the Raiders with a good defense, but I don't know how that's going to happen. I am also surprised at the large dropoff for the teams mentioned in post 20, except for the Giants and Vikings. It seems that for the most part that this site was fairly high on some of the top-ranked defenses last year. What accounts for the the difference?

26
by zlionsfan (not verified) :: Wed, 09/06/2006 - 9:42am

"Cool projection"? Figure of speech, or is he one of the guys from coolstandings.com?

27
by Mr Shush (not verified) :: Wed, 09/06/2006 - 9:43am

I know the pundits do the "pick too few extremes" thing every year, but I think this season might be a rare case in which they're actually right. Last year, everyone except the Dolphins, the Falcons, the first half of the Raiders and the second half of the Vikings was either really good or really bad. This year, everyone will regress to the mean with the exception of the Patriots, who will get much better thanks to players returning from injury, the Eagles, who will regress not just to but well through the mean for the same reason, and the Bills, who will actually manage to be even worse. We really are going to see an awful lot of teams in the 6-10 to 10-6 range this year. No one will suck to the extent that the '9ers and Texans did last year. Equally, though, the gulf between good and bad last year was so vast that it is less likely than usual that last year's bad teams will make the playoffs or last year's great teams fall out of them. Damn, there's nothing I hate like stupid pundits being right for the wrong reasons.

28
by Duff Soviet Union (not verified) :: Wed, 09/06/2006 - 9:54am

"The projections seem to be pretty high on Oakland. They like Aaron Brooks (I think." No, actually they like the Raiders defense and, to a lesser extent, special teams. As a Raiders fan, that #3 ranking for the defense seems ridiculously high to me. Can you offer any elaboration on this or is it one of those "this is what the numbers say, but I don't believe it" things? How are their special teams so good too? They were horrible last year. Regression to the mean from Seb Janikowski?

29
by ernie cohen (not verified) :: Wed, 09/06/2006 - 10:19am

re 22: The easiest way to think about this is, suppose the winner of every game was determined by a coin flip (i.e., perfect parity and no home field advantage). Every team would project to 8 expected wins. However, a team will get exactly 8 only about 1/5 of the time, 7/9 wins 1/6 of the time, 6/10 wins 1/8 of the time, 5/11 wins 1/16 of the time, 4/12 wins about 1/35 of the time.

Similarly, teams that projecting to 4-6 wins have a good chance of getting fewer (or more). But you can't predict which one(s) it will be.

30
by someone (not verified) :: Wed, 09/06/2006 - 10:33am

re.24/25.

Thanks, but I'm still not getting it. So FO has a projected win total of 12 games for the Colts, but FO is not therefore predicting that the Colts will win 12 games? If the Pats' projected win total is 9.6, doesn't rounding it up to 10 simply get to the prediction?

Is it possible for Football Outsiders to have a projected win total for a team, of a certain number, but for their actual prediction to be a number of games different? I just checked the 2005 pre-season forecasts and the divisional standings appeared to be predicted based entirely on the concept of "mean wins", which is partially responsible for my confusion on this point.

If "projected wins" are not in fact a projection of the number of wins a team will have i.e. a prediction, I'm not understanding what it is or what the point of it is. But I am not especially statistically conversant.

If it is a prediction, which to me in my ignorance it seems to be, and it only has three teams (when records are rounded up or down) with 12+ records, and there's never been a season with

31
by someone (not verified) :: Wed, 09/06/2006 - 10:34am

Sorry, truncated by a symbol.

and there's never been a season with less than 6, what does that mean re. the modelling? "Statistics hedge their bets. Pundits aren’t supposed to." I don't get it. I thought the statistics were supposed to give a truer picture of likely outcomes, hence their use on this site in lieu of the methods of the likes of King/Zimmerman et al. You seem to be telling me that they are inherently unreliable for the purpose for which they are being used. But like I said, I'm confused.

32
by Eric P (not verified) :: Wed, 09/06/2006 - 10:37am

On Minnesota's "improved coaching staff"...

When you have a head coach who's never been a head coach before, an offensive coordinator who's never been an offensive coordinator before, and a defensive coordinator who's never been a defensive coordinator before, isn't it a little premature to call the staff improved? I know Mike Tice was one of the worst, if not THE worst, HCs in the league, but it's entirely possible that these guys will be worse.

33
by ToxikFetus (not verified) :: Wed, 09/06/2006 - 11:28am

I'm glad that FO hasn't partaken in the Cowboys Kool-Aid like the rest of the NFL punditocracy. Two gimpy hammys at WR, a colinderesque OL, and a paper-mache statue at QB do not a great offense make.

34
by DWL (not verified) :: Wed, 09/06/2006 - 11:30am

#1 – I think you’ve answered your own question(“hasn’t played up to his potential,” “is serviceable,” and “when healthy”). So basically, the Hawks can count on one guy in the secondary.

35
by Will Allen (not verified) :: Wed, 09/06/2006 - 12:03pm

Eric, Minnesota's coaching problems last year extended well past Mike Tice, who, for all his faults, was not the biggest problem in this area. Previous ownership was in such a short term bottom line mentality, given the desire to sell, that the Vikings had their offensive line coach also taking on the duties of offensive coordinator and quarterback coach. Has any other team in the NFL EVER done this in the past 20 years? I can't prove it, but I highly suspect that defensive coordinator Ted Cottrel had the job because nobody else with any experience would take it for as low a salary as what Cottrel was working for. During his entire time in Minny, the Vikins' defense had trouble with the most basic execution, like lining up in the right place. The guy could not communicate, and he is now drawing on his coach's pension, and no longer working, and it is not because he wanted to retire.

The Vikings last year were among the worst-coached teams in recent NFL history, particularly in the first half of the season. It got so bad that the new owenership imported two retired coordinators, Foge Fazio and Jerry Rhome, at mid-season, in a desperate effort to improve basic execution. The new coaching staff would practically need to show up drunk every day to not be an improvement. From what I can see, however, which admittedly is a distant vantage point, defensive coordinator Mike Tomlin has a better than 50% chance to be a head coach before he turns 40; he really seems to be straight out of the Tony Dungy/Lovie Smith/Monte Kiffin/ school of teaching the proper execution of the cover-2, and the Vikings are not without talent on defense. Childress' offense will fit Brad Johnson's skill set better.

Aaron and company have downgraded the Vikings' mean win projection to some degree due to their injury history, especially Brad Johnson's injury history. This is eminently reasonable, and to me it it the biggest reason to think that 6 wins may be the outcome of their schedule. On the other hand, they will not be running a system that is extremely quarterback dependent by NFL standards, and there is much reason to think that the offensive line will be far better this season. Combined with a healthy defensive line (which is certainly not a given either) and better coaching, the Vikings could easily surpass six wins by a comfortable margin.

Aaron's comments regarding the Vikings schedule are spot-on. Even a 2-5 start will not rule out the Vikings' chances for a winning season to the degree it normally does in the NFL. A 3-4 start will give them a better than 50 percent chance for nine wins, and a 4-3 start will likely make it a 90 percent certainty.

Matthew's comments above regarding the intriguing nature of the match-up with the Redskins on Monday are also spot-on. As a Vikings fan, I 'd gladly swap a home game with the Redskins for the one with the Panthers, since the Panthers are just going to be very difficult match-up for the Vikings no matter where the game is played. If the Vikings go into FedEx on Monday and get a win, however, they will have an excellent opportunity to put some pressure on the Bears at the Metrodome shortly thereafter, where the Bears always have trouble.

36
by coltrane23 (not verified) :: Wed, 09/06/2006 - 12:13pm

#34:

. . . and that guy is a rookie.

Seattle's secondary is definitely an Achilles' heel for the defense. The front 7 should be able to generate a heck of a pass rush when needed, but if it can't then Trufant, Herndon, et al will resemble burnt toast. They're not atrocious, mind you, but they are the weak link on this team. (Well, apart from KR/PR.)

Off the top of my head, not a lot of teams on the sked will be able to exploit the secondary, because a team will require strong O-line play, a QB who can throw the deep ball accurately, and receivers who can get deep quickly. The Giants are the first team up to concern me in that regard.

37
by Pat (not verified) :: Wed, 09/06/2006 - 12:22pm

I know the pundits do the “pick too few extremes” thing every year, but I think this season might be a rare case in which they’re actually right.

If it would be, it would be the first time in the history of the 16-game NFL season that there was as much parity as they thought. Think about that - that's almost 30 years of history. And that includes times when there were fewer teams.

So, from another thread, Zimmerman and Banks are disgraces to punditry for not predicting more 12+ win / 12+ loss teams.

And in that exact same thread, when I pointed this out, I said "I could understand if they were statistical projections, but these aren't." Statistical projections are going to have some error, and since you can't project over 16 wins or under zero, you'll truncate the high end. Since you've got a truncated high end, you'll pull averages toward the middle.

The spread on those projections is usually huge - like 4-6 games or more. But they're still accurate, even if they're not precise.

38
by Walt E (not verified) :: Wed, 09/06/2006 - 12:25pm

Someone in #30 and #31,

You have to remeber that what you are looking at is the average number wins per team over 5,000 simulated seasons. If you were to look at one simulated season at a time, you'd find more 12 plus wins and 12 plus loss teams in each season then what you see by looking at the average. It's just that, from simulated season to simulated season, it won't be the *same* teams over 12 each time. The average "smooths" out the extremes on a team by team basis.

So when you see the Colts listed at 12.0 projected wins, don't read it as "The Football Outsiders say the Colts will win 12 games this year". It's really, "If we could play the 2006 NFL season 5,000 times, based on Football Outsiders' number crunching, the Colts would average 12 wins per season." And since we can, in reality, only play the season once, it's best to read it as I do: "It is likely that the Colts will win around 12 games this year. I won't be suprised if they win only 10, or if they win 14. It would be somewhat suprising if they only won 8 games. And I would be shocked if they finished under .500".

39
by Loose On the Lead (not verified) :: Wed, 09/06/2006 - 12:33pm

Regarding the apparently low number of 12-win teams...

To come up with these projections, 5000 simulated seasons were run. It may very well be that almost all of those simulations came up with a bunch of 12-victory teams. But when you average over all 5000 simulations, you iron out the variability that leads to many of those gawdy records.

In the projections you see in the article, there are a number of teams with around 10 projected wins. Some of those teams will have things go their way and will end up with 12 or more victories. If we could run this season over again, the same thing would happen, but different teams projected for 10 wins would end up being the 12-game winners.

40
by Tom (not verified) :: Wed, 09/06/2006 - 12:36pm

Re #31 and others
It would seem to me that at least part of the lesson of DVOA is the best teams don't always win the most games. Go back to the GUTS & STOMPS--some team that is projected to win 10 will likely win 13+, but that doesn't make them a better team. And I don't believe there's any evidence that that's any more than luck. Thus, the FO projections are: "Here's what it could look like, if things are as we project them and assuming even luck." You would consequently expect them to have fewer extreme predictions, since those essentially are not predictable. Humans like Dr. Z and TMQ, though, get no such leeway, because they have nearly 3 decades of history to show them otherwise.

41
by Pat (not verified) :: Wed, 09/06/2006 - 12:53pm

Incidentally, in the article, Aaron says this regarding the Redskins preseason performance:

Over the past decade, five teams have gone winless in the preseason after a winning record the year before. For the 2005 Colts, the preseason losses turned out to mean nothing.

The Colts might actually be a good example in terms of the games played. The Colts had an abnormally hard preseason schedule last year, facing Atlanta, Denver, Cincinnati, Chicago, and Buffalo: only one of which ended up being really bad. Buffalo really sank over the year, too, and finished very strong the previous year, so them being stronger in preseason is understandable.

The Redskins preseason is a bit abnormally strong as well: Cincinnati, New England, Baltimore, and the Jets. The Jets are really the only weak ones there.

The difference, though, is that Indianapolis's first string played much better in those games than the Redskins did. The Denver game, in particular, was only lost in the second half, as it was tied 17-17 at the half. And the only reason the Bengals blew them out was that the first string offense didn't play at all the last game. The Redskins did.

Some part of me thinks that the coaches were holding back due to the Portis injury in the first week: but then again, that injury came on an interception return after Brunell was under heavy pressure. So... got me.

42
by mactbone (not verified) :: Wed, 09/06/2006 - 1:03pm

Re Minnesota:
IIRC, one of the factors in the offseason DVOA is coaching experience. Teams with new coaches, especially coaches with no experience are downgraded. It's the same with new QBs I think. The more a team is consistent the more likely it is to do well, while teams that are constantly changing usually don't do very well. Just think of it as hedging a bet - most first year coaches don't have great records for various reasons like going to a bad team, hiring first time coordinators or needing experience - so instead of evaluating each new coach, there's just a blanket "New Coach" value.

43
by Independent George (not verified) :: Wed, 09/06/2006 - 1:09pm

#30-31 - Think of it this way. There is always a certain amount of uncertainty around each projected win total. So for each projected win total, there are a range of possibilites within range of that projection. So even though only Seattle and Indy are projected to win 12+ and SF to lose 12, there are several teams projected for 10-6 and 6-10. That means that we can be pretty sure that some of those 10-win teams will win 12 games, while others will win 8; some of those projected 6-10 teams will claw their way up to 8-8, some will sink to 4-12. Seattle could just as easily go 15-1, or 11-5. If Philip Rivers turns out to be Chad Pennington 2002, SD could easily win 12 games. If Larry Johnson breaks his ankle in week 1, the Chiefs might go 6-10. The football gods do play dice; random happens. The projected wins are just where we expect things to fall on average.

To put it another way, in any 14-2 season, there are usually 2-3 tight wins where they could easily have lost if someone didn't miss a kick, or the ref called it the other way, or the opposing QB didn't get hurt, etc. The difference between 12-4 and 8-8 is often just dumb luck. Because there are only 16 games in a football season, luck doesn't always balance out during the season the way it does in baseball's 162 games. Sometimes it does (hence the FOMBC, aka regression to the mean), but sometimes it means that the playoff game between the 12-4 division winner and the 9-7 wild card may really be an evenly matched battle of 10-6 teams (How many coaches do you think breathed a sigh of relief when SD was eliminated last year?)

44
by M (not verified) :: Wed, 09/06/2006 - 1:33pm

I can't recall precisely anymore, but aren't standard deviations given for the teams in the "Official" projections for Estimated Wins? This would give a figure that illustrates what Walt E is saying in #38.

BTW, here is a question that may be best answered by Aaron, but anyone else is free to pipe in as well. Is the distribution for Estimated Wins similar to a Binomial distribution? (i.e. the standard deviation is highest for teams with EW = 8, and declines as teams deviate from the mean) Also, do any Stats people out there know precisely how to test this given several years of EW projections? This could be interesting in looking at teams that have much lower or higher Std Dev's than expected given a binomial distribution.

45
by Wanker79 (not verified) :: Wed, 09/06/2006 - 2:08pm

Re: M (44)

I am by no means a statistician, so I'm not quite sure that I'm even responding to the question you asked. But intuitively I'd assume that the distribution would be Gaussian with the apex being the particular team's Estimated Win total. In other words, the apex of Indy's win distribution would be 12 and the probability of them winning or losing a game would follow typical bell curve. But that's just this chemist's best guess.

46
by Basilicus (not verified) :: Wed, 09/06/2006 - 2:23pm

Re 30/31:

To reiterate a bit on what some others have said (38 puts it very well) those projected win numbers (maybe they should consider giving the stat a new name) are a stat in and of themselves. They are not predictions, but rather a pointer to be considered alongside the unit and schedule rankings (and later, the DVOA rankings, for which I cannot wait). This is why the Broncos (9.8 wins) are projected over the Chargers (10.4 wins), the Giants (7.8 wins) are projected over the Falcons (9.0 wins) and Jaguars (8.5 wins), and so on. I like the way FO treats rankings because they break things up and provide you the stats - it's not telling us what to think but rather purposefully makes things very open to interpretation by the reader.

Re 20:

"I must say a lot of the defensive rankings seem counterintuitive to me. Seahawks 2nd, Raiders 3rd (!!!), Chiefs 4th, Falcons 5th, Chargers 8th- these teams don’t strike me as defensive powerhouses.
Panthers 12th, Giants 13th, Steelers 16th, Broncos 17th, Vikings 21st, Patriots 27th- I think these are probably going to be some of the strongest defenses in the league."

Most of those make some sense to me. Panthers 12th and Giants 13th aren't exactly a bad prediction for them - it simply predicts the possibility for a slightly down year for both of them. Actually, didn't the Giants struggle at times defensively last year? The Steelers with a down year would not terribly surprise me, and since when have the Broncos or Vikings been defensive powerhouses? I don't expect the Pats to be as bad as 27th, but they are entering the season with the most questionable defensive line-up and depth they've had since the Belichick era was first getting underway. The Chargers (8th) do have one of the best defenses in the NFL, the Falcons added a lot of new guys including a top run-stuffer (DT Grady Jackson) and a top pass-rusher (DE John Abraham). OLB Ed Hartwell is also coming back from spending 11 games out due to injury. The Chiefs and Seahawks I expect to be in the top half of the league, but not in the top five, and from the Raiders I do expect a huge improvement in all facets, but especially D.

47
by Chris Owen (not verified) :: Wed, 09/06/2006 - 10:45pm

I liked the dig at NFL Europe in the 49ers preview. Lawrence Phillips was NFL Europe's MVP and earned a spot on the 49ers' roster. Against Arizona, he blew his blocking assigment and allowed the hit from Aeneas Williams that ended Steve Young's career. Does that qualify as "an important role on an NFL team"?

48
by Crushinator (not verified) :: Wed, 09/06/2006 - 11:58pm

46

Denver fields a top 10 defensive unit most years.

49
by Jerry (not verified) :: Thu, 09/07/2006 - 2:35am

Pat,

Should I be concerned about the Steelers' 0-4 exhibition record or relieved by their 6-3 halftime lead over Carolina in last week's atrocity?

50
by hwc (not verified) :: Thu, 09/07/2006 - 2:40am

"I don’t expect the Pats to be as bad as 27th, but they are entering the season with the most questionable defensive line-up and depth they’ve had since the Belichick era was first getting underway."

--------

Many "pundits" keep saying that, but it's just pure baloney.

D-Line:

Nobody in the NFL has the kind of depth the Pats have for their 3-4 front. The three starters are ALL young 1st round draft picks with a lot of experience. They've added a 1st rounder to play backup NT. How many NFL franchises have a 1st rounder backing up a 1st rounder at NT? Now add Jarvis Green as a backup -- a guy who has several years of productive experience as a role player and spot starter. Plus, two more backups who have been in the system for a couple of years now. How could you have more depth? This is less depth than Bobby Hamilton staring at DE?

Linebacker:

Yeah, they lost an aging Pro Bowler in McGinnest. But, they've added an aging Pro Bowler in Seau. I expect Belichick to find a producive role for him. On the plus side, they have Bruschi back for the entire season, rather than tentatively returning mid season with no training camp from a stroke. Colvin is now in his second year back from a hip injury that takes two years back. Honestly...how can anyone complain about a starting lineup of Vrabel, Bruschi, Seau, and Colvin. I can think of a lot of NFL teams who would kill to have those kinds of problems. First backup is Banta-Cain, who has seen enough action to be veteran depth. Heck, half the teams in the NFL have rookies as their depth.

Secondary:

Everyone seems to forget that the Pats had six DBs on injured reserve last year. They were signing stiffs off the street on Tues and starting them on Sunday...and then those stiffs would get hurt.

The upshot is that they played about 4 dozen people in the secondary over the last two years and the best of them currently make up an experienced roster. Harrison is back, so Hawkins (who played VERY well in the second half last year) is now the backup. Wilson is back with several productive years as the starting safety. Samuel, Hobbs, and Gay have all been productive starting corners on a SuperBowl team. The reason the secondary sucked last year is because all of these guys were HURT...often at the same time. Yeah, put 6 DBs on injured reserve and your depth will be tested. I'm not sure that people really understand that most NFL teams would have gone 5-11 with the Pats injuries last season.

51
by Dr. Evil (not verified) :: Thu, 09/07/2006 - 4:02am

#34 - Which NFC secondary would you rather have? I can think of 3-4 at most that are better. They are one of the most talented secondaries in the league no matter how you dice it.

The injury qualifier applies to EVERYONE so lets stop acting like they can't be counted on. NO ONE can be counted on with the current rate of injuries.

52
by Jake (not verified) :: Thu, 09/07/2006 - 5:01am

How can you force a simulation to project more wins than it expects?

Besides, only 4 teams had 12 or more wins last year. 3 were in the AFC. In 2004, only 5 teams had 12 or more wins. 4 were in the AFC.

53
by FinFanUK (not verified) :: Thu, 09/07/2006 - 5:37am

Would it be worth publishing an extra stat of the percentage times the projection system predicts each team to get at least 12 wins or losses? That would give an idea of how likely each team is to be elite or awful. I seem to remember a similar thing being done last year with bands of wins, as it was suggested that Philly had virtually no chance of a losing season, or something like that.

Also, Aaron, the pecentages don't add up in the final table showing the percentage of projected division wins etc. At the moment the NFC West adds up to about 85% division wins, and the AFC East about 92%. Also, just having looked at the decimal places in the other divisions, they don't add up to x.0%. Is this a problem to do with projecting ties for each division, although if that were the case, I'd expect the percentage to be over 100% with both teams in the tie credited with the division win?

54
by g (not verified) :: Thu, 09/07/2006 - 7:06am

It is interesting to see that there is an almost 15% chance that no one will win the NFC West.

55
by ZasZ (not verified) :: Thu, 09/07/2006 - 8:19am

And there's a 110% chance that someone will win the AFC West. These tables are where that cliche percentage comes from?

56
by noah of the ark (not verified) :: Thu, 09/07/2006 - 2:46pm

Jerry, IMO, fans of teams that have just won the Super Bowl should never be worried about the coming season. Enjoy the championship! That's what the players do most of the time, and that's why so many champions have lousy follow up seasons. And let's face it, Cowher is more likely to retire after this year than to start a dynasty. After all, he just won a long battle against Schottenheimer's disease, and it's only natural to relax after something like that. Big Ben's injuries are the perfect excuse to take the year off, but we'll know more after Miami's D crashes the party tonight. If I'm wrong , I'll start to swear by DVOA, I promise.

BTW, I had already posted my thoughts on this topic, but apparently the post got deleted? Sorry for not using DVOA much... yet.

57
by princeton73 (not verified) :: Thu, 09/07/2006 - 3:00pm

link to last year's DVOA predictions (click my name)

pretty good, except for picking Philly first & Da Bears last

58
by Jeremy Billones (not verified) :: Thu, 09/07/2006 - 3:11pm

#53 - Buy the book :)

[Though I have no idea how there is less than 100% division winners.]

59
by ggb (not verified) :: Thu, 09/07/2006 - 3:17pm

Why do the Win Division %'s not sum to 100% in any of the divisions?

60
by Greg (not verified) :: Thu, 09/07/2006 - 3:28pm

I really like the work done at FO, but it is important to do simple checks. For instance, do the divisional winner probababilities add up to 100? Do the playoff probabilities add up to 600 for each conference? Making these obvious mistakes, makes one wonder if there aren't worse mistakes in the DVOA system (ones that aren't so easily spotted)

Looking at the numbers, the divisional probabilities add up to roughly 400 for each conference. I suspect Dr. Ben Alamar's code credits a team with a division win if the finish in the top 4 in the conference (whether or not their is a better team in their division).

61
by Diane (not verified) :: Thu, 09/07/2006 - 3:31pm

Perhaps the division %s don't take into account any tiebreakers, and therefore reflect merely "outright" divisional titles?

62
by Trevor (not verified) :: Thu, 09/07/2006 - 3:33pm

^ i assume ties don't double count toward both teams, hence no one wins the division.

63
by Jay (not verified) :: Thu, 09/07/2006 - 3:33pm

To #44 and 45-

The distributions will not be Gaussian, I imagine. They're almost certainly skewed towards 8 wins- as teams get further from 8 wins, it becomes easier and easier for them to regress to the mean, and thus are skewed n that direction, unless there are unnatural forces that cause the distribution to be much smaller than normal. For instance- look at the Saints, who have an Expected Wins of 3.6 with a SD of 4.5. It's awfully tough to win less than zero games.

64
by princeton73 (not verified) :: Thu, 09/07/2006 - 3:38pm

look at the Saints, who have an Expected Wins of 3.6 with a SD of 4.5. It’s awfully tough to win less than zero games.

then you've never seen the Saints play

65
by Pat (not verified) :: Thu, 09/07/2006 - 3:39pm

But intuitively I’d assume that the distribution would be Gaussian with the apex being the particular team’s Estimated Win total.

I highly doubt it, actually. Teams can't win more than 16 games nor fewer than 0, and so Seattle and New Orleans, for instance, will have a nasty tail.

Should I be concerned about the Steelers’ 0-4 exhibition record or relieved by their 6-3 halftime lead over Carolina in last week’s atrocity?

You know, it's funny. I tried taking out final week games. I thought that'd improve the preseason/regular season correlation. It doesn't. In fact, using only week 4 games gives you a weaker (which you'd expect from fewer statistics), but still present correlation.

I wouldn't take any happiness from the 6-3 lead, though. A 3-point win is nothing. :)

Still, though, all of Pittsburgh's preseason games were pretty much within the margin of error. So not much information there.

66
by Wanker79 (not verified) :: Thu, 09/07/2006 - 3:53pm

Re: 63

I was thinking that the distribution would just truncate at 0 / 16 wins. So the Saints would have an equal chance or having zero wins as they would have of reaching 7.2 wins? Which would mean that New Orleans has a better chance of zero wins than it does of having an 8 win season.

67
by Larry (not verified) :: Thu, 09/07/2006 - 3:55pm

I'm really curious what the formula is for the probability of a win given a team's DVOA splits and their opponent's DVOA splits. I know that the staff here doesn't want to give away all their secrets, but this seems like a really valuable tool, if there's research to back up how its done.

Also, given that the predicions come from estimating three ratings for each team, all of which have a spread, 5000 seasons seems like it might be low to get the full variation of possibilities.

Last, there are 257.1 wins total. It could be round-off error, since there is up to 0.05 wins error, and if every team rounded-off upwards, 32*.05 is 1.6 wins off. But the liklihood of this seems very, very low if the round-off error is random and uniform as expected.

I love the idea of this kind of prediction, but there seems to be something wrong with the execution here. I'm sure there are a few commenters (myself included) who'd be happy to spend some time helping to fix the remaining bugs. This is the sort of thing where perhaps a little more open-source-ness would help, though Aaron and the staff are of course the ultimate arbiters of how open they want to be on their methodological details.

The total playoff percentages are (within reasonable round-off) 600% as expected.

68
by Silent Dibs (not verified) :: Thu, 09/07/2006 - 3:56pm

Someone remind me again... Isn't there a rule-of-thumb converting DVOA percentages to actual bonafide points, so that +8.9% DVOA (for example) = 3 points, or whatever?

69
by Andrew (not verified) :: Thu, 09/07/2006 - 3:58pm

Another math query.

The average wins total up to 257.1, so somebody is getting credit for an additional 1.1 wins that aren't going to happen.

70
by Fnor (not verified) :: Thu, 09/07/2006 - 4:03pm

Average wins is the average number of wins over a large number of simulated seasons. It should be close (and it is), but there's no reason they should add up exactly.

71
by Larry (not verified) :: Thu, 09/07/2006 - 4:16pm

No, they must add up exactly, if these are simulations of actual seasons. Each will have exactly 256 wins. The total will be n_seasons*256. Each teams average is n_wins/n_seasons. So, the total is the sum of n_wins/n_seasons, which is Sum(n_wins)/n_seasons. but, Sum(n_wins) must be 256*n_seasons. So, 256*n_seasons/n_seasons = 256 exactly. Some round off error allowed. But not as much as is hereas I say above.

72
by Walt E (not verified) :: Thu, 09/07/2006 - 4:24pm

Silent Dibs in #68,

Rule of thumb is 0% DVOA = 21.5 points. I beleive Aaron mentioned that in a DVOA mailbag last year, because the 2005 PFP had a misprint which said 0% DVOA = 23.5 points. That was based on 2004 scoring averages, though, so your milage may vary in 2006.

73
by Pat (not verified) :: Thu, 09/07/2006 - 4:32pm

Some round off error allowed. But not as much as is hereas I say above.

Roundoff error allowed: .05 * 32 = 1.6 wins. So I'm pretty sure that yes, it is as much as is allowed. Looks like someone might not have rounded properly, though, but now people are just quibbling.

74
by Diane (not verified) :: Thu, 09/07/2006 - 4:57pm

SF has a worse chance than STL of winning their division, but a better chance of making the playoffs?

I'm gonna assume that's just statistical "noise". Or is it something in the distribution of expected wins within the division, conference, head-to-head (IOW tiebreakers).

75
by DavidH (not verified) :: Thu, 09/07/2006 - 5:09pm

Re: #74

A similar thing happened in lats year's projections, and Aaron attributed it to the difference in standard deviation. In this case, I would assume that because STL has a larger standard deviation than SF, the positive tail of its wins distribution reaches up into "division-winning" territory more often. But since SF has a higher average win total, it wins the wild card more often.

An extreme example that illustrates this would be if TeamA won exactly 10 games every (simulated) season, while TeamB alternated between 6 and 12 wins. (And the rest of the league had normalish records) Their stats would look something like this:

TEAM...AVG W...DIV%...PLAYOFF%
TeamA...10...20...85
TeamB...9...40...50

76
by DavidH (not verified) :: Thu, 09/07/2006 - 5:11pm

oh, plus the standard deviations

TeamA...0
TeamB...3

i hope i calculated those right, or i'll feel foolish

77
by ferarapan (not verified) :: Thu, 09/07/2006 - 5:12pm

So I'm wondering why the division percentages for the NFC west do not add up to 100%. That has to be an error. Other than that, I'm intrigued to know how the predictions pan out!

78
by ferarapan (not verified) :: Thu, 09/07/2006 - 5:14pm

Also, the AFC west division winner percentages add up to over 100%. What's up with that? Is there something I'm missing?

79
by DavidH (not verified) :: Thu, 09/07/2006 - 5:15pm

Are this year's standard deviations way lowert than last years, or did I not understand how you calculated the Std Dev number last year?

80
by Larry (not verified) :: Thu, 09/07/2006 - 5:27pm

Yes, Pat, I mentioned it was theoreticaly possible, but really, really unlikely given reasonable assumptions (the direction of the error is 50/50 and the size is also random with uniform distribution).

The real problem is, now that I recall last year's article, I don't think individual games are simulated. Just a team, its ratings, and its overall schedule strength based on the ratings of scheduled opponents. Maybe my first question in #67 is moot since that isn't done. If that's the case, 256 wins isn't required mathematically, but still should be the result. It also means some of the natural correlation in records due to the specific schedules different teams have (especially head-to-head games) is not as well reflected as it should be.

Let's get the season started.

81
by Oswlek (not verified) :: Thu, 09/07/2006 - 5:34pm

hwc #50,

I thought I was in bizarro-world for a minute there. That post is almost exactly what I have been saying/thinking for the past offseason and I could have written it myself, right down to the 5-11 comment. :)

I find it ridiculous that FO thinks that NEs D will be bottom 6. I will be shocked if they fall out of the top 10 in points allowed this year.

82
by Wanker79 (not verified) :: Thu, 09/07/2006 - 5:53pm

Re: 78

What’s up with that? Is there something I’m missing?

That something would be what seems like the majority of this discussion thread. ;-)

83
by Will Allen (not verified) :: Thu, 09/07/2006 - 6:08pm

Yeah, I gotta say that that the projection of the Patriots defensive rank at 27 seems absurd, unless it contains some sort of injury prediction that isn't being made clear. The defensive line simply has too much talent, and is too well-coached, for that to happen. How often has a team with one of the league's top two or three defensive linemen (Seymour, who, in my opinion is the BEST defensive lineman in the league), played on a bottom-six defense? When (if?) it has happened, did that team have anything close to the talent that the rest of the Patriots defensive line has?

84
by Pat (not verified) :: Thu, 09/07/2006 - 6:14pm

I find it ridiculous that FO thinks that NEs D will be bottom 6. I will be shocked if they fall out of the top 10 in points allowed this year.

Uh, NE, by DVOA, was in the bottom 6 last year. They're actually predicting NE's defense to improve from 10.5% to 6.0%.

85
by Will Allen (not verified) :: Thu, 09/07/2006 - 6:26pm

Yeah, Pat, and that ranking was in large part due to injury, especially Seymour's, as evidenced by what happened when Seymour returned. Have you ever read Aaron saying what role injury projections, if any, comprise a DVOA projection?

86
by B (not verified) :: Thu, 09/07/2006 - 6:34pm

"The projections here are updated from Pro Football Prospectus 2006 based on some variables related to player experience, injury history...."
Does that answer your question?

87
by Pat (not verified) :: Thu, 09/07/2006 - 6:50pm

Well, looking at the final DVOA rankings (playoffs included), NE was still at -1.5%, 18th in the league, and at that point, weighted DVOA barely included any games without Seymour.

Obviously he made a pretty big difference coming back in, but over a multiyear stretch, you're now looking at -22% to -11% to maybe -1.5% only looking at Seymour's last games. I don't think it's a huge stretch to say they might struggle defensively this year.

It's not like predicting defense actually works all that well, though, anyway.

88
by Jerry (not verified) :: Thu, 09/07/2006 - 6:55pm

I wouldn’t take any happiness from the 6-3 lead, though. A 3-point win is nothing. :)

There was, sadly, no happiness to be taken from watching the game.

Still, though, all of Pittsburgh’s preseason games were pretty much within the margin of error. So not much information there.

Thanks, Pat. That confirms how I felt watching.

And, noah, I appreciate what you're saying in #56, but if Miami is romping in the third quarter tonight, my reaction is not going to be "At least they won the Super Bowl last year."

89
by hwc (not verified) :: Thu, 09/07/2006 - 7:22pm

"Well, looking at the final DVOA rankings (playoffs included), NE was still at -1.5%, 18th in the league, and at that point, weighted DVOA barely included any games without Seymour."

--------

It wasn't just Seymour. It was Seymour, Bruschi, and Harrison. Plus FIVE more defensive backs on injured reserve. The guy who finally stabilized the battered secondary was Hawkins -- he was signed on a Tuesday and started on Sunday as was the case at just about every position in the Pats secondary last year.

Hawkins is now the backup at safety. Guys who were emergency starters last year are now either backups or weren't good enough to make the final roster as backups.

What made the defensive problems even worse is that ALL of the Pats opening day RBs were injured at the same time. Remember when they were signing RBs on Tuesday and starting them on Sunday? Plus three starters from the O-line were on injured reserve. The complete lack of running game made it impossible for the Pats to hide their defensive injuries by playing smallball, slowdown, low scoring games.

I understand that Aaron's whole deal is statistics based. But, honestly, predicting the Pats to have the 27th best defense in the league is patently absurd.

Look it up, I believe that 10 of the defensive starters for the Pats have been starters at their position on a Superbowl team. The only exception is 2nd year CB Ellis Hobbs. His backup Randall Gay, however, did start for a Super Bowl winning team. The other backup at CB has started 88 NFL games, mostly for a playoff team (the Steelers).

Ask yourself, wouldn't most teams like a backup corner with 88 NFL starts? Wouldn't that be considered pretty good depth by today's standards?

90
by DavidH (not verified) :: Thu, 09/07/2006 - 7:28pm

just as a reference point for the "projecting defenses doesn't really work" idea...

teams ranked 24-30 in defense in the preseason DVOA projections, and their actual end-of-year DVOA rank

2005
24 indianapolis ... 8
25 new orleans ... 25
26 st louis ... 29
27 cleveland ... 24
28 san diego ... 16
29 detroit ... 21
30 green bay ... 22

2004
24 san francisco ... 31
25 san diego ... 12
26 houston ... 18
27 philadelphia ... 16
28 kansas city ... 30
29 arizona ... 15
30 new york giants ... 21

what i really want to look at is teams that have been ranked well defensively for a couple years, then fell apart one year due to injuries, and how they did the following season. unfortunately i don't know enough about injury history to know which teams that applies to...

91
by underthebus (not verified) :: Thu, 09/07/2006 - 8:06pm

so what you are saying is pick up either KC or OAK defense? Wha?

92
by Pat (not verified) :: Thu, 09/07/2006 - 8:39pm

teams ranked 24-30 in defense in the preseason DVOA projections, and their actual end-of-year DVOA rank

That's... a lot better than I thought it would be, actually, especially last year. 5/7 of the teams predicted to finish in the bottom third were at least clearly in the bottom half. Add in 32 (NYG, big whoops on that one) and 31 (OAK) and that's 66% of them finishing basically in the ballpark.

93
by Deb from Dorchester (not verified) :: Thu, 09/07/2006 - 9:13pm

DeVoidOnArrival

94
by Larry (not verified) :: Thu, 09/07/2006 - 9:13pm

Thinking more, if the individual games aren't simulated, 257.1 is pretty close to 256, especially if a plausable half of that difference is round-off error. Also, the schedule correlations exist to some degree, through the calculation of schedule strength.

There was a quesiton awhile back about the distributions. I'd expect them to look like the distribution for number of heads when flipping a loaded coin 16 times, which is neither gaussian nor poisson, but pretty simple to calculate, though hard to show in an ASCII comment. SD wouldn't be a good statistic to discuss, but mean would be complete to describe it.

95
by BillWallace (not verified) :: Thu, 09/07/2006 - 9:19pm

"someone" is right in that it's incorrect to take other pundits to task for not predicting more extreme records when the statistics don't. Why hold humans to a different standard than the computer?

If mean projections are the most accurate, why should humans be obligated to be less accurate? I think that's instinctively why the human predictors don't pick more 12 win/loss teams, and I think they're right to do so.

96
by Will Allen (not verified) :: Thu, 09/07/2006 - 9:44pm

Yeah, #86, I kinda knew Aaron was using some sort of injury factor to come to his projections, although I missed that specific line. What I am curious about, although I certainly don't expect Aaron to release all his methods, is specifically how he derives his injury predictions, and what value he assigns this prediction when making his overall projections. Given a specific injury prediction, just about any projected mean win total becomes reasonable.

97
by Pat (not verified) :: Fri, 09/08/2006 - 12:11am

Why hold humans to a different standard than the computer?

Because the humans are predicting one season. The computers are predicting thousands. Thousands of seasons will converge to the mean. One won't.

But even if each individual predicted season was still closer to the mean, all that would say is that there's something still slightly missing in predicting high-win and low-win teams. To me, that's fine. If you can't predict the tails, there's no reason to sacrifice the middle yet.

But the pundits are just going by gut. If they're going to be silly enough to do things like predict tiebreakers, you should be bold enough to predict the extremes.

98
by John (not verified) :: Fri, 09/08/2006 - 12:16am

Hey guys, im new to this DVOA. But i saw in the WR section that chris chambers was ranked 65th behind studs like wes welker and sammie parker. now i understand chambers is overrated. But he isnt the 65th best WR in the league. How can we take this system seriosly?

thanks guys

99
by Jake (not verified) :: Fri, 09/08/2006 - 1:08am

#97

I wouldn't list to Pat. He seems to be wrong much of the time.

100
by Basilicus (not verified) :: Fri, 09/08/2006 - 1:17am

98:

DVOA is based on getting first downs when they're needed, effectiveness on down-and-distance, that sort of thing. Chris Chambers is a very boom-or-bust sort of receiver. While he caught 82 passes last year, he was thrown a catchable ball 166 times. That's only a 49% catch percentage. He's great at getting the deep ball but he's not always great at getting a needed first down.

Now let's look at Wes Welker and Sammie Parker, ranked 44th and 45th. While neither is greatly known for the deep reception, they're both very reliable in clutch situations. Something that Chris Chambers isn't. Welker caught 56% of intended passes, Parker caught 62%. Both of them are better at getting needed short yardage, but not as good at pulling in the deep bomb.

The numbers are also representative of something else. Chambers was thrown at 166 times. That's tied for second in the league with Plaxico Burress, both receivers far and away the only real home run threat in their respective passing games (Anquan Boldin was first with 171 passes, but that's cause the Cardinals just abandoned the running game last year and he's an insane possession receiver - 60% of 171 passes caught is very good.) It means Chambers was probably double-teamed a lot more often than most receivers, probably a reason that he missed some of those intended passes. So you have to read into the numbers a little bit. The rankings don't measure certain things, but once you get used to what the statistics do measure and the way they do it, they're actually very helpful to understanding the game better.

Would I rather have Chambers or Parker on my team? Definitely Chambers. Welker's another case, though. It'd be close, but I'd actually rather have Welker, as I think he's yet to hit his stride and will be factored into the Miami passing game a great deal more this year (he was in tonight's game: 4 catches, 67 yards) and because he's one of the best return men in the game.

101
by Kibbles (not verified) :: Fri, 09/08/2006 - 2:31am

In response to your Power Rankings:
Gibran Hamdan, the NFL Europe Player of the Year, was cut this week by two teams (Seattle and San Francisco) and hooked up on the 49ers practice squad. Does anyone remember the last player to star in NFL Europe and use that as a springboard to play an important role on an NFL team? Shouldn't we stop pretending that NFL Europe is actually serving a developmental purpose for young ballplayers?

Ben Hamilton - Offensive Guard - Denver Broncos

Also, the scheduling gods officially HATE the Denver Broncos. Their SoS has ranked 1st, 30th, 1st, 14th, 15th, 3rd, and is now projected to 1st again. That one 30th ranked schedule was tossed in during the Brian Greise era just to mock us, because the schedule gods knew we wouldn't do anything with it. Seriously, isn't Denver due for a real cupcake schedule sometime?

102
by BillWallace (not verified) :: Fri, 09/08/2006 - 2:34am

"Because the humans are predicting one season. The computers are predicting thousands. Thousands of seasons will converge to the mean."

This is sort of true, but makes absolutely no sense in the context of what we're talking about. The computer isn't predicting thousands of seasons, it's predicting one season 2007. It's using a methodology of predicting thousands of versions of 2007 and averaging them. And that has no impact on what I'm talking about.

If a computer can use that methodology to come up with 07 predicts that include no 12 win teams and you say it's a good prediction, then by whatever methodology a human uses to predict 07, you can't say it's bad because it doesn't have 12 win teams. The human is being smart by hedging and improving the overall accuracy os his predicts by keeping them closer to 8-8, just like the averaging process does the same thing for the computer.

In both cases you know neither predict will be 100% accurate because there are always 12 win teams, but the overall accuracy will be higher for not having predicted them.

103
by BillWallace (not verified) :: Fri, 09/08/2006 - 2:36am

#100

Welker looked excellent tonight. He's got great hands, runs great routes, is deceptively quick, is a real hard worker, and is sure to become a fan favorite soon.

Hopefully for him Culpepper will return Kordell Stewart's passing accuracy back to him sometime before the end of the season.

104
by centrifuge (not verified) :: Fri, 09/08/2006 - 2:52am

Those 12th-ranked special teams for Pittsburgh sure looked awful tonight.

105
by Dan (not verified) :: Fri, 09/08/2006 - 3:50am

M (#44),
Binomial seems plausible, but I ran the correlations and it turns out to go the opposite way: teams with projected records closer to 8-8 have smaller standard deviations (r=.474 between SD & |Win-8|). Looking at the graph, which I can't show you here, you can see that it's almost entirely driven by teams projected to be below average. Standard deviations are larger for teams projected to be much worse than average than for those predicted to be only slight below average, r=.720 (look at New Orleans & St. Louis, for instance - worst projected records, highest projected SD). For projected above average teams standard deviation has almost no relationship with projected wins (r=.100). That fits with the idea that it's harder to predict which previously bad teams are going to turn things around and play decently than to predict which previously good teams are going to get worse.

I'm not sure if the standard deviation numbers are meaningful, though. The idea is that teams with lower sd are more likely to end up with records that are close to their projected record, but that doesn't hold up with last year's projections. The distance between projected wins and actual wins (|projected wins - wins|) doesn't correlate with much of anything. With the stat they reported as standard deviation, r=-0.053 (and you'd expect a positive relationship). When I transformed their weird-looking "standard deviation" stat into something that looked more like standard deviation, I got r=0.061 (which is basically 0). When I looked at how extreme the projections were (|projected wins - 8|), I got r=-.051 (when you'd expect a positive relationship). When I only looked at teams projected to be below average, I got r=-.218 (again, you'd expect a positive relationship). Although maybe 2005 was just a weird year.

106
by Basilicus (not verified) :: Fri, 09/08/2006 - 4:44am

104:

To sort of echo something John Madden was saying tonight (the booth analysis was really good tonight, actually) special teams see the most coverage and blocking breakdowns early in the season. Keep in mind a few of the guys playing special teams might not have been sure of a roster spot until just last week, and may not have that much experience in their particular spot on the coverage units. Plus Wes Welker may be the best return man playing right now.

107
by Basilicus (not verified) :: Fri, 09/08/2006 - 4:45am

Oh, and not suprisingly, Miami has the 1st ranked special teams.

108
by Starshatterer (not verified) :: Fri, 09/08/2006 - 10:15am

Seriously, isn’t Denver due for a real cupcake schedule sometime?

Cupcake schedules only happen, when at least two teams in your same division suck.

So Denver will get a cupcake schedule, if one of the other teams collapses (likely KC, for aging O-line reasons) before Oakland recovers.

109
by Andrew (not verified) :: Fri, 09/08/2006 - 10:18am

Okay, I've read through this whole thread and I still have no idea how it is possible that all 4 NFC West teams can only have a combined 90% chance of winning the division. Who wins it the other 10% of the time? And how do divisions like the NFC East or AFC West get won 110% of the time by one of the 4 teams?

What gives?

110
by Andrew (not verified) :: Fri, 09/08/2006 - 10:18am

Okay, I've read through this whole thread and I still have no idea how it is possible that all 4 NFC West teams can only have a combined 90% chance of winning the division. Who wins it the other 10% of the time? And how do divisions like the NFC East or AFC West get won 110% of the time by one of the 4 teams?

What gives?

111
by TomC (not verified) :: Fri, 09/08/2006 - 11:25am

Perhaps the easiest way to clear up the confusion on the seemingly low number of 12+ win/loss teams is to show the standings from a couple of the 5000 simulated seasons. Looking at the (fairly large) standard deviations, I'm guessing that in most of the simulated seasons, enough teams fluctuate beyond the 12 win/loss mark to make the number of 12+ win/loss teams right around average (i.e., 8).

In fact, if I do my own little "simulation" approximating the win distributions as Gaussians, I get an average number of 12+ win/loss teams per season of 8.2.

112
by Miles (not verified) :: Fri, 09/08/2006 - 12:12pm

Is it really so much of a stretch to think that Dallas' OL has improved from atrocious to average with 4 of 5 starters changing?

2005 2006
LT Tucker F. Adams
LG Allen Kosier
C Johnson Gurode
RG Rivera Rivera
RT Petiti Columbo
More than half of last year's OL either were cut or not resigned.

113
by Insancipitory (not verified) :: Fri, 09/08/2006 - 12:15pm

Re: Kibbles, 101,
That's what you get for kicking the Seahawks out of the AFC West.

114
by kibbles (not verified) :: Fri, 09/08/2006 - 1:19pm

Re #113: We didn't kick them out! We asked them very very nicely to stay, but they bolted and never looked back, and now Shaun Alexander is mocking us mercilously with comparisons of how difficult the NFC West is compared to the AFC West. :(

115
by Greg (not verified) :: Fri, 09/08/2006 - 1:24pm

I think the projection for Philadelphia seems abnormally high. With all the others, I can at least see why DVOA would project out the way it did. But I'm interested to know why Philadelphia projects so well in such a strong division.

As a Raiderfan, I'm very interested in the number that DVOA projects for their defense. Ranking 3rd overall seems high, but I guess thats why you guys crunch the numbers.

116
by Chuck (not verified) :: Fri, 09/08/2006 - 1:31pm

I really have to disagree about Minnesota being placed so low -- they have 6 games within their division (Bears, Lions, Packers), and a number of other easier games. In fact, even by your own analysis, they have the 26th easiest schedule in the NFL this year.

I noticed that every team that you have with a lower DVOA than the Vikings has a harder schedule, even by your own calculations, that's just BIZARRE.

8 or 9 wins might be enough to win their division, considering that their division is the Bears, Lions, and Packers. Notice that 6 of the Vikings games are against these teams, as well.

I love your site, and I think you've done so much original and creative thinking about the league, and I really respect what you are doing... but I just gotta disagree with this one...

I see the Vikings as a stronger team than you've indicated.

Can't wait for Kick-Off...!

Thanks for what you guys are doing, by the way... very interesting work.

117
by Basilicus (not verified) :: Fri, 09/08/2006 - 1:39pm

112:

A couple of things on that point. You're right on Adams, although he was in the line-up for six games last year. Interestingly enough, the 'Boys were 4-2 with him, 5-5 without him, though I'm sure the rest of the team is more responsible for those records than just Adams.

The rest of the line, however, is kind of yucky. While Larry Allen was slowing a bit, Kyle Kosier is still a huge step down. He's solid, but this is a guy Detroit was happy to replace with a three-way competition between Rick DeMulling, Courtney Van Buren, and Barry Stokes. Johnson being replaced with Gurode isn't anything to be terribly happy about either. Gurode was Johnson's backup last year, and for a reason: he's a stronger player but he makes a lot more mistakes. Marco Rivera, well, he'd better start living up to expectations sooner or later or else he's a huge free agent bust. Colombo also isn't anything to write home about after being a bust with the Bears. I fail to see where the optimism is here. If anything, the 'Boys line has traded strength, quality, and great technique with athletic potential on the left guard, smarts and technique for strength at the center, and solid potential for a historical bust at the right tackle. The only upgrade is at left tackle and that's not a full upgrade because the guy played a third of the season last year. I think the 'Boys traded in working through some problems in their O-line to develop quality in the future for something that right now is solid at best and is very questionable in some places.

118
by Sid (not verified) :: Fri, 09/08/2006 - 1:57pm

Haha, DVOA really loves Kansas City. I don't see it. That team won't make it to .500, IMO.

119
by DavidH (not verified) :: Fri, 09/08/2006 - 2:06pm

Re: 116
I noticed that every team that you have with a lower DVOA than the Vikings has a harder schedule, even by your own calculations, that’s just BIZARRE.

Why is that bizarre? The same is also true for Chicago, Green Bay, and the New York Jets. The DVOA projection and the schedule projection have NOTHING to do with eachother. (the schedule does have to do with OTHER teams' DVOA though). Is it so weird that the 6 easier schedules all fall in the group of 25 teams above MIN, and none fall in the 6 below? Sure, the odds say there should be one below, but 0 is pretty close to 1.

120
by Greg (not verified) :: Fri, 09/08/2006 - 2:33pm

Re: 118

I really do see that. Notice the key is that the defense is ranked 4th and the offense is ranked 6th. If Herm takes that defense, improves the tackling, and makes it the 4th ranked defense in the NFL by DVOA, how does KC not win the AFC West?

121
by DavidH (not verified) :: Fri, 09/08/2006 - 2:37pm

Re: 118

I'm having trouble believing in the projection of KC as the number 4 defense, so I took a look at how the projections have done the last couple years at recognizing improving defenses. Unfortunately the projections only go back 2 years, so not much to look at, but here's what I found:

I looked at all teams that met all 3 criteria:
1) ranked OUTSIDE the top ten one year
2) were projected to rank INSIDE the top ten the next year
3) projected rank was HALF or less of last year's rank

All numbers are defensive rank - I didn't want to waste time converting last year's "pt value" projections insto DVOA-style percentages):

year team ... previous/projected/actual/improvement
2005 phi ... 16/2/14/2
2005 dal ... 25/7/13/12
2005 min ... 32/9/19/13
2005 car ... 13/4/3/10
2004 chi ... 18/8/10/8

All of the teams improved at least a little, and most of them by around 10 spots. But most were not as good as the projected, either. If KC conforms to this group, they ought to end up as the 6th or 7th ranked D.

I am NOT saying I believe this will happen. Or maybe I think it will, but that DVOA will be overrating them. One thing to note is that last year's DVOA is -5.7%, and they are projected at -5.6% this year. So maybe it's not as crazy as we think. Or maybe DVOA overrates them for some reason.

122
by Riceloft (not verified) :: Fri, 09/08/2006 - 3:05pm

Re:121
Also, DVOA might be ranking others much lower than they were last year, thus artificially making it seem like they think KC will be significantly improved. With such a similar projected DVOA, its basically saying that they will be about the same as last year.

123
by Riceloft (not verified) :: Fri, 09/08/2006 - 3:06pm

Re:121
Also, DVOA might be ranking others much lower than they were last year, thus artificially making it seem like they think KC will be significantly improved. With such a similar projected DVOA, its basically saying that they will be about the same as last year.

124
by Andrew (not verified) :: Fri, 09/08/2006 - 3:14pm

Greg #115:

I think the projection for Philadelphia seems abnormally high. With all the others, I can at least see why DVOA would project out the way it did. But I’m interested to know why Philadelphia projects so well in such a strong division.

Well, last year they went 6-10 only by dint of throwing away at least 4 winnable games in the division (both Giants games and the 2nd Dallas and Redskins games) and a winnable game against the Cardinals, mostly via stuff like blocked Reggie Hodges punts, Mike McMahon pick sixes, and Mike McMahon sack-fumbles returned to the 2 yard line. Its not so hard to think that had those handful of mistakes never occurred, the Eagles would have been a relatively weak 11-5 team last year, but the supposedly vastly improved Giants and Redskins both would have been 9-7 and out of the playoffs, while the Vikings would have been in. Then the NFC would have looked mediocre mostly, rather than strong, and we wouldn't be talking about how strong the NFE East looks now. And this is to say nothing of the idiocy of the Trotter ejection in the first game, or claiming any sort of competitiveness with Seattle despite spotting them 28 points directly via turnovers returned for scores to start the game.

As the abysmal Mike McMahon is not QBing the Eagles this year, and Reggie Hodges is definitely not punting, and McNabb is healthy, and the Eagles pass-rush is clearly improved, and the Eagles kickers are healthy and actually able to kick, it seems reasonable to think things will look up for the Eagles, especially with their cream-puff schedule.

If other teams in the NFC East had to rely upon outlandish mistakes by Mike McMahon to be able to barely beat the Eagles and move above 9 wins for the season, its not unreasonable to think they may have a tougher time of it this year without McMahon tossing touchdowns to their cornerbacks and fumbling the ball to them in the backzone.

Does that seem like a reasonable explanation?

125
by Wanker79 (not verified) :: Fri, 09/08/2006 - 3:25pm

I'd like to apologize ahead of time. The wife is already getting sick and tired of hearing about football (Yes...already. She's in for a long season), so you're going to have to be subjected to my ramblings. Here's my (admittedly biased) two cents on the NFC East.

If you compare the coaches in a historical perspective, Gibbs and Parcells are clearly on one tier, followed by Reid and then Coughlin. But weighing recent history more heavily I would contend that Reid is right up there with Parcells and Gibbs. I'd in no way even try to argue for Reid over either of them (well, I may try to argue him over Gibbs v2.0, but I won't), but he's definitely in the same ballpark as the two of them.

I don't think anyone could make a strong case against McNabb as the best QB in the division. No rational person could argue for Bledsoe or Brunell (especially at this point in their career), and Manning (while his potential is definitely a point of contention) has yet to show enough to overtake McNabb.

Looking at the offenses as a whole, all four teams have huge question marks. Which of these questions will be answered positively are anybody's guess. Westbrook's health? Dallas and NY's o-line? He Who Shall Not Be Named? Eli's accuracy? Philly's WR corp? Tiki after such a heavy load last year? When will the Campbell Experience begin? There are just way too many "if's" associated with all four teams to make an argument for or against any of them. So I won't. I'll say this much, though. Philly and Washington have clearly the best o-lines, and Philly has the advantage in both age and depth.

That leaves just two units, defense and special teams. I would be pretty shocked if all four of them finish at least very close to the top third in the league. Last year in weighted DVOA Washington finished 3rd (-17.1%), NY finished 8th (-10.9%), Philly finished 10th (-5.7%), and Dallas finished 12th (-5.2%). If anything, I could see Philly, Dallas, and NY closing the gap on Washington.

Finally, special teams. If this division is as defensively driven as I think it's going to be, field position and FGs could have a huge effect on the outcome of the division. And in that case, I'll gladly take my chanced with possibly the best special team's coach (John Harbaugh) and (at least) one of the best kickers (Akers) in the game.

126
by Wanker79 (not verified) :: Fri, 09/08/2006 - 5:20pm

That leaves just two units, defense and special teams. I would be pretty shocked if all four of them *DID NOT* finish at least very close to the top third in the league.

127
by Rich Conley (not verified) :: Fri, 09/08/2006 - 5:57pm

re:117

That argument always peaves me. It is based on the idea that Detroit is an accurate judge of talent, which, from what we've seen over the last 4-5 years, is patently untrue.

128
by DW (not verified) :: Fri, 09/08/2006 - 8:35pm

OK, here's what I don't get about the Patriots defense being ranked 27th. In FO on Fox ratings of NFL teams' defensive lines, defensive backs and linebackers, NE ranked 7th, 13th and 13th. How does that add up to an overall 27th defense?

129
by Riceloft (not verified) :: Fri, 09/08/2006 - 10:25pm

Re: 128

Those FO on Fox articles were based partly on the personal evaluation of the writers, not just on DVOA stats. This is based 100% on what DVOA spits out.

130
by Mike (not verified) :: Sat, 09/09/2006 - 12:43am

Re: 109 No kidding, half of the division probabilities don't add up to 100%! Something's not working there.

Try coolstandings.com - there the percentages at least add up!

131
by Kibbles (not verified) :: Sat, 09/09/2006 - 2:15am

Also, one final point. Everyone who thinks that the NFC East is the toughest division in football might want to check the projections for the AFC West. Despite playing the most brutal 6-game divisional slate in the NFL, the AFC West teams are so good that, according to projected wins, both wildcards will come from there. And the 4th best team in the AFC West is almost as good as the 3rd best team from the NFC East.

I know they're predictions and they really don't mean squat, but what's the point of rooting for the best division in football if I can't thump my chest and point out that it really is the best division in football?

132
by Paul (not verified) :: Sat, 09/09/2006 - 11:35pm

I just can't see the skins winning more games than the Giants. can somebody clue me in on what the Giants supposed weaknesses are? the NY media love them, other than age, what's the issue? They look like a 10 win team to me, easy.

133
by Bencoder (not verified) :: Sun, 09/10/2006 - 3:04am

#132

Strength of schedule for Giants is considerably higher than the other NFC East teams.

134
by Budman (not verified) :: Mon, 09/11/2006 - 9:39am

Knowing it is only week one but watching the SEA/DET game today I would say that the 25th rank for DET DEF is going to be wildly off. Especially if today was an indication of capability rather than just a fluke run into a bad SEA day.

Looking at DET's schedule I am going to predict the former. And also going to be my #1 waiver pickup this week.

135
by SEA ickyicky (not verified) :: Tue, 09/12/2006 - 3:03pm

Seattle is good..........good like chocolate milk.

136
by Yaguar (not verified) :: Tue, 09/12/2006 - 6:00pm

132: Eli's low completion percentage, Plaxico's inconsistency and Toomer's age, less-than-special secondary, age regression from Tiki, a line that loves false starts, and difficult out-of-division games including Seattle, Chicago, Tampa Bay, Atlanta, Jacksonville, Carolina, and the L they just got against Indianapolis. That's just off the top of my head. They're plenty good, sure, but every team in the division has at least a decent chance of winning more games than they do.

137
by Chris (not verified) :: Fri, 09/15/2006 - 8:54pm

These meaningless crappy projections have the Titans winning 7 games this year. How's that workin out so far???

This stuff is absolutely worthless.

138
by John Morgan (not verified) :: Tue, 09/19/2006 - 5:13pm

Geeze, the a-hole level of this thread is off the charts.

Who are you people?