Writers of Pro Football Prospectus 2008

Most Recent FO Features

BradyTom00.jpg

» The Week In Quotes: December 19, 2014

It's a Tom Brady-centric edition of TWIQ. What does he say about a potential rematch with Denver? Why does he like to headbutt people? And why do his teammates compare him to a Clydesdale?

09 Jan 2007

Postseason DVOA I

by Aaron Schatz

Hey kids, time for postseason DVOA ratings. Last year, we ranked just the 12 playoff teams. This year, based on reader requests, we've ranked all 32 teams, whether they are in the playoffs or not. Teams which did not play in the wild card round are treated as if they had a bye week. (That includes both the 20 non-playoff teams and the four teams with byes.)

All numbers are weighted DVOA. That means that Weeks 1-4 are not included, while Weeks 5-10 are somewhat discounted. If you want to see the regular total season ratings, you can just use the Just the Stats menu above.

The FOXSports.com commentary is finally up as of 4:00pm EST Wednesday. It includes just the 12 playoff teams, and the order has been changed so that teams still alive are listed above teams that were eliminated this weekend.

I didn't get to do as much game charting breakdown on the playoff teams as I wanted last week, so I hope to do more this week. Check the FO FOX blog for that. It will start later this afternoon with defensive backs from the four teams that did not play in the wild card round.

(Update: That post is now up. Check out the New Orleans stats for Weeks 1-6 vs. Weeks 8-15. Super weird.)

For those wondering, there's no Any Given Sunday this week, as all four home teams won and Ned has nothing left to say about the Cowboys.

* * * * *

To save people some time, we remind everyone to put their angry troll hatred into the official zlionsfan angry troll hatred Mad Libs form:

<team> is clearly ranked <too high/too low> because <reason unrelated to DVOA>. <subjective ranking system> is way better than this. <unrelated team-supporting or -denigrating comment, preferably with poor spelling and/or chat-acceptable spelling>

I'm not going to bother to run the whole DVOA explanation; if you are new to the website, you can read about it here. As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.


TEAM WEI.
DVOA
LAST
WEEK
W-L WEI OFF
DVOA
OFF.
RANK
WEI DEF
DVOA
DEF.
RANK
WEI S.T.
DVOA
S.T.
RANK
1 BAL 39.2% 1 13-3 10.9% 8 -25.2% 1 3.0% 8
2 NE 30.1% 2 13-4 12.5% 7 -13.4% 5 4.1% 7
3 IND 26.4% 4 13-4 33.7% 1 5.0% 19 -2.2% 24
4 SD 24.9% 3 14-2 23.4% 2 4.1% 17 5.6% 4
5 JAC 21.8% 5 8-8 10.0% 9 -11.6% 8 0.2% 17
6 PHI 15.2% 6 11-6 16.3% 4 -0.7% 11 -1.9% 23
7 CHI 13.6% 7 13-3 -8.8% 24 -16.5% 3 5.8% 3
8 NO 12.7% 8 10-6 15.3% 5 1.9% 16 -0.7% 20
9 CIN 9.7% 11 8-8 19.2% 3 10.5% 28 1.1% 14
10 GB 6.7% 13 8-8 -7.2% 21 -17.3% 2 -3.5% 30
11 CAR 4.9% 12 8-8 -8.0% 22 -15.6% 4 -2.7% 28
12 PIT 4.7% 9 8-8 8.4% 11 1.3% 13 -2.4% 25
13 DAL 4.5% 10 9-8 8.8% 10 7.2% 22 2.9% 9
14 TEN 4.3% 15 8-8 1.2% 16 1.6% 14 4.7% 5
15 NYG 4.3% 14 8-9 6.4% 12 1.7% 15 -0.4% 18
16 BUF 2.4% 18 7-9 -8.8% 23 -4.1% 10 7.1% 1
TEAM WEI.
DVOA
LAST
WEEK
W-L WEI OFF
DVOA
OFF.
RANK
WEI DEF
DVOA
DEF.
RANK
WEI S.T.
DVOA
S.T.
RANK
17 MIA 0.8% 17 6-10 -10.3% 25 -10.2% 9 0.9% 16
18 NYJ -1.9% 19 10-7 3.0% 15 9.5% 26 4.5% 6
19 KC -2.1% 16 9-7 6.4% 13 7.0% 21 -1.5% 21
20 ARI -6.5% 20 5-11 5.7% 14 10.6% 29 -1.6% 22
21 STL -10.2% 23 8-8 14.2% 6 18.3% 31 -6.0% 32
22 DEN -10.6% 21 9-7 -11.6% 26 0.8% 12 1.9% 13
23 MIN -10.6% 22 6-10 -18.2% 30 -11.7% 7 -4.2% 31
24 ATL -11.0% 24 7-9 1.0% 17 9.7% 27 -2.4% 26
25 HOU -12.1% 25 6-10 -1.0% 19 7.7% 23 -3.4% 29
26 SF -14.1% 29 7-9 -6.2% 20 8.9% 25 1.0% 15
27 SEA -14.6% 26 10-7 -13.0% 28 4.2% 18 2.6% 10
28 WAS -16.8% 27 5-11 0.7% 18 19.8% 32 2.3% 11
29 DET -18.1% 28 3-13 -12.1% 27 12.1% 30 6.1% 2
30 TB -23.1% 30 4-12 -17.1% 29 8.2% 24 2.2% 12
31 OAK -23.9% 32 2-14 -33.5% 32 -12.1% 6 -2.6% 27
32 CLE -26.7% 31 4-12 -19.2% 31 6.9% 20 -0.6% 19

Here are the one-game DVOA ratings for the first round of the playoffs. Remember that these include opponent adjustments, so the Eagles and Giants having the same rating means that the Eagles outplayed the Giants by a tiny margin.

TEAM TOT OFF DEF ST
IND 100% 17% -76% 7%
KC -82% -75% -5% -12%
SEA 5% -7% -24% -12%
DAL 3% -17% 0% 19%
NE 45% 12% -16% 17%
NYJ -13% -1% 14% 2%
PHI 11% 14% 7% 3%
NYG 11% 18% -6% -13%

The Colts' rating, while high, isn't even close to the highest playoff rating we've ever tracked. Last year alone there were two higher ratings: Carolina over the Giants in the wild card round (132%) and Seattle over Carolina in the NFC Championship (105%).

Posted by: Aaron Schatz on 09 Jan 2007

144 comments, Last at 15 Jan 2007, 12:20am by Pat

Comments

1
by David (not verified) :: Tue, 01/09/2007 - 3:57pm

Single-game DVOA ratings! Sweet!

I think you only got half the chart in, though...

2
by Ch V Kalyan (not verified) :: Tue, 01/09/2007 - 3:58pm

Aaron, Postseason DVOA for the Pats-Jets and Cowboys-'Hawks game are missing. Are you adding them?

3
by Rich Conley (not verified) :: Tue, 01/09/2007 - 4:02pm

2 Theyre fixed now. They werent there the first time I looked

NYJ -13%

So, the Jets were replacement level?

4
by stan (not verified) :: Tue, 01/09/2007 - 4:03pm

Aaron,

Do you have QB numbers? I am amazed at the national sportswriters take on Peyton Manning's game. I don't know how much the INTs hammer DVOA, but I have to believe that 78% completion percentage and 7 ypa ought to grade out reasonably well.

5
by Ch V Kalyan (not verified) :: Tue, 01/09/2007 - 4:05pm

I am surprised to see a huge difference between NE (45%) and NYJ (-13%) as i felt that the game was close through the first three quarters and only opened up in the fourth quarter.

Is there any particular reason why the Pats were able to build this huge difference.

What is the DVOA ratings for Pats and Jets at the end of third quarter?

6
by Not saying (not verified) :: Tue, 01/09/2007 - 4:07pm

Well, Baltimore moves up in offense to join the Pats as the only teams top 10 in all three categories. I guess that's a function of the weighting, as the Ravens offense seemed to get much better toward the end of the season, although I didn't watch them closely.

Given their closeness on offense and the article about Polian's excrement, I'd say this definitely makes them the favorite.

7
by B (not verified) :: Tue, 01/09/2007 - 4:07pm

stan: QB DPAR numbers are in the Quick Reads column.

Heh, five AFC teams, one of whom isn't even in the playoffs, are all ranked above the highest NFC team. the NFC really is the inferior conference this year.

8
by Not saying (not verified) :: Tue, 01/09/2007 - 4:09pm

re: 4 Do you have QB numbers?

See Quick Reads. At least by DPAR, Peyton wasn't very impressive.

9
by Rich Conley (not verified) :: Tue, 01/09/2007 - 4:09pm

5

Kalyan, the pats destroyed the Jets the first half. I think the Jets had 2 first downs in the first half. The problem was, one of those first downs was a 75+ yard completion for a touchdown. Between that completion, and the dillon fumble at the 10, you go from a game that should have been 28-0 at the half, to 17-10 at the half.

The Jets basically got all of their production in the first half out of 2 plays, and then got manhandled for the rest of the half.

10
by Becephalus (not verified) :: Tue, 01/09/2007 - 4:17pm

5 Game was not close.

11
by Digit (not verified) :: Tue, 01/09/2007 - 4:23pm

You know, I was telling Sid the same thing- that even though the score of NE-NYJ was close, I didn't feel like the game was -close-. Thanks for quantifying -how-.

12
by Bobman (not verified) :: Tue, 01/09/2007 - 4:23pm

Aaron, How about the biggest one-game differential? 100% vs -82% looks mighty... well, mighty.

I am surprised Indy is above SD, but I assume that's simply because they played one more game (and is happened to be a blowout over a half-decent team). Jac, on the other hand, still confounds me. I suspect I am not alone there.

Hey, where the hell is Seattle on that chart... oh wait, I didn't scroll down far enough. Yikes!

13
by stan (not verified) :: Tue, 01/09/2007 - 4:26pm

Wow, DVOA really emphasizes the picks! I have to wonder about that emphasis, just as the emphasis on TD passes in passer rating seems misplaced.

In the end, there isn't any stat to adequately assess Peyton's offense against the Chiefs. The Colt running game was entirely a function of the Chiefs' obsession with stopping the pass. How do you measure the value of a threat that is so feared that the other team freaks out in its defensive gameplan to stop it?

It would be as if an offense decided to double team a pass rusher on every play and chip with a receiver as well. That defender might never get in on a tackle, yet he would be the dominating influence on the game.

14
by Pat (not verified) :: Tue, 01/09/2007 - 4:27pm

#7: Eh. It was worse in 2004. The difference was that in 2004, there was a team that was comparable to the top AFC teams - Philly (if you exclude weeks 16-17, Philly was within a few percent DVOA of New England by the end of the season). It's doubtful anyone else in the NFC would've had a winning record in the AFC.

This year, there are three teams that are AFC-quality (PHI, NO, CHI) but they're all a clear tier below.

So it depends if you think one Baltimore/New England team is worth three Jacksonville-ish teams.

15
by kibbles (not verified) :: Tue, 01/09/2007 - 4:29pm

Re #7: Heh, five AFC teams, one of whom isn’t even in the playoffs, are all ranked above the highest NFC team. the NFC really is the inferior conference this year.
It's been that way for a while. Last year, the top 4 teams were all AFC teams, one of which (KC) didn't make the playoffs.

If you really want to blow your mind, check out the 2004 stats. Ten of the top Eleven teams were all in the AFC- meaning that there were FOUR AFC non-playoff teams that were all better than the second-best NFC team. If the playoffs were based on DVOA instead of wins and conferences, only two NFC teams would have made it (compared to 10 AFC teams), and one of those would have been the last team in. Craziest conference disparity in the history of the NFL.

16
by Pat (not verified) :: Tue, 01/09/2007 - 4:30pm

#13: That's called The Tennessee Run Defense Problem (well, an offensive equivalent). See the Glossary in About the Site above.

17
by Randy S. (not verified) :: Tue, 01/09/2007 - 4:32pm

Philly had +7% defensive DVOA? That doesn't exactly bode well for playing against New Orleans.

18
by bsr (not verified) :: Tue, 01/09/2007 - 4:33pm

#13 I don't really see what the problem is with including the picks. Its not like they were really fluky tipped balls or anything. I would assume that most are a result of missed reads or bad passes like those were. It happens. The same critiques that you bring up about this one performance is true of many games and performances. LT faces probably more 8 man fronts then anyone. Chad Johnson faces an inordinate amount of double teams. And KC isn't the first team devoted to stopping the pass...

19
by Pat (not verified) :: Tue, 01/09/2007 - 4:35pm

#15: It should also be pointed out that that year gave us a close Super Bowl even when everyone was wanting to call it a blowout due to the conference disparity.

I wouldn't go crowning the AFC the winner of the Super Bowl yet, but unless Chicago/New Orleans/Philly step it up a bit in the next weeks (which they certainly could do), this year might be the year where the "conference disparity" prediction comes true.

20
by bsr (not verified) :: Tue, 01/09/2007 - 4:35pm

#15 - I bet you the 80s would be just as bad if not worse towards the NFC. AFC teams had no chance back then. Nowadays I still wouldn't be suprised to say an NFC champion any given year.

And what the heck are you doing here anyways. Shouldn't you be out celebrating.

21
by MDD (not verified) :: Tue, 01/09/2007 - 4:37pm

The walking wounded Seahawks are a couple lucky bounces away from the NFC Championship! (Cue maniacal laughter) AH HA HA HA HA HA HA HA!

22
by Pat (not verified) :: Tue, 01/09/2007 - 4:37pm

Philly had +7% defensive DVOA? That doesn’t exactly bode well for playing against New Orleans.

Money says that was rush DVOA, and New Orleans's strength is passing. Not so terrible.

23
by TG (not verified) :: Tue, 01/09/2007 - 4:38pm

Wow, Seattle has a game that is nearly 20% better than their weighted season DVOA and drops one spot in the season rankings. Must have something to do with improving SF's strength of schedule as a result of this game.

24
by Rich Conley (not verified) :: Tue, 01/09/2007 - 4:43pm

12

Here is the end of SD's schedule:
Nov 26 Oakland Won 21-14
Dec 3 @Buffalo Won 24-21
Dec 10 Denver Won 48-20
Dec 17 Kansas City Won 20-9
Dec 24 @Seattle Won 20-17
Dec 31 Arizona Won 27-20

Other than the win over Denver, they've basically barely beat a whole bunch of crappy teams. I'm sure thats got something to do with the low weighted DVOA,and the drop in their total DVOA. Plus, Indy gets a huge boost for that last game.

25
by PantsB (not verified) :: Tue, 01/09/2007 - 4:43pm

Doesn't it seem a bit ridiculous for GB and Car to have better rated Defenses than NE?

SD at 17th makes sense because they've actually given up more points than their opponent's average points.

26
by Becephalus (not verified) :: Tue, 01/09/2007 - 4:46pm

19 That is a very good point. People need to realize that even with the disparity I would still give any random AFC playoff team at minimum a 1/4 chance of beating any random NFC playoff team.

Parity rules all.

27
by Setzer (not verified) :: Tue, 01/09/2007 - 4:50pm

Wow, so the Giants' poor special teams was the difference in the NYG/PHI game, no?

28
by Pat (not verified) :: Tue, 01/09/2007 - 5:01pm

#26: And more important than that, the NFC team is likely to cover. :)

(In NE/PHI, NE was favored by 7 points, and by DVOA, it should've been more like 1 or 2.)

29
by doktarr (not verified) :: Tue, 01/09/2007 - 5:03pm

stan,

DPAR's emphasis on picks is statistically accurate. Of course not all picks are created equal, and some are more the fault of the receiver than the QB, but overall, their negative impact tends to grade out at where DPAR sets it. So DPAR puts Peyton's performance at only slightly above replacement level.

Of course you're right that DPAR fails to measure the indirect impact a player has on the game. Given the current level of play-by-play information going into the DPAR engine, it's more or less impossible for it to do so. Perhaps if the PBP data contained information on defensive formations, how deep the safeties played, who was double teamed, how much time the QB had before he was pressured, and so on, then we could get a better picture of the indirect impact players have on the game. But currently, that's beyond the scope of our tools.

Bottom line, DPAR is just a tool of evaluation and can only tell part of the story. I tend to put more stock in team DVOA because it is looking at things from an 11-on-11 perspective and all these indirect effects are encapsulated in that. That said, DPAR is probably a more accurate individual measure than anything else out there.

30
by B (not verified) :: Tue, 01/09/2007 - 5:07pm

21: More like the walking wounded Seahawks are a couple Grossman interception/fumbles from the NFC Championship game.

31
by DB (not verified) :: Tue, 01/09/2007 - 5:14pm

Maybe I'm reading the chart wrong, but how does Pittsburgh drop in the rankings from 9 to 12 when the teams that hopped them are all out of the playoffs as well? Is it all opponent adjustment changing based on the playoffs or something?

Or is the "last week" column not what I think it is?

32
by Pat (not verified) :: Tue, 01/09/2007 - 5:17pm

#31: The weighting moves forward a week too. Teams not in the playoffs just act like they had a bye. So Pittsburgh's late season is becoming more important.

33
by Rich Conley (not verified) :: Tue, 01/09/2007 - 5:28pm

31.

In more important news, Oakland is no longer the worst team in football.

34
by DB (not verified) :: Tue, 01/09/2007 - 5:38pm

RE 32 : Wouldn't this move Pittsburgh up? Sure, the Baltimore losses count more, but so do the wins; I thought their DVOA later in the season was better than earlier, but I could be mistaken. By the same token Cincinnati didn't play well down the stretch, yet moved up three spots. I haven't looked at the week to week DVOA, but my general impression was that it's weird to see Cincinnati hop up that much while the Steelers moved down that much.

35
by jonnyblazin (not verified) :: Tue, 01/09/2007 - 5:39pm

So if Samari Rolle's stop rate was at 30% at some point this year and now is 43%, then his weighted stop rate should be pretty good, right? I think he certainly has been playing better, and the Ravens secondary has been pretty sharp the second half of the year.
I just find it odd that the Ravens have the best pass D in the league and yet the game charting data shows they have 2 major liabilities in coverage in Reed and Rolle. Their pass rush is great but I don't see how it could make up for such poor play. But if opponents have only throw to Reed 21 times this year that probably means they have been avoiding him.
And I know passes defensed is an arbitrary stat, but how is it that the charters have Reed getting beat 18 out of 21 times and the NFL has 9 passes defensed and 5 picks? Missing a game or two I don't think would make up for this discrepancy.

36
by Richard (not verified) :: Tue, 01/09/2007 - 5:39pm

33: Indeed, it's a sad day.

37
by paytonrules (not verified) :: Tue, 01/09/2007 - 5:46pm

What about non-weighted DVOA.

38
by Lance S. (not verified) :: Tue, 01/09/2007 - 5:46pm

#20

There really wasn't that much disparity in the '80s. NFC always won the SB but intraconference records were pretty even for the most part. These last few years intraconference play has been very uneven.

39
by Pat (not verified) :: Tue, 01/09/2007 - 6:03pm

34: Kansas City's their best game of the year, and it's now seriously dropping in importance. Wins don't necessarily move you up - their wins came against pretty bad teams.

40
by MCS (not verified) :: Tue, 01/09/2007 - 6:14pm

In #15 Kibbles said, "Craziest conference disparity in the history of the NFL."

It is pretty crazy, but craziest? Forever is a mighty long time. FO has only been around recently. Prior to FO we don't really have a good metric for comparison purposes.

Remember when the NFC won 13 Super Bowls in a row? Every year, the AFC would struggle to find someone to send to the SB and then that team would usually get their a$$ handed to them. Year after year.

The average margin of victory was 21 points. Over 13 Years

I really don't feel like researching more to see how the best on the AFC stacked up against the NFC also-rans during that era. I just rememember wondering when the AFC was gonna come around.

Now, it looks like we're on the flipside of that.

41
by James G (not verified) :: Tue, 01/09/2007 - 6:16pm

23 - Perhaps it's the weighted formula as Seattle's demolition of the NY Giants gets farther back in the rear view mirror (or was that game already gone from weighted DVOA?)

42
by James G (not verified) :: Tue, 01/09/2007 - 6:23pm

Another thing I once kept track of the - the interconference standings (click my name for that data from 1970-1998).

A couple NFC top heavy years really do stand out: 1991 was 33-19, 1992 was 30-22, and 1995 was 33-27.

When SF mauled Denver, on year I thought the NFC was totally dominant, the series was only 28-24 in favor of the NFC. When Chicago crushed New England, it was actually 27-25 in favor of the AFC. I remember that year Denver was denied a playoff spot at 11-5, but I think the best AFC team was only 12-4.

43
by navin (not verified) :: Tue, 01/09/2007 - 6:31pm

Check out the premium stats on Green Bay. They had a ridiculous third down defense--and I'm guessing it got more ridiculous as the season improved.

That's one thing I'd be worried about as a Green Bay fan--they're going to be worse on defense next year than they were this year because that kind of a third down stop rate is unsustainable.

44
by Pat (not verified) :: Tue, 01/09/2007 - 6:33pm

Personally, I think that has everything to do with coaching stability - which is, in turn, produced by winning. So you get this situation where winning produces winning, and losing produces losing, as long-term coaches are forgiven for short stretches of problems (Jeff Fisher), and short-term coaches are quickly tossed at the first sign of trouble (Jim Mora, Jr) seemingly regardless of early success, which means clawing out of the morass of ineptitude is really hard.

I mean, how many competent front offices does the NFC have? Philly. Maybe Seattle, though I'd even put that down a notch (losing Hutchinson, paying Alexander that much, paying Branch that much). Maybe Carolina. Uh. OK, now I'm out of teams. Anyone?

Maybe New Orleans will be a new, solid franchise (but fear the Jim Mora, Jr effect). Maybe Chicago will (but defensive brilliance is hard to maintain).

But other than those five teams, what team in the NFC doesn't have a rookie or second year coach, an open head coaching job, or a coach that people aren't sure is going to stick around, and/or be fired?

45
by Moltar (not verified) :: Tue, 01/09/2007 - 6:35pm

23: Seattle's drop can also be attributed to Houston climbing the ranks, in part perhaps because of their victory over Indy, who are now ranked #3.

Not sure what's happening with Arizona, though. A #20 ranking, up 6 spots from last week. Maybe Dennis Green wasn't such a bad coach after all.

46
by The Grizzle (not verified) :: Tue, 01/09/2007 - 6:36pm

In addition to DVOA, I also like to use a system I've created. Basically, on offense, a stat I call PVOO (point value on offense) a team gets six Value Points for every touchdown, three for every field goal, two for every two-point conversion and one for every extra point. This adds up to get a PVOO for the season, or you can break it down by week, quarter, whatever. The flip side is a stat I call PVOD (point value on defense), where a defense get charged six Value Points for every touchdown, three for every field goal, two for every two-point conversion and one for every extra point. The do however, get do subtract defensive value points in the amount of two value points in the event of a safety. I'm still working on special teams.

Essentially, I use this system along with DVOA to gauge offense and defensive team success...

47
by navin (not verified) :: Tue, 01/09/2007 - 6:36pm

Back in the 80s/90s the top NFC teams were dominant over their AFC counterparts, but the weak NFC teams were also a lot worse. Think of two bell curves with the NFC one having a higher variance. So the end result was the two conferences played even--and the AFC even won more Pro Bowls, if that even matters.

48
by The Grizzle (not verified) :: Tue, 01/09/2007 - 6:39pm

re: 46

any suggestions for special teams?

49
by M (not verified) :: Tue, 01/09/2007 - 6:42pm

AFC vs NFC discrepancy

The NFC was dominant in 1970, 1971 - then the next 9 seasons were AFC-dominated, with the worst being 1979 (36-16). During the NFC's "dominant" era, the had the edge only a few times: 1981, 1989, 1991, 1992, 1994, 1995. There were quite a few years when the conferences were tied, and a few small discrepancies toward the AFC. The AFC's run actually started in 1996, as they have had the edge every year except 2000 & 2001, when the conferences were tied. Of all the seasons, 1979 & 2004 are the most lopsided toward the AFC. 1980, 1999 & 2006 were lopsided as well, and the NFC had lopsided edges in 1970 & 1991. #15 has it very close, though. One last thing - weren't there ALOT of blowouts by AFC teams over NFC teams that year? I don't remember it swinging the other way often.

50
by Gerry (not verified) :: Tue, 01/09/2007 - 6:42pm

"Remember when the NFC won 13 Super Bowls in a row? Every year, the AFC would struggle to find someone to send to the SB and then that team would usually get their a$$ handed to them. Year after year.

The average margin of victory was 21 points. Over 13 Years"

That was domination in the SB, no question. But oddly, back then the worst in the AFC was better than the worst in the NFC. I just looked at most of the 80s and early 90s, and more times than not the AFC had the better overall interconference record than the NFC. I was very surprised at that, actually.

51
by pawnking (not verified) :: Tue, 01/09/2007 - 6:45pm

I wonder how each team would look if their best and worst DVOA games were removed?

52
by Yaguar (not verified) :: Tue, 01/09/2007 - 6:50pm

That NFC dominant era had a lot to do with a handful of teams, not overall conference superiority. It was basically just the Gibbs Redskins, the Emmitt Smith Cowboys, the Lawrence Taylor Giants, and the Jerry Rice 49ers.

If you looked at the typical NFC team and the typical AFC team from that era, there wouldn't be that much of a difference. It was just the very top of the NFC that was unstoppable.

53
by B (not verified) :: Tue, 01/09/2007 - 7:02pm

45: You have it backwards, Green was such a bad coach that firing him retroactively raises them 6 spots.

54
by Pat (not verified) :: Tue, 01/09/2007 - 7:18pm

#51: Been tried. Doesn't help things - just makes them worse.

55
by Wanker79 (not verified) :: Tue, 01/09/2007 - 7:20pm

Re: 22
Money says that was rush DVOA, and New Orleans’s strength is passing. Not so terrible.

*cough*Lito*cough* sonofa...

56
by Kachunk (not verified) :: Tue, 01/09/2007 - 7:53pm

RE 44: Um... Steelers? You have to count them based on past record although with Cowher leaving I can see how future competence would be in doubt.

57
by PantsB (not verified) :: Tue, 01/09/2007 - 7:53pm

Check out the premium stats on Green Bay. They had a ridiculous third down defense–and I’m guessing it got more ridiculous as the season improved.
>>>>>
New England was #2 in opponentent's ppg and were within .5 ppg of Baltimore on purely offensive scores (discounting returns). Putting them behind GB (25th) and Car (8th- tied), suggests that the formula is flawed. It is failing to accurately reflect what has happened.

New England also was 2nd in takeaways, and picks (and 4th in fumble takeaways). They were superior to GB on 3rd down % and while Carolina has a better 3rd down %, the Panthers gave up more total first downs.

This can't be explained by weighting since New England is #8 in total season (unweighted) defense.

58
by Bobman (not verified) :: Tue, 01/09/2007 - 8:02pm

Pat #16, thanks for the tip. I had no idea about the Glossary. Most useful, and I've been coming here since the beginning. I love the Tenn D conundrum in concept.

But, for offense, it doesn't quite describe how teams play Indy every once in a while (Chicago in 2004 when Edge had a 210 yard game, Philly earlier this year, KC last week.): they sell their souls to the devil to stop the pass at the expense of the run--and get ripped.

As I read the definition in the glossary, the complete analogy would be that when the D sells out to stop the pass and Indy runs, they're so surprised to have holes to run through and other opportunities that they screw up a bit. No?

59
by Nathan (not verified) :: Tue, 01/09/2007 - 8:07pm

Really want someone to answer #35

60
by Pat (not verified) :: Tue, 01/09/2007 - 8:12pm

#58: Not quite. The idea is that you get a false sense from another unit because the first unit is so dominant. So the Indy run offense isn't quite as good as it looks because teams are so scared of their pass offense.

#57: Points per game isn't an unbiased measure of a defense. The basic unit of a defense is a drive, not a game. And New England's points per whatever (including points per drive) is so good because of their ridiculous red zone defensive DVOA (they're third in red zone defensive DVOA, behind BAL and MIN).

But a defense's job isn't to prevent points. It's to stop drives. Preventing points is better than allowing points, but stopping a drive right away is best. Allowing a drive to continue (giving up yardage) hurts your offense, even if it doesn't affect your points allowed stat.

I think a question should be added to the "FO FAQ" regarding "why isn't X ranked higher on defense? they rank really high in points allowed per game!"

See also: Denver, beginning of the year.

61
by Pat (not verified) :: Tue, 01/09/2007 - 8:13pm

#56: The Steelers are shifting into the NFC? Well, I would agree that would balance things a bit more...

62
by doktarr (not verified) :: Tue, 01/09/2007 - 8:14pm

RE: 58,

I think the analogy better applies to Indy's pass offense. That is, the D sells out to stop the pass, and as a result Indy passes less often and is less effective doing so, and (on this day) looks rather average in passing offense. As such, it's hard to tell whether teams are justifiably afraid of Indy's passing offense, or whether Indy's passing offense is actually average and teams are overreacting by allowing them to run.

In this case, I think we have enough alternate data points to know that the fear of the Indy passing attack is well justified.

63
by Pat on the back (not verified) :: Tue, 01/09/2007 - 8:40pm

Coaches salaries being high:

The issue isn't one of demand or supply, or even training, per se. The economic issue is that of hidden information. Basically, it is incredibly hard to exhibit ones' true ability as a coach.

Since the line between greatly effective and grossly incompetent at the NFL level is apparently so thin, and the track to proving yourself is so hazy (how does one prove or become educated enough to be a great coach), any ability to distinguish oneself from the pack is highly valued.

The issue isn't "can we raise or lower a coaches salary and get a different quality of coach". Most anyone in the coaching ranks would accept just about any head coaching job for $100,000 a year. The problem is that owners are not willing to hand over their franchise to a 3rd year TE coach on the cheap unless the person had demonstrated somehow that it might be a good idea to do so. If they had done somehow shown that they might make a solid head coach, well, some other owner would also probably be able to see that and bid up his price tag.

So, "established" coaches with good track records are few and far between, and the opportunity cost of a bad coach is so high, that wages for those "hot" coaching prospects is through the roof.

Also, this is why retreads are so common, because either 1) the coach had some success at some point in the past and the owner hopes he can recapture it or 2) the coach displays again what he did that got him his first job and makes him attractive even if he still is going to fail like he did the prior two times (See: Carrol, Pete).

Finally, the dynamic nature of the sport as it pertains to strategy muddles the whole issue up even further. Just ask Joe Gibbs if his strategies that won him 3 SBs are still working.

So, basically, coaches are really unable to demonstrate with any reliability their "true ability", if you will. As such, owners throw a lot of money and the limited few coaches that distinguish themselves from the crowd and outbid each other for the in demand guys that seem to be a good bet.

Really, the best tactic might be to do what the Eagles did Reid, where they found a coach they liked before he was noticed by other teams and acted definitively. Then again, you could end up with the next Mora Jr doing that.

64
by Nick (not verified) :: Tue, 01/09/2007 - 10:27pm

I love it. The Lowest ranked team out of the lowest ranked conference is still in the playoffs. It's a crazy world.

65
by Dean from Oz (not verified) :: Tue, 01/09/2007 - 11:08pm

Somewhat off topic (weekly rankings) but on topic (weekend games);

A lot is being made of Gramatica's feeble attempt to make a block, notwithstanding his size and usual role.

But what if instead of trying to block legally, he did it illegally - I dont particularly care how, pull his head off if necessary - so much so that Romo gets the first down, and the penalty gets called.

If Seattle accept, its still 4th down and a short FG. If Seattle decline, its 1st down and a short FG. Either way I dont see Dallas stuffing it up a 2nd time.

What am I missing here?

66
by Will Allen (not verified) :: Tue, 01/09/2007 - 11:14pm

The most important thing an NFL head coach does is evaluate talent, both in players and assistant coaches. Until somebody gets the job, it is really, really, really, hard to determine whether they will perform those roles well. A coordinator usually has expertise on only one side of the ball, and it can be hard to tell whether the coordinator or his boss was evaluating the talent on that side. Until a coach hires and manages a staff, it is even harder to know whether he will be any good at that task. There really is a lot of luck involved, and then, even if the right guy is hired, he still may not succeed, due to factors beyond his control. Marv Levy was a pretty darned good coach, but nobody was beating down his door to give him another head job, before the Bills did so.

Of the current coordinators I am familiar with, I think Mike Tomlin of the Vikings has an outstanding chance to be a very good head coach, if he gets some luck with regards to who hires him. I kinda' wish he was the Vikings' head coach, actually.

67
by Will Allen (not verified) :: Tue, 01/09/2007 - 11:20pm

Dean, that's an interesting point, and perhaps a crazy for detail coach will make it part of his preparation one day; tell his kicking unit that if a hold is botched on a field goal, everybody should just tackle the nearest defender!

68
by B (not verified) :: Tue, 01/09/2007 - 11:46pm

35: Pass is more important than coverage in pass defense. Samari Rolle stinks, but Reed is a safety not a corner back, so when he's fingered as the guy who blows coverage he's more likely to be cleaning up somebody else's mess.

69
by Pat on the Back (not verified) :: Wed, 01/10/2007 - 12:09am

#63

Hmm... my post seems to have ended up in the wrong discussion thread...

70
by Zac (not verified) :: Wed, 01/10/2007 - 12:11am

Dean reminds me of something that happened in the Capital One Bowl. The Badgers had a punt blocked, and it bounced back to the punter. He threw a pass to someone who got a first down. But they were penalized for having an illegal man downfield (no surprise there) so they got a 5 yard penalty and the chance to rekick.

I still don't know what the point of having a QB as the holder is if he isn't going to throw the ball in that situation.

71
by saneiac (not verified) :: Wed, 01/10/2007 - 12:52am

70:
The point is that the QB is the guy who handles the ball more often than anyone else on the team. In theory, this should make him less prone to bobbling or dropping the snap.

Green Bay moves into the top 10 by beating Chicago's scrubs (including Rex Grossman) in a meaningless game. I watched every Green Bay game this year, and believe me, they are not a top 10 team.

72
by RichConley (not verified) :: Wed, 01/10/2007 - 1:39am

re:71
Chicago played their starters, not the scrubs.

73
by Brett (not verified) :: Wed, 01/10/2007 - 1:47am

#71 and 72:

Bears played most of their starters for a half, while resting anyone with anything more than a hangnail for the whole game.

The Packers kicked the Bear's tail, but you can't tell me the lack of motivation, and a first string lineup, for the whole game didn't make a difference. DVOA should take that into account.

74
by tanner (not verified) :: Wed, 01/10/2007 - 1:59am

Am I missing the link to the Fox commentary? (It's late, I'm stupid.)

75
by Scott de B. (not verified) :: Wed, 01/10/2007 - 2:02am

The Packers kicked the Bear’s tail, but you can’t tell me the lack of motivation, and a first string lineup, for the whole game didn’t make a difference. DVOA should take that into account.

Going through past seasons shows that 'meaningless' games are just as predictive as any other game.

76
by David Ferrier (not verified) :: Wed, 01/10/2007 - 2:06am

The Packers did better in the first half against the Bears' starters than the 2nd half, so that confuses things a bit more. If you want to disregard the 2nd half of that game it would help the Packers ratings even more I would think. They were up 23-0 at halftime.

77
by BillWallace (not verified) :: Wed, 01/10/2007 - 2:28am

re: Dean

I think anyone who's making much of Gramatica's lack of a block is being really unfair. First you have to realize that he's not thinking about anything but kicking the ball and he had to be stunned that it wasn't there. Then he was headed forward, and had to stop, turn a bit, and try to get a block on an already running safety from not a great angle. he didn't have time to set himself, just reach out his arms. Expecting anything there is stupid.

You make a good point, but who's going to think to do that?

78
by Yosi Scharf (not verified) :: Wed, 01/10/2007 - 2:47am

The biggest problem the AFC had during the years of NFC domination was that the #1 seed in the AFC was consistently defeated in the playoffs. That is, a cinderella from the AFC was playing the best team in the NFC. The results were predictable. The '85 Bears against the Patriots is perhaps the most egregious of these.

79
by kc (not verified) :: Wed, 01/10/2007 - 3:19am

Regarding manning's picks, why not just think of the rating the same way you would think about a QB who had 48 picks in a season? Now it makes more sense huh?

80
by Bobman (not verified) :: Wed, 01/10/2007 - 5:11am

#79 KC... and is much more applicable to Manning, given his 9 picks this year. (sarcasm)

36 picks vs Baltimore should put him right at that 48 mark for tghe year.

81
by Rick (not verified) :: Wed, 01/10/2007 - 9:16am

re: 78

The Buffalo Bills had the best (or were tied for the best) record each of their four Super Bowl seasons.

Going year by year

year SB team record best AFC record
1984 Dolphins 14-2 Dolphins
1985 Pats 11-5 Miami, Raiders 12-4
1986 Broncos 11-5 Browns 12-4
1987 Broncos 10-4-1 Broncos
1988 Bengals 12-4 Bengals, Bills 12-4
1989 Broncos 11-5 Broncos
(notable no other team had more than 9 wins)
1990 Bills 13-3 Bills
1991 Bills 13-3 Bills
1992 Bills 11-5 4 teams at 11-5
1993 Bills 12-4 Bills, Oilers 12-4
1994 Chargers 11-5 Steelers 12-4
1995 Steelers 11-5 Chiefs 13-3
1996 Patriots 11-5 Broncos 13-3

You've made the mistake of thinking "data" is the plural of "anecdote". What you call the "biggest problem" only happened 5 times in 13 years, and 3 of the 5 times it was an 11-5 team going instead of a 12-4 team (not a huge difference IMHO). Only in the last two years of the run were there serious upsets in the AFC, and those two years saw Super Bowls that were relatively close by comparison with the games in the previous decade.

The biggest problem was that the best AFC teams were nowhere near as good as the best NFC teams for this period. It was not that the best AFC teams kept being upset in the playoffs.

82
by James G (not verified) :: Wed, 01/10/2007 - 9:46am

78, 81 - And we also saw that the NFC didn't necessarily send their top team either. In '88, the 10-6 49ers went instead of the 12-4 Bears, in '90, the 13-3 N.Y. Giants made it in place of the 14-2 49ers, in '92, the 13-3 Cowboys made it over the 14-2 49ers. That's 3/13, although certainly the 49ers-Benglas and the Bills-Giants were the tightest Super Bowls during that stretch.

83
by Andrew (not verified) :: Wed, 01/10/2007 - 10:51am

1 BAL 39.2%
2 NE 30.1%
3 IND 26.4%
4 SD 24.9%
6 PHI 15.2%
7 CHI 13.6%
8 NO 12.7%
27 SEA -14.6%

(singing) One of these is not like the others, one of these is not the same.(/singing)

84
by MCS (not verified) :: Wed, 01/10/2007 - 11:03am

Back to the Pack.

The GB defense has also been helped by two TD-less games in the weeks prior to the Chicago game. They did not give up a touchdown against either Detroit or Minnesota.

85
by Polaris (not verified) :: Wed, 01/10/2007 - 11:44am

#83

That's exactly the point I've been making off and on for most of the last part of the season. Now consider Seattle's WC opponent: Dallas.

Dallas rates 13 at 4.5% overall DVOA which while out of synch is at least somewhat in line with the lower part of that progression.

I am not trying to say that DVOA is useless. Far from. Instead I am trying to suggest that Seattle may be an example of (to borrow a term from particle physics) an "exotic", i.e. a single case instance where the conventional DVOA wisdom/assumptions might not apply fully. I would suggest that Seattle merits deeper analysis by F.O. to figure out why...not to invalidate the model but to strengthen it (to find out WHY seattle is a legit playoff contender....a simple fact now...in spite of a bad DVOA).

-Polaris

86
by Not saying (not verified) :: Wed, 01/10/2007 - 11:54am

Re: 35 So if Samari Rolle’s stop rate was at 30% at some point this year and now is 43%, then his weighted stop rate should be pretty good, right?

That's an interesting point. I think overall the Raven's Defensive DVOA has slipped a bit in terms of absolute numbers from the middle of the season (-32.8% in week 9), but not in ranking. That might be a problem with the defensive adjustments though (not enough data earlier).

If Rolle is playing better, that seems to take away, or at least reduce, their one glaring weakness. It would be interesting to see if coverage schemes have changed to help him or not.

Their pass rush is great but I don’t see how it could make up for such poor play.

If you have only 2 seconds to throw, it seems like even the worst cornerbacks can cover for that long.

if opponents have only throw to Reed 21 times this year that probably means they have been avoiding him.

Aaron mentioned this, small sample size, and also the difficulty in saying whether Reed was the primary defender or not. (I think the safety and linebacker numbers from the charting project are shakier than the cornerback numbers, since corners are much more likely to be in obvious man coverage.)

how is it that the charters have Reed getting beat 18 out of 21 times and the NFL has 9 passes defensed and 5 picks

First, 29% stop rate on 21 passes would seem to be 15, not 18, successes. Second, are picks counted separately from passes defensed or does a pick count for both? Because then we could account for 6 out of 9, and 1 or 2 games could easily make up that difference.

Third, I think the charting numbers are only about times when Reed is the primary defender. That's where I think Aaron's quote from above comes into play. I'm not sure how charting works when a safety comes to help a CB who's in man coverage.

87
by Not saying (not verified) :: Wed, 01/10/2007 - 12:04pm

Re: 85 to find out WHY seattle is a legit playoff contender

Depends what you mean by legitimate playoff contender. If you mean a team that with homefield advantage (in Seattle's case, one of the best in the league) and a few lucky breaks can pick off other low ranked (or unlegitimate, if you like) playoff foes, DVOA already shows that. Homefield is usually looked at as about 17% DVOA (I think), which puts the difference between Seattle and Dallas at 2.1% (Any Given Sunday).

If you mean a team with that should be considered the equal of the top playoff teams, with a strong chance of winning the Super Bowl, I don't anyone really thinks that's true about Seattle, DVOA-based thinking or not.

88
by Polaris (not verified) :: Wed, 01/10/2007 - 12:30pm

#87

Seattle has not had very good homefield advantage this year. You are deliberately trying to cook a number to avoid the point....shame on you.

Seattle is a legitatmate contender because Seattle WON in the playoffs and did so against a team that was (by DVOA anyway) better. Moreover, it was NOT luck. Check the box scores and single game DVOA. Seattle was the better team. Period.

Given that, this indicates that the DVOA model while useful may be incomplete because it fails to account for an "exotic" like Seattle (and during the regular season there were other exotics). In fact per DVOA Seattle is the worst team in the NFC West. I am suggesting (and this is all I am suggesting) that there may be factors that the DVOA model doesn't include and perhaps should.

-Polaris

89
by james (not verified) :: Wed, 01/10/2007 - 12:47pm

looking at single game data for the first time

I now know the reason DVOA doesn't pick winners....

Games are not treated as a zero sum event.

No wonder

(banging head against wall)

90
by Polaris (not verified) :: Wed, 01/10/2007 - 1:03pm

#89

Your point is?

MY point is that the DVOA model needs to be examined to determine WHY some exotics (like Seattle) occure. If the game DVOA is completely out of synch with predicted DVOA then something is going on. In the case of Seattle, you will find that the DVOA for the last three games has been on an upward tick (because of a much improved defense) ranging from about 1-5%.

Not stellar, true, but more than you'd expect from the overall weighted DVOA. Smells like an exotic.

-Polaris

91
by Pat (not verified) :: Wed, 01/10/2007 - 1:12pm

#89: VOA is. DVOA isn't because of opponent adjustments.

Seattle is a legitatmate contender because Seattle WON in the playoffs and did so against a team that was (by DVOA anyway) better.

Can we drop this stupid "look! Seattle won! they must be better!" argument? That game (and the Eagles game) was as close as you could possibly imagine the game being. All teams have variances. Seattle's variance is roughly 10%. They had a DVOA of 5% this game, and a season average DVOA of roughly -15%. That's within two-sigma. You've been making similar arguments all year, and statistically, they're just wrong. The Seattle-Dallas game was not that surprising at all. You want surprising? Try the KC/IND game. That's a game that doesn't match with season-long DVOA. But all the other ones are easily consistent (PHI/NYG within 1 sigma on both, NE/NYJ within 1 sigma on both, DAL/SEA within 1 sigma/2 sigma, and only KC/IND with both outside 2 sigma).

Make the same argument when Seattle beats Chicago, the winner of NO/PHI, and the Super Bowl, okay?

92
by Polaris (not verified) :: Wed, 01/10/2007 - 1:20pm

Pat,

Temper, temper.

In fact my arguement is dead on. Seattle's DVOA was 5% and weighted DVOA was -14.5%. That's more than two sigmas and that's an exotic.

NEXT!

-Polaris

P.S. Look at the stat-lines as well. The score was closer than the game. Remember the any-given-sunday article w/r/t Arizona and the Bears. Same, same.

93
by Pat (not verified) :: Wed, 01/10/2007 - 1:35pm

In fact my arguement is dead on. Seattle’s DVOA was 5% and weighted DVOA was -14.5%. That’s more than two sigmas

Um.

No?

The difference between -14.5% and 5% is... 19.5%. Seattle's standard deviation is 10.4%. 19.5% is less than twice 10.4%.

and that’s an exotic.

Roughly 10% of a distribution will typically lie outside of two sigma. I don't know why you think things are so cut and dry.

94
by Pat (not verified) :: Wed, 01/10/2007 - 1:51pm

Oh, and this is starting to oh so much smell like a FOMBC. Weren't you saying the exact same things when Seattle was 8-4, and ranked 24th - right before their 0-3 run?

checks Week 14 DVOA thread

Yup. Ooh, playing with fire.

95
by Polaris (not verified) :: Wed, 01/10/2007 - 2:44pm

Pat,

Yes, and IT STILL DOES.

If anyone is playing with fire, you are. You can't admit that the system isn't perfect. Seattle has been an exotic all year long. Deal. [So was Pittsburg]

Seattle did indeed go on a 0-3 run but that does not invalidate my point especially when one of those losses was to San Diego by three points (which is in of itself a DVOA upset...look it up).

So yes, it's worthy of an "Any Given Sunday" article. You are just blinded by your slavish devotion to a system.

-Polaris

96
by zip (not verified) :: Wed, 01/10/2007 - 2:54pm

Seattle's game DVOA: 5%
Seattle's Weighted DVOA: -14.5%

Standard Home DVOA adjustment: 17%
Thus a team at home should play +8.5% DVOA,
and a road team should play -8.5%.

So Seattle would have been expected to play around -6%, since they were at home. So they deviated by barely a sigma.

Also, if Seattle hadn't willed Romo to botch that snap, we could've avoided all this.

97
by Andrew (not verified) :: Wed, 01/10/2007 - 2:57pm

Polaris:

to find out WHY seattle is a legit playoff contender….a simple fact now…in spite of a bad DVOA

Because they play in the freaking NFC West, and NFL rules dictate at least one NFC West team must be in the playoffs and hold homefield advantage in their first game.

Homefield advantage is worth about 17% in DVOA (which also translates to about 3 points), so Seattle at -14% and Dallas at +5% are not that far apart when playing in Seattle.

This isn't to say that they can't go on a hot streak for a few games and make some postseason upsets, like the 9-7 Colts in 1995 or the 9-7 Rams in 1979 or the 9-7 Jaguars in 1996, just that projecting them to do so for 3 or 4 games in a row is unreasonable, especially when they haven't won more than 3 games in a row all season (at the beginning, when they had perfect health).

98
by Polaris (not verified) :: Wed, 01/10/2007 - 2:59pm

Zip,

You are factoring in a "fudge factor" which AFAIK DVOA doesn't recognize (it's certainly not put in the weighted average) and then claiming luck and you STILL admit that Seattle is over a sigma.

You are making my case.

In addition, you are wrong about the Botch. Even IF Gramatica makes the FG, Seattle gets the ball back with about 1:14 left to play to return the KO and get in Josh Brown's FG range and Josh Brown is the most clutch kicker in the NFL. Seattle likely wins the game anyway especially given that Dallas was unable to stop Seattle's 2 minute offense (at least until the red zone which Seattle would not have to reach in this case).

In addition, let's also not forget that Dallas only got 13 offensive points, 5 less first downs, about 80 less total yards, 4 minutes less in T.O.P.....shall I go on??

-Polaris

99
by Polaris (not verified) :: Wed, 01/10/2007 - 3:01pm

Andrew,

WRONG! All NFC West teams play essentially the same schedule. Going by seasonal weighted DVOA, if what you claim was correct, then Arizona would have made the playoffs.

-Polaris

P.S. You do realize that the NFC West has the 2nd most wins in conference (behind the NFC East only). That suggests that the NFC West may not be as weak as you think.

100
by Polaris (not verified) :: Wed, 01/10/2007 - 3:05pm

Andrew,

At the beginning Seattle was not perfectly healthy. Review your facts. For starts, Alexander broke his foot the first game of the season. Also, Seattle had to put their initial starting safety (Green from Chicago) on IR. Injuries aren't an excuse, but the fact is Seattle hasn't been healthy all season long including their initial three win streak.

Given that, it's not unthinkable that Seattle could go on another streak. Personally I don't expect that. I do think that Seattle can beat Chicago (much more likely than you might think) because Chicago has played bad ball going into the playoffs. I don't think Seattle beats either the Eagles nor the Saints.

-Polaris

101
by Pat (not verified) :: Wed, 01/10/2007 - 3:07pm

#95: You are just blinded by your slavish devotion to a system.

Yes, that slavish devotion to "math", which you apparently don't have such a devotion to (see #93).

There's absolutely no evidence that Seattle's a significant outlier. Seattle was projected at midseason to finish with 9.1 wins. They finished 9-7. They bounced all year between +15% and -45%, which leads to a mean of about -15%, and a standard deviation of about 10-15% or so, and posting a 5% DVOA is entirely consistent with that.

I have no idea whatsoever why you think they're an outlier. Other than some bizarre desire to believe it.

102
by Polaris (not verified) :: Wed, 01/10/2007 - 3:11pm

101

Actually I am better at math than you think. Seattle was projected at 9.1 wins because when the projections were made, Seattle had already won six games! (Forgot to include that teeny, weeny, inconvenient fact)

Fact is that if you removed all the team names and W-L record, and had to pick the 9-7 NFC West champ based on DVOA, you'd get the Arizona Cardinals.

That tells me that the system is incomplete.

-Polaris

103
by Pat (not verified) :: Wed, 01/10/2007 - 3:23pm

#98: What in the devil are you talking about? What fudge factor? HFA is roughly +/-8.5% home/away. Aaron's mentioned that a half-dozen times. It's also entirely consistent with a 3-point HFA swing, which is what betting lines have known for years.

and you STILL admit that Seattle is over a sigma. You are making my case.

A team posting a number ~1 sigma from its season average is making your case?!? That's a heckuva weak case! Do you have any idea how little one sigma is?

104
by James G (not verified) :: Wed, 01/10/2007 - 3:26pm

The 17% is not a fudge-factor. It's been recognized as pretty standard for a while now. Also, Pittsburgh last year was not "an exotic." Take out the games that Roethlisberger missed, and the Steelers had a much better DVOA.

Interestingly, it's not just DVOA that has Seattle rated low - Sagarin has him at #17 in his overall ratings, and #20 in his predictive ratings. Seattle is also #20 in Dolphin predictive ratings and 18th in the Dolphin standard ratings.

Finally, there isn't anything all that exciting about these ratings anyway - Seattle was outscored on the season (341-335) while playing in the worst division in the league. They won one playoff game over a team that they'd likely beat about 40% of the time on a neutral field, but they played them at home.

105
by Waverly (not verified) :: Wed, 01/10/2007 - 3:30pm

Enough already.

Is it possible for a team to host a playoff game with just a single win?

All teams in division lose all games against non-division opponents, and they all tie each other within the division, except for one winner.

106
by zip (not verified) :: Wed, 01/10/2007 - 3:40pm

and you STILL admit that Seattle is over a sigma

DUDE. Are you sure you know what a sigma is?

Assuming a normal distribution, 32% of your data will be outside +/- 1 sigma. Seattle's performance in no way statistically proves, or even strongly suggests, that their DVOA isn't centered around -14.5.

Furthermore, your insistence that "you'd have to pick Arizona to win the NFC West, but Seattle won it, so DVOA must be flawed" is proof that you don't understand probability or expected outcomes. If you flip a weighted coin (60% heads, 40% tails) and get tails, does that mean it isn't weighted? How often do you need to get tails before you can say that the coin isn't weighted 60/40?

107
by James G (not verified) :: Wed, 01/10/2007 - 3:41pm

102 - What the heck? You would not get the Arizona Cardinals, who finished lower in DVOA than the Sehawks. You would get the St. Louis Rams, who finished 1 game behind Seattle. 1 game here or there for either team, and it's not that surprising at all.

108
by Pat (not verified) :: Wed, 01/10/2007 - 3:42pm

#102: That's weighted DVOA. Just the latter half of the season.

Strangely, if only the second half of the season is considered, Arizona did win the NFC West - all teams went 4-4, and Arizona swept everyone.

What's your next argument?

109
by Pat (not verified) :: Wed, 01/10/2007 - 3:45pm

#106: The Chernoff bound!

The answer to that is 150 coin flips (with 95% accuracy). Or a little over 9 full NFL seasons.

110
by Not saying (not verified) :: Wed, 01/10/2007 - 3:48pm

Re: 88 Seattle has not had very good homefield advantage this year. You are deliberately trying to cook a number to avoid the point….shame on you.

Just because they haven't won a lot of games at home does not mean that they didn't have a good homefield advantage. They're not a great team, so they didn't win that many, but the stadium is designed (acoustics) to provide a good advantage. I am not cooking any numbers. I don't think that was about any numbers.

Seattle was the better team. Period.

I was not disputing that. the point is just that they can play better some days then others and it is not outside their DVOA.

Re: 90 #89

Your point is?

Not everything that is written after your post is in response to what you said.

Re: 95 it’s worthy of an “Any Given Sunday� article. You are just blinded by your slavish devotion to a system.

1. If so, that would seem to indicate that Seattle is the worse team. That is the point of the AGS articles.
2. From the top: For those wondering, there’s no Any Given Sunday this week, as all four home teams won and Ned has nothing left to say about the Cowboys.

Re: 102 Fact is that if you removed all the team names and W-L record, and had to pick the 9-7 NFC West champ based on DVOA, you’d get the Arizona Cardinals.

1. This article is about weighted DVOA, meaning it doesn't even count the first 5 games. If you look at the final full season DVOA, Seattle's above Arizona. If you look at the numbers from the first 5 weeks, Arizona was terrible, which doesn't show up here at all.
2. Arizona is not that far ahead of Seattle, so if the teams played thousands of seasons as they are now, yes, on average Arizona would win more often, but it is not completely crazy that Seattle would do better.

111
by zip (not verified) :: Wed, 01/10/2007 - 4:03pm

#109 -- A solid answer to a rhetorical question :)

112
by Kulko (not verified) :: Wed, 01/10/2007 - 4:33pm

Re 102:
First: If I would guess winners, I would choose Total DVOA because having an easier schedule also makes for more wins.
Second: The best Team in the NFC West is STL. So when I see DVOA I would expect something like STL 9, SEA 7 ARI 7 SF 5.
Third: Seattle has gone 4-1 in games which were a draw (less than 3 points diff). So they got a bit luck here, but not much.
Especially they got 2-0 in games which were basically draws against STL, which probably gave them the division.

There are also other categories, where they got lucky, like opposing FG percentages but most look minor to me.
Forth: Lets take a look at these draws with STL.
In STL they obviously played well better then DVOA suggests, but that happens often in single games (e.g. NE-NYJ, JAX-HOU etc.).
In Seattle all was business as usual a slightly better team was held to a draw on the road and lost. If Seattle were a TOP 15 Team as you suggests, this would actually be a disappointment.
Fifth: DVOA doesn't include injuries, so one would hope the hawks are a bit better now that their season average. We all agree, that this is a fault of dvoa, but nobody came ever up with a good way to fix it.
Sixth: DVOA doesn't measure who won games, but who probaly wins more games, when you play often enough. This means, that every year, there are 1-3 teams which overplay their DVOA and the same vice versa. Ok the Hawks overplayed their DVOA a bit, but so did the Jets and Chi. Every Year some teams does and some doesn't. That is a simply effect that there are random nonpredictable events, which have relative great influence on the W/L Result combined with a small sample size.
Seventh: How about making some suggestion, what is so different about the Hawks?

113
by IsaiahC (not verified) :: Wed, 01/10/2007 - 4:36pm

Wow, I just read that Mike Zimmer has left dallas (link my name) How will that change the Cowboys?

114
by Travis (not verified) :: Wed, 01/10/2007 - 4:39pm

Is it possible for a team to host a playoff game with just a single win?

All teams in division lose all games against non-division opponents, and they all tie each other within the division, except for one winner.

0 wins actually (and a 0-10-6 record). Everybody in the division ties all 6 divisional games, loses in the rest, leaving the division winner to be determined by strength of schedule (the two uncommon conference games).

115
by Kulko (not verified) :: Wed, 01/10/2007 - 4:41pm

Re 110 AGS:
I would say IND-KC would be well worth an AGS, but EPC handled that. Although I must admit, I am still not really smarter, how Things could have gone so bad.

116
by Waverly (not verified) :: Wed, 01/10/2007 - 4:56pm

Re: 114:

Right. I bet such a team would be proud to enter the playoffs in such a manner.

What's the limit on how far negative a team's DVOA can go, for a game or for a season. I assume it's not a hard limit, but there must be some practical limit.

Then my real question is: how low a DVOA can a division-winning team have?

117
by jonnyblazin (not verified) :: Wed, 01/10/2007 - 5:14pm

re: 86
"Their pass rush is great but I don’t see how it could make up for such poor play.
If you have only 2 seconds to throw, it seems like even the worst cornerbacks can cover for that long."

Thats my point, though. The CBs #s should reflect that. Look at SD, who has the second best adj. sack rate. Their 3 CBs look pretty good stop rate-wise (Florence and Cromartie do, Jammer is just ok), they pressure the QB a ton, and yet their passing DVOA is just -3.2%. Baltimore's LBs must have insanely high stop rates to get -25.2% DVOA.

My bad on the Ed Reed numbers, I saw the 14.2yds per pass and confused it with the 29% stop rate. And you are right about INTs being counted as passes defensed, even though it seems counterintuitive. If a QB overthrows his WR and the ball goes into the arms of a deep safety, he's not defending anything, he's just catching the ball.

118
by Doug (not verified) :: Wed, 01/10/2007 - 5:18pm

Just to jog some memories here - I ran a guest column back in week 9 or so called "The Bears Trend"...I'm seeing a lot of comments on this thread about AFC dominance, NFC dominance in the 80's and early to mid 90's, etc. Read over the post again and check out the site (which can also be linked to through my name). Sounds like you guys have a lot to contribute as far as data and ideas, so let's have it.

119
by James G (not verified) :: Wed, 01/10/2007 - 5:31pm

As a charter, I know we have to mark a reason for an INT just like we do for an incompletion. If the interceptor really made a nice play, he would get credit for pass defensed, whereas if it's underthrown/overthrown/thrown ahead/thrown behind, there is no credit for a pass defensed.

120
by Not saying (not verified) :: Wed, 01/10/2007 - 5:57pm

Re: 117 Baltimore’s LBs must have insanely high stop rates to get -25.2% DVOA.

If you look at their splits, I think that shows up. Their weakest spot (-7%) (in terms of lowest absolute DVOA) is vs. #1 receivers. Their strongest spot (-42.9%, more than 12% better than the #2 D) is vs. RBs who are much more likely to be covered by linebackers.

You can also look at the #2 WRs (-29.5%) and see that there's a good chance that Rolle is dragging one number down significantly.

Re: 119

Yeah, but that's FO and charters, and has nothing to do with the official stats. I think there was an Fo blog post about this a while back that showed how arbitrary the passes defensed stat was. I think the Philly stat guy showed up as especially crazy.

121
by admin :: Wed, 01/10/2007 - 6:19pm

Hey, look kids. The DVOA commentary finally went up: http://msn.foxsports.com/nfl/story/6351676.

By the way, regarding Rolle's charting stats, better charting stats now doesn't necessarily mean getting better later in the season. Remember, we're still filling in holes from early in the season, so those new games in the data may have been a few weeks ago.

122
by B (not verified) :: Wed, 01/10/2007 - 6:47pm

121: I think these can be filed under the troubles with long posing delays.
Reasons for Giants Pessimism: Tom Coughlin will be back. He signed a 1 year contract extension, for reasons I am unable to comprehend.
Reasons for Jets Optimism: Bryan Thomas signed a 5 year, $25 million deal.

123
by Not saying (not verified) :: Wed, 01/10/2007 - 6:51pm

Re: Giants commentary

You probably wrote this yesterday, but it was announced I think today that Coughlin will be back next year. Not so much optimism.

In related news (also to the Cowboys commentary), apparently Parcells asked to be the new Giants GM, but was turned down. Raises questions about his return to Dallas, although he denied the report.

124
by Goathead (not verified) :: Wed, 01/10/2007 - 7:11pm

Sigh. As the G's lost on Sunday I was happy in the knowledge that this was the end of the Coughlin era. Not sure I can stand another season of shots of him looking like he just ate a bad oyster.

125
by thad (not verified) :: Wed, 01/10/2007 - 10:36pm

Polaris,
yes
please go on...

126
by Mike W (not verified) :: Wed, 01/10/2007 - 11:11pm

Under the Chiefs' 'reasons for pessimism' you forgot: Herm Edwards

127
by Pat (not verified) :: Thu, 01/11/2007 - 12:06am

#125: Bears fan?

128
by NF (not verified) :: Thu, 01/11/2007 - 12:49am

Philadelphia is clearly ranked too high because the Saints will destroy all who stand before them in the Superdome in January. Not being unfortunate enough to be an Eagles fan is way better than this. get backorz on teh active roster soon d-mac!!!

129
by NF (not verified) :: Thu, 01/11/2007 - 12:51am

Does the angry troll mad-libs work as a reverse jinx?

130
by mactbone (not verified) :: Thu, 01/11/2007 - 11:30am

Re 129:
Not really, it's the impassioned arguments that are clearly cogent for the author that work better. Also, multiple trolls are more powerful than one. Not sure if that works when it's really just one person with multiple aliases.

131
by Pat (not verified) :: Thu, 01/11/2007 - 1:37pm

that are clearly cogent for the author that work better

I like how you slipped 'for the author' in there. Well done. :)

132
by NF (not verified) :: Thu, 01/11/2007 - 3:36pm

Attempts to create a reverse jinx to help the Eagles aside, since 1993 the Saints are undefeated at home in the playoffs. Seriously.

If any Saints fans think that comment was completely uncalled for, I agree with them.

133
by Buzz (not verified) :: Thu, 01/11/2007 - 5:22pm

Just curious as to the predictability success on the weighted DVOA compared to the season DVOA. Seems like 16 games is already a low sample size (yes I understand 180 possessions, is still large) so does it make sense to eliminate games and create a smaller sample size, or does weighted DVOA typically predict better?

134
by thad (not verified) :: Thu, 01/11/2007 - 8:00pm

uhhh, no Pat, still pissed off Cowboys fan.

135
by Pat (not verified) :: Thu, 01/11/2007 - 8:52pm

#134: You're a week late, then! You needed to encourage him last week to guarantee a victory.

136
by Sid (not verified) :: Fri, 01/12/2007 - 2:32am

RE: 22

When you're missing your best cover corner, very terrible.

137
by Peter (not verified) :: Fri, 01/12/2007 - 11:51am

Dean,
Ironically, I felt the Dallas ST did a great job committing an illegal penatly in the DAL/Was fiasco. I forget the name, but someone tried to tackle Sean Taylor by the face mask to prevent the run back (since the clock was at :00). Unfortunately, the guy let go of the helmet and taylor subsequently ran into field goal range.

Also, is anyone else fearful that the current gang violence in NO may come to a head if the Saints lose at home tomorrow? Hopefully the NFL will have even additional security on hand, or else the eagles bus may not make it out of the lot!

138
by Pat (not verified) :: Fri, 01/12/2007 - 12:34pm

#136: Eh. The step down from Sheppard to Hood isn't that large. Hood's clearly got the skills to be a starting NFL corner, although this year he's been a little tentative. Given that he's a free agent after this year and there might not be a game after this one, I doubt he'll be tentative.

Besides, Lito's strength is really in deep passing (speed, mainly). Brees won't be throwing deep versus Philly - the Saints offensive line won't give him enough time versus Philly's pass rush (that's not an insult to the Saints offensive line - Philly's pass rush is very good). Last game, Brees had all of three deep passes the entire game.

139
by PantsB (not verified) :: Fri, 01/12/2007 - 4:07pm

Pat, #60
"But a defense’s job isn’t to prevent points."

No offense, but thats such a ... strange... position I almost don't know how to respond. A defense's job is exactly that - to prevent points. Stopping drives is only important so far as it stops points from being scored. The game hinges on points, not drives. A drive that is stopped at the 40 that results in 3 points is no more valuable than one that goes to the 5 and results in 3 points in the final measure.

This fundamental assumption would obviously penalize defenses that were bend-but-don't-break defenses for no appreciable reason. In terms of winning the game, a team that allows 8 long drives that result in 7 points altogether is no more or less valuable in any truly significant way than a team that has 7 3-and-outs and one long TD pass.

140
by Pat (not verified) :: Fri, 01/12/2007 - 4:36pm

#139: I love it when people challenge that position. A lot of people think it sounds odd, but it's absolutely correct. You're forgetting that field position is fluid.

Stopping drives is only important so far as it stops points from being scored.

No way. You're not realizing something - if you stop a drive, you've stopped points. Period. If you stop every drive from gaining a first down, the only ones which will gain points are when the opposing offense gets the ball already in the red zone. And that is not the defense's fault.

In terms of winning the game, a team that allows 8 long drives that result in 7 points altogether is no more or less valuable in any truly significant way than a team that has 7 3-and-outs and one long TD pass.

Wrong. Field position is fluid. A team that allows 8 long drives, yet only 1 touchdown, has put the offense in poor field position on 7 drives, assuming average special teams. In all likelihood, the score will be 7-0, and the team will lose.

In the other case, however, the defense has given the offense great field position on seven drives, assuming average special teams.

See Denver, early 2006. Denver's offense was not nearly as bad as people thought, because they were constantly in awful field position due to long drives by the defense which rarely resulted in points. Denver won most of those games, but far too much credit was given to the defense when in fact it was the combination of the defense and the offense that did allow them to win those games.

A drive that is stopped at the 40 that results in 3 points

A drive that is stopped at the 40 doesn't result in 3 points. It results in less than 1 point, since less than 1 in 3 field goals of 55+ yards are made.

A drive that is stopped at the 5 yard line results essentially in 3 points, considering short field goals are essentially never missed (save with Tony Romo).

You're presuming the outcome of the field goal, which the opposing defense has no control over. In any specific outcome, they might be worse, or better, in terms of points, but a defense that forces a 62-yard field goal attempt can't be considered equal to one that allows a 10-yard field goal attempt.

Otherwise you're just judging the defense by the opposing special teams.

A defense's goal is to stop drives. Stopping points comes as a result of that.

When people forget that, what happens is they assign blame to a defense when it should go to an offense, and vice versa.

141
by PantsB (not verified) :: Sat, 01/13/2007 - 4:47am

#140 I disagree. Case in point with this quote:
A drive that is stopped at the 40 doesn’t result in 3 points. It results in less than 1 point, since less than 1 in 3 field goals of 55+ yards are made.
This presumes that expected result is more important than actual result in terms of evaluating a team. In terms of determining who will win a particular game, it is not. Field position is important yes, and so a team that gives up statistically insignificantly more points but gives its offense better field position independent from the ST unit should be rated higher if takeaways are even.

However, if two teams over the course of a 16 game season give up essentially the same number of points and an overall defensive metric says one defense is much better than the other, unless SoS is widely divergent or this is a major outlier, the metric is broken. Baltimore gave up 21 defensive TDs, 17 FG and scored 5 times on defense. New England gave up 21 defensive TDs, 23 FGs and didn't score on defense. Their unweighted DVOA are -25.6%(1) and -8.4% (8). Clearly the Ravens should be rated higher, but the degree and the number of teams ahead of the Patriots indicates either a metric problem or an outlier.

Green Bay was 25th in scoring defense, giving up 37 defensive TDs, 26 FG and scoring 5. They are ranked ahead of New England in unweighted DVOA.

Carolina gave up 32 Defensive TDs (more passing than the Patriots total), 24 FG and scored 2 TD. They are also ahead of New England.

Miami, 29, 22, 3 against an almost identical schedule.

I'm not at all alleging an anti-New England bias. However, the metric clearly isn't representing actual conditions in the case of a bend-don't-break defense like New England has.

In 2001, New England was 6th (2nd in defensive scores) in scoring D but 13th in defensive DVOA. In 02, the metric was reasonable, perhaps even overvaluing the performance.

In 03, NE was far an away #1 in scoring D against a ridiculously difficult schedule (4 10 win teams, 3 12 win teams) but were a distant 2nd in DVOA to the #6 scoring D (Ravens).

The same pattern happened in 04 - 2nd in scoring, 6th in DVOA.

One can either assume the metric is correct and the Patriots have just been repeatedly lucky, or acknowledge that the metric is not fully or accurately measuring something about the Patriots style of defense.

142
by Pat (not verified) :: Sat, 01/13/2007 - 12:41pm

This presumes that expected result is more important than actual result in terms of evaluating a team.

The outcome of the field goal is not controlled by the defense. It's controlled by the opposing special teams.

If you have a defense which faces the worst special teams in the world, ranking by points allowed in that case would make them look far better than a defense that faces the best special teams in the world. This is silly.

The defense isn't even on the field when they try the field goal attempt.

Two defenses, team A, and team B. Game is tied in both cases. Opposing team has the ball in both cases. Team A's defense lets the other team drive all the way to the 5 yard line. Team B's defense forces a stop at the 40. Both kick a field goal and win.

You honestly mean to tell me you'd say "oh, that's the same defensive performance in both cases?" That's insane.

Field position is important yes, and so a team that gives up statistically insignificantly more points but gives its offense better field position independent from the ST unit should be rated higher if takeaways are even.

In other words, a defense's goal is to stop drives.

However, if two teams over the course of a 16 game season give up essentially the same number of points and an overall defensive metric says one defense is much better than the other, unless SoS is widely divergent or this is a major outlier, the metric is broken.

This statement completely contradicts your previous statement. Really, it does. In the first one you say "two teams, even in points, one who gives the opposing team better field position should be ranked higher." Then you say "ranking a team higher when they've given up the same points is broken."

However, the metric clearly isn’t representing actual conditions in the case of a bend-don’t-break defense like New England has.

Why? Because you don't believe it is?

I've had this discussion a lot. It always comes down to the same things - some people would prefer to rank defenses by points allowed first and foremost, because they've heard the phrase "bend but don't break!" in the press. It's a crap phrase. A "bend but don't break" defense is an average defense with a great red zone defense. The defense usually has a weak unit (DL, LB, or secondary) which gets compensated for when the field shrinks.

Ranking units purely by points allowed is incredibly silly. You're not taking into account the field position that they started with - if the opposing offense turns the ball over already in field goal range, and the defense stops them... what, you're going to blame the defense for giving up three points? You're not taking into account the field position they give up - they force a stop at the one, and their own offense takes over and then gets hit with a safety - how is that not in good part the defense's fault?

I'm not saying points allowed isn't important. Giving up a touchdown is worse than not giving one up. But points allowed isn't the end-all, be-all of ranking defenses.

143
by PantsB (not verified) :: Sat, 01/13/2007 - 7:04pm

Ranking units purely by points allowed is incredibly silly. You’re not taking into account the field position that they started with - if the opposing offense turns the ball over already in field goal range, and the defense stops them… what, you’re going to blame the defense for giving up three points? You’re not taking into account the field position they give up - they force a stop at the one, and their own offense takes over and then gets hit with a safety - how is that not in good part the defense’s fault?

I’m not saying points allowed isn’t important. Giving up a touchdown is worse than not giving one up. But points allowed isn’t the end-all, be-all of ranking defenses.
Thats unfortunate because games measure defenses in terms of points allowed. If the metric doesn't translate to points allowed at some point, its not measuring what a defense does.

What is a "stopped drive"? A drive that results in no points? A drive that scores a TD is worth more than two drives that score FGs. "Stopped drives" is only an approximation of how many points a defense should have given up given the assumptions of the methodology. When reality consistently does not meet those expectations, adjustments must be made.

The New England Patriots in 01, 03, 04 and 06 have given up much fewer points and took the ball away much more than would be predicted by the DVOA value (compared to their peers). Its very unlikely this is due to random chance.

144
by Pat (not verified) :: Mon, 01/15/2007 - 12:20am

What is a “stopped drive�?

A first down which does not result in another first down (or touchdown). Field goals are omitted because the opposing offense is no longer on the field for the play. In some sense you have to consider basically every 10 yard chunk a "potential first down" so if a drive goes 30 yards, that's 3 unstopped first downs, regardless of if it happened in one play or in 6. But that's just my opinion.

Thats unfortunate because games measure defenses in terms of points allowed.

No, they don't! Games measure teams by point differential. There's no conventional, simple stat which separates offense from defense.

Its very unlikely this is due to random chance.

You're right. It's not. It's due to a high red zone DVOA. But giving up yards, and not points, requires a better offense and special teams to generate a similar point differential.

Defensive DVOA doesn't correlate to points allowed. It correlates to point differential.

Teams don't win by limiting their opposition to few points. They win by scoring more than their opposition.