Writers of Pro Football Prospectus 2008

Most Recent FO Features

ManningPey98.jpg

» Game Preview: DEN-CIN

Maybe the Bengals are a paper tiger, but are they really that bad in prime time games? Is Peyton Manning struggling in Denver's new offense? We detail the Monday night clash.

16 Jan 2007

Postseason DVOA II

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. Any week where a team did not play in the postseason is treated as a bye week. All numbers are weighted DVOA. That means that Weeks 1-5 are not included, while Weeks 6-11 are somewhat discounted.

Let's be perfectly straight here: by this point, this is resulting in some goofy rankings. The weighted DVOA formula is not designed to be used like this. Teams that played a couple strong games at the end of the year show up very high, because they have two empty weeks at the end of the season. Green Bay and Buffalo in the top ten? Not if they had actually played games the last two weeks.

If you want to see the regular total season ratings, you can just use the Just the Stats menu above.

Commentary now up on FOXSports.com as of Wednesday morning. It includes just the 8 playoff teams, and the order has been changed so that teams still alive are listed above teams that were eliminated this weekend.

* * * * *

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 39.5% 1 13-4 6.1% 12 -29.1% 1 4.3% 6
2 IND 32.4% 3 14-4 32.8% 1 0.3% 12 -0.2% 19
3 SD 32.1% 4 14-3 24.6% 2 -3.2% 11 4.2% 7
4 JAC 23.7% 5 8-8 11.8% 8 -11.8% 6 0.1% 17
5 NE 22.1% 2 14-4 6.2% 11 -11.9% 5 4.0% 8
6 NO 16.6% 8 11-6 18.0% 5 1.2% 13 -0.2% 18
7 PHI 14.1% 6 11-7 18.2% 4 2.6% 17 -1.5% 24
8 CIN 10.0% 9 8-8 19.3% 3 10.7% 29 1.3% 16
9 GB 9.0% 10 8-8 -7.1% 21 -19.8% 2 -3.7% 29
10 BUF 7.0% 16 7-9 -6.8% 20 -5.8% 10 8.0% 1
11 CHI 6.6% 7 14-3 -11.2% 26 -12.5% 4 5.2% 3
12 DAL 5.8% 13 9-8 11.8% 7 9.4% 26 3.4% 10
13 CAR 4.8% 11 8-8 -7.4% 22 -14.7% 3 -2.5% 26
14 PIT 4.6% 12 8-8 7.8% 10 1.3% 14 -1.8% 25
15 MIA 3.1% 17 6-10 -10.0% 25 -11.4% 7 1.6% 14
16 TEN 2.1% 14 8-8 1.1% 17 3.8% 19 4.7% 5
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 NYG 1.6% 15 8-9 5.6% 14 3.4% 18 -0.6% 20
18 NYJ 0.2% 18 10-7 3.4% 15 8.0% 20 4.8% 4
19 KC -3.7% 19 9-7 5.9% 13 8.8% 23 -0.9% 22
20 ARI -4.7% 20 5-11 8.5% 9 12.4% 30 -0.8% 21
21 SEA -6.6% 27 10-8 -8.3% 24 1.9% 15 3.6% 9
22 ATL -10.7% 24 7-9 0.1% 18 8.2% 22 -2.6% 27
23 STL -11.2% 21 8-8 15.2% 6 19.6% 32 -6.8% 32
24 DEN -12.1% 22 9-7 -12.4% 28 1.9% 16 2.2% 11
25 MIN -14.8% 23 6-10 -19.1% 30 -9.2% 9 -4.9% 31
26 DET -15.6% 29 3-13 -11.8% 27 9.9% 28 6.1% 2
27 WAS -15.7% 28 5-11 1.5% 16 19.4% 31 2.1% 12
28 SF -16.0% 26 7-9 -7.9% 23 9.4% 27 1.3% 15
29 HOU -16.4% 25 6-10 -3.6% 19 9.0% 25 -3.7% 30
30 TB -25.6% 30 4-12 -18.4% 29 9.0% 24 1.8% 13
31 OAK -26.0% 31 2-14 -34.3% 32 -11.3% 8 -3.1% 28
32 CLE -28.7% 32 4-12 -19.3% 31 8.0% 21 -1.4% 23

Here are the one-game DVOA ratings for the first round of the playoffs. Remember that these include opponent adjustments. Yes, those really are the ratings for San Diego and New England, even though the Patriots won the game. That's why the Patriots drop from second to fifth in the table above.

TEAM TOT OFF DEF ST
IND 67% 6% -50% 11%
BAL 10% -56% -53% 13%
NO 33% 35% 5% 4%
PHI -2% 16% 16% -2%
CHI -30% -19% 9% -2%
SEA 37% 17% -11% 9%
NE -25% -45% -13% 6%
SD 71% 15% -59% -2%

The Colts had a defensive DVOA of -76% against Kansas City and -50% against Baltimore, which has moved their weighted DVOA on defense all the way up from 26th to 12th over the last two weeks. Is there any precedent for this? Actually, it turns out there is.

I went looking for every team with a below-average defense that had two straight very good defensive games in the playoffs, as far back as the DVOA ratings go, which is 1997. There are a few of them, but none of them turned things around quite like the Colts.


Year Team Season Change 18 19 20 21
2006 IND 11.3% -74.6% -76.2% -50.4% ? ?
2005 NE 10.5% -40.3% -47.9% -11.6%    
1998 DEN 0.4% -40.0%   -31.5% -47.8% -27.8%
1999 WAS 2.5% -31.6% -20.8% -37.4%    
2006 SEA 4.5% -17.4% -24.1% -10.7%    

"Change" represents the difference between the team's defensive DVOA during the season and the average of its first two games in the playoffs. The 2005 Patriots actually come the closest, but unlike the Colts, the Patriots' defense had been improving during the second half of the season -- the change didn't show up suddenly in the playoffs. This list also doesn't help us figure out if the Colts can keep this going. Only one of these teams survived to play three games in the postseason, and the 1998 Broncos defense was far better during the regular season than the 2006 Colts defense was.

What about on offense? There isn't even one team with an offensive "change" above 30%. Here are the three mildly similar teams:

Year Team Season Change 18 19 20 21
2005 CAR -3.2% +26.1% 28.5% 23.8% -62.0%  
2001 PHI -3.4% +24.3% 21.9% 26.7% 7.4%  
2003 CAR -7.2% +12.6% 4.8% 20.4% 19.0% 39.8%

The 2005 Panthers did play a third game -- and their postseason offensive improvement completely disappeared. The 2001 Eagles weren't that spectacular in their third game either. The 2003 Panthers did keep their offensive improvement going all the way through the Super Bowl. Unfortunately, their defense -- which had been good all year -- had a bad game in the Super Bowl and they lost to... wait for it... the New England Patriots.

Posted by: Aaron Schatz on 16 Jan 2007

421 comments, Last at 25 Jan 2007, 9:02pm by Pat

Comments

1
by ChrisFromNJ (not verified) :: Tue, 01/16/2007 - 4:18pm

Wouldn't a solution to the weighted DVOA goofiness be to just count every team's last game of the season as being part of their "most recent week"? So, for example, teams that missed the playoffs would count as if Week 17 was last week.

2
by B (not verified) :: Tue, 01/16/2007 - 4:19pm

Is the difference of 91% DVOA for the losing team the most backwards a game has been in DVOA history?

3
by Ch V Kalyan (not verified) :: Tue, 01/16/2007 - 4:19pm

wow! that DVOA for Pats-chargers game was mind boggling ...

no wonder LT was upset!

4
by throughthelookingglass (not verified) :: Tue, 01/16/2007 - 4:20pm

Color me unsurprised about the Pats-Chargers DVOA.
Is there any hope of getting unweighted DVOA including playoff games in the future?

5
by Doug (not verified) :: Tue, 01/16/2007 - 4:22pm

Detroit is clearly ranked too low because Matt Millen told me I had to say it or else he'd hurt my momma! Trend Analysis is way better than this. Please Mr. Millen, please let my momma go!

6
by Rich Conley (not verified) :: Tue, 01/16/2007 - 4:23pm

More in support of what I've thought all year: The patriots offense isnt nearly as good as DVOA seems to think it is. Last year, everyone expected the Patriots D to not be able to hold Denver, and the offense to play well. Exact opposite. D played great, offense blew it. Same thing this year, D played great, offense blew it, but they got lucky and won. Looks like ANOTHER year of the Pats drafting offensively.

7
by B (not verified) :: Tue, 01/16/2007 - 4:26pm

4: I don't know how to get DVOA from the future (If I did, I'd be rich), but here is the total unweighted DV
OA including playoffs from the:
http://www.footballoutsiders.com/stats/teameff.php

8
by GBS (not verified) :: Tue, 01/16/2007 - 4:27pm

I believe the Colts season opening win over the Giants had them losing the VOA battle by more than 91%, but there are no opponent adjustments in week 1.

9
by MDZ (not verified) :: Tue, 01/16/2007 - 4:28pm

The 67% difference in DVOA between the Bears and Seahawks really sticks out to me. Watching the game I thought the two teams were pretty close and I wouldn't have been surprised if DVOA said Seattle slightly outplayed the Bears, but wow. I would never have guessed that Seattle may have outplayed the Bears by that much.

10
by Randy S. (not verified) :: Tue, 01/16/2007 - 4:28pm

6 - So who should they take to fix things? This year's WR crop is very deep, so they don't really need to spend a first rounder there. Line prospects usually go later in the draft, outside of left tackle; the only elite tackle in the draft is going in the top five, so the Pats are out of luck.

I would imagine they'd take a linebacker in the first round. The offense should be able to be addressed later.

11
by throughthelookingglass (not verified) :: Tue, 01/16/2007 - 4:28pm

7
At the top of the page it says Regular Season DVOA Postseason Not Included, and revised as of 1/2/07, so i don't think that's it.

12
by Will Allen (not verified) :: Tue, 01/16/2007 - 4:31pm

Aaron, offhand (in other words, don't expend energy finding an answer) has any team played three straight playoff games without having faced an opponent with a defense that did not finish in the top 12 in either regular season DVOA, or regular season weighted DVOA?

13
by admin :: Tue, 01/16/2007 - 4:36pm

Sorry, I should have pointed this out: the Seattle-Chicago thing is in large part, although not entirely, opponent adjustments, since Chicago has been a much higher-rated team.

14
by Charles Jake (not verified) :: Tue, 01/16/2007 - 4:37pm

I'm a Bears fan. Somebody, please, talk me off the ledge.

15
by Rich Conley (not verified) :: Tue, 01/16/2007 - 4:37pm

AAron, how are the Patriots coming up with such great special teams numbers? Its been now atleast 3 straight weeks where they've gotten absolutely gashed by kick/punt returners. Are thier return teams really that good that theyre outweighing the coverage teams?

10. They dont need a linebacker really. Theres some good free agent talent out there... including one San Diego middle linebacker, and one baltimore OLB, allthough they need a MLB more, Banta-Cain seems to be playing pretty well.

Aaron, I think single game VOA would be better than DVOA. I think in the case of single games, DVOA just kind of muddies the waters.

16
by JJcruiser (not verified) :: Tue, 01/16/2007 - 4:37pm

10: They should take the best player available at any position (but QB) that is available when they draft.

17
by Randy S. (not verified) :: Tue, 01/16/2007 - 4:39pm

14 - I think Aaron just did.

As a Seahawks fan, I'll cheer against the Bears, but best of luck anyways. Just think about this: the Seahawks basically dared Rex Grossman to win that game. And he did! That's something, right?

18
by MWH (not verified) :: Tue, 01/16/2007 - 4:40pm

re: 9
I thought so as well, is there an easy explanation as to why the Bears DVOA was so low? I know they were inconsistant, but they did move the ball and its not like the defense gave up a lot of yards to the 'hawks. The bears are trending - like a rock.

19
by Will Allen (not verified) :: Tue, 01/16/2007 - 4:41pm

By the way, the reason I ask is that if Chicago and Indy win, the Bears will end up having playoff opponents which were ranked, in final regular season DVOAs, 20,19, and then 27, defensively. This strikes me as being extraordinary good fortune for a team with the Bears' personnel, but pehaps my perception is inaccurate.

20
by MWH (not verified) :: Tue, 01/16/2007 - 4:41pm

re: 13 Aaron responds before I can even ask the question . . .

21
by Todd (not verified) :: Tue, 01/16/2007 - 4:45pm

re 15

I think DVOA needs to be tweaked if seattle outplayed chicago by 67%. watching the game...simply not an accurate number. Tweaking is definetly needed.

22
by Rich Conley (not verified) :: Tue, 01/16/2007 - 4:47pm

21. Its the opponent adjustments playing havock. Bears defense is rated well, seattle's is rated poorly. If they put up the same VOA, it makes it look like the Bears had a poor game, and seattle had a great one.

23
by b-man (not verified) :: Tue, 01/16/2007 - 4:51pm

15: When did they get gashed for a kick/punt return against SD. The only bad one I remember is the KO out of bounds. Saurbee's punting was very good including a muff and they went 3/3 on FGs and one was a 51 yarder. They didn't return any punts but KOs got good fp. I would say the ST helped keep them in the game.

24
by BillWallace (not verified) :: Tue, 01/16/2007 - 4:55pm

Statistical forumla designed by Pats fan shows Pats extraordinarily lucky to beat Chargers.

What else can you say?

25
by Rich Conley (not verified) :: Tue, 01/16/2007 - 5:01pm

23 Punts were great, forgot about that.

Now, they didnt get gashed by SD, but they did get comparitively gashed, and maybe thats what I'm remembering. The Patriots average kickoff return that game was 17.5 yards. The chargers was about 23 yards. (if you discount the Mkinnie squib recovery)

They certainly got gashed by Tennessee, and got hit pretty bad by NYJ.

26
by MJK (not verified) :: Tue, 01/16/2007 - 5:02pm

Stupid question, but for a single game, shouldn't the VOA of one team's offense be exactly equal to that of the opponent's defense? I.e. if my offensive plays "succeed" 15% more often than average, doesn't that mean that the other guy made me fail 15% less than average?

In which case, is the difference between, say, the Pats offensive DVOA and the |Chargers' defensive DVOA| entirely due to the fact that NE gets a better adjustment for facing the Chargers defense than the Chargers get for facing the NE offense? Or am I missing something?

27
by Pat (not verified) :: Tue, 01/16/2007 - 5:03pm

#22: Well, not exactly 'havoc'. It depends on whether or not Seattle really wasn't that good (hard to tell, they were playing better the past four games), and whether or not the Bears aren't tanking horribly (again, hard to tell).

If New Orleans just pounds the crap out of the Bears next week, then the DVOA of that game will pull closer together as the Bears adjustments get more minor.

28
by Rich Conley (not verified) :: Tue, 01/16/2007 - 5:03pm

Bill, there was a good deal of luck, and a good deal of Marty's usual issue: Allowing a clearly inferior team to keep the game close.

29
by navin (not verified) :: Tue, 01/16/2007 - 5:03pm

Re 22
Exactly right on the adjustment explanation. That's why we can't use difference in DVOA to show how much one team outplayed another in a single game. You'd have to used difference in VOA for that.

For SD-NE, since the teams are close together the two should be close. This is not true for SEA-CHI.

30
by dbt (not verified) :: Tue, 01/16/2007 - 5:07pm

The Bears at this point have a merely above-average defense. It really is a shame they couldn't pull off the opening-round win last year.

I'd love to see the raw single game VOA, given the up-thread hand-wringing.

31
by Wanker79 (not verified) :: Tue, 01/16/2007 - 5:13pm

I'm really surprise by the difference between NO's & Philly's defensive DVOA considering their offenses are relatively even (so it shouldn't be opponent adjustment), and subjectively it didn't seem like NO had any better chance of stopping Philly's offense than there was of Philly's defense stopping NO's offense.

I'm even more surprised by the difference in total DVOA between Philly and NO. By comparison, it was the closest of the 4 games, but that still feels like an awfully big differential between the two teams. But that's just a completely subjective feeling.

32
by Wanker79 (not verified) :: Tue, 01/16/2007 - 5:17pm

Consider me among those who think that the one-game ratings would be more useful as VOA. DVOA is great for analyzing the relative strength of the teams, but that's not really important anymore. What is important is who-is-outplaying-who, and for that VOA or (I believe?)PAR would be a better indicator.

33
by Grouchy Bills Fan (not verified) :: Tue, 01/16/2007 - 5:22pm

#7: Future DVOA is easy to calculate. I explained how in post #174 on this thread.

(sorry, I just couldn't resist)

34
by Sean (not verified) :: Tue, 01/16/2007 - 5:33pm

So is there such a thing as an embarrassing victory?

35
by Buzz (not verified) :: Tue, 01/16/2007 - 5:40pm

RE 32 and others
I disagree that VOA would be better than DVOA even at this point in the light of single game results showing the Bears play poorly. The most important aspect of DVOA is the predictability aspect, it doesnt necessarily always say who is the best team/player (like standard rankings based on wins/losses) but who will do the best in the future. You are kidding yourself if you didnt watch the bears game and realize they would have gotten beat soundly by anyone other than SEA this week. Hence the low ranking for this game shows they would have lost to a playoff caliber team but got lucky in the fact that they were playing a below average team. This means a lot more to me than a rating that says these two teams were about equal, because how good is it if they are the same? We need to know how good the game was compared to other teams throughout the year.

36
by TBall (not verified) :: Tue, 01/16/2007 - 5:40pm

Re: 15
Banta-Cain is playing the pass well, but he disappears on running plays. An mid-game graphic on Sunday showed LDT running to the right side (Colvin's side) 4 times for about 15 yards, up the middle 5 times for about 20 yards, and to the left 8 times for about 70 yards. The announcers were gushing over Marcus for handling Seymour, but Banta-Cain was doing nothing to provide containment.

37
by b-man (not verified) :: Tue, 01/16/2007 - 5:43pm

34: How about your team up 21 points and the QB kneels for the last play of the game. All the defenders come rushing in and then pull down the pants of offensive players like in the old Meatballs movie when the camp councelors pulled down the shorts of the opposing camp's basketball team. Your center then trips backwards knocking down your QB and then falls down on him squatting over his head and giving him a proper tea-bagging.

38
by Zac (not verified) :: Tue, 01/16/2007 - 5:44pm

Re: 26. There are some situations which affect one team's VOA but not the other, so two team's VOA's don't necessarily add to zero. Opposing FG % and Kickoff Distance, for example, and maybe Fumble Recovery % as well. They affect the outcome of the game, but there is nothing a team can do (presumably) to affect them.

You can see the value of special teams related ones by looking at the special teams stat, which has a column called "VOA w/ Hidden".

39
by Boots Day (not verified) :: Tue, 01/16/2007 - 5:50pm

Trying to figure out this Bears-Seahawks thing... The Bears clearly outplayed them, outgaining the Seahawks by 65 yards as well as, you know, winning the game. The one clear advantage Seattle had was, surprisingly, on special teams.

Does DVOA account for where on the field turnovers occur? The one thing that sent this game into OT (rather than an outright Bears win) was the fact that both Seattle takeaways occurred near the ends of the field (one at either end), where turnovers are more significant, while the Bears' one takeway was right near midfield.

And of course, DVOA doesn't know this, but Grossman's interception was a fluky deflection to the loan officer, while Hasselbeck's INT was a truly bad throw. (Grossman's fumble, on the other hand, resulted from an incredible play by Julian Peterson.)

40
by CaffeineMan (not verified) :: Tue, 01/16/2007 - 5:52pm

Re: 37

Excellent...

41
by navin (not verified) :: Tue, 01/16/2007 - 5:56pm

Aaron,
I've got a question on how you account for turnovers on 4th down.

Brady threw his infamous interception on fourth down. Depending on what the expected outcome of a pick is, shouldn't it be roughly the same as an incompletion? I'm not sure what the percentages of picks returned for TDs and picks eventually fumbled back are.

Also, NE recovered one of the fumbles on San Diego's 4th and 11 attempt. Shouldn't that count about the same as a sack on 4th down. Does New England get "lucky" points for recovering that fumble even though if San Diego had recovered it then New England would have gotten the ball in 99.9% of those scenarios.

I checked the FAQ again but saw no mention of turnovers on 4th down.

42
by Not saying (not verified) :: Tue, 01/16/2007 - 5:56pm

Re: 30 The Bears at this point have a merely above-average defense.

Well, if you go by weighted DVOA they're still #4 and significantly above average. If you go by last week's game, they are below average (9% below [or above, or whatever]).

It really struck me when watching the Bears game that their D wasn't that great. It was such a sharp contrast to the Raven's D. I guess losing Harris was a big deal. Although I didn't see much of Chicago before, they had to have been getting a better pass rush or their DVOA that low would make no sense to me.

43
by Sean (not verified) :: Tue, 01/16/2007 - 5:57pm

Re 37: No, I think the Pats were worse.

44
by Wanker79 (not verified) :: Tue, 01/16/2007 - 5:58pm

Re: 35

That may all be true, but since there are only a maximum of 8 teams playing in a given playoff weekend, I don't see how it's all that difficult to list the VOA's of (for example) Chicago & Seattle and then apply the context of the overall quality of the two teams.

In other words, at this point in the season I'd much rather have data that tells me exactly who played better in a particular game and then have to mentally adjust that data for the regular season DVOA ratings.

45
by CaffeineMan (not verified) :: Tue, 01/16/2007 - 5:59pm

If I recall, the Bears also lost one or more DB's, right? I think that makes Harris' loss more in the "last straw" category, rather than the "one guy" category.

46
by Wanker79 (not verified) :: Tue, 01/16/2007 - 6:07pm

Re: 39

I'm pretty sure the turnover field position is taken into consideration, but don't quote me on that.

If I had to guess, I would think that the VOA for Seattle and Chicago was pretty close. The reason Seattle looks so good in DVOA and Chicago looks so bad is (as Aaron mentioned in #13) because of the opponent adjustments. They both played pretty equal, but Seattle hasn't been very good (to put it politely), and Chicago has been one of the best teams in the league. So if Seattle's performance against one of the best is about equal to Chicago's performance against one of the worst, Seattle gets a whole crap-ton of credit in DVOA while Chicago's rating just gets craped on.

47
by Not saying (not verified) :: Tue, 01/16/2007 - 6:07pm

RE: 41

Not exactly sure about the 4th down thing (although, I'd have to imagine that the comparison to league average makes some sort of difference).

Does New England get “lucky� points for recovering that fumble No one gets points for recovering fumbles. They get some points for causing the fumble, although I can't remember the adjustment made for fumbles on sacks. They're either significantly more likely to be recovered by the O or D (I think D, but I'm not sure).

I did note at the time of the fumble that in real life the fumble hurt the Pats with a few yards of field position rather than Rivers just going down.

48
by Andrew (not verified) :: Tue, 01/16/2007 - 6:10pm

JJcruiser #16:

They should take the best player available at any position (but QB) that is available when they draft.

So if there is Tight End who is theoretically the BPA, they should take him? Despite being overstocked on Tight Ends already?

49
by Pat (not verified) :: Tue, 01/16/2007 - 6:11pm

#39: Yes. Search for "how many points is a turnover worth."

Believe it or not, turnovers are, to first order, worth basically the same wherever they occur on the field. It's only a minor correction that turnovers near the edges are worth more.

It makes sense, once you think about it - when the Bears turned the ball over at the 5 yard line, it cost them a chance to score. But it also put the Seahawks in poor field position (with the return, average starting field position). Those two factors basically cancel each other out (a little).

50
by stan (not verified) :: Tue, 01/16/2007 - 6:13pm

6,

Rich -- you have seen far more of the Pats this year than I (and I know you and I are the only people on the planet who seem to have noticed that Brady missing wide open receivers is not all that unusual). However, even though the season-long DVOA has plenty of bad passes by Brady factored in, the SD game was bad even by his standards. The Pats offense had two open TDs that he missed. He had people open all day. Once NE adjusted the pass pro, he had a lot of time on most pass plays. NE could have put up 40+ points easily if he'd just made the routine passes.

It wasn't like the line was getting abused and the receivers couldn't get open. Ten people on the offense were executing really well.

51
by D (not verified) :: Tue, 01/16/2007 - 6:17pm

#2
No. The gap in the Bears "only against the Cardinals" victory in week 6 was 108.8%.

52
by Buzz (not verified) :: Tue, 01/16/2007 - 6:17pm

re 44
Good point. But if you are looking forward to next weeks game vs NO and only had VOA to look at and it showed that CHI played pretty decently, wouldnt you have to make the same adjustment showing that they played decent but against a bad team in order to figure out how they will match up with NO? Either way you are going to have to do some math either to get to how "close" of a game it was or how "well" a team played.

53
by Ben B. (not verified) :: Tue, 01/16/2007 - 6:21pm

51: That game had major opponent adjustments to separate the DVOA's even more. This game was between two pretty evenly matched teams, so I'd imagine the VOA would be pretty similar to the DVOA.

54
by Rich Conley (not verified) :: Tue, 01/16/2007 - 6:21pm

"It wasn’t like the line was getting abused and the receivers couldn’t get open. Ten people on the offense were executing really well."

I completely agree. I was ranting in the open discussion about how Matt light was just utterly dominating Merriman, and then they moved Merriman over to the other side, and Kazcur pretty much shut him down. IIRC, the sack/fumble was an OLB on Kevin Faulk, which isnt a matchupI like. Brady had a good amount of time, and he threw a LOT of passes out of bounds when he had recievers open on the sideline (not throwing them away, just missed the guy by 3 feet, and the ball hit a coach or something.) Hes just not playing all that well, and hasnt been for weeks.

I think that Jabar Gaffney play where Gaffney came across and broke open in the flat, and had to dive back for the ball, for a loss of 2 was pretty telling. He was open, but by the time he managed to corall the misthrown ball, there were two linebackers on him. If that ball is thrown to him, thats a 10 yard gain, instead, its a 2 yard loss.

55
by Eddo (not verified) :: Tue, 01/16/2007 - 6:26pm

53: See post 13 by Aaron himself. DVOA was largely affected by opponent adjustment.

56
by Joseph (not verified) :: Tue, 01/16/2007 - 6:28pm

Re: 31

DVOA relates to the situation. The Saints D stopping the Eagles on the goal line early in the 4th Q after the Eagles having 2nd & 1 near the goal line helps their rating, and (IMO) the Eagles inability to stop Deuce in the Saints last two drives of the 4th hurts the Eagles rating. IMO, that prob accounts for most of the difference.

57
by Rich Conley (not verified) :: Tue, 01/16/2007 - 6:35pm

56

The problem is, DVOA makes a close game (in the case of bears/seattle) look like a blowout. The issue is that the opponent adjustment for the bears is no longer accurate, because theyre no longer really an elite team.

In the Case of Philly/NO, Philly is a 25% team, and NO is a 10% team. Them playing each other straight up, should result in a DVOA spread of 30%, which is exactly what happened. If they played even, NO outplayed it's prediction by 15% keeping up with the Eagles, and the Eagles underplayed their prediction by falling to the Saints level.

It looks like a blowout because the opponent adjustments make it about not who played better, but about who played better as compared to the dvoa perception of the team.

58
by Scott de B. (not verified) :: Tue, 01/16/2007 - 6:40pm

I would imagine they’d take a linebacker in the first round.

Don't forget, the Patriots have two first-round picks this year.

59
by Charles Jake (not verified) :: Tue, 01/16/2007 - 6:44pm

#45

Yeah, the Bears lost starting SS Mike Brown for the season (again). He's instrumental in lining the secondary up properly. His loss left them starting a rookie and a 2d year player at safety. They've been caught out of position on several big plays.

60
by Pat (not verified) :: Tue, 01/16/2007 - 6:45pm

#57: Well, you presume they aren't an elite team anymore. If next week they blow the crap out of the Saints, then maybe this week they were just "shaking the rust off" or something, and Seattle was starting to actually live up to their potential, and the DVOA for those games would be accurate.

It really is tough to say.

61
by navin (not verified) :: Tue, 01/16/2007 - 6:52pm

The reason I ask for VOA is because I want to see how each team played relative to each other, ie. I want to see if the Seattle-Chicago game was pretty even. It's hard to tell with DVOA when two teams are not matched evenly.

Yes, DVOA will show how the Bears played in context to the rest of the year, but I want to know that in addition to knowing about how they played relative to the Seahawks in just that one playoff game.

62
by Wanker79 (not verified) :: Tue, 01/16/2007 - 6:53pm

In the Case of Philly/NO, Philly is a 25% team, and NO is a 10% team.

I'm not really sure where those numbers are coming from. According to last week's DVOA numbers, NO was a 12.7% team and Philly was a 15.2% team. The offenses were only 1.0% different, the defenses were only 2.6% different, and the special teams were only 1.2% different. That's why the difference in the game DVOA has me so confused. By any measure, these two teams are almost dead-even, subjectively the game was pretty damn close, but DVOA (while still saying that it was the closest game of the weekend) gives a fairly decent advantage to NO.

63
by Alex (not verified) :: Tue, 01/16/2007 - 6:57pm

#55:"See post 13 by Aaron himself. DVOA was largely affected by opponent adjustment."

That was the Chicago-Seattle game, not the San Diego-New England game that we're discussing. The Chargers and Patriots are fairly close according to regular season DVOA, so the VOA should be fairly similar.

64
by Boots Day (not verified) :: Tue, 01/16/2007 - 6:59pm

It's all very circular. Rich Conley presumes the Bears aren't an elite team based on their final three games of the season -- before that, they were definitely an elite team. Now, beating the Seahawks reinforces that, since based on their DVOA the Seahawks were a bad team -- except that based on their last three regular-season games plus their earlier playoff game, they are a decent team, better than the Bears by that criteria.

So the Bears are a bad team (based on their late-season swan dive) who beat a bad team (based on Seattle's DVOA) and thus get negative credit for it. But nobody ever says the Bears are a very good team (based on their yearlong performance) who beat a pretty good team (based on Seattle's performance in its last four games), so they must be outstanding by now.

65
by Rich Conley (not verified) :: Tue, 01/16/2007 - 7:00pm

Wanker, you're looking at the weighted ones, I was looking at the Season DVOA.

66
by Randy S. (not verified) :: Tue, 01/16/2007 - 7:01pm

58 - Good point. I was curious and checked out the mock draft over at nfldraftcountdown.com, and Scott has the Pats taking a CB and WR in the first round.

I still think a WR in round one would be a waste, since it's not necessary. The Pats seem to be a "system" team when it comes to wideouts, they can probably draft a receiver late that fits their system instead of drafting on talent early. Maybe a Jeff Samardzija flyer at the end of day one.

Corner makes sense, especially if Asante Samuel bolts in free agency. Are the other Pats DBs ok? I suppose they could consider drafting an apprentice for Mr Rodney Harrison, esq. Definately not a replacement, I would never suggest such a thing.

67
by LnGrrrR (not verified) :: Tue, 01/16/2007 - 7:03pm

I won't be surprised if we hear about Brady getting shoulder surgery again in the offseason...

68
by Rich Conley (not verified) :: Tue, 01/16/2007 - 7:12pm

Randy, they've drafted about 5 safeties in the last 2 years, and moved a guy who was a pretty good corner (Eugene Wilson) to free safety. If they resign Samuel, theyre in great shape, assuming Hobbs finally gets healthy (hes been playing hurt all season... people forget, he was better than asante last year). Hawkins actually does a good job when hes not facing Jericho Cotchery. I assume they'll take a LB, but I can't see them taking one in the first, more of a developement prospect (of which they've got a few right now).

I dont really see them drafting a WR, as they've got Jackson, who hopefully will be healthy, and has maybe learned the system, and Reche and Jabar Gaffney seem to have stepped up quite a bit to fill in the void. They also aren't doing so well drafting WRs high as of late.

I'd guess they'd look at a premier tackle, or o-line guy, but they just locked up the entire oline till like 2009. Thomas is supposed to replace Graham, so I can't see them drafting another TE high.

They've got Defensive Ends and Defensive tackles coming out their ears, so not there.

So really, I have no idea...maybe a MLB prospect in the first? Maybe a WR...maybe an elite SS prospect if one is out there...

69
by Rich Conley (not verified) :: Tue, 01/16/2007 - 7:13pm

re:67

I dont think I'd even blink if I heard that. I'd honestly be more surprised if he isnt.

70
by Dan R. (not verified) :: Tue, 01/16/2007 - 7:14pm

I'd like to see VOA tables of the games. That would tell us more about who outplayed whom.

71
by Eddo (not verified) :: Tue, 01/16/2007 - 7:37pm

63: My mistake, I must have got the lines crossed.

72
by MDZ (not verified) :: Tue, 01/16/2007 - 7:38pm

Here's how I see it. Single game DVOA is useful when you want to see if a win might carry over to the next week, or if you want to compare two teams that did not play each other. So when looking at the conference championship matchups, the single game DVOA from the last week or two might help. However, VOA is more appropriate when looking back at a game to determine if the team that "played better" won.

73
by Rich Conley (not verified) :: Tue, 01/16/2007 - 7:43pm

"Single game DVOA is useful when you want to see if a win might carry over to the next week, "

Mike, I disagree. If you look at the season DVOA graphs, the variance for most teams is way too high to take much from the single game DVOA score for a team.

74
by TracingError (not verified) :: Tue, 01/16/2007 - 7:45pm

Aaron, since you said that weighted DVOA corresponds worse to playoff performance than unweighted full year, why not give the full year DVOA? And I agree that when weighting you should just count the most recent games and forget about byes for non-playoff teams.

75
by TracingError (not verified) :: Tue, 01/16/2007 - 7:49pm

Also, as long as we're showing one-game results, how about VOA too. --Eric

76
by Dan (not verified) :: Tue, 01/16/2007 - 7:51pm

The Bears D was hurt both by losing Harris (Tommie) and having to play Harris (Chris) in place of Mike Brown. I noticed two Alexander runs where missed tackles by C. Harris at the line allowed him to break through for 10+ yards, and those are just the cases where the drop-off in safety play was most obvious.

And add me to the chorus of people who would like to see VOA for individual games rather than (or in addition to) DVOA.

77
by Brett (not verified) :: Tue, 01/16/2007 - 7:52pm

DVOA is a great tool during the year, and I think it really does a good job measuring a team's overall performance and ranking; however, it's not useful at this point in the playoffs. It just doesn't do a good job of measuring playoff intangibles.

The Bears were #1 or #2 in DVOA most of the season, but, by squeaking-out a victory at home (where they bore the burden of thier well-documented run of playoff futility) against a well-coached, defending NFC Champion Seahawks team facing them a second time after being blown-out during the season (over 40% of the time in the NFL, the blown-out team wins the rematch), their rating drops like a rock and it looks like they won't have a chance against anybody left in the playoffs. That just simply isn't the case.

With the monkey off the Bear's back going into this week's game, I expect a looser, more aggressive Bear's team against the Saints. It will be a helluva game, but I'm more inclined to side with the Vegas oddsmakers on this one (whose job it is to measure intangibles), rather than DVOA. Bears win by 3, 27-24.

78
by NF (not verified) :: Tue, 01/16/2007 - 7:56pm

I'm actually curious how DVOA handles the interception on 4th down that was fumbled to the Patriots. First there is an interception, and with DVOA 5.0 interceptions are treated as if they had a return that was average for the line of scrimmage and position of interception. But the ball is fumbled on the return, and most fumbles close to the line of scrimmage that aren't fumbled at the snap or on a sack have a 50% chance of being recovered by either team. But does the location the fumble happened at matter for DVOA? I'm a little confused because two non-predictive events occur in sequence, an interception return and a fumble recovery.

Related, is it possible that DVOA underrates interceptions on 4th down that are not expected to be returned to the line of scrimmage, because there is a possibility that it will result in better field position than a turnover on downs would give? To put the whole thing in terms of predictive value, does a defense intercepting a pass on 4th down well upfield from the line of scrimmage show that a defense is more likely to intercept long passes on 4th down if not any other down, and if so, does frequently intercepting it in those situations result in the team eventually having a return gain enough field position relative to the field position from a turnover on downs that would offset the field position lost from previous interceptions in that situation? My hypothesis is that interceptions on 4th down are like blocks on special teams, too rare to make predictions from.

Of course, the practical value of this sort of gambled field position resulting from defense on drive stops is inversely proportional to how well the offense can move the ball, with bad offenses needing more help from the defense and special teams to put up points (the Raiders) and good offenses needing less help (the Colts). For the Colts often getting bad field position due to an interception of the other team on 4th down and occasionally getting scoring-drive field position will be less help than consistently getting average field position from drives stopped on 4th down, while the Raiders offense's only hope of scoring may be to start with the ball in field goal range.

79
by PantsB (not verified) :: Tue, 01/16/2007 - 8:15pm

I don't mean to be the anti-DVOA guy but the difference between the Patriot's actual results and the DVOA suggests a problem with the metric.

They allowed the 2nd least amount of points and got the 2nd most INTs - yet were 8th in DVOA (without a schedule that would justify that kind of difference). The same essential aberration occurred in 01 and 04 (and to a lesser extent in 03, when DVOA has them 2 when they had an infamously difficult schedule and won the opponent ppg going away) regarding their defense.

Similarly, if a DVOA has a rating has the team that won down by 100% does not that suggest a flaw in the methodology?

I'm not doing this to be a jerk, but merely to point out that an adjustment is probably called for.

80
by MDZ (not verified) :: Tue, 01/16/2007 - 8:22pm

Rich, I agree with what you're saying about variance, and I'll try to convey what I was thinking better. I guess a better way to say it would be, to paraphrase what Buzz wrote in #35, that DVOA is more likely to indicate if a team played down to it's opponent, and maybe give a better indication of whether or not they could have beaten a better opponent. If the Bears play the way they did last week when they play the Saints they are in trouble. Of course, who knows which Bears will show up because of the team's variance.

81
by Pat (not verified) :: Tue, 01/16/2007 - 8:22pm

#79:

I've said this elsewhere, and I'll repeat it here: check out the DVOA splits. New England has the 3rd highest red zone DVOA defense in the league, and by far the highest red zone DVOA passing defense in the league.

You can't rank defenses purely by points allowed. Points allowed completely ignores any field position surrendered/saved by defense that contributes to offensive points. And that's just the simple version of it.

If you think that preventing points is the end all, be all of a defense, feel free, but it certainly doesn't appear to be true.

82
by Boots Day (not verified) :: Tue, 01/16/2007 - 8:34pm

No. 80: Like I said before, this is all circular. You can't know if the Bears played down to their opponent's level until you know what that level is. And it's awfully hard to estimate that level for a single game.

Seattle was obviously not very good considering the season as a whole, but they were just as obviously better very late in the year. They looked fairly good to me against the Bears, for whatever that's worth.

Maybe if the Bears play against the Saints the way they did last week, they will lose. Then again, the Eagles' DVOA for last week was much worse than the Seahawks, yet the Saints only beat them by three points. Does that mean the Seahawks were significantly better than the Eagles? If so, maybe the Bears will end up slapping around the Saints.

83
by calbuzz (not verified) :: Tue, 01/16/2007 - 8:46pm

#39. I've had to correct this several times: The Rex inteception was by the Border Patrolman, not the Loan Officer, even though the story being told sounds better than the truth.

Does this make Grossman's VOA better?

84
by Marko (not verified) :: Tue, 01/16/2007 - 8:49pm

Actually, the Border Patrolman and the Loan Officer are the same person. He is a loan officer who is training to be a border patrolman. He is not, however, a hunter, although his name is Pete Hunter.

85
by Chris M (not verified) :: Tue, 01/16/2007 - 8:56pm

Re: 77

The Bears are down about 10% in DVOA to the Saints, but it's been suggested that homefield is worth about 17%. So your prediction of the Bears squeaking past the Saints is roughly what DVOA would predict.

86
by Dave (not verified) :: Tue, 01/16/2007 - 9:17pm

Hey, anyone else think John Clayton had really weird analysis of the Pats-Chargers game? He claims over and over that the Bolts played not to lose, like he's taking a page right out of the complaints about Martyball playbook without actually having watched the game. Seems like they'd have given LT the ball more if they were playing Martyball. I don't know how he could possibly have been watching the same game I was watching.

87
by B (not verified) :: Tue, 01/16/2007 - 9:48pm

86: Indeed. Marty ball=Running the ball early and often, and never attempting 4th down conversions. Marty's game plan this weekend was to try on 4th and 11, and to throw the ball early and often. It was as though he purposefully went against type, pulling a Constanza, as they say. As TMQ might say, dance with what brung ya.

88
by Pat (not verified) :: Tue, 01/16/2007 - 10:14pm

#85: Well, let's be fair - DVOA wouldn't predict anything. It would say "are you freaking nuts? With HFA, these two teams are dead-on equal. Flip a freaking coin."

"Predicting" a single-score victory is just silly.

89
by fungus (not verified) :: Tue, 01/16/2007 - 10:29pm

"(I)f a DVOA has a rating has the team that won down by 100% does not that suggest a flaw in the methodology?"

No, it just means Marty Schottenheimer happened. HTH.

90
by paytonrules (not verified) :: Tue, 01/16/2007 - 10:44pm

Bears fan - and I'm considering getting my mail delivered to the ledge.

If I was Vegas - and I'm not - I'd favor the Saints by a touchdown. The Bears just aren't very good anymore. It's damn depressing.

And if Aaron prints VOA and DVOA people can stop kvetching.

91
by B (not verified) :: Tue, 01/16/2007 - 10:50pm

68: I think the Patriots top needs are an elite SS and LBs. But they have lots of cap space next year, so I'm hoping they will make a big splash in free agency, filling other needs through the draft.

92
by ElTiante (not verified) :: Tue, 01/16/2007 - 10:55pm

#86- I agree totally. My first thought was that Clayton woke up, realized he hadn't written his piece, had 10 minutes to get something in to his editor... Not only doesn't his analysis fit the game as it was played, but he seems to forget at times the point he was trying to make. Very strange unfocused writing.

93
by Dan Riley (not verified) :: Tue, 01/16/2007 - 11:05pm

I love FO. I really do. Visit here most every day. Love to hear from the peanut galleries from Denver and Indy and Pitt pissing and moaning about how FO is all about the Pats and the Pats are all about luck and the NFL is all abut making the Pats happy. Love to see Rich Conley, classic NE sports fan, running around like Chicken Little--Brady's arm's hanging, it's swollen, it's in a sling (I'm betting if we could get a Webcam in Rich's bedroom we'd see an autographed Drew Bledsoe poster). I love it when the ultimate mensch Aaron goes all Rodney King on us and begs us all to get along. But I got to admit the numbers, the whole DVOA thing makes my head numb. I admit I'm not a numbers guy. But if I saw a system that actually, accurately predicted games from week to week, I'd eat pi for breakfast, lunch and dinner. But DVOA can't even tell us who won last week.. I mean just look at that NE/SD break. Do numbers lie? Numbers crap their pants, my friends. And here's why. You have no way...and will never have a way of measuring the intangibles. And for all the Pats Haters out there, this one's for you: The Pats always lose to teams with better players, but they never lose to better teams, because they are always the better team. And what makes them better is BB preaching situational football and guys like Brown & Brady playing situational football. McRee didn't fumble that ball; Brown stripped him. Brady wasn't on all damn game, he was just on when he absolutely, positively had to be--at the end of the half and with 6 minutes to go in the game. How do you measure that? The loonies can howl all they want about luck and conspiracies and, oooh (a new one) immoral, godless heathenism. The Pats win because of professionalism. I thought for sure they were going to lose on Sunday (and there you go, the numbers bear me out). They SHOULD have lost on Sunday, but every time the camera went to the sideline, they looked like a bunch of firemen fighting a raging 60-story blaze, calmly studying the building blueprints...trying to locate the exits and the gas-lines. Crunching numbers can't account for that level of cool and (uh-oh) calculation, and it's hopeless to think that fans who've been raised on Martyball or the greatest surf on turf or whatever for most of their lives will ever get it (aside to Jets fans, your coach gets it, and you guys are officially dangerous).

94
by MFurtek (not verified) :: Tue, 01/16/2007 - 11:29pm

Consider the Pats crowned. I congratulate you Dan on having 3-4 hours of free time this Sunday, and you can even avoid the hype of the Super Bowl. You don't need to watch the game to know the Patriots will win.... better team, more professional, screw the numbers, blue-collar, coached by Jesus incarnate.

Thanks for the insight... you're going to make a lot of money the next two weeks, I assume you're willing to bet the house on the Pats right now.

95
by dbt (not verified) :: Tue, 01/16/2007 - 11:32pm

Can we make a new troll template out of that? No? YOU ARE USELESS TO ME, DAN RILEY.

Also, I hope you show up if your pats get beat on Sunday and take your lumps like a man. I know I'll be here crying in my iBeer if the Bears lay an egg.

96
by Charger_Fan_In_Reclusion (not verified) :: Tue, 01/16/2007 - 11:35pm

93(Dan): I'm sorry to go ad hominem but I don't know if there is another way to school an irrational Pats fan (that's you). Your comments would make an epsilon amount of sense if you knew how names are written. Moreover, you are not worth listening until you sound at least halfway intelligent. At present, you, sir, do not.

97
by Jon Jansen (not verified) :: Tue, 01/16/2007 - 11:35pm

I think that the colts are going to win the superbowl. and in this post i will not mention "you know who" or "he who should not be named".

the colts defense is playing great. while baltimore does not have a particularly potent offense, kansas city sure does. the reason why the colt's defense is playing so well is

a. they are fed up with everyone talking about how they cant tackle.
b. they have gotten much better at tackling
c. the cover 2 defense that they use, keeps everything in front of them. which means that while they dont give up the big play, they give the opponent's offense short completions and some good runs.

in a cover 2, the players who are covering someone or playing zone, always have their eyes on the quarterbacks eyes, so that they can react quickly when the ball is thrown, if they can do this successfuly, then that means that they may give up 3 yards on a completion, but they YAC after the completion will be close to ZERO.

however, since the colts use a cover 2, it is utterly important that their defensive line plays well and that their linebackers can plug holes and tackle the running back, otherwise, they will be run all over.

apart from the colts defense, their special teams is excellent because they signed vinatieri. they can rely on him to kick the winning field goal, as long as its in his range.

as far as the colts offense goes, they are playing phenomenal. every year, the colts offense is one of the best. and this year, it has struggled a bit more than in other years because of its inability to maintain a solid and consistent running game.

however what people dont realize, is that the colts have been so excellent on offense in the past couple of years, that because they are not doing as excellent is last year, or the year before, that people do not look at the colts offense as their way to win.

they look at the colts defense as the way to lose. but if the colts defense plays against the patriots like they did against the ravens and the chiefs

HOW CAN THE COLTS NOT WIN THE SUPERBOWL, EASILY?

the colts have an outstanding offense, led by "you know who" they should be able to put up more points than their defense allows to the kind of struggling patriots offense led by "he who should not be named", who threw 3 interceptions on sunday.

so if the colts prevail at INDY, which they will, they most likely will be battling the saints. and i know "you know who" will not let his chance at a superbowl slip by. i know he will throw for 300 yards and 3 tds in a 31-17 rout of new orleans.

98
by mm (not verified) :: Tue, 01/16/2007 - 11:41pm

Didn't someone say the Pats recovered 4 of 4 fumbles this week? So the Pats got lucky (which DVOA accounts for), and may have gotten lucky in other areas of the game as well.

As to the Saints/Eagles game, I'd expect the VOA to indicate it wasn't as close as the score indicated. Every so often on FO they'll talk about how a steady offense is much better than a 'boom and bust' offense. (ie, an offense that gets 4 yards every play is preferrable to an offense that gets 0 yards 2 straight plays then gets 12 yards on the third play).

The Eagles offense was a boom/bust offense...much of their output came on 2 plays. Lots of Eagles runs seemed to go for little or no gain. Meanwhile the one pass Devery Henderson caught was the only 'big' gainer the Saints had, and it wasn't as long as the 2 Eagles plays. And Deuce's carries were about as bust-free as you can get (check out this weeks Quick Reads).

99
by NF (not verified) :: Tue, 01/16/2007 - 11:54pm

93:"The Pats always lose to teams with better players, but they never lose to better teams, because they are always the better team. And what makes them better is BB preaching situational football and guys like Brown & Brady playing situational football."

Uhhh...

Propositional logic time:

(1) Patriots ALWAYS LOSE TO teams with better players.
(2) Patriots NEVER LOSE TO better teams.
(3) Patriots ARE ALWAYS the better team.
(4) Proposition 3 is true BECAUSE Belichick preaches situational football and players like Brown & Brady play situational football.

Therefore:

If (1), given (2): (5)Teams with better players ARE NEVER better teams.
If (3) given (5): (6)Patriots ARE NEVER the team with better players.
If (1), given (6): (7)Patriots ALWAYS LOSE.
If (7) is true because of (6) and (1), and (6) is true because of (5) and (3), given (4) : (8)Patriots ALWAYS LOSE BECAUSE Belichick preaches situational football and players like Brown & Brady play situational football.

Admit it Dan, you have an unconscious hatred for all things Belichick, Brady, and Brown.

100
by B (not verified) :: Wed, 01/17/2007 - 12:08am

98: Patriots never lose in the playoffs, except when they do.
Colts, on the other hand, only lose in the playoffs to the Superbowl Champs.

101
by PantsB (not verified) :: Wed, 01/17/2007 - 12:08am

#81
I’ve said this elsewhere, and I’ll repeat it here: check out the DVOA splits....
You can’t rank defenses purely by points allowed. Points allowed completely ignores any field position surrendered/saved by defense that contributes to offensive points. And that’s just the simple version of it.
>>>
Points allowed is the be all of actual results. Only two statistics are allowed in the final measurement of who won or lost a game - the points each team scored. All other statistics are only worthwhile in that they either a) approximately predict how many points each team will score in a future game, b) measure how those points were scored/what led to points being scored c) quantify how many points should have been scored if luck was eliminated. If a statistic does not contribute to this it is useless.

DVOA has undervalued the Patriots actual results - the amount of points scored by opponents and the number of turnovers (which are noteworthy because they often lead to extra offensive points) - this season and in each of the last three Championship level Patriot teams. This suggests a failure in reflecting actual performance because the degree of error is too great to dismiss as statistical variance.

The Offensive DVOA ranking of the Patriots has not been too high compared to their offensive points scored. Field position is only valuable so far as it helps obtain or prevent points from being scored. Football has a memory but its still a discrete system that doesn't have play by play granularity.

If we look at the DVOA, the Chargers destroyed the Patriots. Yet, the biggest lead in the game was 11 and the Patriots outscored the Chargers over the 2nd half, 3 out of 4 quarters and - in the end the only thing that matters - the course of the game. One can say the Patriots got "lucky" but combined with an undervaluing of actual results in past Patriot years....

102
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 12:24am

Data points, people, data points. One data point (or three) does not DVOA make or DVOA break.

103
by Peter Libero (not verified) :: Wed, 01/17/2007 - 12:42am

101: "DVOA has undervalued the Patriots actual results"
Huh? They were 5th in DVOA this year with the 8th defense. In 2004 they were the #1 team, with the #6 defense. In 2003 they were the #3 team, with the #2 defense. How much higher do you really think they need to be?

104
by ToxikFetus (not verified) :: Wed, 01/17/2007 - 1:00am

103: sqrt(-1)? I can't imagine a ranking better than that! *rimshot*

Thanks, I'll be here all week!

105
by me (not verified) :: Wed, 01/17/2007 - 1:04am

Dan Riley: Nice post!

And the proof is in the responses you received.

People don't understand what they don't understand.

You speak the truth.

106
by Led (not verified) :: Wed, 01/17/2007 - 1:15am

#99 is awesome.

107
by DGL (not verified) :: Wed, 01/17/2007 - 1:35am

#93: What, nothing about swagger?

108
by hwc (not verified) :: Wed, 01/17/2007 - 1:37am

For a good points based ranking system, the Sagarin rankings in USA Today are pretty good. See link at my sig.

Sagarin is a synthesis of two rankings: PURE POINTS is a ranking based on points totals and opponent strength. ELO CHESS is a ranking based on wins/losses and opponent strength.

It's flaky early in the season, but starts to look pretty good later in the year.

I try to look at Sagarin's rankings and Aarron's DVOA rankings each week. Both add a piece to the puzzle.

109
by zip (not verified) :: Wed, 01/17/2007 - 1:45am

#99 wins

110
by Fat Tony (not verified) :: Wed, 01/17/2007 - 1:48am

Will #93 trigger the FOMBC or are the Pats still exempt?

111
by nat (not verified) :: Wed, 01/17/2007 - 1:51am

Zowie!

The NE-SD game is truly weird. I know that the fumble recoveries account for a lot of the discrepency between the score and the DVOA, but it's hard to believe that the play "success" was as lopsided as the numbers say.

I was curious about how the Pats pulled it off. So I counted the "successful" plays (using 4 yrds for 1st down, half way for second down, and all the way for 3rd and 4th down) for each team. I also counted the field goals and extra points. The number of successes was within 1 or 2, and I think it favored the Pats.

Then I thought that play success leads to first downs, yards, and thus scores. SD had more first downs (by 3). NE had more scores (by 2). And SD's average play outgained NE's average play by 5.2 to 4.4 yards. But NE had more plays (74 to 68, give or take penalties) and the total yards with penalties included favored SD by about 10 yards.

So here's the question: HOW did the Chargers rack up such an incredible DVOA without it showing up in first downs, scores, or yards? Is it all about NE going 5 for 5 on fumbles? Did SD get a lot of worthless yards (12 yrds on 3rd-and-15, 17 yrds on 3rd-and-1?) that look way better than average, but in reality aren't way more useful? Were the NE unsuccessful plays that much more awful? (If so, why didn't that translate into first downs or scores?)

Was SD somehow optimized for getting a high VOA in the least productive way?

Weird. Weird. Weird.

112
by craig (not verified) :: Wed, 01/17/2007 - 2:04am

Using a simple pass/fail score like in post 111, you can easily have a high pass/fail be less successful than a lower one, based on when they occur.

Theoretically, a team that goes from 1st and 10 to 2nd and 5 to 3rd and 2 to 4th and 1 (and punting), succeeded on 2/3 of their plays.

That seems like what I saw of SD early when they were squandering good field position.
If the other team then goes 1st and 10, 2nd and 8, 3rd and 5, 1st down, they only succeeded on 1/3 of their plays, but are more successful overall.

113
by Huggy (not verified) :: Wed, 01/17/2007 - 2:08am

I think the guy who posted 93 is really Peter King. That bit about the Jets gave it away.

114
by MFurtek (not verified) :: Wed, 01/17/2007 - 2:13am

nat,
Brady had 50 more yards on 19 more attempts. The Chargers rushed for 80 more yards.

I think it was the fact that San Diego was successful on a greater percentage of plays.

Special teams is essentially a wash due to Scifres dropping 5 inside the 20.

DVOA doesn't account for positive field position of recovering a fumble, not sure about interceptions. There are some hidden yards there... and also I'm not sure how they account for a kickoff from the 15.

115
by Pat (not verified) :: Wed, 01/17/2007 - 2:16am

Points allowed is the be all of actual results.

No, it's not! Unless you want to be the one going to tell a team that loses the Super Bowl 7-0 "Congratulations, you held the other team to just 7 points! That's way below the average!"

the points each team scored.

Yes. But "points scored = offense" and "points allowed = defense" is far, far too simple a model of football. (For one thing, what do special teams do? Twiddle their thumbs?)

New England's been one of the top 5 teams in the NFL for the past 3 out of 4 years. They've had a fantastic offense for 2 out of those 3 years. A fantastic offense coupled with a great red zone defense makes a good defense look fantastic.

116
by hwc (not verified) :: Wed, 01/17/2007 - 2:46am

The problem with the DVOA system is that individual football games are usually determined by a few critical plays at critical times in a game.

For example, a key play in this game was that idiot Charger headbutting the Patriot and giving the Pats a free first down when they had been stopped.

Those things tend to even out over time, so season long DVOA rankings start to approximate the results we see on the field. Even then, there's still a shortage of data compared to "moneyball" in baseball where the whole sport is a statistical excercise played out over 162 days.

All systems in football statistics have real problems with strength of schedule adjustments. Every season is different so using prior season data to identify strong and weak teams is useless. It isn't until you get 12 games into the season that you have enough games to even know what is a strong team and a weak team. We've all seen teams get on a flukey win streak early and then prove to be frauds down the stretch.

The statistical systems really get test in the postseason tournment. I don't believe there is a datapoint for pressure or choking in the DVOA formula. Or a datapoint for teams that take the field in the postseason sky high on adreneline, only to burn themselves out and fade in the fourth quarter (see the Eagles QB Donovan McPuke, the Chargers on Sunday, etc.).

Franky, after watching the Ravens/Colts game, I've decided that football prognostication is a fool's errand. There can't be a single person on the face of the earth who expected that game to unfold in that manner.

117
by Charger_Fan_In_Reclusion (not verified) :: Wed, 01/17/2007 - 2:51am

Shaun Phillips will have (at least) two shots against the hated (for what they have done now; besides, Colts are my second team) Patriots. Just look at the schedule. NE comes to SD in 2008 and if both win their division, which is an almost certainty for NE considering how weak AFCE has been, NE will play at SD in 2009 as well.

Tom, meet grass; Grass, meet Tom. Well, not really. In 07, Tom, meet field turf; field turf, meet Tom.

118
by Scott de B. (not verified) :: Wed, 01/17/2007 - 2:55am

They allowed the 2nd least amount of points and got the 2nd most INTs - yet were 8th in DVOA (without a schedule that would justify that kind of difference).

They played one of the easiest schedules in the league, so yes, it does justify the difference. Anyway, they were 2nd in WDVOA until this weekend.

119
by jebmak (not verified) :: Wed, 01/17/2007 - 2:57am

Wouldn't VOA be more appropriate for single games?

120
by nat (not verified) :: Wed, 01/17/2007 - 3:06am

#114 Good point. NE had as many successful plays, but more failures. VOA is a per-play statistic, so that gives SD an edge per-play. Somehow (bounce, bounce) that didn't translate into more plays. It's just hard to believe that six extra NE plays can overcome a 71% to -25% DVOA drubbing.

I did not count punts and kickoffs, because I have no idea what counts as a successful result. I did include the penalty yardage - including the 15 yards before the kickoff.

Still, this was not one of those games where the winning team gains half as many yards but wins on interception and fumble returns. Despite the lopsided VOA, they produced about the same number of "successful series of downs".

#112 Timing of the successes is probably a big part of it. 28 out of 33 of the Pats' successful offensive plays contributed to scoring drives. 16 out of 31 of the Charger's successful plays did not lead to scores.

121
by hwc (not verified) :: Wed, 01/17/2007 - 3:21am

which is an almost certainty for NE considering how weak AFCE has been

That's a misperception. The Sagarin rankings had the AFC East as the toughest division in football this year, followed by the AFC South. All four AFC divisions were ahead of the toughest NFC division.

Remember, the AFC East finished the year with a 12-4 team and a 10-6 team. Buffalo finished a repectable 7-9 against the toughest schedule of any team in the NFL according to Sagarin (Pats twice, Jets twice, Colts, Jags, Bears, Chargers, Ravens). The Bills going 7-9 was actually a very strong season.

122
by Yaguar (not verified) :: Wed, 01/17/2007 - 3:32am

99: Brilliant

123
by hwc (not verified) :: Wed, 01/17/2007 - 3:33am

Timing of the successes is probably a big part of it. 28 out of 33 of the Pats’ successful offensive plays contributed to scoring drives. 16 out of 31 of the Charger’s successful plays did not lead to scores.

That's reflective of the game plan the Pats ended up with on Sunday. Belichick said that, by the end of the first half, they had already concluded that they were not going to get the Chargers into nickle or dime packages, no matter what. And, as a result, running the ball or dink and dunk passing wasn't going to work.

At the same time, they saw that they could handle the blitz packages and decided to go spread the field and go exclusively to a downfield passing game. By it's very nature, that kind of offensive attack will produce a lot of nothing, interspersed with the occasional successful scoring drive.

In a way, the Chargers were very successful. They forced the Pats to try to beat them with their worst stuff (and abandon their best stuff...the ball control run and sort pass game). Unfortunately for the Chargers, the Pats were able to beat them with their worst stuff. What are ya gonna do?

124
by Richard (not verified) :: Wed, 01/17/2007 - 3:35am

123: Unfortunately for the Chargers, the Pats were able to beat them with their worst stuff.

Recovering 100% of fumbles is "their worst stuff?"

125
by hwc (not verified) :: Wed, 01/17/2007 - 3:58am

RE: #124

The Pats were only +1 in the turnover category on Sunday. I don't think that was the determining factor. Don't forget that the Chargers picked Brady 3 times.

Don't get me wrong. I'd rather be +1 in turnovers than -1. But, we aren't talking about -3 or -4 where it is simply impossible to win.

126
by Rick (not verified) :: Wed, 01/17/2007 - 7:16am

re 77, 101

Yes, DVOA has been rather punishing to the Pats' defense all season. After an early season game against the Jets that the Pats dominated for three quarters, only to give up a few meaningless TDs toward the end of the game, DVOA dropped the Pats D to the bottom 3rd in the NFL. The Pats defensive game plan is less geared to stopping yardage than it is to minimizing the number of points scored. Part of this is based on their weaknesses in the secondary. The biggest threat to the Pats is the possibility of a long TD run by a WR matched against non-Assante Samuel (e.g. Cotchery in the Jets' game). Thus, the Pats' D tends to give up yards to prevent the longer passes.

Having said all of that, I think the DVOA of the Pats/Chargers game shows the limitations of the system. DVOA measures what most of us who were watching the game thought: that the Chargers were dominating the game. The problem is that DVOA doesn't measure the gap between the play-to-play domination and the final result. For example, the now-infamous 4th-and-11 call. The Chargers had driven to the 30 yard line and got zero points from it. I suspect that DVOA is assuming that the Chargers on-field production would actually lead to points on the scoreboard.

Should DVOA be tweaked to reflect the actual results? Nah - if we want to know what the actual results are we can simply look at the scoreboard.

re: 125
The bizarre 4th down play would count as a net zero in the turnover category, but it felt like a turnover in the Pats' column. The interception really would not have made a possession difference as opposed to simply hitting the ball to the ground. OTOH, the fumble recovery really did change the possession. 4th down turnovers (and very long 3rd down interceptions) often do not feel like turnovers.

127
by Eagles Fan (not verified) :: Wed, 01/17/2007 - 9:03am

Since there is no separate thread for this, I thought I'd ask here.

What's been the reaction to there being a game in London next season?

Naturally we're delighted over here.

128
by Insancipitory (not verified) :: Wed, 01/17/2007 - 9:10am

The NFL's reaction is fans love giving up regular season home games. As a season ticket holder, even assuming the price of the tickets decreased by one game, I'm betting the per game ticket prices would just go up, I'd be pretty pissed. But since it's not my team... (yet) I'm neutral on it. Sounds like you're getting Giants at Dolphins. Seeing the Giants implode in person is a lot of fun, well worth the price of admission.

129
by Rick (not verified) :: Wed, 01/17/2007 - 9:23am

re: 127
I'm guessing that if there were a regular season game in London, it would involve the Patriots, since they are the closest team (in addition to being quite successful and having lots of fans in London). In addition to their AFC East rivals, their opponent could be the Colts, Chargers, any AFC North team or any NFC East team. Which team would lose a home game? Would said team be compensated?

From the current perspective, the obvious matchup to pick would be Pats/Colts, esp. if Sunday's game is good (I expect it to be) and the winner wins the Super Bowl (which I also expect to happen).

130
by Eagles Fan (not verified) :: Wed, 01/17/2007 - 9:26am

You're right. I saw them doing exactly that against the Chargers in San Diego last year. Priceless...

Losing a home game is a huge deal for both the fans and the team's prospects (though I guess the Fins have pretty much zero prospects next year anyway).

Look at the outcry when the Saints had to play at Meadowlands last year.

That said, whoever comes over here will make plenty of cash.

131
by Eagles Fan (not verified) :: Wed, 01/17/2007 - 9:30am

#129

Apparently it's going to be be Giants @ Dolphins in late September or early October. Both are pretty well supported over here, especially the Dolphins.

132
by admin :: Wed, 01/17/2007 - 10:18am

Commentary now up:

http://msn.foxsports.com/nfl/story/6376778

Just to make sure people understand, we've given up entirely on policing the comment threads this week. Pats-Colts makes otherwise intelligent, reasonable people into blithering idiots -- no matter which team they root for -- and the irrational nonsense has taken over. We're trusting that normal sanity will return next week and hope that longtime readers will bear with us until then.

That also means that if you've made a suggestion or a request in any comment thread related to the Pats or Colts, I'm not going to read it. Sorry about that. Send an e-mail.

133
by Homer S. (not verified) :: Wed, 01/17/2007 - 10:30am

Whoa, whoa....there's a New England now?

134
by steelberger (not verified) :: Wed, 01/17/2007 - 10:32am

Re 121: Sagarin rankings just dont work for the NFL. Here is a simple question...who do you think a team would rather play... the Jets or the Bengals...the Dolphins or the Steelers...

The AFC East was very weak this year. The Pats and Jets played against a TOTAL of 9 winning teams between them (only 5 if you take away the games they played against each other), and they lost 6 of those games.

That doesnt sound like an exceptionally strong division to me.

135
by Pat (not verified) :: Wed, 01/17/2007 - 10:48am

#126: Yeah, those sure were meaningless touchdowns. I mean, the Jets couldn't possibly have come back, right? I mean, it wasn't like they made it a one-score game with two-thirds of the fourth quarter left, or got the ball back with a minute left or anything.

The Pats defensive game plan is less geared to stopping yardage than it is to minimizing the number of points scored. Part of this is based on their weaknesses in the secondary.

That's not a gameplan. That's a weakness in the defense. I guarantee that Bill Belichick in the offseason didn't say "You know what we need? A bad secondary. That way, we can build a gameplan around stopping points more than stopping yardage. I think that's really the way to win in the NFL."

It's not insulting, punishing, or whatever else to say that a defense has a weakness. If that weakness doesn't ever cause them to lose a game, that's good coaching, but it doesn't actually fix the problem on defense.

Why do some people have some bizarre desire to see their teams ranked at the top in defense/offense? Last time I checked, it is actually a team game, right?

Incidentally, if you want a list of teams that defensive DVOA has "shorted" over the years, I can give you a long list. And every single one of those teams have similar features - they all have a 'meh' normal DVOA, and a fantastic red zone DVOA, and they all have a unit (defensive line, linebackers, secondary) that's markedly worse than the rest. That unit gets hidden as the field shrinks.

And usually, those teams have underperforming offenses as well, unless the team has great special teams also. This is because they're losing field position on each drive, and they need a way to get it back.

Remember Denver, beginning of the year? How everyone was fawning over the fact that they had allowed so few touchdowns? No one noticed that their offense was starting with the worst starting field position in the league.

I suspect that DVOA is assuming that the Chargers on-field production would actually lead to points on the scoreboard.

It was a 49-yard field goal. That would have led to points on the scoreboard about half the time. The Patriots defense has no control over either the Chargers playcalling nor the Chargers field goal percentage.

Note that I'm not criticizing the Patriots at all (God forbid). But fans are putting way, way too much bizarre man-love on the defense, and ignoring the significant contributions the offense and special teams are having. End result is still the same - Patriots are a top 5 team in the league - but the credit is going to the wrong unit on the team.

136
by J.D. (not verified) :: Wed, 01/17/2007 - 11:28am

#67: I've been saying there is something wrong with Brady's arm ever since they were shut out by Miami earlier this season. He's struggling to get zip on the ball and he has to do a full windup for little 10 yard passes. I'm looking forward to shoulder/arm/Tommy John surgery this offseason, followed by the decline into Tom "Chad Pennington" Brady.

137
by Starshatterer (not verified) :: Wed, 01/17/2007 - 11:37am

Steelberger (#134 )--
The AFC East was very weak this year.
Combined DVOA for the divisions in 2006:

AFC East: +12.3%
AFC North: +27.5
AFC South: +8.7
AFC West: -1.0

NFC East: +33.7
NFC North: -5.6
NFC South: -17.4
NFC West: -57.1

If, by "Very weak," you mean "Second strongest in the AFC and third overall," then DVOA agrees with you.

138
by Joe (not verified) :: Wed, 01/17/2007 - 12:05pm

re: 93

Numbers lie when they are intentionally misused by people with agendas (agendae?). The good people here at FO are trying for a predictive system. If one team beats another by 100% in DVOA, that means that the higher team will win something like 9/10 times. Just because the 1/10 time happened, it doesn't mean the system is wrong. Last week's win-loss results are unimportant when it comes to predicting who will win next week. I give FO credit for sticking their necks out and saying the Pats played a much worse game than the Chargers DESPITE the result. Remember earlier this year, when the Eagles were ranked in the top 5 despite their 6-5 record? Even a McNabb-less Eagle team proved to be the 3rd best in the NFC. It's easy to be swayed by winning percentages and results, but those do not predict outcomes.
As they caution every week, DVOA is a tool. If you think the Pats' situational savvy is worth 3 extra points per game, then use DVOA and then tack on 3 points to the Pats.

139
by mactbone (not verified) :: Wed, 01/17/2007 - 12:14pm

Re 93:
I'll remember this. When I'm a coach I will spike the gatorade with Prozac. My players will be the most calm, cool and collected individuals ever which means they will surely win the game.

140
by Rich Conley (not verified) :: Wed, 01/17/2007 - 12:31pm

"That’s not a gameplan. That’s a weakness in the defense. I guarantee that Bill Belichick in the offseason didn’t say You know what we need? A bad secondary. That way, we can build a gameplan around stopping points more than stopping yardage. I think that’s really the way to win in the NFL."

Pat, I would agree with you, if the Red Zone defensive DVOA wasnt so much higher. Its pretty damn obvious at this point that bellichek really doesnt care what happens in the middle of the field.

141
by Rich Conley (not verified) :: Wed, 01/17/2007 - 12:35pm

"Incidentally, if you want a list of teams that defensive DVOA has shorted over the years, I can give you a long list. And every single one of those teams have similar features - they all have a ‘meh’ normal DVOA, and a fantastic red zone DVOA, and they all have a unit (defensive line, linebackers, secondary) that’s markedly worse than the rest. That unit gets hidden as the field shrinks."

Right, but theres an assumption here, that DVOA makes, that the rest of the field is as important as the red zone, and I frankly disagree. To make that point, you have to prove that DVOA between the 20s correlates to wins better than red zone DVOA

142
by Rich Conley (not verified) :: Wed, 01/17/2007 - 12:37pm

"Remember Denver, beginning of the year? How everyone was fawning over the fact that they had allowed so few touchdowns? No one noticed that their offense was starting with the worst starting field position in the league."
Yeah, adn DVOA (at the time) said their Defence was better than their Offense, so this goes counter to your argument.

143
by Gus (not verified) :: Wed, 01/17/2007 - 12:46pm

The Ravens are clearly ranked too high because they are no longer playing. Post 93 is way better than this. Teh Patriots are so clutchy that it's uncanny. Tom Brady is the messiah, Bill Belichick is a saint, Asante Samuel is Allah incarnate, and Vince Wilfork is Buddha.

Having just read this entire thread, I want to thank Peter King for making a personal appearance under the name Dan Riley (echoing earlier sentiments). As always Peter, I would ask you to shut up. You give us Pats fans a bad name.

144
by steelberger (not verified) :: Wed, 01/17/2007 - 1:09pm

137: The AFC East, regardless of what numbers you want to use, was just not very good this year.

Sure Sagarin and DVOA say they were good, but the fact is that they have accumulated all of those wins because they were scheduled to play against the NFC North (crap) and the AFC South (not much better than crap).

Lets look at it this way...against teams with winning records at the end of the season, OUTSIDE of the AFC East....

The Pats were 1-2

The Jets were 0-2

The Fins were 2-2 (the class of the division, by the way)

The Bills were 0-4

That is a combined 3-10....wow....

Which brings up another point...the AFC East played a total of 40 games against teams in other divisions...which means that 75% (30) of those games were against teams that were 8-8 or worse. That brings up another...wow.

So basically, the AFC East beat up on bad teams this year...but couldnt cut it against the good teams....does that pretty much sum it up?

145
by steelberger (not verified) :: Wed, 01/17/2007 - 1:20pm

Sorry, my math was off...it is 68% (27 games). That changes everything.

146
by Fat Tony (not verified) :: Wed, 01/17/2007 - 1:23pm

re: #143 Okay, using your advanced metrics, please rank the divisions 1-8. Tell us which divisions were clearly better than the AFC East and why. Bonus points for using the zlions template.

147
by Pat (not verified) :: Wed, 01/17/2007 - 1:29pm

Yeah, adn DVOA (at the time) said their Defence was better than their Offense, so this goes counter to your argument.

No, it doesn't. Their defense was better than their offense. Their offense was bad. It wasn't "worst in the league" bad, though, which is what they looked like by points scored, or at least close to it.

To make that point, you have to prove that DVOA between the 20s correlates to wins better than red zone DVOA

No, you don't. If that were true, then you wouldn't use red zone DVOA at all. What you meant to say was "to make that point, you have to prove that DVOA between the 20s correlates to wins at all." But, uh, that's what Aaron did. What you would need to show is that performance between the 20s doesn't correlate to wins at all. Good luck with that one.

Keep in mind red zone plays are already boosted by 20% (it's in 'Our Stats Explained'). So red zone performance already counts more. Boosting it more just makes it correlate with points less.

Of course, you could try to enhance DVOA by boosting red zone offense/defense even more, and correlate it with wins, right? Oh wait. That's already done, here, with "estimated wins" (again, it's in 'Our Stats Explained'. Really, just read through it and read through all the links.)

And red zone offense/defense does improve a correlation with wins. Minorly. About a factor of 6 less than offense/defense in general.

One of the things that Aaron's said many times here is that if there's something that people have thought would improve DVOA, it's probably already been done. This is one of them. It's been done.

148
by Pat (not verified) :: Wed, 01/17/2007 - 1:47pm

Pat, I would agree with you, if the Red Zone defensive DVOA wasnt so much higher. Its pretty damn obvious at this point that bellichek really doesnt care what happens in the middle of the field.

No, it's not. He just doesn't have a choice right now.

Their red zone defensive DVOA isn't that much higher. You're making it sound as if New England is thunderously better, and no other team is like that. Minnesota's the same way. Hey, guess what? Minnesota has exactly the same problem New England has! Denver's pretty similar too, although Denver's problem is the defensive line.

Oh, and hey, look there, Philly's up there as well. They're even #1 in Goal-to-Go defense. And I know a little about that team. Philly, also, has had a strong red zone DVOA defense for the past few years. And Philly also has a striking weakness that gets hidden in the red zone as well (Philly's red zone rushing defense: -10.6%. Philly's rush defense elsewhere - not so good).

So what happens to Philly's defense when a team closes in on the red zone, and especially when they get in goal-to-go situations? The field shrinks. The safeties move closer to the line. And suddenly, those safeties - like Brian Dawkins and Mikell are able to keep teams from gaining a yard with one yard to go. Even with multiple attempts.

Philly has crappy linebackers. New England has a weak secondary. Denver has a weak defensive line. Minnesota has a weak secondary.

This isn't me being biased against New England or something. Philly fans have been talking about a "bend but don't break" defense for a while now. They call it a 'philosophy'. I call it "Jim Johnson can't find decent linebackers."

Now, maybe JJ and Bill Belichick aren't looking to have great linebackers/a great secondary. That's fine. And they probably aren't, because they believe they can compensate for them (and they can). But if they got them, they'd be a much, much better defense.

149
by hwc (not verified) :: Wed, 01/17/2007 - 1:47pm

Here is a simple question…who do you think a team would rather play… the Jets or the Bengals…the Dolphins or the Steelers…

I can tell you without a doubt that the Patriots would rather play the Bengals than the Jets and the Steelers than the Dolphins.

Pats and Jets played against a TOTAL of 9 winning teams

Can you get any more convoluted that that?

As far as Sagarin's ratings "not working", I have not noticed that they are any worse than DVOA at predicting future games. One could argue that the Sagarin ratings are more valid.

At least Sagarin's rating don't move the Patriots DOWN in the rankings after beating the winningest team in the NFL on the road in a divisional playoff game like DVOA does! Common sense tells us that when your #2 ranked team (DVOA) beats your #4 ranked team on the road, the losing team doesn't actually move ahead of the winning team the following week! But, that's exactly what happened with the DVOA rankings. I think that legitimately earn the moniker "doesn't work".

150
by Waverly (not verified) :: Wed, 01/17/2007 - 1:56pm

Re: #120, #123: about the timing of the successful plays

That reminds me of the scoring system in bowling, where getting multiple strikes in a row is worth a lot more than the same number spread out over the game.

Does VOA account for which play number it is in a drive? Should gaining a first down on 3rd and 4 at mid-field be worth more if it's the third first down than if it's the first first down of that drive?

151
by hwc (not verified) :: Wed, 01/17/2007 - 1:56pm

BTW, when I say "doesn't work", I am perfectly aware that Aarron make no claims for DVOA's ability to predict future games. I happen to think DVOA works very well for what it attempts to do.

152
by steelberger (not verified) :: Wed, 01/17/2007 - 1:59pm

RE 145: I sense a little sarcasm...but you asked for it, so here it is.

1. AFCN 9-10 .474
2. AFCS 8-12 .400
3. AFCW 4-7 .364
4. NFCS 4-10 .286
5. NFCN 4-11 .267
6. AFCE 3-10 .231
7. NFCE 2-8 .200
8. NFCW 3-16 .158

153
by steelberger (not verified) :: Wed, 01/17/2007 - 2:06pm

"I can tell you without a doubt that the Patriots would rather play the Bengals than the Jets and the Steelers than the Dolphins"

I guess you just cant reason with a homer.

"At least Sagarin’s rating don’t move the Patriots DOWN in the rankings after beating the winningest team in the NFL on the road in a divisional playoff game like DVOA does!"

Letting a little anger show are we? I didnt say DVOA was better or worse.

154
by steelburglar (not verified) :: Wed, 01/17/2007 - 2:06pm

144
The AFC East, regardless of what numbers you want to use, was just not very good this year.

Sure Sagarin and DVOA say they were good.

Contradictory?

The AFC East had a better record (23-17) outside their division than any other division in football. Sure, if they'd lost one more game to the Jags, Bengals, Packers, or Titans, they'd be able to say they played more teams with "winning records".

Carry that forward, they only played 14 against losing teams outside their division, which means 26 games (65%) were against teams 8-8 or better. Wow.

Taking it a little further, the AFC East was matched up against the AFC South, which was 22-18 outside the division, second best in the NFL and the NFC North, which had an NFL best 26 wins against NFC teams. And, interestingly, three of the four teams still practicing hail from these three divisions.

And the important matter, when you do get to play against losing teams (like the Raiders) is to go out and record the W.

155
by Pat (not verified) :: Wed, 01/17/2007 - 2:07pm

#151: Why is there this wacko fascination with predicting games? We know football can't be predicted. When teams play each other a few weeks apart, you can get completely different outcomes.

When the game itself can't predict the outcome, no ranking system is ever going to.

156
by slo-mo-joe (not verified) :: Wed, 01/17/2007 - 2:09pm

#148: That makes sense, but to a point. My question is, why doesn't the same happen to a defense that is already good between the 20's? Shouldn't that be an even better red zone defense, after all, than one that only manages to compensate for its deficiencies in one unit because of the short field? There still must be something the NE D does in the red zone that makes it take advantage of the short field better than other D's, even those that are better elsewhere on the field.

157
by b-man (not verified) :: Wed, 01/17/2007 - 2:11pm

110: It is already the ultimate FOMBC. It is delicious irony that the Patriots, the very team that brought about DVOA (since the site’s creator wanted to determine why the 2001 Patriots succeeded, flying in the face of the talking heads and standard NFL statistics), is the same team that is now causing DVOA great pains by defying it.

158
by steelberger (not verified) :: Wed, 01/17/2007 - 2:14pm

Re 154: Well, I guess you win. I dont even know why any team would show up to play such a superior division....

And I guess when you play a losing team (like the Dolphins) it is good to go out and record a point.

159
by Fat Tony (not verified) :: Wed, 01/17/2007 - 2:17pm

re: 152 Thanks for responding in good humor. Now, is that really how you would rank the divisions?

160
by b-man (not verified) :: Wed, 01/17/2007 - 2:17pm

158: Are you going to just respond to your own posts all day? Do you look in the mirror and argue with yourself too?

161
by steelberger (not verified) :: Wed, 01/17/2007 - 2:19pm

160: Read before you write.

159: No, I would rank them in that exact order personally, but it would be pretty close.

162
by hwc (not verified) :: Wed, 01/17/2007 - 2:20pm

“I can tell you without a doubt that the Patriots would rather play the Bengals than the Jets and the Steelers than the Dolphins�

I guess you just cant reason with a homer.

It has nothing to do with homer, except that I've watched the Pats play all of the teams you mentions. The Pats blew out the Bengals and have been successful against the Steelers. The Dolphins give the Pats fits, every time. If you recall, the Dolphins totally destroyed the Pats 21-0 in December and the game wasn't as close as the score indicated. If you want to stop the Pats offensive, put the Dolphins defense on the field. The matchup is horrible for Brady and the Pats. Always has been. The combination of pressure on the QB, effective man press coverage, and good run stopping just shuts down the Pats.

The Jets were a good football team this year and a team that also gives the Pats fits. It's all about matchups.

163
by MJK (not verified) :: Wed, 01/17/2007 - 2:29pm

The problem with arguing about which division was best or which schedule was easy is that people make up their minds about that after about week 6 or so, and then color all other opinions of later happenings based on their pre-concieved notions. Buffalo and Miami and New York all started slow (partly because they had to all play teams that turned out to be pretty good fairly early), so everyone decided that the AFC East was weak.

steelberger, I bet almost every team, at best, beats up against bad teams but struggles against good teams, especially when you consider an entire division at a time, since your're including good and bad teams in your sample. How about putting up similar stats for some other divisions so you're comparing apples to apples? How did the AFC West do against teams with winning records at the end of the season that were'nt in the AFC West? How about the AFC South? How about any NFC division?

I've played with using Maximum Liklihood Estimation to rank teams (linked to my name)--it gives a good picture of schedule strength taking all games into account. Click my name for the link. By my estimate, leaving out the "floating games", the #2 team (so you minimize the effect of only looking at the best or worst team) in the AFC East (Jets) had to play 8 games against above average teams and six against below average teams. This is typical (for the AFC at least). The AFC North: 6 above average teams and 6 below (Pittsburg was almost exactly average). AFC South: 8-6. AFC West: 7-6-1. So each AFC division had similar schedules. Wins?

AFC East: 35
AFC North: 33
AFC South: 34
AFC West: 34

I just don't see the argument that the AFC East is weak...

164
by steelberger (not verified) :: Wed, 01/17/2007 - 2:30pm

Re 163:

"How about putting up similar stats for some other divisions so you’re comparing apples to apples?"

See post 152.

165
by Rich Conley (not verified) :: Wed, 01/17/2007 - 2:33pm

"No, you don’t. If that were true, then you wouldn’t use red zone DVOA at all. What you meant to say was 'to make that point, you have to prove that DVOA between the 20s correlates to wins at all.' But, uh, that’s what Aaron did. What you would need to show is that performance between the 20s doesn’t correlate to wins at all. Good luck with that one."

Saying that dvoa correlates to wins doesnt meant that specific parts dont correlate better.

You have to prove that DVOA between the 20s correlates better than DVOA in the redzone to prove that its more important, and is indicative of anything.

166
by steelberger (not verified) :: Wed, 01/17/2007 - 2:33pm

162...I do see your point...but consider a team outside of both the AFCE and AFCN...what do you think they would pick?

167
by Pat (not verified) :: Wed, 01/17/2007 - 2:34pm

My question is, why doesn’t the same happen to a defense that is already good between the 20’s?

Why would it? Those defenses might not be as imbalanced. They wouldn't be as good as New England is in the red zone, but they'd be better elsewhere. (*)

Let me give examples so it's clearer, although all these numbers are made up, okay?

Imagine you could separate DVOA by linebackers/D-line/secondary. Then Philly might have a DVOA of 50%/-10%/-50% (i.e. the linebackers are really bad, the D-line is above average, and the secondary is fantastic). In the normal portion of the field, all three of the units are important, and so the DVOA of the defense is ~-3% or so. In the red zone, however, the weakest unit gets hidden, so their DVOA jumps to -30%.

Another team (team 'B') might be -20%/-20%/-20% - that is, all three units are significantly above average. The DVOA of the defense anywhere on the field would be -20%.

In the end, team B's defense will contribute more to their wins than Philly's will. Philly might even allow less points than team B's defense, but team B's offense will, on average, score more from the better field position than their defense allows.

(*: You might then say 'well, why does it matter? if they give up more points in the red zone, that's all that matters!' That's not true. Giving up yardage reduces your ability to score points on offense. Just like defenses shouldn't be blamed when they give up 3 points if the ball is turned over in field goal range, offenses shouldn't be credited when the ball is turned over to them already in field goal range. Defenses score points, too.)

168
by MJK (not verified) :: Wed, 01/17/2007 - 2:34pm

Oops, wrong link. Ignore that first one unless you really want to read my resume... ;-)

169
by steelberger (not verified) :: Wed, 01/17/2007 - 2:37pm

168...I do actually.

I work closely with MIT Lincoln Labs. Did you ever do any work there?

170
by Rich Conley (not verified) :: Wed, 01/17/2007 - 2:39pm

"Why would it? Those defenses might not be as imbalanced. They wouldn’t be as good as New England is in the red zone, but they’d be better elsewhere. (*)"

So a defense that has no weeknesses is going to be weaker in the redzone than a team that has a glaring weakness? Thats crazy.

171
by hwc (not verified) :: Wed, 01/17/2007 - 2:40pm

There still must be something the NE D does in the red zone that makes it take advantage of the short field better than other D’s, even those that are better elsewhere on the field.

I think it two things, and Belichick has talked about both.

First, the Pats get aggressive in the red zone (blitzes, etc.) while typically playing very conservatively between the 20s. They get aggressive because there is less field to cover and no threat of the deep ball in the red zone. There's no reason to be conservative in the red zone.

Second, the Pats sell out to stop tendencies in the red zone. Belichick said it this week in so many words, "you can't stop everything in the red zone." So you have to take your best shot and gamble on what you think the offense wants to do. I think it's fair to say that the Pats coaching staff is probably better than league-wide average at identifying opponent tendencies.

Actually, plenty of teams have a number one priority of not giving up big play TDs between the 20s. The Colts are a good example. The difference is that the Colts don't have the horses to shift gears and play gambling aggressive defense in the red zone as effectively as the Pats do. The Pats also have size on defense. In the red zone, that's valuable. Speed is nullified down there because you are only defending within 20 yards of the line of scrimmage (on average).

I can tell from listening to Belichick weekly for many years that he places huge emphasis on red zone defense and red zone offense. For example, red zone offense is clearly why he favors a road grader running game (Corey Dillon) over flashier breakaway backs. The threat of the power run is THE secret to red zone offense. Worrying about a Corey Dillon ties your hands on pass defense in the red zone. (See above, "You can't stop everything". See Jerome Bettis!)

BTW, Belichick has also stated that once he defends the long play, he would prefer to stop drives between the 20s, too. (3rd down efficiency). He cites the whole field position issue as discussed previously in this thread. However, if he had to choose (and with multiple starting DBs on IR in recent years, he has had to choose), he would rather stop the deep ball. Live to play another down. Give the opponent an opportunity to self-destruct.

172
by Pat (not verified) :: Wed, 01/17/2007 - 2:40pm

Saying that dvoa correlates to wins doesnt meant that specific parts dont correlate better.

Yes, it does (in fact, that's what a multiple regression does). Read the article I linked to. The contributions from each part are right there.

You have to prove that DVOA between the 20s correlates better than DVOA in the redzone to prove that its more important, and is indicative of anything.

Are you actually reading what I'm writing? It's not more important. I never said it was more important. In fact, I specifically said that red zone plays were more important! Like, three times!

Let me say it loud, straight, and clear: red zone plays are more important than plays between the 20s. Is that clear enough?

Now, let me say what I have been saying: plays between the 20s still mean something, and therefore you can't just ignore them. And in order to show that, all you need to do is show that DVOA between the 20s correlates at all to winning. Which it does. Therefore, saying "red zone plays are all that matters" is wrong.

I don't know why you keep saying that they would have to correlate better. There is this thing called a "multiple regression" - where more than one factor can influence something.

173
by Rich Conley (not verified) :: Wed, 01/17/2007 - 2:44pm

Pat, again, your discounting strategy. When you have a good offense, defense starts being more about points than field position.

I think the Patriots could get their middle of the field DVOA up (and they do, when its the end of the half, close games, etc, their middle of the field defensive DVOA goes way up). Its just not a priority. They know their offense can move the ball, so they concentrate on cutting off big plays, and preventing POINTS, and not yards.

Yes, its important to keep teams from moving the ball, but its way more important to keep them from scoring.

174
by Rich Conley (not verified) :: Wed, 01/17/2007 - 2:44pm

"Therefore, saying red zone plays are all that matters is wrong."

Please dont quote me on things I didnt say.

175
by Pat (not verified) :: Wed, 01/17/2007 - 2:46pm

So a defense that has no weeknesses is going to be weaker in the redzone than a team that has a glaring weakness? Thats crazy.

Why? I gave an example right there.

It's not like teams are purely described by "SUCK/NOT SUCK" at all positions. Tedy Bruschi can be a great linebacker without being as good as other linebackers, for instance.

I really don't understand your logic here. A defense, on average, can be better than another defense, but the other team can have better units than the first. See also: the Minnesota defense.

176
by Rich Conley (not verified) :: Wed, 01/17/2007 - 2:46pm

If the Patriots defense is hiding a weakness in the secondary, and thats why they give up yards in the middle of the field, how do they deal with that weakness in the fourth quarter? How do they deal with it in close games? How do they overcome that weakness REGULARLY in these situations?

177
by Pat (not verified) :: Wed, 01/17/2007 - 2:50pm

Please dont quote me on things I didnt say.

You did say it. In different words. Quoth I:

You have to prove that DVOA between the 20s correlates better than DVOA in the redzone to prove that its more important, and is indicative of anything.

The only way this is true is if the only factor that matters is the factor that correlates most - that's just statistics. The factor that correlates most is red zone performance. Stating this is equivalent to saying red zone performance is the only thing that matters, and you have to prove that something else correlates more for it to be indicative of anything.

Pat, again, your discounting strategy. When you have a good offense, defense starts being more about points than field position.

I'm not discounting strategy! You're trying to give a "double boost" to the Patriots defense because they have a good offense. They get credit for that offense. That's what makes them a top 5 team!

Do you honestly think the Patriots wouldn't be a better team if their defense gave up, say, 7 fewer yards on average?

178
by slo-mo-joe (not verified) :: Wed, 01/17/2007 - 2:52pm

I am not sure I understand. I can see why the sort field can mask the deficiency of a unit: e.g. weak DB's can get support from LB's dropping back, or whatever. However, the short field heps all defenses, so if your DB's are already good, the short field can only make them better, and so on.

I totally agree that Belichick would rather stop every opponent drive at the 40 rather than in the red zone, but it still seems to me that the fact that the NE D improves so disproportionately in the red zone compared to other D's deserves a more specific explanation than just "the short field masks their deficiencies". It could simply be that since the coaches know they can't stop much between the 20's, they put more emphasis in playing the red zone, where they have a chance because of the short field. But that still requires specific coaching, not just a yardstick.

179
by Pat (not verified) :: Wed, 01/17/2007 - 2:54pm

If the Patriots defense is hiding a weakness in the secondary, and thats why they give up yards in the middle of the field, how do they deal with that weakness in the fourth quarter?

Because teams no longer have the freedom to attack all points of the defense freely. Knowing that, you can play more nickel and play your linebackers in coverage more, assuming you're ahead. (Assuming you're behind, the secondary matters less anyway).

Same deal.

And, oh, look! When you rank them by "late and close", the same teams show up - PHI, MIN, NE.

180
by Starshatterer (not verified) :: Wed, 01/17/2007 - 2:54pm

Steelberger (#143 )--

It may be just me, but I'll take DVOA over "non-division wins against teams with regular-season winning records" as a metric. After all, the AFC divisions ended the regular season with 35-29 (East), 33-31 (North), 34-30 (South), and 34-30 (West) records. (Since they all finished with excess wins, the NFC had all the excess losses, so I'm not bothering with them for now.)

If the East was worse against winning teams, it must follow that all the other divisions were worse against non-winning teams. Which proves exactly what your point does: nothing.

As a side note, maybe the fact that Indianapolis and New England struggled a little more within their divisions, gave them an edge in the playoffs that Baltimore and San Diego lacked? Crazy talk, I know, but since you're all about disregarding the numbers, maybe crazy talk is the way to reach you.

In any case, if it makes you feel better about the AFC North to note that Cleveland beat the Jets and Chiefs, but ignore the fact that they coughed up six wins in their own division, then knock yourself out.

181
by B (not verified) :: Wed, 01/17/2007 - 2:57pm

If Patriots fans want to know what a defense that's good between the 20s and in the red zone is like, watch tames of the 2003 Patriots teams. Which, not coincidently, was the last time the Patriots had a completely healthy secondary. That year, teams weren't getting any yards at all on them. Unfortunately, with no running game, their offense was unbalanced and couldn't move the ball at all, so they won a lot of games with scores like 12-0 or 12-6.

182
by Scott de B. (not verified) :: Wed, 01/17/2007 - 3:03pm

Sure Sagarin and DVOA say they were good, but the fact is that they have accumulated all of those wins because they were scheduled to play against the NFC North (crap) and the AFC South (not much better than crap).

DVOA already adjusts for opponent.

183
by Pat (not verified) :: Wed, 01/17/2007 - 3:07pm

#178: I'm not sure I agree. A DB's speed can be minimized on a short field, for instance.

But anyway, it doesn't matter. Philly's DBs can be better than Baltimore's DBs even though the overall defense in Baltimore can be better. If you get into a situation where the DBs can be emphasized, suddenly Philly's defense looks much better.

I don't see why it's so confusing.

but it still seems to me that the fact that the NE D improves so disproportionately in the red zone compared to other D’s

No, it doesn't. Denver's improved more than that during the regular season. Minnesota's improves almost as much. Philly has improved that much in previous years as well.

It could simply be that since the coaches know they can’t stop much between the 20’s, they put more emphasis in playing the red zone, where they have a chance because of the short field.

Oh, I agree. I think that's a huge part of it. See post #148. I think a huge part of coaching is being able to be better in the red zone than between the 20s.

184
by Ryan (not verified) :: Wed, 01/17/2007 - 3:07pm

"Saints will try to become the first dome team to ever win a playoff game outdoors on its way to the Super Bowl." If the Colts win, won't they also be a dome team to win a playoff game outdoors (last week @Baltimore) on their way to the Super Bowl?

185
by Pat (not verified) :: Wed, 01/17/2007 - 3:12pm

#181: Good point, B. Same goes for Philly back in 2001. Which, also unsurprisingly, is the last time Philly had great linebackers as well. The next year they lost Trotter, in his prime. There was a nice resurgence in 2004 and 2005 (Trotter's over the hill now), but 2002, 2003, and 2006 all had exactly the same theme.

186
by PantsB (not verified) :: Wed, 01/17/2007 - 3:13pm

"Points allowed is the be all of actual results.

No, it’s not! Unless you want to be the one going to tell a team that loses the Super Bowl 7-0 “Congratulations, you held the other team to just 7 points! That’s way below the average!�

the points each team scored.

Yes. But “points scored = offense� and “points allowed = defense� is far, far too simple a model of football. (For one thing, what do special teams do? Twiddle their thumbs?)

New England’s been one of the top 5 teams in the NFL for the past 3 out of 4 years. They’ve had a fantastic offense for 2 out of those 3 years. A fantastic offense coupled with a great red zone defense makes a good defense look fantastic. "
>>>
Its simple because its true. Special teams can accounted for - the Patriots defense is even better if you only include offensive TDs.

The Patriots offense has been good in its run. The strength of the team has always been the defense (and 01 the Special Teams). Belichick has said the only defensive stats worth noting are points allowed and takeaways because those are the ones that matter. Yards don't matter unless they translate to points. Super Bowl teams are always near the top in points allowed - the same can't be said for yards.

I'm sorry but you're assuming DVOA is correct and dismissing the actual results as irrelevant in the measurement of a game. Thats not the mark of a rationalist, but rather one who is clinging illogically to a belief system.

Every metric says the Patriots Defensive DVOA is undervaluing them - Pythagorean wins is always closer than DVOA estimated wins, points allowed and points/Drive are all available and cited on this website. Each has the Patriots defense higher.

The actual results can't be dismissed. If three teams give up 100 more points a year than another team, but a statistical measure of defense ranks them higher... there better be a very big extenuating circumstance.

A 7-0 Super Bowl loss is not the fault of the defense. Clearly the defense has done its job. So yes, I would love to have a D that could guarantee giving up a single TD at the highest level of pro football.

Points are what matter not "successful plays." If "successful plays" translate into points they are worthwhile. If they do not, they are not. The basis of this type of statistical analysis depends on being able to correlate how likely each success or failure will lead to points. If this analysis is leading to variations that can't be dismissed as statistical noise, then the correlation must be examined and either refined or scrapped.

187
by Pat (not verified) :: Wed, 01/17/2007 - 3:21pm

I’m sorry but you’re assuming DVOA is correct and dismissing the actual results as irrelevant in the measurement of a game.

I love this. I've said nothing about DVOA in most of the responses that I've had to you. I'm not talking about DVOA in this. I don't care about it.

All I'm trying to say is that treating "points allowed = defense" and "points scored = offense" is wrong. Even if you limit it to just offensive points allowed/scored, it's still wrong.

I'm not 'assuming' anything. I'm just talking about the game itself. Period.

A 7-0 Super Bowl loss is not the fault of the defense. Clearly the defense has done its job.

Really? So, even if the other team drove the length of the field, and was stopped on 4th and 1 on every single play except one, turning the ball back over to the same team's offense on the 1-yard line, it's the offense's fault?

Even if the game, completely scoreless for four quarters, and in the last two minutes, the defense allows a 99-yard drive for a touchdown which kills the clock and prevents the offense from recovering the ball, it's still the offense's fault?

188
by PantsB (not verified) :: Wed, 01/17/2007 - 3:24pm

BTW
DVOA has matched the "great" offense that is covering up for the New England defense in 03 (DVOA two spots lower), 04 and 06 with points per game quite well. Its never been as highly rated as the New England defense in terms of pts/game, points/drive, DVOA etc in their Super Bowl years.

The Patriots have had less than 1 drive a game less than Green Bay (the most). #1 Defensively DVOA ranked team Baltimore had 2 more.

Stop with making excuses up about why DVOA is reality and actual results are not.

189
by Rich Conley (not verified) :: Wed, 01/17/2007 - 3:33pm

"Really? So, even if the other team drove the length of the field, and was stopped on 4th and 1 on every single play except one, turning the ball back over to the same team’s offense on the 1-yard line, it’s the offense’s fault?"

Not a logical argument. Its a strawman. No one does that.

Good offenses score points. Period. Yes, a good defence helps, but even with a bad defenese, they'll still score lots of points. (See colts, Chiefs 2 years ago, etc)

"Belichick has said the only defensive stats worth noting are points allowed and takeaways because those are the ones that matter. Yards don’t matter unless they translate to points."

This is exactly my point. Its why Red Zone DVOA, IMO, is way more important than middle of the field DVOA. Notice the teams that have great Red Zone DVOA seem to be really good teams? Philly? NE? BAL? MIN could be a great team if their offense wasnt abysmal.

It doesnt matter if you drive 50 yards down the field and punt, if you dont score points. If the other team's offense is good, giving up yards no longer hurts them, and is no longer worried about.

Yes, Belichek doesnt want you to get any yards, but if letting you move 20 yards downfield takes away the chance of a long play with a quick score, he'll take that every time. Why? Because it doesnt really hurt his offense.

190
by hwc (not verified) :: Wed, 01/17/2007 - 3:33pm

I totally agree that Belichick would rather stop every opponent drive at the 40 rather than in the red zone, but it still seems to me that the fact that the NE D improves so disproportionately in the red zone compared to other D’s deserves a more specific explanation than just “the short field masks their deficiencies�.

I think it may simply be coaching philosophies. For example, it is quite possible that the Pats defense could post better DVOA numbers between the 20s if they game-planned to do so. Specifically, third down stops would increase if the Pats played more aggressively. More blitzing, more press coverages, etc. On average, those things will stop opponents drives.

HOWEVER, that same aggressiveness also creates big play opportunites for the offense. It only takes one big play per game to turn the Pats 14.8 points per game allowed into 21.8 points per game allowed. It doesn't matter how great your defensive DVOA stats look if one extra big play per game means you are giving up 21.8 points per game.

At the end of the day, football boils down to how many points you score and how many points you allow. Lose sight of that, whether coaching a team or devising a statistical model, at your own peril.

191
by PantsB (not verified) :: Wed, 01/17/2007 - 3:37pm

"Really? So, even if the other team drove the length of the field, and was stopped on 4th and 1 on every single play except one, turning the ball back over to the same team’s offense on the 1-yard line, it’s the offense’s fault?"
Yes. It really is that simple. Holding an opposing offense to seven points means the D has done its job. The offense has failed its job because it did not score. They were not in an ideal situation for most of its drives but the degree that the opposing defense outplayed our offense exceeds that of the degree that their offense outplayed our defense because of the expected number of points in a game. Our offense has completely failed. Our defense allowed a very low score for nearly any game.

Furthermore, this argument is an extension to the absurd. The reality is that the Patriots field position hasn't been notably bad for the offense, nor have they been deprived of drives on offense. Allowing a longer drive means less drives for both teams not just the Patriots offense. Shortening the game can be advantageous or disadvantageous depending on team style. One less drive per game (from the league maximum) doesn't truly equate with shortening the game in any significant way but even if we assume thats the case, it reduces the points equally. In fact, if that drive doesn't result in points the team is now in a more beneficial position (given field position is neutral) because the score is the same, there is less time and our team has the ball. This is beneficial in every situation except being down multiple scores.

192
by Pat (not verified) :: Wed, 01/17/2007 - 3:37pm

Stop with making excuses up about why DVOA is reality and actual results are not.

Ooh! I can play this game too!

"Stop making excuses as to why points allowed are purely from a defense, and not partly from an offense."

I dare you to explain to me how it's the defense's fault when the offense turns the ball over, or the special teams gives the other offense field position, already in field goal range, and the other team kicks and makes a field goal.

193
by B (not verified) :: Wed, 01/17/2007 - 3:40pm

PantsB, do you think the Current Patriots defense is as good as the 2003 version was?

194
by Brad (not verified) :: Wed, 01/17/2007 - 3:41pm

"A 7-0 Super Bowl loss is not the fault of the defense. Clearly the defense has done its job. So yes, I would love to have a D that could guarantee giving up a single TD at the highest level of pro football"

It sure can be lets just look at 4 drives

Team A recieves the ball at their 20 drives it to the 50 punts to the 20
TEam B takes the ball and drives it back to the 50 punts to the 20
Team A drives it 40 yds to the 40 and punts inside the 10.
Team B drives it 90 yds for a TD.

Score 7-0 for Team B, who has been worse team A's defense or offense. Tam A's defens has been given great starting field position by the offense and special teams and they have allowed 7 points. The offense has been given crappy field position and changed it for the defense on each drive.

195
by PantsB (not verified) :: Wed, 01/17/2007 - 3:42pm

I think it may simply be coaching philosophies. For example, it is quite possible that the Pats defense could post better DVOA numbers between the 20s....At the end of the day, football boils down to how many points you score and how many points you allow. Lose sight of that, whether coaching a team or devising a statistical model, at your own peril.
------
Exactly. My contention is that it is not accidental that there is a disconnect between points/game or points/drive and DVOA. Field position is not remarkable enough to explain the issue.

The SD-Patriots game was an example of this. SD seemed to move the ball at will and thats one reason everyone (including me) had the emotional reaction that the Patriots were down by 30. In reality, the Patriots strengthened when you reached the 35-45 yard line (when they were less worried about the long pass) and the Chargers were forced to punt from just outside FG several times. Since this was a fairly extreme example of the Patriots underlying scheme, it might partially explain the huge DVOA advantage by the team that lost.

196
by Rich Conley (not verified) :: Wed, 01/17/2007 - 3:43pm

188

"BTW
DVOA has matched the 'great' offense that is covering up for the New England defense in 03 (DVOA two spots lower), 04 and 06 with points per game quite well. Its never been as highly rated as the New England defense in terms of pts/game, points/drive, DVOA etc in their Super Bowl years."

I loved hearing about how good NE's offense was last year, when it was blatantly obvious watching that they were a paper tiger. They go into Denver, and the offense just absolutely craps the bed. That awful defense manages to pretty much shut down denver all day though, and then Brady throws some more picks/the offense fumbles it away on route to 5 turnovers, and they lose.

DVOA gives a lot of points for offenses that move the ball in small chunks methodically because they have so many successful plays compared to failed plays. The problem with these offenses (See Pats/Jets) is that all it takes is one failed play to completely blow a drive. Teams that tend to move the ball in more of a boom or bust fashion, dont rate as highly, but are more likely to be able to convert after a failed play because theyre more likely to be able to pick up big chunks of yards.

Thats what we saw against SD. They went into the game playing the consistent ball control style offense. Unfortunately, SD was blowing up atleast one play on each series, and they weren't holding onto the ball. They changed at the end of the half, and suddenly they werent' succeeding any more often, but they put up points around some failed drives. Sometimes, Lower success rate with higher returns works out better.

197
by Pat (not verified) :: Wed, 01/17/2007 - 3:43pm

Yes, a good defence helps

If a good defense helps the offense score points, then points scored is not a measure of an offense. It's a measure of offense+defense. Why is this difficult to understand?

Yes, Belichek doesnt want you to get any yards, but if letting you move 20 yards downfield takes away the chance of a long play with a quick score, he’ll take that every time. Why? Because it doesnt really hurt his offense.

Exactly!

Defensive DVOA is a measure of a defense independent of the offense.

Now, it's entirely possible that New England's defense, with a worse offense, would play even better. But that's the Tennessee Run Defense Problem. You can only measure what you see.

198
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 3:46pm

“Really? So, even if the other team drove the length of the field, and was stopped on 4th and 1 on every single play except one, turning the ball back over to the same team’s offense on the 1-yard line, it’s the offense’s fault?�

"Not a logical argument. Its a strawman. No one does that."

And nobody has won a Super Bowl 7-0, either ;)
I'd make the logical argument, too, but Pat's already done that.

199
by PantsB (not verified) :: Wed, 01/17/2007 - 3:47pm

PantsB, do you think the Current Patriots defense is as good as the 2003 version was?
----
It is comparable. The current team gave up less points but took the ball away less and had a weaker SoS. Do you believe the current Patriots defense is weaker than Green Bay's?

200
by hwc (not verified) :: Wed, 01/17/2007 - 3:47pm

Another premise not to be forgotten is that NFL football is a zero sum game.

That's what makes the sport interesting. The salary cap makes personnel decisions zero-sum. You can't be better across the board. You must decide where you want to be better and where you want to give something up.

Same thing with defense. There is no defense that will stop middle runs, edge runs, short passes, and long passes equally well. Every defensive alignment requires tradeoffs. Bring the eighth guy up into the box and...voila, you only have two CBs and a safety in pass defense.

Blitz six guys? OK, now you're down to five guys covering the entire field. You better hope the blitz gets to the QB before he has time to set and throw.

Defend the long pass as your top priority? OK, but that means you are vulnerable to the run and the underneath stuff.

There's no way around these tradeoffs. They are inherent in the game, just like there are inherent tradeoffs in pulling out the driver on a US Open golf course or playing serve and volley in tennis.

BTW, the really great QBs in this game are great because they understand and can efficiently attack those inherent tradeoffs. See Peyton Manning.

201
by Pat (not verified) :: Wed, 01/17/2007 - 3:50pm

#198: You know what the funniest thing about this whole argument is?

On this very site, just a few bits down, actually, there's a really interesting article. Shows that "hey, the Patriots - their defense, you know, they kinda seem to be built exactly in the right way to win the Super Bowl. More than any other team, in fact."

And people are trying to complain that DVOA ranks their defense poorly.

Yah. Makes no sense to me.

There's a difference between a defense built to dominate offenses (Baltimore) and a defense built to win Super Bowls (New England).

202
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 3:51pm

One more thing. The difference, in football situations, between the 'absurd' and 'normal' is usually a matter of degree.

203
by hwc (not verified) :: Wed, 01/17/2007 - 3:53pm

I loved hearing about how good NE’s offense was last year, when it was blatantly obvious watching that they were a paper tiger.

I totally agree with that. The inability to run the ball made last year's Pats offense fatally flawed. They were forced to play all pass/all the time. IMO, that is a recipe for disaster over the long haul...and especially in the tournament.

A great team needs the ability to win an occassional all pass/all the time game (see NE over SD last Sunday), but it is really a flawed long-term approach. Sooner or later, it will get your QB killed in today's NFL.

Everyone points to NE's defensive injuries last year...and indeed that was huge obstacle. But, the real reason the Pats didn't go further in the tournament was three OL starters on IR and an entire squad of injured RBs.

204
by Rich Conley (not verified) :: Wed, 01/17/2007 - 3:54pm

"Defensive DVOA is a measure of a defense independent of the offense.

Now, it’s entirely possible that New England’s defense, with a worse offense, would play even better. But that’s the Tennessee Run Defense Problem. You can only measure what you see. "

Right, I completely agree. I just think that saying theres a weakness in the defence based on the fact that they give up the middle of the field, when their coach has made a point of saying that only points are important doesnt paint an accurate picture. Theres no evidence that its a weakness, and not a strategy (that obviously is working pretty well). If its a weakness, it should continue to be a weakness, which it doesnt. Their DVOA goes way up in the 4th quarter, especially in close games. In the 4th quarter in close games, everything is game.

Now, you keep saying the weakness is the secondary. Again, theres no evidence of that. According to FO's game charting project, Samuel is one of the top 5 corners in the game, and Hobbs is one of the better corners in the game.

Again, as has been said a couple of times in different threads here. The Patriots have been in the top 5 teams in stopping long plays for the last 4 years. No other team has been in the top 10 more than once (IIRC). I can't help but think thats a scheme thing. They're scheming specifically to stop quick scores.

205
by MJK (not verified) :: Wed, 01/17/2007 - 3:54pm

169 (oops, I hate it when I accidentally post personal information to public message boards...)

Re Lincoln Lab: Yes, I worked there off and on for a few years. But now I live far far away, surrounded by Raiders fans. :-( Did you know that the sports talk radio in the Bay area is still complaining about the tuck rule?

Regarding weaknesses on defense:

I'm not sure Jim Johnson can't find good linebackers, just as it's not true that Belichick can't find good cornerbacks. The problem is that no team can afford to pay good linebacker AND good defensive backs AND good linemen, unless you want to have a bad QB, RB, WR corps, or O-line, etc. You have to pick a weakness. I guess what makes some teams built better than others is picking the right kind of weakness.

206
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 3:56pm

201
I know, and that's the reason I think that the Patriots will win this week. That, and I think Indy's lack of a power running game will really hurt them in the red zone, where NE's biggest weakness can be camoflauged, and where the Colts' biggest strength is somewhat neutralized.

207
by PantsB (not verified) :: Wed, 01/17/2007 - 3:57pm

"There’s a difference between a defense built to dominate offenses (Baltimore) and a defense built to win Super Bowls (New England)."
The point is that the former (Baltimore) and the latter (New England) didn't perform much differently this year. Same number of offensive TDs (21), 2 FG difference the entire season, .15 pt or .007 TD/drives... Baltimore had 6 more INT. Is that what constitutes "domination" as opposed to "Super Bowl style"? Yet on this sight if we assumed linear values Baltimore is considered three times as good, or in terms of places Baltimore is first and NE is 8th.

208
by Rich Conley (not verified) :: Wed, 01/17/2007 - 3:57pm

"There’s a difference between a defense built to dominate offenses (Baltimore) and a defense built to win Super Bowls (New England)."

And thats exactly what we're saying. That the weakness is not a lack of personell or ability, it is a scheme decision, and is no way indicative of how the defense will perform in certain situations.

209
by Pat (not verified) :: Wed, 01/17/2007 - 3:58pm

#204: Except, they haven't always played like that. They beat the crap out of everyone defensively in 2003. They had a fantastic defensive DVOA in 2003. This team is different than the one in 2003.

According to FO’s game charting project, Samuel is one of the top 5 corners in the game, and Hobbs is one of the better corners in the game.

I seem to remember this position called "safety". It might make up one half of the secondary. It might also be disrespectful to ignore this position.

210
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 4:01pm

204
NE's Defensive DVOA on "Deep Zone" is 19th. That suggests to me the lack of a good nickelback, and problems with safeties in coverage, as well as the occasional toasting of a cornerback (which happens to everyone except, seemingly, Champ Bailey).

211
by Rich Conley (not verified) :: Wed, 01/17/2007 - 4:03pm

"Deep Zone" is when the offense is starting inside their own 20. Their DVOA is bad because they, again, play thier corners and safties back and let teams take 5-7 yard plays.

Its not about teams throwing deep on them.

212
by Pat (not verified) :: Wed, 01/17/2007 - 4:04pm

#208: I just don't agree. Belichick is the same guy he was in 2003. We'll just have to disagree.

The problem is that no team can afford to pay good linebacker AND good defensive backs AND good linemen, unless you want to have a bad QB, RB, WR corps, or O-line, etc. You have to pick a weakness. I guess what makes some teams built better than others is picking the right kind of weakness.

Well, not really. You can pick them up in the draft, and keep them under rookie contracts. That's what I mean by "JJ can't find linebackers" - he can't find linebackers in the draft. Then again, Philly basically hasn't drafted linebackers at all in the draft (though the ones they did draft, they sucked at). So I guess I mean he can't find them in the draft, or as a UDFA.

With Belichick, I think it's more "he can't keep them healthy."

213
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 4:04pm

Pat, I think Rodney Harrison would definitely agree with you.

214
by Rich Conley (not verified) :: Wed, 01/17/2007 - 4:06pm

Seriously, look at the charts for each zone. NE's defence gets better the further you move down the field. This competely meshes with what I've seen watching them: The closer they are to their own endzone, the further the corners come up, the further away, the further back theyre playing. Kind of like Baltimore was playing this weekend, with the safties back 30 yards off the line.

This team is always trying to protect itself against long plays.

215
by Rich Conley (not verified) :: Wed, 01/17/2007 - 4:10pm

Watch a NE game. At their own 40 yard line the coverage schemes all change. They go from 7-10 yard cusions to bump and run. The safeties move more to the outside. Its not the personell, or the weakness that changes at that point, its the defensive scheme.

And yeah, bellichek is the same guy as he was in 2003, but none of his defensive staff is. You dont think that could make a difference, do you? Hes lost his coordinator, his defensive backs coach, and a couple of other guys. A change in scheme isnt surprising at all.

216
by Pat (not verified) :: Wed, 01/17/2007 - 4:11pm

#214: Right, exactly. But they weren't always like that. They certainly weren't like that in 2003. Because they didn't need to be - because Harrison can play closer to the line, and still stop a deep play.

217
by PantsB (not verified) :: Wed, 01/17/2007 - 4:13pm

204: Except, they haven’t always played like that. They beat the crap out of everyone defensively in 2003. They had a fantastic defensive DVOA in 2003. This team is different than the one in 2003.
---
Based on? Oh DVOA.

The 2003 defense was better - but it wasn't much better. And it furthers my point.

New England gave up the fewest points by far that year. They had more takeaways than Baltimore. They scored more times on defense. Who was #1 in DVOA? Baltimore. Despite 3+ more points a game, less interceptions, less defensive scores, less per drive... the only thing Baltimore led in was yards/drive and defensive DVOA.

218
by White Rose Duelist (not verified) :: Wed, 01/17/2007 - 4:15pm

#184: Yes, but since the Saints are in the early game, they get the first shot. :)

219
by hwc (not verified) :: Wed, 01/17/2007 - 4:18pm

NE’s Defensive DVOA on “Deep Zone� is 19th. That suggests to me the lack of a good nickelback, and problems with safeties in coverage, as well as the occasional toasting of a cornerback

It may also be situational. Over the years (and this year) I can think of many Pats games when they so comprehensively stuffed the run game that the opponent resorted to a "hail mary" offense -- heaving the ball downfield on every play.

That's basically how the SuperBowl against the Panthers unfolded. In fact, Belichick pondered after the game that they might have been too successful in stopping the run.

With the current NFL rules (can't touch the QB, can't play aggressive pass defense), the Hail Mary offense can be pretty effective. Just heave up jump balls.

Having said that, the Pats are definetly fighting uphill against injuries to the secondary again this year. They have three safeties on IR plus Rodney Harrison has missed essentially the entire season. They also have their starting nickleback on IR. Troy Brown played quite a bit of DB earlier in the season.

On top of that, Ellis Hobbs has played much of the season with a broken wrist/cast. There were games when he was clearly playing one handed, even in tackling.

220
by Pat (not verified) :: Wed, 01/17/2007 - 4:18pm

#217: Based on? Oh DVOA.

Nope. Yards. In 2003 they were near the top of the league in yards/drive, as well. They haven't been since then.

The team was different in 2003.

221
by Rich Conley (not verified) :: Wed, 01/17/2007 - 4:21pm

"That’s basically how the SuperBowl against the Panthers unfolded. In fact, Belichick pondered after the game that they might have been too successful in stopping the run."

That game unfolded the way it did because Harrison, Wilson, and someone else all got hurt in the 2nd quarter. The second half was essentially against the Pats second string.

222
by B (not verified) :: Wed, 01/17/2007 - 4:22pm

212: Indeed, the difference between the current Patriots, and that 2003 version is the lack of a healthy Rodney Harrison. They're no forced to scheme around Hawkins, and give up some yards in order to prevent long plays. It's a good strategy, but the fact is, if the defense was playing better, the offense could then score more points.
And, to answer PantsB's question, I don't think Green Bay's defense is better that the Patriots, but the weighted DVOA rankings don't work for teams that didn't make the playoffs.

223
by Rich Conley (not verified) :: Wed, 01/17/2007 - 4:30pm

"It’s a good strategy, but the fact is, if the defense was playing better, the offense could then score more points."

Watching the Pats, I'm not so sure about that. Their offensive DVOA splits leave an interesting trend
1-20 20-40 40-40 40-20 red
7.9% 12.7% 8.4% -4.6% 37.4%

With those splits, it makes me believe that the most common result, no matter where the Pats offense starts, is to stall in the opponent's maroon zone. They seem, to me atleast, just as likely to score when they get the ball on the 10, as they do when they get the ball on the 40.

224
by PantsB (not verified) :: Wed, 01/17/2007 - 4:31pm

Nope. Yards. In 2003 they were near the top of the league in yards/drive, as well. They haven’t been since then.

The team was different in 2003.
---
But as we've said over and over, yards are irrelevant. In 2001, New England was 23rd in y/drive and 6th in points and points/drive. In 2003, New England was 4th yards/drive, and 1st in ppg and points/drive. In 2004, New England was 15th in yards/drive and 4th in points/drive (and 2nd in ppg). In 2006 they were 13th, 2nd in p/g and p/d.

The best defensive minds (Ryan, Belichick) have dismissed yards as a measure of a defense because in the end its tertiary. Points are the measure of success.

225
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 4:33pm

Going by points, the Bears and Colts had equally good offenses this year.

226
by Pat (not verified) :: Wed, 01/17/2007 - 4:35pm

But as we’ve said over and over, yards are irrelevant.

I don't agree. They need to be put into context.

The best defensive minds (Ryan, Belichick) have dismissed yards as a measure of a defense because in the end its tertiary.

If something is 'tertiary', then it's third-most important. Which means it's not unimportant. Which means it has some importance.

The only question is "how much".

227
by PatsFan (not verified) :: Wed, 01/17/2007 - 4:36pm

Re: #221

Not quite.

In SB38 Wilson was injured late in the game -- on the 85-yard bomb to (IIRC) Muhammed. If you watch the replay, you'll see Wilson grabbing his leg as he hits the ground after his failed tackle attempt. Harrison was injured even later in the game -- on Carolina's tying drive.

In SB39 Wilson was injured on (IIRC) the second half kickoff -- he broke his arm in kick coverage.

228
by PantsB (not verified) :: Wed, 01/17/2007 - 4:38pm

If something is ‘tertiary’, then it’s third-most important. Which means it’s not unimportant. Which means it has some importance.

The only question is “how much�.
----
Actually tertiary means that its third tier, not third most important. When there are things of primary importance such as points and takeaways available, resorting to something that does not correlate well with points OR wins is grasping at straws.

229
by Rich Conley (not verified) :: Wed, 01/17/2007 - 4:39pm

225.

Thats lazy as hell. Look at points per drive.

230
by PantsB (not verified) :: Wed, 01/17/2007 - 4:42pm

Re 225
Actually, Indy has 48 offensive TDs and Chicago has 38. The points difference comes from Chicago's ridiculous number of return TDs.

231
by Pat (not verified) :: Wed, 01/17/2007 - 4:44pm

Actually tertiary means that its third tier, not third most important

Actually, tertiary means "third."

When there are things of primary importance such as points and takeaways available, resorting to something that does not correlate well with points OR wins is grasping at straws.

That's not true at all. If you've got two teams that are very closely ranked in points, for instance, putting more stock in a separation of 0.08 points/drive than over 2 yards/drive is silly.

Again - if something has any importance, then it's not unimportant, and the question is "how much." You can't just throw away the yardage a defense gives up.

232
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 4:47pm

229, 230
exactly. now apply that logic to the field position argument, and voila!

233
by PantsB (not verified) :: Wed, 01/17/2007 - 4:52pm

Tertiary: Third in place, order, degree, or rank.

Yardage is less important than the primary defensive measures: points/unit, and takeaways, and less important than mitigating factors such as SoS and starting position.

You're trying to justify huge DVOA rankings through yards/drive when points/drive and takeaways don't justify those differences. It doesn't work.

234
by Rich Conley (not verified) :: Wed, 01/17/2007 - 4:52pm

Pat, we're not saying to throw it away. What we're saying is that the correlation isnt even. DVOA is used because it correlates to points across the league. That doesnt mean it correlates well with each individual team, or each individual defensive scheme. Yards is a decent measure, but its hardly great.

Has there ever been any work to see if defensive DVOA correlates better to points against the 3-4 or the 4-3? Whether it correlates to points on certain coaches over the course of their career?

The fact that something correlates well across a group (and DVOA doesnt correlate all that well) doesnt mean it is considering the correct factors. Teams that allow more yards per play generally allow more points. That doesnt mean that teams allow more points BECAUSE they allow more yards per play.

235
by Rich Conley (not verified) :: Wed, 01/17/2007 - 4:53pm

232

Except field position doesnt win games. Points do. Field position doesnt mean Jack if you can't put the ball into the endzone.

236
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 4:58pm

235
Anecdotally, Steelers-Jaguars. Week 2 or 3. Jacksonville won that game with field position.
If the 2nd half of Chargers-Patriots played out like the 1st half, field position would've decided that game.

237
by Pat (not verified) :: Wed, 01/17/2007 - 5:01pm

Third in place, order, degree, or rank.

There's another kind of "third"?

The fact that something correlates well across a group (and DVOA doesnt correlate all that well)

The correlation factor between VOA and wins is 0.89. By any measure of statistics, that's a very high correlation. It means that the spread in DVOA explains 80% of the spread seen in the data.

The correlation factor between defensive DVOA and wins is 0.60. Considering the correlation factor between offense and wins is 0.68, that makes sense (the spread you're not explaining is the spread explained by offense - the sum of the two is the total).

DVOA should correlates less well, because it's adjusted for opponent, and wins aren't.

238
by hwc (not verified) :: Wed, 01/17/2007 - 5:02pm

Field position is critical in games with little or no scoring!

Hmmm. Would that make it tertiary or secondary to scoring in importance?

239
by B (not verified) :: Wed, 01/17/2007 - 5:12pm

I think if points are primary, then first downs are secondary, and yards tertiary. Of course, if yards weren't important at all, then teams would never punt. Maybe we can get back to that argument again.

240
by Pat (not verified) :: Wed, 01/17/2007 - 5:12pm

#238: When you're talking about differences between teams of less than a tenth of a point per drive, "secondary" effects become easily the deciding factor.

Oh, and can we drop Green Bay from the discussion? Green Bay's bizarre. They mixed great games on defense and awful games on offense with great games on offense and awful games on defense.

On average, Green Bay worked out to have a better defensive DVOA, but they also had the highest variance in the league, along with Jacksonville.

What is it with bipolar teams this year? Jacksonville - crazy mix of great and awful. Minnesota - bipolar. Carolina - bipolar. Green Bay - bipolar AND crazy mix of great and awful. Miami - bipolar.

241
by PantsB (not verified) :: Wed, 01/17/2007 - 5:12pm

237:
And Pythagorean wins correlates with actual wins better than DVOA with only points scored and allowed. Pythagorean wins (by the nature of the Pythagorean method) weights defensive points allowed more heavily than offensive. DVOA is much more complex and thus should be better than a simpler method over a large enough sample size. We can see with - at least with the Patriots - this is not the case.

Fixing the issue with the Patriots and DVOA might make DVOA a better predictor than simple Pythagorean predictors.

242
by B (not verified) :: Wed, 01/17/2007 - 5:13pm

235: Just a week ago there was a team that won a game with field position despite being unable to put the ball in the end zone.

243
by PantsB (not verified) :: Wed, 01/17/2007 - 5:14pm

"Oh, and can we drop Green Bay from the discussion?"
Its another clear outlier. But fine - how about Minnesota? A mediocre-to-bad team with an easy schedule with mediocre points/game or drive.

244
by Pat (not verified) :: Wed, 01/17/2007 - 5:15pm

Incidentally:

It should be noted that the Patriots are third in defensive VOA. They also happen to be third in points/drive.

Most of the disagreement is explained by opponent adjustments - and points, wins, etc. - none of those are opponent-adjusted.

245
by B (not verified) :: Wed, 01/17/2007 - 5:16pm

241: Pythag Wins will always be a better predictor of past success than DVOA because DVOA adjusts for opponent strength, where Pythag wins does not. However, because DVOA adjusts for strength of opponent, it's better at forcasting the future.

246
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 5:20pm

DVOA does not correlate with actual wins as well as other measurements because wins are not opponent adjusted. Also, DVOA is a predictive model, and does not use random elements of point creation.

247
by Peter Libero (not verified) :: Wed, 01/17/2007 - 5:22pm

Man, you know this argument became a full-fledged stupid message board argument when the competitors disputed the meaning of "tertiary."

Okay, I'm not sure why Rich and PantsB insist that nothing matters except points. Do you really disagree that certain things LEAD to points? Couldn't strategic decisions to limit points against above all else (as you argue Belichick does this season) harm other aspects of the game, like offense?

If you've read Aaron's "field position is fluid" article, you would not pooh-pooh the importance of letting your offense take the field on the 30 instead of the 10. When an offense starts on the 10, it is actually more likely that the opposing team will score first. When an offense starts on the 30, it is more likely to register the next score.

DVOA does not correlate with victory as well as points, primarily because it is opponent adjusted (VOA is almost identical to points) and secondly because it attempts to eliminate aspects of the game which are non-predictive (i.e., lucky) like fumble recoveries. This is why DVOA is more important to understanding why teams win games than points are.

248
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 5:22pm

I really should refresh my browser before posting.

249
by Rich Conley (not verified) :: Wed, 01/17/2007 - 5:23pm

245 Are we sure about that? Has anyone checked with 4 games left in the season whether teams with more pythagorean wins win more consistently than DVOA picks?

"Field position is critical in games with little or no scoring!

Hmmm. Would that make it tertiary or secondary to scoring in importance? "

EXACTLY. Field position is only important if the offense can't move the ball well. Jax/Pitt was a perfect example of that. Once your offense can move the ball well, field position no longer matters, and points are what matters.

250
by Pat (not verified) :: Wed, 01/17/2007 - 5:24pm

#243: Mediocre points/drive? It's 1.45 points/drive. That's not mediocre. And their starting field position was pretty bad - about 2 yards/drive worse than New England. That's like 350 yards of field position over the season. That's about the equivalent of 0.26 points/drive, which puts Minnesota down at 1.19 points/drive.

Basically where New England is.

251
by Rich Conley (not verified) :: Wed, 01/17/2007 - 5:26pm

I'm going to VOMIT on the next person who says
"Okay, I’m not sure why Rich and PantsB insist that nothing matters except points."

I never goddamn said that. What I'm saying is that a defenses ability to stop the opponent from scoring points is vastly more important than its ability to stop them from moving the ball. Field position is important, and giving up field position is bad, but its not nearly as bad as giving up points.

252
by Pat (not verified) :: Wed, 01/17/2007 - 5:29pm

#249: Predicting game outcomes is not the way you measure the success of a metric. The way you measure the success of a metric is seeing how well it correlates with future versions of itself.

The reason why predicting game outcomes isn't the way you measure a metric's success is because the quantization (you can't win less than 1 game) dominates everything.

253
by Rich Conley (not verified) :: Wed, 01/17/2007 - 5:30pm

"When an offense starts on the 10, it is actually more likely that the opposing team will score first. When an offense starts on the 30, it is more likely to register the next score."

Yes, across the league that is the average result. My whole point is that when you start taking different strategies into effect, and different weightings of team Defense/Offense, that may not necessarily be true.

Between the 10 and the 30, theres a big difference, but again, thats the extreme. How about between the 30 and the 40? Is the average point differential in those 10 yards high enough to justify taking greater risk to keep those 10 yards?

254
by Pat (not verified) :: Wed, 01/17/2007 - 5:30pm

#251: What's 'vastly' mean?

Seriously, you're saying that field position matters, right? How much? How much is giving up 20 yards worth? And how do you justify that?

255
by throughthelookingglass (not verified) :: Wed, 01/17/2007 - 5:33pm

249
If a team can't move the ball, then they'll suck no matter what the field position. If they can move the ball at will, then they'll score on every position.
However, every team except the Raiders is in between these two extremes. They can move the ball, but not at will, and will be stopped sometimes. That's why field position is so important for real teams. The longer the field, the more chance of coming away without points. Except if you're the Raiders.

256
by Rich Conley (not verified) :: Wed, 01/17/2007 - 5:35pm

"When an offense starts on the 10, it is actually more likely that the opposing team will score first. When an offense starts on the 30, it is more likely to register the next score."

Also, thats true of a 0% DVOA offense. Is it still true of a 15% DVOA offense? Where is that crossover point for each DVOA rating?

What I'm saying, is that if a team has a better offense, giving up field position means less. The colts dont care if they get the ball on the 10. Why? BEcause theyre still more likely to score next than their opponent. In most games, the colts could allow thier opponents 3 points on every drive, and they'd still win, so for them, keeping opponents out of the endzone becomes more important than keeping them from moving down the field.

257
by Rich Conley (not verified) :: Wed, 01/17/2007 - 5:37pm

254

What I'm saying is that the value of moving between the 40s is lower than the cost of trying to stop teams from moving between the 40s.

IE those 20 yards aren't worth the potential of giving up 7 points.

258
by Pat (not verified) :: Wed, 01/17/2007 - 5:39pm

#255: Ah, Oakland. No defense has ever been screwed so mightily by their offense as Oakland has.

259
by Pat (not verified) :: Wed, 01/17/2007 - 5:44pm

#257:

How much potential? Clearly, the 20 yards aren't worth giving up if they make the chance for a TD go from 0.0001% to 0.0002%.

That's kinda my point - you can hem and haw and say "yards are less important" but you've admitted they have some importance. The question is "how much"?

260
by hwc (not verified) :: Wed, 01/17/2007 - 5:47pm

Pythag Wins will always be a better predictor of past success than DVOA because DVOA adjusts for opponent strength, where Pythag wins does not.

The Sagarin rankings are based on wins/losses, points, opponent strength, and home/away game location.

Just from following their rankings over time, my impression is that they are pretty good at predicting. I wonder how the actual results stack up against Pytharoian wins, VOA, or DVOA?

261
by Rich Conley (not verified) :: Wed, 01/17/2007 - 5:56pm

Exaclty

I feel DVOA puts too much of an emphasis on yards, and disregards points. It overvalues some teams because of that, and undervalues others.

In the case of a team that has said that they dont care about yards, and points are all that matter, I feel that it is very likely to undervalue theyre defensive quality. In the case of the Patriots, their Defense has, in the playoffs, consistenly outplayed their DVOA. The real reason that the Patriots defense, IMO, is rated lower than it should be, is that they allow bad teams to move the ball in the middle of the field. Its funny though, that the good teams they play dont move the ball any better than the bad teams do. Its the opponent adjustment with weak opponents thats killing their rating.

For instance, when the patriots were getting worse and then better and blah blah blah at the beginning of the season, their defensive VOA was remarkably consistent. Their DVOA was swinging all over the place. That tells me, in that case, the defensive adjustments were hurting more than helping.

262
by Pat (not verified) :: Wed, 01/17/2007 - 5:57pm

And as a measure of how much opponent adjustments matter:

New England's average opponent scored 1.66 points per drive. That'd be 17th in the league. They gave up an average of 1.20 points per drive. That's 0.46 below average.

Jacksonville's average opponent scored 1.84 points per drive. That'd be 12th in the league. They gave up an average of 1.28 points per drive. That's 0.56 below average.

263
by hwc (not verified) :: Wed, 01/17/2007 - 6:06pm

I don't pay any attention to oppponent rankings until the second half of the season. It's just impossible to know who the good and bad teams are.

For example, early in the season, I would have said that Denver and Pittsburgh were "tough" opponents and the Titans and Jets were cakewalks. But, lo and behold, it turns out that all four were just the opposite.

264
by Pat (not verified) :: Wed, 01/17/2007 - 6:11pm

#260: About the same, actually. If you search for "NFL Prediction Tracker" you can find comparisons.

It's actually amazing how little difference there is between picking the higher Sagarin rating and picking the higher pythagorean win rating. Both of them only pick at about a 60% clip.

I think, though, it's relatively easy to show that 60-65% is the upper bound on being able to predict NFL games. The games themselves simply aren't separated by much more than that.

I feel DVOA puts too much of an emphasis on yards, and disregards points.

Er? But they didn't hold those crappy teams to few points either. See above.

I think you're actually saying that you don't agree with the opponent adjustments at all. An opponent adjustment simply based on points/drive, etc. shows pretty much the same thing as DVOA.

Miami, for instance, only played 4 different teams than New England - NE, NE, KC, PIT as opposed to MIA, MIA, DEN, CIN. That's 1.97 points per drive vs 1.57 points per drive. Assuming about 48 drives in those four games, you'd expect New England to have allowed about 20 less points than Miami. Adjusting for that, and the yard of starting field position advantage, that puts Miami and New England neck-and-neck.

265
by Rich Conley (not verified) :: Wed, 01/17/2007 - 6:17pm

Er? But they didn’t hold those crappy teams to few points either. See above.

Umm, they had the lowest points allowed in franchise history, so yeah, they did.

266
by Rich Conley (not verified) :: Wed, 01/17/2007 - 6:20pm

"Miami, for instance, only played 4 different teams than New England - NE, NE, KC, PIT as opposed to MIA, MIA, DEN, CIN. That’s 1.97 points per drive vs 1.57 points per drive. Assuming about 48 drives in those four games, you’d expect New England to have allowed about 20 less points than Miami. Adjusting for that, and the yard of starting field position advantage, that puts Miami and New England neck-and-neck."

Right, but you can't just say that, because part of the reason those teams had lower points scored is because they played NE, and not Miami.

My issue isnt with opponenet adjustments, or anything specific. My issue is, again, people thinking DVOA is better than "moderately close".

267
by Pat (not verified) :: Wed, 01/17/2007 - 6:23pm

#265: Umm, they had the lowest points allowed in franchise history, so yeah, they did.

Yes. But they also faced worse offenses than the past few years (though 2003 was even worse). Great defense facing crappy offenses = not surprising that they had the lowest points allowed.

Like I said, your problem isn't "yards/points" - it's the opponent adjustments at all.

268
by Pat (not verified) :: Wed, 01/17/2007 - 6:30pm

#266: Um. Except I just showed that Miami and New England's points per drive weren't significantly different when compared to the average and adjusted for the better starting field position that New England had.

Which means that facing Miami or New England doesn't matter. They both have the same adjusted defense.

Even if you pretend that all teams are average (which... seems odd considering we're trying to -rank- them), that's still only a 1% effect, in any case.

My issue isnt with opponenet adjustments, or anything specific

Before you said you thought DVOA put too much emphasis on yards, not points. Then you said that their VOA was 'remarkably consistent'.

Their VOA puts their defense at third in the league. Dead in line with points/drive. So it's not 'yards vs points'. The only difference between the yards per drive ranking and the DVOA ranking are the opponent adjustments.

269
by Rich Conley (not verified) :: Wed, 01/17/2007 - 6:40pm

You know what theyre also 3rd in? Red Zone defence.

270
by Pat (not verified) :: Wed, 01/17/2007 - 6:50pm

And they're also tied with New England in DVOA, as well! So, which teams in specific do you actually disagree with the rankings on?

Because based on points/drive, adjusted for opponent, it certainly looks like Baltimore makes sense, Chicago makes sense, Jacksonville makes sense (see #262), Minnesota makes sense (see #250) - maybe Carolina and Green Bay are odd, but both of those two seriously came on strong at the end of the season, so depending on how "bad" they were at the beginning, it's understandable that two different methods of 'averaging' (points vs. DVOA) would come up with two different measures. Oakland makes sense (check out their horrific starting field position!), Pittsburgh makes sense (see Oakland).

271
by Paul (not verified) :: Wed, 01/17/2007 - 7:03pm

I was just thinking...how about if we take a team composed of the NE offense and the Indy defense and line 'em up against a team of the Saints offense and the Bears defense. Would we get a more competitive Super Bowl?

272
by hwc (not verified) :: Wed, 01/17/2007 - 7:33pm

Well, the Pats offense is the highest scoring offense of the postseason so far. And, the powerhouse Colts defense is the stingiest defense of the postseason.

So, on paper (and DVOA), that would be a pretty potent combo.

Certainly makes for an interesting game on Sunday. The Pats offensive machine against the Colts defensive juggernaught! And, that folks, is why I don't predict NFL football games!

273
by Richard (not verified) :: Wed, 01/17/2007 - 7:46pm

Marty not fired. Offered extension. Extension turned down.

274
by Eddo (not verified) :: Wed, 01/17/2007 - 8:09pm

Here's my $0.02 into this conversation:
Consider two defenses, A and B. For simplicity's sake, let's say each defense had to take the field for 100 opponent's drives, and that neither defense ever causes a turnover.
Defense A's team has a poor offense/special teams combination that enables their opponents to start every single possession at Team A's 20 yard line. On 80 of the possessions, Defense A forces three straight zero-yard plays, and the opponent kicks a field goal. On the other 20 possessions, Defense A gives up a touchdown on the first play from scrimmage. Overall, they have given up 380 points (3.8 per drive), but only 400 yards (4.0 per drive).
Defense B's team has a very good offense/special teams combination that forces their opponent to start every drive at their own 20 yard line. On 80 possessions, Defense B allows their opponent to move 60 yards downfield, to Team A's 20 yard line, before the drive stalls. From there, the opponent kicks a field goal every time. On the other 20 possessions, Defense B allows their opponent to move 30 yards, to the 50 yard line, before the drive stalls and the opponent punts. Defense B has given up only 240 points (2.4 per drive), but 5,400 yards (54.0 per drive).
So here's your choice: which defense would you rather have? Well, if you look at points given up per possession, you would choose Defense B--they allowed zero touchdowns! However, you could also argue that Defense A had more success actually minimizing their opponents success--after all, they held the other team to zero yards on 80% of their possessions!
Obviously, these are two extreme cases. However, they illustrate the problem with using only one statistic to measure the quality of a defense. It is true that points and points alone determine a win or a loss. However, as seen in my admittedly extreme examples, other factors (primarily field position), affect the amount of points that a defense gives up.

275
by Peter Libero (not verified) :: Wed, 01/17/2007 - 9:00pm

Thanks, Eddo, though it seems like Pants has disappeared and Rich has admitted that yards matter. He just thinks that points are "vastly" more important in a way he cannot quantify, but he feels that DVOA ought to.

I think this is a point that Pat addressed a long time ago in this thread; total DVOA DOES think that red zone performance is more important, but it sees that the bigger picture of stopping progress INTO the red zone is also important. I don't know what else to say about the Pats being "underrated" if we acknowledge that yards do matter. The formula just isn't quite what we think it ought to be? I promise, if Aaron could just shift things around a little to make it correlate with wins better, he would. The Pats system of preventing big plays and allowing a march to the red zone might work for them, and be the best thing they can do with the options they have on defense, but that doesn't mean it's a good way to play defense in general, or that it leads to wins at a higher rate than if they DID stop them before the 20 yard line.

In defense of my earlier post, A) PantsB unabashedly declared that points were all that mattered, and B) Rich suggested that Belichick literally did not care about field position (only that it would be nice), and has defended mythical teams that "decide points are all that matter" as being undervalued by DVOA.

I think what is really kind of ridiculous is the argument that field position matters less to good teams. Most drives, even from GREAT offenses, end without scores of any kind, let alone touchdowns. Field position ALWAYS matters, and EVERY YARD matters; that's the point of the article I mentioned. The difference between 20 and 10 is less than 30 and 10, but it's still significant, particularly over an entire season. No offense is so powerful that it can "score from anywhere" or has no need to start further up the field. It doesn't matter where the "crossover point" is for various offenses, because it increases your chance of scoring no matter how good your offense is.

Anyway, you've basically yielded this point, so we're back to the coaching decision to prevent big plays overriding the decision to work harder on stopping people before the red zone. Like Pat, I dispute the degree to which this is a coaching decision by Belichick, and find the evidence of the 2003 Pats (who could, and did, try to stop them earlier), but really that's immaterial.

Pat's point in 259 is far more important: do we really know the degree to which the supposed conservative pre-redzone playcalling actually reduces points scored? Are we really sure that it's in general a good decision to allow marches down the field in order to have a slightly better defense against big plays? I suppose this requires some evidence which I don't have handy.

276
by Ryan (not verified) :: Wed, 01/17/2007 - 9:08pm

Why does everyone talking about the Colts D seem to overlook that fact that both KC and BAL rushed a combined 37 times? Jacksonville ran 42 times in that infamous game and Houston also had 42 in theirs. Does anyone honestly think that Dillon and Maroney aren't going to have at least 15 carries each?

277
by Eddo (not verified) :: Wed, 01/17/2007 - 9:13pm

Great post Peter, and I have been agreeing with those points and just about everything Pat has said this entire thread. I think what Rich may not have realized that if Belicheck is employing the bend-don't-break strategy, it is showing up in DVOA. It always seems to me that DVOA's detractors often don't realize that skill and strategy show up in DVOA (a.k.a. the "games are won on the field, not on a computer/paper" view). However, those players and coaches on the field are the ones that are determining the teams DVOA in the same way they are determining their points per game and other statistics.
Pat's post #259 is hugely important. And correct me if I'm wrong, but doesn't FO use a more refined form of Virgil Carter's old graph of how much each expected yardline is worth in expected points (the one that goes from -2 to 6 and is linear between the 20's)? That graph was/is based on actual results from years and years of games, not theoretical results. That system should give a value to allowing an opponent to move from their own 20 to your 20 (or from any X yard line to any other Y yard line).

278
by Ben (not verified) :: Wed, 01/17/2007 - 10:08pm

re:276 More likely the reason the rush totals were so low was that the Colts were stopping it and the KC and Baltimore were behind. Jax and Houston both got up 14 points early in their games. I doubt either KC or Baltimore went into their games planning on rushing so few times...

279
by Rick (not verified) :: Wed, 01/17/2007 - 10:08pm

Lots of argument today.

Thought I'd chime in to agree with an earlier poster, that it's unclear who would be the "first dome team to win a playoff game outdoors on the way to the Super Bowl", if both the Saints and Colts won on Sunday. While the Colts victory outdoors (against Baltimore) would have occurred earlier, they would not be "on the way to the Super Bowl" until they beat the Pats, which would happen after the Saints game.

Or would they?

280
by B (not verified) :: Wed, 01/17/2007 - 10:16pm

Well, if they win, the Saints will be the first dome team to win a conference championship in cold, outdoor stadium.

281
by Peter Libero (not verified) :: Wed, 01/17/2007 - 11:29pm

277: Yes, they do use that graph... it so happens that it hits 0 at around 20, so that's why the 10/30 example was particularly meaningful, even though the point of the article is that field position is linear.

I also agree with Ben, from watching the game I never thought "Damn, why aren't the Chiefs/Ravens running more!" It was more "Wow, they keep running, but it never works! How??" In the Chiefs game they only had 24 passing plays to 17 runs, and in the 3rd quarter they were down 16-0, so it's not like running more would've helped. The Ravens might be a little different, since they actually had a comparable number of first downs and weren't horribly far behind at any point, and it's not like the passing was doing great (5.6 y/a vs. 4.2 for the run). So maybe you're right about them.

282
by Pat (not verified) :: Wed, 01/17/2007 - 11:45pm

#277: Yeah. Search for "how many points is a turnover worth." They basically use a third-order polynomial rather than a straight line. When you do that, you find that the 'edges' of the field are worth slightly more than the middle. But it's a small correction.

283
by Pat (not verified) :: Thu, 01/18/2007 - 12:03am

I don’t know what else to say about the Pats being “underrated� if we acknowledge that yards do matter.

Honestly, you don't even need to go that far. I just didn't look at VOA to realize that the Patriots were being shoved down a lot due to adjustments.

Patriots are below Chicago and Baltimore in basically every pace-free defensive metric, so being third in VOA makes sense. Add the opponent adjustment (Jacksonville gets a huge boost), take into account starting field position (Oakland gets a massive boost), and the only defensive rankings which look a bit odd are Carolina and Green Bay. But like I said, both of those two defenses went through massive changes over the year (both Green Bay and Carolina were positive in defense at midway through the season!) and so it's not that surprising that they look a bit odd.

284
by hwc (not verified) :: Thu, 01/18/2007 - 12:07am

Richard Seymour, defensive captain of the New England Patriots, weighed in on our little discussion today in his press conference:

Q: Do you feel like to some degree your defense has fallen under the radar this season? You set a franchise record for fewest points.

RS: Well, I think that's the most important step for us, points allowed. At the end of the day, it doesn't really matter how many first downs you give up, how many rushing yards, all of that stuff, for a personal reason, you never want to have a lot of yards but the most important step is keeping the other team off the scoreboard and we have done a pretty good job of that several times in my career. If we can keep the Colts off the scoreboard and our offense put up a lot of points, ultimately that's what we want.

Q: Do you feel that this defense compares favorably to the championship defenses that you have been on? Statistically it's there on the bottom line.

RS: I think we definitely have a lot of talent. But those teams, like I said, won a championship. Until we win a championship, it still remains to be seen.

285
by Peter Libero (not verified) :: Thu, 01/18/2007 - 12:55am

283: To take a step back from the numbers, Carolina is certainly loaded with talent on the defensive side of the ball, so it's not altogether surprising that they are at least capable of pulling it together into being a good defense. Pre-season they were supposed to be one of the best. Green Bay, uh... yeah...

286
by hwc (not verified) :: Thu, 01/18/2007 - 1:20am

Rereading Seymour's comments, it seems that we can rank defensive goals in order from least important to most important:

Yards allowed
Points allowed
Winning games
Winning the championship

287
by Eddo (not verified) :: Thu, 01/18/2007 - 1:47am

286: That seems rather obvious even without Seymour's comments, no?
Pat, Peter Librero, and others' whole point is not that points allowed is less important that yards allowed, just that yards allowed also has importance. Just like saying that having good special teams is important is not the same as saying that having good special teams is more important than having a good defense.
DVOA even takes points into account by weighting the red zone heavier than the rest of the field. DVOA doesn't (and probably shouldn't) flat-out use points scored in the formula because, as we all know, points can come in from non-predicitve events.

288
by Pat (not verified) :: Thu, 01/18/2007 - 2:30am

because, as we all know, points can come in from non-predicitve events.

Field goals, for one. I mean, when a defense stops a team in field goal range, they have absolutely no control whatsoever as to whether or not that team hits the field goal. Hell, most members of the defense aren't on the field - and even if they were, that's still 'special teams', and not 'defense.'

Any sensible ranking system needs to treat a field goal attempt given up by the defense independent of the result of the attempt itself.

It's like interceptions for an offense - you don't "credit" the offense for the interception if the opposing offense does nothing with it. Same deal.

289
by Pat (not verified) :: Thu, 01/18/2007 - 2:35am

#284: Oh, and regarding the 'Richard Seymour/Bill Belichick/Buddy Ryan/Whoever Else said this to the media, therefore it must be true' idea? C'mon. These guys also talk about 'swagger' and other cliches constantly. It's media. They're not going to go into a detailed discussion of basic football with the media.

Not saying that's what you're putting forth, but similar ideas have been mentioned in the thread before. Football players and coaches say the simplest possible thing to the media. I'm not sure the media could understand more.

290
by hwc (not verified) :: Thu, 01/18/2007 - 3:21am

Pat:

Depends on the coach, I guess.

Belichick routinely goes into quite detailed mini-lectures on professional football during his Friday news conferences -- if a media member is smart enough to ask a good question.

I've heard him talk for 20 minutes on the history of the screen pass in professional football with special emphasis on Tom Landry.

I've heard him talk forever on Paul Brown and his contributions to the game.

I've heard him talk about the biggest change in pro football with the salary cap (having to treat your roster like a college squad with a special emphasis on teaching young players and replacing them every four to five years).

I've heard him go into considerable detail on the differences between punt returns and kickoff returns and why the same guy often can't handle both.

Today, he spent ten minutes talking about his first job in coaching 30 years ago -- charting defenses on punch cards in a cinder block closet with the Colts. Sharing an appartment with Ted Marchabroda and three other assistant coaches.

He has done all day film breakdown sessions for the media in the offseason explaining nuances of the game using film breakdowns from recent Pats games.

And does a weekly film breakdown segment on his TV show. (YouTube link below and in sig)

http://www.youtube.com/watch?v=VBLx8GPzBpQ

I've heard him talk specifically about his philosophy of stopping the big play first between the 20s. To such an extent that he instructs his players that, if they are ever in doubt about their responsibilities on a play, cover deep. He'd rather screw up and have four guys in deep coverage and give up 20 yards underneath than let a play get behind the defense.

One of the key things I've learned from him is that football can never be played perfectly. You can't cover everything. You can't execute without mistakes. You can do everything right and get caught in the wrong play call against the wrong coverage and get killed. So, the key is trying to just shift the odds a little bit in your favor over the course of a game or a season. The "prevent" between the 20s shifts the odds in your favor because not only can the defense make a stop, but so can an offensive holding call or a dropped pass. Give most offenses an opportunity and they'll stop themselves.

291
by Peter Libero (not verified) :: Thu, 01/18/2007 - 5:33am

290: I think it shifts the odds in your favor... if you've built a team around that premise, diverting strengths to other places. Much the same way a team that's built on ace forms doesn't need a good fullback (i.e., the Colts), a team that follows Belichick's philosophy can have a whole section of their team be basically incompetent, like their safeties. It's actually a very smart way to manage the cap, I suppose.

292
by hwc (not verified) :: Thu, 01/18/2007 - 6:05am

a team that follows Belichick’s philosophy can have a whole section of their team be basically incompetent, like their safeties

Well, it's a little hard to be strong at safety when both of your starters and two of your backups miss the whole season with injuries.

That comes on the heels of having six DBs on injured reserve last year.

They are actually quite adequate in coverage. They miss the very physical run stopping safety play of Rodney Harrison. Who wouldn't?

They had the lowest points allowed in franchise history this year (14.8 per game), so it wasn't like the secondary was getting routinely torched.

They don't play as much press coverage as they did with Ty Law, but the rules outlawing pass defense make that a risky proposition regardless of your DBs. When you aren't allowed to even look at a receiver wrong beyond five yards, you've got to be very careful. Play tight coverage and any decent WR can draw a pass interference penalty.

293
by Pat (not verified) :: Thu, 01/18/2007 - 11:26am

They had the lowest points allowed in franchise history this year (14.8 per game)

Yeah, as I pointed out elsewhere, that's because they faced worse-than-average offenses, they had relatively short games (few drives = few points), and, for the first time in a long time, they had a good kickoff kicker. They had the best starting field position on defense in the league.

In other words, the odds were stacked heavily in the defense's favor in terms of 'total points allowed'.

I’ve heard him talk specifically about his philosophy of stopping the big play first between the 20s. To such an extent that he instructs his players that, if they are ever in doubt about their responsibilities on a play, cover deep. He’d rather screw up and have four guys in deep coverage and give up 20 yards underneath than let a play get behind the defense.

Yes, but that's what most defensive coordinators say as well. It's pretty normal. Of course, most of them would prefer to not give up any yardage at all, but it's very tough to do that.

Pretty much, it's just good coaching to do the best with what you have. Sometimes it works. Sometimes it doesn't - when Philly played Indianapolis, they specifically tried to take away the big play way, way too much - and promptly got run over.

But obviously, if you have to make a choice between giving up yards and giving up points, you don't have a world-breaking defense. Just a very good one.

294
by b-man (not verified) :: Thu, 01/18/2007 - 11:46am

292: I think losing Seau was probably the biggest hit to the run D. They have been below average ever since the Chicago game.

295
by PantsB (not verified) :: Thu, 01/18/2007 - 12:55pm

Chicago, Minny and Jacksonville gave up ~120% of the points NE gave up.
Oakland gave up 150%. Green Bay gave 200%.
Nearly all gave up within 3 Int of each other. Oakland is 19th in Pt/Dr and 27th in TO/dr. GB is 21st and 7th. Jax is 4th and 27th. Car is 11th and 29th. New England is 3rd and 4th. New England's D is in the top 5 of every single drive category except yards/drive (which as we've repeatedly contended is by design) and punts/drive.

The metric is not accurately measuring the Patriots D. The number of drives, LoS/Dr (top 5 in both O and D) or any other field position argument. The consistent disparity between DVOA and the Patriots Defense when comparing the metric to pts/game (or drive or whatever) and turnovers indicates that Strength of Schedule and random chance can not explain the difference.

Yards/game and Yards/drive have been shown to not correlate well with wins. Points/game and Points/drive have been shown to correlate with wins. The entire point of DVOA is to correlate with on field performance.

From the explanation of DVOA:

Does it work? Absolutely. The only stat that is a more accurate indicator of how many wins a team will have is points scored and allowed. And DVOA is more consistent from year to year than points or yards. Over the past few years, DVOA has been a better predictor than wins or points when it comes to figuring out how many games a team will win the following year, and the same is true when it comes to using first half stats to predict wins over the second half of the season

Points are superior for predicting wins in season. We have seen that DVOA consistently ranks the Patriots lower on defense than their points and turnovers would indicate they should be. A reasonable assessment would indicate an adjustment is necessary.

296
by PantsB (not verified) :: Thu, 01/18/2007 - 1:04pm

"they had relatively short games"
So? How is that a weakness on D? Because they made the decision to:
"But obviously, if you have to make a choice between giving up yards and giving up points, you don’t have a world-breaking defense. Just a very good one."
it makes the game shorter. You are making assumptions about how football should be played and valuing that more than actual results. Keeping the offense in front of you and stiffening when the situation is more advantageous but before points can result is not inferior to a big play defense unless the latter has superior results. In any reasonable terms, many of the teams ranked above New England both this year and in past years had inferior results.

297
by VarlosZ (not verified) :: Thu, 01/18/2007 - 1:07pm

#295: "Points are superior for predicting wins in season."

Huh? The text you quoted contradicts this: "Over the past few years, DVOA has been a better predictor than wins or points when it comes to figuring out how many games a team will win the following year, and the same is true when it comes to using first half stats to predict wins over the second half of the season."

298
by Rich Conley (not verified) :: Thu, 01/18/2007 - 1:09pm

"And correct me if I’m wrong, but doesn’t FO use a more refined form of Virgil Carter’s old graph of how much each expected yardline is worth in expected points (the one that goes from -2 to 6 and is linear between the 20’s)? That graph was/is based on actual results from years and years of games, not theoretical results."

Again, you're missing my point. The problem is, if you say that league average, a team starting from the 20 averages 4.2 points, that doesnt mean that against X team it does. Thats my whole problem. Using an average ignores the fact that you're never playing an "average team."

I'm gonna go back and look at the Patriots drives this weekend, and set up a chart of expected points from drives starting at a certain field position.

Just because the average comes out to a number, doesnt mean thats what usually happens. The average carry is about 4 yards. The median is about 2 yards.

If your team has higher expected points further back on the field, than field position becomes worth less. You can not compare across the league for the same reason you can't say that teams should pass every down because passes get more yards than runs: you're ignoring too many factors.

299
by Pat (not verified) :: Thu, 01/18/2007 - 1:12pm

Chicago, Minny and Jacksonville gave up ~120% of the points NE gave up.

In points per drive, Chicago was better than New England. Minnesota and Jacksonville I addressed elsewhere. Jacksonville faced ridiculously better offenses than New England, and Minnesota had more drives as well.

So? How is that a weakness on D? Because they made the decision to:

It's not a weakness. It's just an effect. In basketball, there are things called "pace-free statistics". In baseball, you adjust a Pythagorean win projection based on the runs/game environment. It's exactly the same thing here. You can't use points per game, because not all games are of equal length in terms of the basic unit of the game - drives.

You don't choose to have shorter games. It's a side effect of allowing opposing drives to be longer than average, and your opponent's play selection (which you have no control over).

Shorter games mean fewer points allowed, but they also mean fewer points scored, and a smaller point margin between the two.

Look, if you want to continually fawn over points per game, feel free. But you haven't shown anywhere that it's a better metric. And if you just want to keep banging the pots and touting it as the end all, be all, well, we'll have to disagree.

300
by Rich Conley (not verified) :: Thu, 01/18/2007 - 1:15pm

Pat, is DVOA's prediction of the Pats being the 27th best defense this year, any better?