Writers of Pro Football Prospectus 2008

Most Recent FO Features


» 2017 Adjusted Games Lost

Two NFC teams were hit hardest by injuries last year. One already set the AGL record in 2016, while the other has a coach with the worst AGL since 2002. Also: the Rams' incredible bill of health in L.A., and Tampa Bay's questionable injury reporting.

30 Oct 2007

Week 8 DVOA Ratings

by Aaron Schatz

Here are the Football Outsiders team efficiency ratings through eight weeks of 2007, measured by our proprietary Defense-adjusted Value Over Average (DVOA) system that breaks down every single play and compares a team's performance to the league average based on situation in order to determine value over average. (Explained further here.) OFFENSE and DEFENSE DVOA are adjusted for opponent and consider all fumbles, kept or lost, as equal value. SPECIAL TEAMS DVOA is adjusted for type of stadium (warm, cold, dome, Denver) and week of season.

Opponent adjustments are currently set at 80% and will increase each week until they are full strength after Week 10.

No more DAVE this week, as all ratings now are based solely on performance in 2007. Instead, we bring back WEIGHTED DVOA, the formula which drops the importance of earlier games to get a better picture of how well teams are playing now. Right now, those ratings aren't much different from regular DVOA, but that will change over time, especially for teams that struggled early like the Giants and Saints.

To save people some time, we request that you please use the following format for all complaints:

<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>

1 NE 70.6% 1 73.6% 8-0 49.0% 1 -16.4% 3 5.2% 5
2 IND 57.0% 2 55.6% 7-0 42.8% 2 -13.4% 6 0.8% 13
3 DAL 35.2% 3 35.1% 6-1 24.9% 3 -10.6% 7 -0.3% 14
4 PIT 29.9% 4 37.9% 5-2 13.3% 7 -18.2% 2 -1.6% 19
5 SD 20.0% 9 18.4% 4-3 11.1% 9 -2.3% 14 6.7% 4
6 TB 18.6% 5 10.9% 4-4 15.9% 4 -3.9% 13 -1.2% 16
7 GB 17.3% 7 19.8% 6-1 11.7% 8 -7.1% 9 -1.4% 17
8 TEN 17.1% 6 17.1% 5-2 -3.3% 19 -25.9% 1 -5.4% 29
9 JAC 17.0% 8 11.3% 5-2 8.2% 12 -6.0% 10 2.8% 10
10 SEA 12.4% 10 20.1% 4-3 -2.5% 18 -5.6% 11 9.4% 2
11 PHI 10.6% 13 16.3% 3-4 14.1% 6 1.1% 17 -2.3% 21
12 NYG 6.7% 12 13.3% 6-2 7.0% 13 -5.3% 12 -5.6% 31
13 WAS 4.1% 11 -6.4% 4-3 -12.2% 25 -14.3% 5 2.0% 12
14 BAL 0.5% 14 14.3% 4-3 -13.0% 26 -15.1% 4 -1.6% 18
15 CLE 0.0% 18 -2.3% 4-3 10.3% 11 17.9% 29 7.6% 3
16 MIN -2.0% 15 -0.6% 2-5 -4.2% 20 0.4% 16 2.6% 11
17 CIN -2.1% 16 -4.0% 2-5 14.6% 5 12.0% 24 -4.7% 27
18 BUF -5.8% 23 -9.8% 3-4 -15.1% 28 0.3% 15 9.5% 1
19 ARI -6.1% 17 3.2% 3-4 -2.3% 17 3.4% 18 -0.4% 15
20 DEN -7.8% 21 -13.9% 3-4 10.8% 10 15.6% 28 -3.0% 22
21 KC -9.3% 20 -2.3% 4-3 -15.1% 27 -8.0% 8 -2.2% 20
22 DET -12.9% 24 -5.3% 5-2 0.8% 15 7.0% 22 -6.6% 32
23 NO -14.0% 26 -18.6% 3-4 0.0% 16 10.7% 23 -3.3% 23
24 HOU -14.7% 19 -18.6% 3-5 -4.6% 22 13.1% 25 3.0% 8
25 CAR -14.7% 22 -7.8% 4-3 -4.2% 21 5.2% 20 -5.3% 28
26 CHI -23.8% 25 -26.2% 3-5 -24.6% 31 4.0% 19 4.7% 6
27 MIA -25.1% 28 -30.7% 0-8 2.9% 14 24.6% 31 -3.5% 24
28 ATL -25.6% 27 -21.1% 1-6 -8.6% 24 13.1% 26 -3.9% 25
29 OAK -30.0% 30 -23.5% 2-5 -17.7% 29 6.6% 21 -5.6% 30
30 NYJ -31.2% 29 -32.9% 1-7 -8.6% 23 27.0% 32 4.4% 7
31 STL -40.4% 31 -41.5% 0-8 -22.4% 30 13.7% 27 -4.4% 26
32 SF -50.0% 32 -44.8% 2-5 -34.1% 32 18.8% 30 3.0% 9

  • ESTIMATED WINS uses a statistic known as "Forest Index" that emphasizes consistency as well as DVOA in the most important specific situations: red zone defense, first quarter offense, and performance in the second half when the score is close. It then projects a number of wins adjusted to a league-average schedule and a league-average rate of recovering fumbles. Teams that have had their bye week are projected as if they had played one game per week.
  • PAST SCHEDULE lists average DVOA of opponents played this season, ranked from hardest schedule (#1, most positive) to easiest schedule (#32, most negative). It is not adjusted for which games are home or road.
  • FUTURE SCHEDULE lists average DVOA of opponents still left to play this season, ranked from hardest schedule (#1, most positive) to easiest schedule (#32, most negative). It is not adjusted for which games are home or road. (Note: This is based on 2007 performance only. In other words, DVOA, not DAVE.)
  • VARIANCE measures the statistical variance of the team's weekly DVOA performance. Teams are ranked from least consistent (#1, highest variance) to most consistent (#32, smallest variance).

1 NE 70.6% 8-0 8.0 1 70.3% 1 -0.6% 19 5.3% 13 5.0% 24
2 IND 57.0% 7-0 8.0 1 57.2% 2 0.2% 16 5.7% 10 5.8% 22
3 DAL 35.2% 6-1 6.0 3 34.4% 3 -2.5% 23 -0.7% 18 8.7% 18
4 PIT 29.9% 5-2 5.0 9 27.3% 4 -7.4% 29 -1.3% 19 29.0% 1
5 SD 20.0% 4-3 5.2 6 23.2% 5 0.3% 15 3.7% 14 16.4% 6
6 TB 18.6% 4-4 5.0 8 17.9% 7 2.7% 10 -18.3% 31 24.7% 3
7 GB 17.3% 6-1 5.5 4 16.2% 9 1.0% 13 -13.8% 29 4.2% 27
8 TEN 17.1% 5-2 5.0 7 17.2% 8 1.0% 12 1.8% 17 4.5% 26
9 JAC 17.0% 5-2 5.3 5 18.0% 6 4.4% 9 5.6% 11 3.6% 29
10 SEA 12.4% 4-3 4.6 11 11.7% 10 -8.0% 30 -18.7% 32 17.9% 5
11 PHI 10.6% 3-4 4.8 10 9.2% 11 -5.2% 26 14.9% 3 14.4% 8
12 NYG 6.7% 6-2 4.5 13 8.2% 12 -8.1% 31 9.5% 4 8.2% 20
13 WAS 4.1% 4-3 4.1 18 4.3% 13 7.6% 4 5.4% 12 9.5% 17
14 BAL 0.5% 4-3 3.8 20 0.9% 15 -17.0% 32 24.1% 1 2.8% 31
15 CLE 0.0% 4-3 4.5 14 1.3% 14 0.4% 14 -8.4% 25 8.7% 19
16 MIN -2.0% 2-5 4.1 15 -2.9% 17 -1.1% 20 -9.5% 26 3.8% 28
17 CIN -2.1% 2-5 4.6 12 -2.9% 18 9.1% 2 -10.0% 28 2.1% 32
18 BUF -5.8% 3-4 4.1 16 -2.5% 16 8.3% 3 7.1% 8 12.7% 13
19 ARI -6.1% 3-4 3.4 23 -7.3% 19 -7.3% 28 -14.2% 30 27.7% 2
20 DEN -7.8% 3-4 3.6 21 -10.3% 21 13.2% 1 -8.1% 24 16.3% 7
21 KC -9.3% 4-3 4.1 17 -7.9% 20 -4.4% 25 2.7% 16 9.5% 16
22 DET -12.9% 5-2 4.0 19 -12.5% 23 -5.8% 27 8.9% 5 18.0% 4
23 NO -14.0% 3-4 3.3 24 -11.0% 22 1.8% 11 -9.9% 27 10.0% 14
24 HOU -14.7% 3-5 3.0 25 -17.3% 25 4.5% 8 7.2% 7 13.9% 11
25 CAR -14.7% 4-3 3.5 22 -13.2% 24 -3.2% 24 3.5% 15 14.1% 9
26 CHI -23.8% 3-5 2.6 26 -23.3% 26 5.8% 6 -1.7% 20 6.1% 21
27 MIA -25.1% 0-8 2.3 27 -25.8% 28 5.1% 7 8.3% 6 5.0% 23
28 ATL -25.6% 1-6 2.2 28 -25.4% 27 -0.6% 18 -2.3% 21 4.6% 25
29 OAK -30.0% 2-5 1.7 30 -29.1% 29 -2.2% 22 6.7% 9 13.0% 12
30 NYJ -31.2% 1-7 1.9 29 -30.7% 30 6.2% 5 15.3% 2 3.2% 30
31 STL -40.4% 0-8 1.5 31 -39.8% 31 -0.5% 17 -4.8% 22 14.0% 10
32 SF -50.0% 2-5 1.1 32 -51.4% 32 -1.4% 21 -7.5% 23 9.7% 15

Best DVOA Ever Watch

2007 NE 70.6% 2007 NE 70.6% 2007 NE 49.0% 2007 NE 49.0%
2007 IND 57.0% 1996 GB 49.4% 2000 STL 45.5% 2000 STL 43.7%
1999 STL 55.9% 2001 STL 46.8% 2000 IND 44.9% 2000 IND 43.1%
1996 GB 55.9% 1999 STL 46.5% 2007 IND 42.8% 2004 IND 40.3%
2006 CHI 47.4% 2001 PHI 46.5% 2004 IND 37.8% 1999 WAS 36.3%
1998 DEN 43.6% 2000 TEN 43.7% 1999 WAS 36.5% 1998 DEN 33.0%
2001 STL 43.5% 1999 JAC 43.1% 1998 DEN 35.0% 2006 IND 32.7%
2005 IND 40.6% 2003 KC 40.4% 2006 IND 34.8% 2002 SF 32.1%
2003 KC 40.4% 2005 IND 40.1% 2002 SF 34.3% 2004 KC 30.9%
2002 TB 39.9% 2002 TB 39.9% 2004 KC 31.6% 2005 DEN 30.4%

* * * * * *

I'm really sick today and I think I've been coherent for about three hours total, so no extended commentary, just a couple notes. No team had ever put up 70% DVOA for the season any later than Week 5 until the current Patriots. That rating is just mind-blowing. My nightmare scenario has the Colts beating the Patriots by a last-minute field goal, which would probably mean the Pats still had the higher DVOA even though the Colts were the last undefeated team.

Mike Harris re-did the playoff odds as if each team won the "Game of the Century." If the Colts win, their chances of going 16-0 improve from 4.7% to 11.4%. If the Patriots win, their chances of going 16-0 improve from 12.3% to 20.0%.

Look for updates soon on these pages:

For (short) comments on every team, look for DVOA on AOL, every Wednesday. (This will be linked on the FO Goes Mainstream page.)

Posted by: Aaron Schatz on 30 Oct 2007

201 comments, Last at 02 Nov 2007, 12:18am by Nat


by fish shure (not verified) :: Tue, 10/30/2007 - 8:17pm

I'm going to go ahead and guess that that isn't really Weighted DVOA in blue on that second chart.

by fish shure (not verified) :: Tue, 10/30/2007 - 8:18pm

Oh, looks like it's fixed now. Nevermind.

Does anyone know what odds DVOA would give (roughly) to the Pats-Colts matchup? I suppose it's wildly complicated, but is a 13.6% difference very large? I assume some of it is negated by Indy being at home.

by Briguy16 (not verified) :: Tue, 10/30/2007 - 8:23pm

fish: I think to pick games, you subtract 13% from the DVOA of the visiting team. That makes NE .6% ahead of Indy--virtually a tie.

by Richard (not verified) :: Tue, 10/30/2007 - 8:26pm

3: I thought the difference home field made was determined to be 17%.

by Kulko (not verified) :: Tue, 10/30/2007 - 8:28pm

Re 2: Exactly its very close to the generic home field advantage of 15.

Probably Somethign like 52%-55% wim chance for the colts.

by AndyE (not verified) :: Tue, 10/30/2007 - 8:31pm

looks like the simulation gives the Pats a 61% chance of winning, based on the odds of them going undefeated.

by david (not verified) :: Tue, 10/30/2007 - 8:34pm

Wow... Even with the adjustments from playing half of the top 10, the Broncos numbers are crap. Sweet deal. I know I said this last week, but the numbers are even crazier now: -21.3% difference between past schedule and future schedule! The road back to mediocrity starts now!

by Gus (not verified) :: Tue, 10/30/2007 - 8:39pm

And...the Jets have the worst defense in football! Get rid of Pennington, that bum! He can't tackle for sh*t!

by slo-mo-joe (not verified) :: Tue, 10/30/2007 - 8:41pm

And Baltimore's future to past schedule differential is now a freakish 41%. That's kinda like having to play an extra top-3 team. Insane.

by Yaguar (not verified) :: Tue, 10/30/2007 - 8:42pm

"My nightmare scenario has the Colts beating the Patriots by a last-minute field goal, which would probably mean the Pats still had the higher DVOA even though the Colts were the last undefeated team."

I don't think we get as many DVOA trolls because it's not posted on Fox this year. I doubt people will complain much. DVOA also hasn't been very goofy this year in general. The only oddity is that DVOA doesn't think 2-5 Minnesota sucks.

by Waverly (not verified) :: Tue, 10/30/2007 - 8:44pm

What's the record for largest before-the-game difference in DVOA for two teams playing each other?

It's a good thing the AFC East isn't playing the NFC West this season. It's bad enough within the AFC East as it is.

by TanGeng (not verified) :: Tue, 10/30/2007 - 8:44pm

I thought home field advantage is something like 10% or 3 pts. No matter what though, the game could go either way. But you do have NE getting Richard Seymour back and the Colts getting Harrison back. I think Seymour is going to be more of a force.

But the game will probably be decided by how effective Dallas Clark will be. Rodney Harrison might be exposed in coverage.

by zenbitz (not verified) :: Tue, 10/30/2007 - 8:46pm

what's the worst offensive DVOA of all time?

by Yaguar (not verified) :: Tue, 10/30/2007 - 8:49pm

13: I'm pretty sure it's the Niners from Alex Smith's rookie year.

by TanGeng (not verified) :: Tue, 10/30/2007 - 8:51pm

But those Baltimore Ravens are in for a rough stretch. They're going from the easiest schedule to the hardest. Yikes!!

Do they have to play NE, INDY, the San Diego and the rest of AFC North or something? I'm not expecting the Ravens to win their division!

by Peter (not verified) :: Tue, 10/30/2007 - 8:55pm

How much does the Patriots running up the score count towards their DVOA?

by Yaguar (not verified) :: Tue, 10/30/2007 - 8:57pm

Oh god no.

by Temo (not verified) :: Tue, 10/30/2007 - 9:02pm

16. Please, can we not?

by RMoses (not verified) :: Tue, 10/30/2007 - 9:02pm

DVOA loves Philadelphia!!!

by AndyE (not verified) :: Tue, 10/30/2007 - 9:08pm

There is no number 14. There is no number 14. There is no number 14. However, the playoff odds reports seems to give the Pats a 63.4 to 34.5 advantage next week, assuming the winner of this week correlate to the winner of the season

by jimm (not verified) :: Tue, 10/30/2007 - 9:08pm

Regarding strength of schedule. Sagarin ratings have the Vikings having played the 4th toughest schedule and DVOA has them 20th. I think I read on this site somewhere that Sagarin ratings are better future game predictors than DVOA.

I just find it odd that the spread would be so big.

by Mike Tanier :: Tue, 10/30/2007 - 9:10pm

Just so you guys know, there is no "official" home field advantage modifier for DVOA. In fact, we like to point out that DVOA isn't a blunt instument for use when handicapping games. There are a lot of individual breakdowns you can use to try to get an edge in games (like checking run defense vs. opponent's run offense) that you can use as tools which, with a dose of your own judgment mixed in, can lead to some accurate predictions.

For amusment purposes only, naturally.

by Sam (not verified) :: Tue, 10/30/2007 - 9:12pm

further evidence DVOA isn't everything:

If the Jaguars can go TO tampa bay missing their starting QB (who may return by the Tennessee game in 2 weeks btw) and win, and be 3 games over .500... they are inferior to .500 Tampa Bay. I'm not buying it.

by Peter (not verified) :: Tue, 10/30/2007 - 9:23pm


Let me rephrase the question. How much did the Patriots' second half performance effect their DVOA? This is a legitimate question in assessing the disparity between them and the Colts.

I'm not going to malign Belichick for his late-game philosophy, but I think it bears considering how it should be regarded as far as FO goes. I understand that there's an adjustment for junk time scoring, but that it doesn't completely eliminate the effects of so called "running up the score."

For example, the Chargers throttled Houston on Sunday and probably could have scored 60 if they pressed. But they turned of the jets during the second half. Is their performance on Sunday worse than the Patriots who had more trouble with the Redskins for the first 30 minutes?

by Nat (not verified) :: Tue, 10/30/2007 - 9:26pm

The playoff odds have both IND and NE near locks to reach the playoffs, favored in the Super Bowl if they get there, and favored to reach the AFC championship game. What's more, they are BOTH favored to win the AFC championship if they get there. (The Pats are more favored, presumably because they would be favored if they meet there.)

This is uncharted territory. I don't trust DVOA-to-points conversions or homefield DVOA adjustments at all at these heights. I doubt there is enough similar data to show whether they apply.

But I seriously doubt the Patriots will get to run their clock-killer offense much, unlike this last game. (Yes, check the play-by-play. Brady's last two TD drives each took more than 7:45 off the clock.) The emphasis will be on scoring until late in the fourth quarter, with no room for biasing the play selection towards runs and short passes.

It ought to be fun.

by Kaveman (not verified) :: Tue, 10/30/2007 - 9:26pm

#7: That schedule difficulty differential is encouraging, huh? :-)

Another couple things I like is the special teams rank... now up to 22 from 32, and that number 10 offense (despite the red zone woes which I put down to youth and inexperience).

It's interesting that nobody in the media seems to have picked up on the fact that Shanahan has quietly rebuilt the Broncos, instead focusing on the difficulties the defense is having picking up Bates' scheme.

For the Packers game, Matt Lepsis and Erik Pears were the only starters on offense who were starting last year (or the year before, in Lepsis' case since he was on IR last year). Injuries have played a part in this of course, but to have the 10th ranked offense in the league with almost all new players says quite a bit about Shanahan as a coach.

I wish I was as optimistic about the defense. :-/ Still, Aaron Kampman of the Packers counsels patience with Bates' scheme, and GB did get pretty good on D. So here's hoping.

by blacksuit (not verified) :: Tue, 10/30/2007 - 9:32pm

I think it's pretty intersting that the Jags are 29th in variance.

And are the estimated wins done out of 8 even for the teams with 7 games?

by Eddo (not verified) :: Tue, 10/30/2007 - 9:33pm

22: For amusment purposes only, naturally.
Off-topic question: why do so many writers (Simmons comes to mind) always make disclaimers like this? Gambling is perfectly legal if done through the right channels, so I don't understand this behavior.
Probably an inapt analogy, but when one of the outsiders discusses good tackling or blocking technique, they mean to do it on the football field, not on the street, where it could be contrued as assault. Therefore, why can't they discuss gambling in the context of a legal casino or sports book?
(P.S. I don't mean this as an insult or a dig to Mike Tanier, just a simple question. If there is some legal precedence for making such a disclaimer, then I completely get it.)

by PatsFan (not verified) :: Tue, 10/30/2007 - 9:36pm

Re: #12

On the other hand, TE Kyle Brady got banged up and missed the entire second half. If he can't go on Sunday that's a big, big minus for NE. He's not much of a receiver these days, but he's an absolute monster in the run-blocking and pass-blocking game. He'll probably be needed to stop Freeney from chewing Matt Light up and spitting him out.

It'll also be interesting to see if TE Ben Watson is able to go. He's not nearly as useful as a blocker as K. Brady is, but he's fast for a TE and is very hard to bring down once he has the ball. Can definitely cause matchup problems. I would really be nice if he's able to play.

And there's also whether or not S Eugene Wilson can play. He (like Watson) didn't practice at all last week. Can't have enough healthy DBs when playing the Colts.

by Starshatterer (not verified) :: Tue, 10/30/2007 - 9:37pm

Peter (#24 )--
I’m not going to malign Belichick for his late-game philosophy, but I think it bears considering how it should be regarded as far as FO goes. I understand that there’s an adjustment for junk time scoring, but that it doesn’t completely eliminate the effects of so called “running up the score.
Understand that the final drive by the Patriots' offensive starters began in the third quarter, and lasted almost eight minutes. DVOA loves long, time-killing drives with multiple first downs and lots of successful plays, even for a team that has a big lead.

That, I think, is what most people have ignored in their rush to praise or blame the Patriots for what they did to Washington. They ran about four extra minutes off the clock after the long pass to Moss, and just shy of two of those, came after the fourth-down conversion.

The Patriots weren't running up the score; they were killing the clock. It wasn't the touchdown that put the game out of reach for the 'Skins; it was that almost half the time remaining after they turned it over on downs late in the third, was gone when they got the ball back again.

by Carlos (not verified) :: Tue, 10/30/2007 - 9:39pm

honest question.

FUTURE SCHEDULE lists average DVOA of opponents still left to play this season

Is average DVOA really the best way to estimate future schedule, especially in a year in which there's a crazy DVOA outlier like NE? Thru week 7, washington had the toughest future sked (or close to it), but now that NE is behind them, their past sked has moved up to 4th toughest and future has dropped to 12th. Getting blasted by the Pats is unpleasant, but it still only counts as 1 loss. Hope this question makes sense.

by Chris (not verified) :: Tue, 10/30/2007 - 9:40pm

So, this Sunday will feature a matchup of possibly the best two teams EVER?

by David (not verified) :: Tue, 10/30/2007 - 9:40pm

The game was already out of reach. That's why it's called running up the score.

by PatsFan (not verified) :: Tue, 10/30/2007 - 9:44pm

For the love of god, please not in this thread, too.

by Jimmy (not verified) :: Tue, 10/30/2007 - 9:48pm

I was really going to avoid this but

They ran about four extra minutes off the clock after the long pass to Moss

that is classic.

The game was already out of reach. That’s why it’s called running up the score.

Stop it, everyone is bored. That goes for pats fans too, avoid this topic.

Disclaimer - non pats fan, non pats hater.

by Peter (not verified) :: Tue, 10/30/2007 - 9:48pm


I'm not here to debate the moral value of those final drives. My opinion about Belichick's tactics has no bearing on how DVOA treats his team for conducting them. In assessing this Patriots v. Colts game, we need to know just how much value is assigned to those final two drives towards their total DVOA.

Why? Because we can't ignore the argument that several other teams would (and have) deliberately handed the ball off three times and punted rather than converting on fourth down twice with the game in hand. This, I suspect, is the primary philosphical disparity that distances the Patriots from the Colts.

Again, I'm not disparaging the Patriots. My opinion of their strategy aside, they're a spectacular team, and there's something to be said for the psychological effect of their cutthroat attitude on their opposition.

by Peter (not verified) :: Tue, 10/30/2007 - 9:52pm

In summary, what I mean is that the Patriots' tactics might reward them in DVOA, but not really represent a gulf in ability between them and the Colts. In effect, 'punishing' the Colts for choosing to go conservative when a game is clearly in hand.

by Purds (not verified) :: Tue, 10/30/2007 - 9:54pm

In looking at the numbers, I chuckled when I saw that Indy is #1 in defense against an opponent's top WR. That's such a misleading stat. Indy is likely #1 versus the WR1 because Hedyon, who is a big hitter and I like him, is not very good at coverage, and his WR (often the #2) gets many, many passes in front of him that go for 6-8 yards. The one good thing is that if he's matched up on Stalworth, he ought to do better in stopping YAC than most of the defenders I have seen (thinking of the 'Ole Dolphins defense here).

(PS: Colts fan here)

by Starshatterer (not verified) :: Tue, 10/30/2007 - 9:55pm

The game was already out of reach.
Not with 2:02 left in the third. At that point, it would take a comeback like the Lions had against the Bears in week 4, scoring something like two points a minute to end the game.

But the Lions *did* score two points a minute against the Bears in week 4, so the precedent was there. Grinding eight minutes off the clock meant that the Redskins had to score something over four points a minute to reach 38. The touchdown was less important than the time the Patriots used on that drive.

Not running up the score, running down the clock.

by Purds (not verified) :: Tue, 10/30/2007 - 10:00pm

Looking at WR DVOA, I was wondering what folks in general thought about this: Who would you want roaming the slot postion, Welker or Clark or Stokley?

Stokley is having a good year, but the fundamental question is this: which type of slot receiver would you prefer, the semi-TE, or the more pure WR? (I think both of them do many of the same things well). Clark fits for the Colts, as they don't sub often so as to keep the defense from subbing, and Clark's passible as a TE in short-yardage. But Welker allows NE to really spread it out, running him underneath the routes of the other WR's.

But, for non-Pats or non-Indy fans, which would you like more?

by Carlos (not verified) :: Tue, 10/30/2007 - 10:01pm

To build on my Q in #31 -- I'm sure you can do the math that prompts my question, but by this methodology, if your remaining schedule is
This methodology yields a 23.8 Future DVOA average. In reality it'll yield 1 or 2 wins for an average team.

yields a 23.1 Future DVOA average, and in reality yields 0 or 1 win for the average team.

With this methodology being skewed by NE (and IND), the top schedule gets rated harder than the bottom schedule, even though I think everyone would prefer the top schedule if what you want to do is maximize wins.

by Carlos (not verified) :: Tue, 10/30/2007 - 10:03pm

But, for non-Pats or non-Indy fans, which would you like more?

Clark, and is it even close? Let's see a lot more performance from Welker over time and then reevaluate.

by Nat (not verified) :: Tue, 10/30/2007 - 10:04pm

Are you serious? OK, I'll give you the benefit of the doubt.

The issue with "garbage time" is that if the losing team stops trying, or the winning team swithces to a pure run-out-the-clock strategy, DVOA could get a bit dicey.

We can safely assume that the Redskins continued to try to score right up to the time of their last score, and possibly beyond that. (After all, they had tried very hard to score earlier and failed.) The Patriots certainly stopped seriously trying to score when they took Brady out, although their backup QB did score -- even a blind squirrel finds a nut once in a while. They may have been in clock-killing mode earlier (possibly the entire half, considering the times of possession and length of the drives). They seemed to be trying to prevent a score after that, at least until their next possession.

That gives about 3:00 of garbage time, with some additional time where Washington was trying on both sides of the ball and New England was trying on defense, and possibly biasing their play-calls towards runs and short possession passes on offense.

Not enough to matter - maybe the Patriots are slightly better than their DVOA says.

If you believe the Redskins were "letting" the Patriots score earlier, or that some other effect was in play, you're still only talking about one quarter out of thirty-two, or one half out of sixteen. Again, not much of a difference.

My own belief is that the Patriots were running their clock-killing offense (which includes longer passes only to keep the defense honest - just one such pass in the Skins game), but that since this offense strives for sustained drives, it doesn't skew DVOA much.

So trust the DVOA, at least as far as this issue goes.

by Peter (not verified) :: Tue, 10/30/2007 - 10:05pm


I'd really urge you not to make this into a debate it's not. Some teams who dominated sat on leads last Sunday (i.e. San Diego). But when the game was in question, they played better than the Patriots.

See, that's thing. Now it becomes a question of choice. A dominant team either A) chooses to take their foot off the gas or B) just keeps humming. So are is the team that deliberately stopped being aggressive worse or less able than its counterpoint? I think we'd all agree the answer is no. So maybe the Colts and Patriots are a lot closer than we think.

by Tom (not verified) :: Tue, 10/30/2007 - 10:06pm

#37: The problem with trying to negate the two Pats scoring drives is that you're not really replacing them with anything else. In essence, the Pats played for 60 minutes, when some other teams might have stopped trying after say, 45 minutes. But the difficulty lies in determining exactly where the other team might have stopped trying.

It might not be fair to compare the offensive accomplishment of a team that tries for a full 60 minutes to one which starts handing off to their 2nd string running back after 45 minutes but then it becomes a judgement call as to when you should stop counting plays towards DVOA.

If, as you suggest, the last half of the game was easily in hand and should be ignored, you're trying to compare 45 minutes of effort to 30 minutes of effort, which is just as bad as trying to compare 45 minutes to 60 minutes. I guess what I'm saying is, it's hard to compare apples to apples if we start asking that question.

by Peter (not verified) :: Tue, 10/30/2007 - 10:11pm


Again, not a question of anything the Skins did, but a postulation on what the Colts might do in the same situation. Clearly, DVOA says Indy could probably beat Washington nearly as badly as the Pats did.

So, up 38-0, and 45-0, does Dungy try for fourth down twice? No, probably not. Does it hurts his teams DVOA relative to the Pats? Very much so, I take it. But not really because the Redskins outplyed the Colts, or the Colts are "worse" than the Pats.

by Starshatterer (not verified) :: Tue, 10/30/2007 - 10:13pm

In summary, what I mean is that the Patriots’ tactics might reward them in DVOA, but not really represent a gulf in ability between them and the Colts. In effect, ‘punishing’ the Colts for choosing to go conservative when a game is clearly in hand.
Excpet that the Colts did the same thing. (Not talking morality here, just tactics.) The Colts got the ball with 2:26 in the third and the lead. They ran just over four minutes off the clock, taking the game into the fourth quarter, mixing runs and passes to get first downs, until they ran out of field and scored.

DVOA should see not much difference between those two drives, other than rewardibng the Patriots for their relatively worse starting field position (the Colts' drive started around their own 45, the Patriots' at their own 13).

by Some Dude (not verified) :: Tue, 10/30/2007 - 10:15pm

It's interesting to compare Washington and Baltimore's chances of making the playoffs, and how it reflects the difference in the strength of their respective leagues. Despite the two teams having identical records and a division with one top-tier team and some strong competitors, Washington's odds are three times better! Part of this may be down to the Ravens' more difficult remaining sked (the fact that the Redskins dropped from 3rd to 12th in that category after playing the pats shows just how much that game looms over the Ravens' schedule) but still, it's crazy how much easier life is in the NFC.

by Peter (not verified) :: Tue, 10/30/2007 - 10:17pm


Very good point, Tom. It does become a subjective call. *hears the running-up-the-score debate begin again*

But, as a reasonable and intelligent guy, you might not be able to quantify the point at which Dungy would start handing the ball of, but you can qualitatively recognize that his philosophy would "hurt" his team's DVOA.

I agree that the Pats and the Colts are the two best teams in the NFL, but we haven't seen very much of the Patriots' late game strategy in the DVOA years, so I'm not sure it represents them correctly.

I suspect that the two teams are probably even, rather than separated by the margin DVOA suggests.

by Nat (not verified) :: Tue, 10/30/2007 - 10:17pm

Since I really don't want to get into the "running up the score" debate, I hereby concede that the Patriots were "running up the first downs" well into the fourth quarter, until they took Brady out. If others think that is unsportsmanlike, I guess I'm okay with that.

Now, let's get back to interesting discussions of DVOA.

Does anyone know if the dataset used to tune DVOA includes any matchups even remotely like next week's IND-NE game? How far out of the range of the analysis is this game? Or is this more common than I think?

In particular, do we have any data that helps us understand how a 70.6-57.0 advantage compares to a 13.6-0.0 advantage?

by Mike Tanier :: Tue, 10/30/2007 - 10:35pm

28; I personally do the disclaimer because I am not really a gambler and think it's a bit of a bad habit. I also do it to be coy, which is why I think a lot of writers do it.

It would really, really be lovely to get through the rest of this thread without reading about the Patriots and running up the score. If you do want to talk about such things, remember that there's a special message board on the site for that.

by bsr (not verified) :: Tue, 10/30/2007 - 10:37pm

#44 - How many minutes have the other teams sat their starters during games? I would guess that the Patriots have sat their starters longer then any other team and as a result have had sub par results from the drives started by those back ups in their data more then other teams. Take Matt Cassell's Int for a TD for an example. I would think this would offset the one or two running up TDs that anyone else wouldn't score.

by Fat Tony (not verified) :: Tue, 10/30/2007 - 10:41pm

Peter, what's your basis for asserting San Diego played better than New England on Sunday while the game was "in doubt."

by TanGeng (not verified) :: Tue, 10/30/2007 - 10:45pm

Discussing running up the score and its effect on DVOA is plain stupid. It only hurts those who are good enough to do it but don't do it. It only matters to NE and Indy and even then it depresses their DVOA.

by PatsFan (not verified) :: Tue, 10/30/2007 - 10:45pm

Re: #51


by MRH (not verified) :: Tue, 10/30/2007 - 10:45pm

I don’t think we get as many DVOA trolls because it’s not posted on Fox this year. I doubt people will complain much. DVOA also hasn’t been very goofy this year in general. The only oddity is that DVOA doesn’t think 2-5 Minnesota sucks.

Yeah, because trolls from Fox would have made the Audibles thread so unreadable.

Actually, from my standpoint there are several goofy things about DVOA. Look at KC. The Chiefs are 3-1 against teams ahead of them in DVOA (SD, JAC, MIN, CIN) and 1-2 against teams worse than them (OAK, CHI, HOU). And if they were to play all of them this week, in the same location as the original game, they might easily go 5-2 (move CHI and HOU to the win column and SD to the loss side).

by Fat Tony (not verified) :: Tue, 10/30/2007 - 10:50pm

".... for amusement purposes only" comes from the standard disclaimer written on football cards, the ones that list spreads on the major college and nfl games. As Simmons might say, it's one of those jokes that writes itself.

by Starshatterer (not verified) :: Tue, 10/30/2007 - 10:51pm

Does anyone know if the dataset used to tune DVOA includes any matchups even remotely like next week’s IND-NE game? How far out of the range of the analysis is this game? Or is this more common than I think?
This year's Patriots and Colts are #1 and #2 for total DVOA after eight weeks. The remaining eight teams in the top ten, are all from different seasons.

I doubt it's even a little bit common.

by flounder (not verified) :: Tue, 10/30/2007 - 10:57pm

Re: 31 It would be interesting to be given a variance on the average DVOA of the remaining schedule. It could give you a sense of what you're driving at without going to the trouble of looking up a team's individual schedule.

by Gil (not verified) :: Tue, 10/30/2007 - 11:00pm

Putting the NE-running-up-the-score question aside completely. The Pats handed the 'Skins their third largest margin of defeat ever. The 'Skins were not even remotely in the game after the first quarter. And yet they're still fifth in the league in defense DVOA? As a 'Skins fan, I hafta take a bit of solace in the notion that the 'Skins aren't really that bad: but the Pats are really that good.

by Nat (not verified) :: Tue, 10/30/2007 - 11:09pm

Good point. So we probably have enough data about 13.6 DVOA advantages for teams that are about half as far above avaerage as these. Which I think means we know almost nothing about what this difference means, if it means anything at all.

Turn up the thermostat a few degrees and pipe in some sound and I'd give the advantage to the Colts.

Actually, it will be hot enough and loud enough without any artificial help.

by Will Allen (not verified) :: Tue, 10/30/2007 - 11:11pm

Yeah, I can't really figure out how the Vikings rank so high, other than they are really tough to run against, and their red zone passing defense has actually been decent. It intuitively seems impossible to me, despite Adrian Peterson, that the Vikings could be better than 12 other teams on offense, or that, for instance, that Kelly Holcomb has a positive DPAR.

The numbers can't really capture how pathetic the Vikings were on offense against the Eagles; I don't think there is an adjustment made for the Eagles defense having nine or ten guys within four or five yards of the line of scrimmage for huge chunks of the game, even on third and ten or more. Yes, perhaps that is slightly balanced by the fact that the Vikings rushed a little bit against such a front, but in terms of actually scoring enough points to win, it was just laughable. The Eagles basically were saying, "We dare you to throw downfield, in fact we'll just barely defend the pass downfield, in order to choke off other aspects of your offense.", and the Vikings could do nothing about it. It was the most noncompetitive 2nd half game with a 6 point margin I've ever seen, and the numbers just don't do it justice, given the Vikings weird ineptitude.

by Karl Cuba (not verified) :: Tue, 10/30/2007 - 11:12pm

Is anyone else interested in working out whether or not there is a reason why the Colts and Pats are the two best DVOA teams ever? Personally, I think that the continual roster and coaching staff turnover on most teams (all in the name of parity) has rendered half the league seriously lacking in major areas. Off the top of my head; the Niners, Rams, Bills, Jets, Dolphins, Bears, Vikings, Falcons, Bengals, Broncos and maybe the Texans all have a big weakness. Many of these sides are abysmal in one phase of the game. Add the dearth of decent quarterbacks to this mess and there seem to be an unusual nummber of teams that can't execute and get outmannned when they do.

by pete (not verified) :: Tue, 10/30/2007 - 11:12pm

i've been on these boards now for weeks listening to pennington defenders. Yes, the jets have the 32nd ranked defense and that has little to nothing to do with pennington. But they still only have the 23rd ranked offense, and that is almost entirely on pennington. That is not to mention the killer pick 6's he's thrown this season.

The defense is not good, but i'm not completely discouraged. They are prone to giving up the big play and missing tackles, but there have been times this year in which they have played like a semi-competent unit.

by Nat (not verified) :: Tue, 10/30/2007 - 11:16pm

Or, the Skins may have had one bad game. But that still gets averaged with a defense that DVOA had ranked highly.

The Skins' problem was as much that their offense didn't stay on the field long. That will wear a defense down. (Although giving up a fumble-TD does let the defense take another drive off.)

I would expect the Redskin's defense (and offense) DVOAs to climb a little bit as the rest of opponent adjustment comes in, and the Pats game fades to one-sixteenth of their season instead of one-seventh.

by Kaveman (not verified) :: Tue, 10/30/2007 - 11:16pm

#56: Yeah, because trolls from Fox would have made the Audibles thread so unreadable.

For amusement purposes only (*cough*), if one were to run the following on a Unix box...

wget -O - http://www.footballoutsiders.com/2007/10/29/ramblings/audibles/5671/ | awk '$1 ~ /::/ { if ($2 == "]"); $2 = t[2]; } posts[$2]++ } END { for (n in posts) { if (posts[n] > 10) { print n, posts[n]; } } }'

... one would observe that 32% of the posts there were made by 5 people.

Just sayin...

by Kaveman (not verified) :: Tue, 10/30/2007 - 11:21pm

Hmm, there were angle brackets in that script which didn't make it and snipped out a bit. If I remember my rusty HTML correctly, this should work... if not, never mind, it's silly anyway.

wget -O - http://www.footballoutsiders.com/2007/10/29/ramblings/audibles/5671/ | awk '$1 ~ /::/ { if ($2 == "]"); $2 = t[2]; } posts[$2]++ } END { for (n in posts) { if (posts[n] > 10) { print n, posts[n]; } } }'

by John (not verified) :: Tue, 10/30/2007 - 11:23pm

Add the dearth of decent quarterbacks to this mess...

Actually, I think there are plenty of decent quarterbacks, although a few teams are obviously lacking.

There aren't many great quarterbacks, but I assume that's a universal truth. There aren't many great anythings.

by B. (not verified) :: Tue, 10/30/2007 - 11:28pm


How come heat aversely affects the Pats but not the Colts? I guess the laws of thermodynamics disrespect the Pats.

by Will Allen (not verified) :: Tue, 10/30/2007 - 11:33pm

Karl, I think the variance in qb quality throughout the league is simply gigantic, which I suppose is not too surprising, given the position gets harder and harder to execute well, given the paramount importance of the passing game, and increasing speed and size of defensive players on a field that stays the same size. The production out of that position dominates team performance so much, relative to other positions, that such a huge variance may be leading to situations where parity starts to recede.

I have great respect for Dungy and Belichik, but I think they would be the first guys to tell you that their most brilliant moves were to sign a contract to coach a team with Peyton Manning on the roster, or to trip over Tom Brady at the bottom of the draft. The best coaching job in the NFL this year may be by Bill Parcells, in nurturing the career of an undrafted nobody for three seasons, before he was ready to play. Then again, if Quincy Carter doesn't relapse, Romo may have been cut, and who knows what happens to him.

by kevinNYC (not verified) :: Tue, 10/30/2007 - 11:34pm

Looking at the huge drop in the Giants' offensive rankings, I guess there's no adjustment for playing on a different continent in driving rain on grass not fit for a pee wee game. Unless 7 fumbles by both teams is considered "normal". Plus, the Giants are the only team in the top half outperforming their estimated wins by a substantial margin. Can one bad offensive performance and 2 horrible defensive performances really do that?

Will we have to hear again how Philly is actually a good team, but very unlucky?

by AndyE (not verified) :: Tue, 10/30/2007 - 11:36pm

Kaveman 66/67 - I've noticed that it only really takes one troll with PDS (Pats Derangement Syndrome) to get all of us Pats homers in a tizzy. Morgana did it there, and Peter is trying it here, and I'm resisting mostly valiantly. But not quite.

by Nat (not verified) :: Tue, 10/30/2007 - 11:40pm

Same reason that playing in cold weather outdoors favors the Patriots, I expect. Practice. Only the Patriots don't have access to a thermostat.

Now go run along and play. The grownups are trying to have a conversation.

by Yaguar (not verified) :: Tue, 10/30/2007 - 11:40pm


"Yes, the jets have the 32nd ranked defense and that has little to nothing to do with pennington. But they still only have the 23rd ranked offense, and that is almost entirely on pennington. That is not to mention the killer pick 6’s he’s thrown this season."

Really? It has nothing to do with Thomas Jones and his suckiness?

by Darrel (not verified) :: Tue, 10/30/2007 - 11:43pm

It's amazing how good of a shot the Saints have at winning the division after being 0-4.

by zip (not verified) :: Tue, 10/30/2007 - 11:44pm

Kaveman, care to show us the output of that script? Not that I'm not a 1337 enough hacker to run it on my own unix box or anything.... ok so I'm not.

by Will Allen (not verified) :: Tue, 10/30/2007 - 11:44pm

Hell, Andy, I'm a Colts supporter, and I fell prey to the lure of the rabid PDS people in that audibles thread. It's kind of like interacting with the most obsessed conspiracy theorists for the first time; you kinda' get sucked in. I'll give myself a break because I never really participated in the Brady/Manning thread from hell.

by Yaguar (not verified) :: Tue, 10/30/2007 - 11:47pm

71: Kevin, you can say you think there are special circumstances that DVOA doesn't account for without turning it into a big thing about how DVOA clearly disrespects the Giants. Nobody is out to get you.

by B. (not verified) :: Tue, 10/30/2007 - 11:54pm

Thanks Gnat. That cleared it up. The Pats don't prepare for heat and exhaustion as well. That must be why they do play as well against all dome / warm weather teams.

by AndyE (not verified) :: Tue, 10/30/2007 - 11:56pm

And in random analysis while we wait for Wanker's pretty charts:

* The AFC gets better while the NFC gets worse. Instead of +4%/-4%, it's now +5.4%/-5.3% (yay rounding errors).

* While the NFC east is the tightest division still (6.3% deviation), the AFC north is catching up (6.6% deviation).

* The AFC East at +2.1% is still the most "average" division (although that 101.8% spread between NE and NYJ is ... comical. Does it mean that the Jets against the Pats is going to be like a 5 on 10 pickup game?).

* The AFC South remains the best division (+19.1%).

And while everyone is chatting about the unprecedented nature of the Pats and Colts, any insight into the unusual nature of STL and SF at -40 and -50?

* Our 2006 AFC losers division (MIA, CLE, HOU, OAK) is still outperforming the NFC West -17.5% to -21.0%. OUr 2007 AFC DVOA losers division (NYJ, CIN, HOU, OAK) also stays ahead at -19.5%.

* The average DVOA of NFC 2006 division losers (WAS,DET,TB,ARI) is higher than the average DVOA of any other NFC division spot.

by zip (not verified) :: Wed, 10/31/2007 - 12:02am

Buffalo is clearly ranked too high because they don't have an offense. A juggs machine throwing to Lee Evans is way better than this. Dammit Trent Y U so smiley?!?!

by hooper (not verified) :: Wed, 10/31/2007 - 12:05am

A couple of comments:

The Broncos are at #5 in offense pass rank, which puts them between Dallas and Pittsburgh. As a Broncos fan, allow me to bask in my rose-colored dream-bubble thoughts of Cutler's future. Mmm... (Hey, after MNF, I need something to smile about.)

The Giants are ranked #2 in rush offense. Man, do they ever miss Tiki...

by Eddo (not verified) :: Wed, 10/31/2007 - 12:06am

51: Thanks Mike...I hope I didn't come off as being snarky, I just wondered if there was a legal/official reason or if it was just being coy.
63: I'd say that the Colts and Patriots have had such great success over the last 5-10 years due to great quarterback play, above all else. Think of all great continuously great teams - the Colts with Unitas, the 49ers with Montana, the Broncos with Elway, the Bills with Kelly, the Vikings with Tarkenton, etc. A great quarterback gives you a consistently good chance to win games, masking any other deficiencies that weak teams succumb to.

by kevinNYC (not verified) :: Wed, 10/31/2007 - 12:06am

#78... Your sarcasm meter is apparently out of service.

It says kevinNYC, not raiderjoe.

by Will Allen (not verified) :: Wed, 10/31/2007 - 12:11am

Hey, where is raiderjoe?

by TanGeng (not verified) :: Wed, 10/31/2007 - 12:11am

Oh yeah, since we have two historically good franchises, it makes sense to have historically bad franchises as well.

How bad are SF and St. Louis? And how is SF worse than St. Louis. St. Louis has looked incomprehensibly bad every time I've watched. Did I miss some quality plays or is SF really that much worse?

by Jason (not verified) :: Wed, 10/31/2007 - 12:14am

This year's Pats are likely going to end with the hihgest ever DVOA ever. Teams like 96 GB and 97 Den used to run the ball up the middle to kill the clock from the middle of the 3rd quarter until the end of the game as opposed to scoring 50+ on everyone. Heck every othher game 97 Den was up so large Davis would barely play after halftime if not at all. If they had wanted Davis could have run for 2,500+ yards that year

by Starshatterer (not verified) :: Wed, 10/31/2007 - 12:14am

Add the dearth of decent quarterbacks to this mess…

Actually, I think there are plenty of decent quarterbacks, although a few teams are obviously lacking.
There are also problems of distribution and talent evaluation: some teams have a decent prospect backing up a good quarterback, and other teams passed up on getting a useable quarteback to keep a stiff they had convinced themselves was good.

(This season, the Browns managed to do both, keeping the apparently quite handy Derek Anderson backing up the near-useless Charlie Frye.)

by Yaguar (not verified) :: Wed, 10/31/2007 - 12:27am

85: At first I was worried that he might be depressed that the Raiders are 2-5. But then it occurred to me that he never gets depressed about the Raiders. That's why we love him so much. Could you imagine being a Raiders fan and not getting depressed that your team spent a pick on Robert Gallery when they could have had Fitzgerald or Rivers?

He's somewhere out there, and soon he'll return to tell us how great J Russell will be.

by Scott de B. (not verified) :: Wed, 10/31/2007 - 12:29am

So, up 38-0, and 45-0, does Dungy try for fourth down twice? No, probably not. Does it hurts his teams DVOA relative to the Pats?

I don't think so. Why would it? Indy's fourth down punt is compared to other punts. New England's 4th and 1 is compared with other 4th and 1's. Two different beasties.

If there's any significant difference, it'd come on 1st-3rd downs. But we don't know enough about how DVOA is calculated to determine the difference.

by Nat (not verified) :: Wed, 10/31/2007 - 12:29am

I was going to say that it was easier to be bad, because any team is just a couple of key injuries away from a 30% drop in DVOA. Only I don't think that explains the 49ers. (Or does it? I don't follow them enough to know for sure.)

I think the real reason is that it is hard to get good talent cheap. You have to be good at evaluating players AND you need to be successful enough to attract good players to come/stay at a discount. If you don't have both of these, you tend to get near replacement-level players surrounding a few overpriced veterans.

by Brian (not verified) :: Wed, 10/31/2007 - 12:31am

Your estimates of the probabilities of IND and NE going undefeated are terribly low. I suspect you have a math error somewhere.

The Colts would have 8 games left after Sunday. To have an 11.4% chance of finishing undefeated would mean they have an average win probability of only 0.76. (0.76^8 = 11.4)

The Patriots would have 7 games left. To have a 20.0% chance of going undefeated would mean they have an average win probability of only 0.79. (0.79^7 = 20.0)

Those kind of average win probabilities are typical of common 12 or 13-win teams. Your own DVOA shows that both NE and IND are significantly beyond typical 12-4 or 13-3 division champs.

Check the link under my signature for more.

by Starshatterer (not verified) :: Wed, 10/31/2007 - 12:49am

Brian (#92 )--

I think you misunderstand the Playoff Odds Report:
The playoff odds report plays out the season 10,000 times.
So the 20.0% in the Patriot example, simply means that the Patriots went 16-0 in 2,000 (rounded, I think) of the 10,000 simulated seasons.

by John (not verified) :: Wed, 10/31/2007 - 12:51am

#92: I can't speak for the Patriots' remaining schedule, but the Colts still have to face the AFC South teams at home and San Diego on the road. Not to mention Baltimore, KC, and Oakland.

We know Oakland will give the fits, right?

Anyway, tough schedule left for the Colts. Regardless of their historic greatness.

On the positive side, if they beat the Patriots, it's highly unlikely they'll get to coast at the end of the season like they've done recently, because the Patriots will probably win out, forcing the Colts to keep playing to win.

by Yaguar (not verified) :: Wed, 10/31/2007 - 12:54am

Brian, I've always thought your projections were a little bit too sure of themselves. For example, your system gave the Eagles an 89% chance of beating the Bears, and the Cardinals a 71% chance of beating the Panthers, and the Seahawks an 82% chance of beating the Saints. It was wrong each of those times. Obviously, I'm cherrypicking, but my impression, just from eyeballing the data, is that your system's projections are good (nothing wrong with picking the Eagles over the Bears) but it places too much confidence in them. When it says a team has a 90% chance of winning a game, it may have only a 60 or 70% chance of winning.

Obviously, it has the Colts and Patriots as favorites for every game on their schedule after this week. And if it's too confident in them, it's going to give an artificially high chance of the teams going undefeated.

As you've shown, predictions are very difficult. I'm more inclined to believe the more conservative DVOA than your system.

by Nat (not verified) :: Wed, 10/31/2007 - 1:06am

A high variance on remaining strength of schedule would not tell you if you're playing one really good team or one really bad team. After all, you only get to beat St. Louis one time, just as you only have to play New England once.

I think once you get past the average, you might as well look at the actual schedule.

by Jimbohead (not verified) :: Wed, 10/31/2007 - 1:07am

86 - the niners are really that bad. on a side note, I think the nfc west is well on its way to destroying their own record for worst overall division DVOA...

by MarkV (not verified) :: Wed, 10/31/2007 - 1:18am

RE 63, 70
Although I agree with Will Allen to an extent that QB parity has something to do with the Colts and Pats, I think he is considering the wrong side of a causal relationship. I believe that a large part of this is that the salary cap is absolutely brutal in the NFL and causes a strong level of parity. This cap impact has changed dramatically, but temporarily with the new CBA extension, and different teams are using this money for different purposes. Some are using it to facilitate a faster rebuilding process, some are buying time from cap hell, some are avoiding immenant cap hell, some are trying to attain a competitive advantage in the coming years, and some are now from it. Naturally in this process the teams with the best front offices can gain a tempory advantage. The Colts and Pats have both good front offices and great teams currently, and used these advantages to turn their teams into fantastic teams.

This is especially obvious with the Pats. With the money they "should" have had they might have only been able to afford Asante Samuel or Adelius Thomas, and almost certainly would not have gotten Moss, Stallworth and Welker, probably settling for one or maybe two of them. Not to say the Pats wouldn´t be great, but maybe they would have been 45-65% DVOA. Thats still top of the league, still within the realm of best season ever (DVOA era), but it is not the level they are at now.

by Mike Harris (not verified) :: Wed, 10/31/2007 - 1:20am

92: No math error

Intuitively, the numbers seem low to me as well. There are two main factors

1. Randomness going forward. The formula implicitly takes into account unforseen events. What if Brady gets hurt, things like that

2. The Pats and Colts are such outliers that there's really no precedent. The formula shades conservative in this unprecedented case.

by emcee fleshy (sd/atl) (not verified) :: Wed, 10/31/2007 - 1:26am

Love DVOA - but I still don't get Variance.

How can STL, at -40% have High variance (10th) and lose every game, while SF is worse (-50%) and more consistent (15th) and has won twice?

by Nat (not verified) :: Wed, 10/31/2007 - 1:32am

Similar schedule issues for the Pats. Baltimore, Pittsburgh, New York, and Philadelphia still to come. Buffalo counts as average. Miami and the Jets are arguably above replacement level. :-)

It will be tough to win all of these... Brady has a bad day now and then where he thinks he is god-like and starts throwing ill-advised balls into double coverage. My hope is that this year he will decide that Randy Moss is the god-like one and start throwing ill-advised balls to him. Or has that already happened?

by JFP (not verified) :: Wed, 10/31/2007 - 1:45am

Just kind of doing some crazy thinking here. What if the Pats & Colts end in a tie? Chances for each of them to go 15-0-1?

by Brian (not verified) :: Wed, 10/31/2007 - 2:11am

95: Keep in mind that an 85%-confident pick will be wrong 15% of the time. With 120-some games so far, that means a few 85%-confident predictions should be wrong. Plus, for every botched prediction it made, it predicted 2 upsets.

All I'm doing is running a simple efficiency model. True, early in the season teams with extreme and/or unrepresentative performances will cause some overconfident estimations, but that's the the nature of a small data set, not of the model itself.

DVOA counteracts this by factoring in pre-season "predictions" which artificially temper the data (DAVE). Nothing wrong with that per se, except when the predictions are way off, e.g. NO, CHI, BAL, STL, etc.

My comment above doesn't really rest on my own model's projection but on the fact that the FO undefeated probabilities require somewhat mediocre individual game win probabilities. IND and NE would have to be, at heart, 12-4 or 13-3 teams for the FO projections to hold. Given the data we have on these two teams, and given the fact they're already 8-0 or 7-0, I'd say that the DVOA projection is too conservative.

But, I'd have to agree with #99 that injury would be a reasonable factor to reduce confidence in either team. But my question would be: how do they do that? You'd have to estimate the probability Brady or Manning is injured, then estimate the effect that would have on their team's performance. Both estimations would be wild guesses.

But they are indeed outliers. If anything, shouldn't that increase confidence?

by Yaguar (not verified) :: Wed, 10/31/2007 - 2:19am

Brian, my point is that I'm not convinced your system's 85% picks are wrong only 15% of the time. It seems like they're wrong more than that.

An example of a prediction that seems overconfident to me: IND 92% over TEN. Tennessee has the league's best defense, and they held the Colts to a game that was decided by a dropped pass on Young's final drive. By all means, I don't expect the next IND/TEN game to be so close, but I do think the Titans probably have a better than 8% chance of beating the Colts.

by Peter (not verified) :: Wed, 10/31/2007 - 2:24am


Did you see the Chargers game? 35 points in 18 minutes? Granted, one was on a gimme long snap shank that went straight into the end zone. But there were 28 legitimate points earned by the Chargers in about 18 minutes, 21 by their offense, and 7 by their defense.

It was a ridiculous display, and probably the most insane stretch I've seen this year--and I've watched a lot of New England games.


Andy, don't be so condescending. I'm asking a legitimate question about the affect of the Patriots' late game aggressiveness on their DVOA, relative to a Colts side (their chief competition and the team they're playing this week) that would likely take a different approach in similar situations. I'm not criticizing their tactics. That is what this message board is for, right? To discuss the standings in DVOA?

PDS? Your arrogance is astounding. You're like a fan site message board, but instead of calling people names outright you make up these little derisive acronyms to belittle them. I guess somehow that makes you feel smarter. Thanks for the diagonosis, Freud. ^^

by Jerry (not verified) :: Wed, 10/31/2007 - 4:25am


(D)VOA is just evaluating the success of the plays that are run. So late drives with successful plays will increase DVOA, and I don't think we can ask the system to treat them any other way.

by Richard (not verified) :: Wed, 10/31/2007 - 4:43am

106: I think Peter's only suggesting that the difference shown by DVOA might not be representative of the actual difference between the teams. That isn't to say that DVOA should necessarily be changed, only to say that in this case it may be somewhat off. That would seem to be an appropriate topic for the comment thread attached to this week's DVOA ratings.

by Alex (not verified) :: Wed, 10/31/2007 - 5:10am

i’ve been on these boards now for weeks listening to pennington defenders. Yes, the jets have the 32nd ranked defense and that has little to nothing to do with pennington.

I don't think you really appreciate just how bad that defense has been. They don't just have the 32nd ranked defense. Their defense, if it continues to play as poorly as it has been, will be the worst defense of the DVOA era. This isn't just regular old "worst defense in the NFL" bad, this is historically bad. To make the Jets an average team, you'd need a top 3 offense. Sorry, but unless you've got Peyton Manning at QB, the Jets aren't going to be a top 3 offense.

And even then, they'd only be an average team. Not a playoff team, not even a good team. Average. That's with Peyton Manning, and the rest of the Colts offense, for that matter.

But they still only have the 23rd ranked offense, and that is almost entirely on pennington.

Right, because those other ten guys that start on the offense really aren't that important. I mean, it's not like the offensive line or running backs are struggling at all. It's all Pennington dragging down the team.

Seriously, if you put Pennington on a team with a great defense, so that he wasn't under such constant pressure to score quickly, and could run an offense that catered to his skills, he could definitely still lead a team to a Super Bowl. All you'd need to do is find a team with a great defense, and a giant black hole of suck at QB.

Fortunately, there are already two such teams in one division! Ah, the NFC North. Still playing football like it's the 1950s, I see. Also, Green Bay will join this list if Favre retires at the end of the year. I'm thinking that some NFC North team makes Pennington a great deal richer next year, assuming the Jets let him leave. Hell, he might end up getting $12 million a year!

Btw, I think that Pennington's benching should give a certain group of NFL fans a different QB to hope their team acquires, instead of constantly trying to get Eagles fans to run McNabb out of town. I'm looking at you, Vikings and Bears fans! You can't have McNabb, so quit trying.

by jm (not verified) :: Wed, 10/31/2007 - 6:45am

Peter is obviously either a charger fan or colts fan who is grinding his teeth because the Pats are so obviously great. As a result he must generate specious arguements to vent. Hang in there Peter there is always a chance Brady will get hurt.

by AndyE (not verified) :: Wed, 10/31/2007 - 7:23am

Peter 104:
Now you are asking a legitimate question about the Pats late game aggressiveness; before you were making an assumption about "running up the score". See the difference?

I thin that the fact that the Pats play 60 minutes of football every week, regardless of the DVOA impact, makes them a better team for it. While some teams in blue and white who shall remain nameless get a slow start, and then rest the last 15 minutes (effectively getting used to a 35 minute game tempo), the Pats are training themselves to compete at a consistently high level all game long. If anything, exposing themselves to more plays should make their DVOA more reliable rather than less so, as well as making their opponents' DVOA more reliable (if a team takes the fourth quarter off, you should expect this to boost the VOA of their opponents).

The other thing DVOA doesn't really account for completely is coaching value, but that's another thread.

by Richard (not verified) :: Wed, 10/31/2007 - 7:29am

109: Is it not possible that someone could think the Colts were underrated relative to the Pats w/o being some kind of bitter Colts/Chargers fan?

by RickD (not verified) :: Wed, 10/31/2007 - 7:57am

re: 105
The Chargers had two TDs created by their defense. When that happens, points accrue rather quickly. What happened in the second half? Well, their defense stopped scoring TDs (which is the more normal state of affairs) and the Houston D started doing a better job against the Chargers.

Also, the Chargers got a very nice bump this week in the DVOA from their impressive game against the Texans. I suspect if they had been playing that well every week, they would be right up at the top with the Colts and Patriots.

Those who made fun of Peter King for putting the Chargers at #3 last week should keep an eye on the Chargers. I agree with King that the Chargers, as well as the Pats and Colts, are better than any NFC team. I'm not as sure about the Steelers.

Mind you, I'm not accounting for the coaching.

by Starshatterer (not verified) :: Wed, 10/31/2007 - 8:51am

I’m asking a legitimate question about the affect of the Patriots’ late game aggressiveness on their DVOA, relative to a Colts side (their chief competition and the team they’re playing this week) that would likely take a different approach in similar situations.
But the difference is not what you think it was.

Both the Patriots and the Colts got the ball late in the third quarter, ran their regular offense with the normal starters into the fourth quarter, and scored. After that, both teams put in the backups. DVOA doesn't treat these situations differently, because they are not different enough to matter.

by b-man (not verified) :: Wed, 10/31/2007 - 9:00am

108: Given the incestuous nature of the AFC East, he will probably end up as a backup on NE.

by Jeremy Billones (not verified) :: Wed, 10/31/2007 - 9:16am

102: In order for the Colts and Pats to tie, two of the four best offenses in the DVOA era would have to play a scoreless 5th quarter. I really don't see that happening :)

by brian c (not verified) :: Wed, 10/31/2007 - 9:31am

#38 mentioned Indy's stats against a teams #1 WR, where do I find this info in the
premium database?

by shannon (not verified) :: Wed, 10/31/2007 - 9:40am

103: Perhaps the games are not independent?

For example, I think going 15-0 increases the chance the pats lose game 16, compared to if they were 14-1.

by BadgerT1000 (not verified) :: Wed, 10/31/2007 - 9:44am

I continue to be surprised at the Packers defensive DVOA. AJ Hawk has been MIA pretty much all season, Bigby has regressed since the season began, Al Harris is reluctant to tackle and Nick Collins thinks catching interceptions is forbidden by the rules.

But Kampmann is a stud. He's healthy now and a d*mn whirling dervish at defensive end. I have heard/read multiple NFL folks describe him as "perfect technician". Combine that with a motor going at about 10,000 rpm and GB has one serious asset on defense.

It should be interesting to compare/contrast him against Allen of KC this weekend.

by Scott (not verified) :: Wed, 10/31/2007 - 9:47am

As a Colt fan, I am pleasantly surprised by the Colts slow and steady rise in regard to their special teams ranking. They started out the year very poorly, The improvement on the field has been noticeable.

by brian c (not verified) :: Wed, 10/31/2007 - 9:47am

per #115, never mind

by shannon (not verified) :: Wed, 10/31/2007 - 9:47am

You've also got the classic variance issue in football - ie, the ball isn't round, etc.

If you take the variance into account, you're numbers should get more conservative. And Monte Carlo simulations could do that, while multiplying probabilities naively would not.

by Brian (not verified) :: Wed, 10/31/2007 - 9:57am

116: That's true. All of the cumulative season outcome models assume independence. But it's not that much of a stretch to assume each team will try its best in each game.

But it would be fair to say that a team that is 15-0 going into its 16th game would not rest its starters with 16-0 on the line, especially a Belichick team.

by BadgerT1000 (not verified) :: Wed, 10/31/2007 - 10:10am

My one and only Patriots comment is that when NE plays the Jets in Foxboro later this year I predict the score to be Cumberlandlike for the pro level.

Say 72-0.

At the end the Pats intentionally and very overtly take a knee near the Jets end zone and proceed to tackle someone on NY in the end zone for the safety in the enusing series.

And thus ends my Pats commentary until much later this season.

by Kaveman (not verified) :: Wed, 10/31/2007 - 10:16am

#76: I didn't post the names because that would be faintly trollish. On reflection however, it is enlightening to know which people need to spend more time contemplating their navels. ;-) So I've posted post counts to both the threads on the front page that have more than 300 posts.

by mawbrew (not verified) :: Wed, 10/31/2007 - 10:20am

I want to echo a comment earlier in the thread regarding Tampa Bay (and Jacksonville). After watching them I just don't see how the Bucs are rated this highly. This may relate to a hypothesis outlined in an earlier DVOA thread about DVOA perhaps undervalueing defenses that allow lots of small gains waiting for an offensive error. The Tampa offense is the flip side of this, an offense that consistenly moves the ball with short gains but (at least occasionally) self destructs before scoring.

by Gerry (not verified) :: Wed, 10/31/2007 - 10:59am

#100- "How can STL, at -40% have High variance (10th) and lose every game, while SF is worse (-50%) and more consistent (15th) and has won twice? "

SF won 1) at home 2) by small amounts 3) against mediocre-to-poor opponents. That will lead to a bad team having low variance.

St.Lou has lost at home to even poor opponents. Even though their overall DVOA ends up higher than SF, their variance will be higher because the games they should have won, they did not. They just did better in the games they lost (relatively, that is).

by CA (not verified) :: Wed, 10/31/2007 - 11:03am

Re: 92 Your estimates of the probabilities of IND and NE going undefeated are terribly low. I suspect you have a math error somewhere.

Brian, you should dial down the attitude. You do some interesting work on your site, although it doesn’t seem to me to be as sophisticated as the work that Aaron et al. do here (no offense). But you frequently come here essentially stating with a cocky and condescending tone, “The Outsiders are wrong. I’m right. Check out my site!� A better approach would be something along the lines of, “Hmm… I came up with something different. Let’s figure out why we differ and how the work at both of our sites can improve.� Instead, you choose aggressive tactics that seem designed more to increase traffic to your site than to enhance readers’ understanding of the game. Also, I agree with Yaguar that you seem to place too much confidence in your prediction system and ignore the unknown, and therefore the probabilities that you produce for the Patriots and Colts winning each game are too high.

by John Kim (not verified) :: Wed, 10/31/2007 - 11:05am


I agree. There's no way our Offense (Titans) should be as high as 19th.
I think we have the highest FG:TD ratio in the redzone in the league, and our offense sputters inside the 30-40 yard line.

Sure they'll gain some first downs (quite a few because they run most of the time) on the way inside the 30-40 yard line, but they just can't finish in there because of our predictable dink-and-dunk passing offense and 2 run plays: "Lendale up the middle" "Chris Henry/Brown up the middle"

by Peter (not verified) :: Wed, 10/31/2007 - 11:41am

RE 111:

Don't worry about 109, his lame brain comment is surrounded by several great posts that really do a good job of answering the question I posed earlier (106, 107, even one by AndyE who accused me of PDS), and affirm my suspicion that the Patriots and Colts are far more evenly matched than DVOA suggests. By the way, I'm a Bronco fan who's distraught because...


RE 112:

As far as the Chargers go, their last three games have been utterly dominating, where they've crushed their opposition by an average score of 35-9. They never had a run like that last year. You want more proof? I think they've gone from something like 23d in DVOA/DAVE to their current ranking of 5th in the last three weeks. I can't recall a rapid rise like that in recent years.

I thought my Broncos had a chance to take the West in what appeared to be a down year for San Diego. Then they were embarrassed 41-3 against the Dolts at home.

I know the Patriots and Colts are kings of the hill right now, but from someone who sees more of the Chargers than he'd like, if they keep playing at this level, it might end up being a different pair of teams in the Conference Championship game than we expected.

by Mike B in VA (not verified) :: Wed, 10/31/2007 - 12:01pm

#102 -

I believe all sports commentators will actually implode if that occurs.

by Wanker79 (not verified) :: Wed, 10/31/2007 - 12:18pm

Back by popular demand, here are DVOA charts broken down by conference for Total DVOA, Offensive DVOA, Defensive DVOA, and Special Teams DVOA, and Total DVOA broken down by division. All the charts are on the same scale (-60% to 80%).

Just a rough guesstimate puts the AFC at about 7% better than their NFC counterparts (not including the top 2 in each conference). A couple weeks ago the difference was only about 4%, and the 3-seeds were also outliers instead of just the 1 & 2-seeds being outliers.

I'd say the AFC South is probably the best division with the NFC East a close second and the AFC North a close third. The NFC West is probably still the worst division, but they're actually 1 decent team, 1 mediocre teams and 2 really bad teams (not 1 mediocre and 3 really bad like I suspect most people would guess).

by bsr (not verified) :: Wed, 10/31/2007 - 12:26pm

#129 I'm not sure you can make that assumption about the Colts and Pats being more evenly matched. I am sure that the Patriots data, for instance, contains far more back up player time, despite their suppossed aggressiveness than either the Colts or Chargers. If you are trying to make the argument that the Colts and Chargers DVOA has been dragged down because they are spending more time in a prevent mode, then I don't think I am buying it just by the simple point that those two teams have been involved in many more close games. Meanwhile every Patriots game involves back up player time.

by Peter (not verified) :: Wed, 10/31/2007 - 12:46pm

Re #129:

I can make whatever assumption I see reasonable. I'm not asking you to make the same assumption, but I'm personally of the opinion that there is some "excessive scoring" going on late in those Pats games that is skewing their data.

By the way, I don't think the Chargers are anywhere near those other two teams right now (though I'd be interested to see how the three teams compare over their last three games, respectively). I only reference last weekend's Houston game because it was over five minutes before the second quarter was up, at which point San Diego completely turned off the jets (whereas I suspect the Patriots might have kept on going into the fourth).

by Al 45 (not verified) :: Wed, 10/31/2007 - 12:47pm


Some teams who dominated sat on leads last Sunday (i.e. San Diego). But when the game was in question, they played better than the Patriots.


Did you see the Chargers game? 35 points in 18 minutes? Granted, one was on a gimme long snap shank that went straight into the end zone. But there were 28 legitimate points earned by the Chargers in about 18 minutes, 21 by their offense, and 7 by their defense.

So, you find what the Chargers did against the 25th ranked defense, by DVOA, to be superior to what the Patriots did against the 5th ranked defense by DVOA?

Come on now.

by Peter (not verified) :: Wed, 10/31/2007 - 12:56pm

Re 134:

Well, that's just my point. A team is comprised of offense, defense, and special teams. The Chargers had a better all around game in the first half on Sunday, by my estimation, and put the game away before the first half was even over.

You don't have to agree. That's OK with me. ;-)

by Dan (not verified) :: Wed, 10/31/2007 - 12:59pm

although they lost, wasn't anyone else impressed that Denver was able to keep Green Bay out of the endzone monday night after they got to within a yard or 2 of the goal line on two different occasions? Does the run defense suck so bad that it offsets what appeared to be a great effort in redzone defense by my boys in Denver???? :broncosfangraspingatstraws:

by Al 45 (not verified) :: Wed, 10/31/2007 - 1:07pm

re: Peter

The Chargers had a better all around game in the first half on Sunday, by my estimation, and put the game away before the first half was even over.

Against a far inferior opponent than the Patriots were facing. You're looking at two games and comparing them equally, saying the Chargers were more impressive, and not accounting for the fact their opponent was much weaker than the Patriots opponent. (Total DVOA rank: Hou 23, Was 13)

by Brian (not verified) :: Wed, 10/31/2007 - 1:09pm

127: You're probably right about my tone. I am seriously deaf in that regard. These topics really interest me and I like to engage in good debates.

Regarding trying to drive traffic, that's not my intent at all. I do not have even the tiniest of ads at my site, and hits from links on FO might account for 3% of my hits at most. I definitely get my best and toughest critiques from FO guys though.

I welcome any and all criticism on my methodology, but please allow me the chance to respond as well. I think it's a good thing to be open about coefficients, significant levels, projection accuracy, etc.

Obviously the posters here accept DVOA already, but I would encourage people to apply the same critical thought to FO stuff they do elsewhere.

But one thing I will take issue with is saying FO is more "sophisticated." It is certainly more complicated and requires a lot more effort though.

by mawbrew (not verified) :: Wed, 10/31/2007 - 1:11pm

Re: 132

I don't think the Pats played any back-ups in the Cleveland game.

by gmc (not verified) :: Wed, 10/31/2007 - 1:16pm

DVOA is not a means of picking individual games. The higher DVOA team is not always the favorite, even if you adjust for homefield.

16: "Running up the score" (and I've said before that it's silly to talk about running up the score in pro football, but I won't belabor the point)would be irrelevant to DVOA, which measures effectiveness on a per play basis. Of course, given how successful the Patriots have been throwing the football, they might have a higher offensive DVOA than if they had run out the clock with Laurence Maroney; but that's because they really are that good throwing the football.

23: I think Mike Tanier's point answers your implied query. Jacksonville may have a lower DVOA, but that was part of their gameplan. Running Fred Taylor 24 times for 68 yards is DVOA poison, but if you convert the 4-5 yard third downs you end up in often enough, it can still be the "Right Thing To Do". I thought that Jacksonville would win - I didn't think it would be that close - just because Tampa Bay's run defense can't stop Jacksonville. Likewise, I would give Cincinnati, Arizona (Kurt Warner Version only), or Detroit very good odds of beating Minnesota even though their DVOA is higher; those teams' standard plan is just too easy to execute.

Pennington: No, he isn't the Jets biggest problem. No, Kellen Clemens won't solve any of the Jets problems. To be honest, neither quarterback is ever going to win a Super Bowl, in all likelihood (but you would have said that about Trent Dilfer!). The decision to bench him, though, probably comes at the right time. He might be tradable and would bring more value than Clemens from another team (it would be funny if Baltimore grabbed him to back up Captain Checkdown; we'd have to get him a nickname... Lance Corporal Lob?), and he needs to get healthy before that happens.

Colts - Pats: I suspect the colts will win, only because the Patriots aremore vulnerable to what the Colts do (Checkdown/stretch running play for a quarter, then bomb to Wayne), than the Colts are to what the Patriots do (short and intermediate passes with a receiver deep on the off chance it's viable). The Colts have a fast set of linebackers and safeties that should be successful. We'll see if Belichick can manage to come up with something new.

by Scott (not verified) :: Wed, 10/31/2007 - 1:27pm

Indy needs a strong 1st quarter Sunday. The last two NE-IND meetings in the RCA dome have seen them get pretty much dominated the first half, and they came up with 21 and 18 pt comebacks in the 2nd half to make the game go down to the last play (both ending 38-34). I don't think anyone can afford to do that to this NE team this year. That's what I said of Dallas, who are not a good 1st half team, and they started way too slow against the Pats a few weeks ago. Even if it's just 10-3 Indy after 1, I think they have a far better shot at winning. The defense will get a boost from the homefield and stuff the Pats early on (much like the playoffs last year), so Indy's offense better cash in early with points, because they're not holding the Pats under 28.

by JordanT (not verified) :: Wed, 10/31/2007 - 1:32pm

"although they lost, wasn’t anyone else impressed that Denver was able to keep Green Bay out of the endzone monday night after they got to within a yard or 2 of the goal line on two different occasions?"

One of those times it required two false start penalties :) I understand rationalizing for your team though, I'm a 49ers fan who rationalized Alex Smith is a good quarterback who would lead the 49ers to a win against the Saints.

by Peter (not verified) :: Wed, 10/31/2007 - 1:33pm

Re 141:

My feeling is that the Colts probably do win this game. I feel like there was a considerable gap between the two teams last year in Indy's favor, one that the Patriots have made a tremendous effort to close in the off-season. The Patriots are probably a slightly better team on a neutral field, but the Colts are improved since last year and are playing at home.

by nat (not verified) :: Wed, 10/31/2007 - 1:35pm

The Patriots took Brady out in 6 of their 8 games, for a total of 24:43 time with second or third-string QBs at the helm. Brady played all of Cleveland and Dallas games.

The Colts, to compare, played Manning 100% of 6 out of 7 games. Carolina was the only game he came out of, for a total of 9:40 time with someone else (nominally) at the helm. No moral judgments are necessary. The Colts philosophy has them leaving starters in where the Patriots would take theirs out. Plus they didn't have as many large leads. No big deal.

So, the Patriots' DVOA gets "penalized" (if that's the word) by a little over a quarter when compared to the Colts DVOA.

As far as I can see, both teams continued to roll up first downs and scores when their starting QB was in, until they could run the clock completely out.

I doubt one quarter of time or about five drives matters too much for comparing the teams. If you think it might, then you should assume that the Patriots are slightly stronger than DVOA says they are.

by AndyE (not verified) :: Wed, 10/31/2007 - 1:38pm

Agains the Browns, on offense, the Pats played six non-QB skill players with 42 or more snaps, as well as Evans (13 snaps), Gaffney (10), Eckel (7), Vrabel (5), and Washington (1). On defense, they played 13 players with 42 or more snaps, and Harrison (32), Wright (17) and Meriweather (2).

by AndyE (not verified) :: Wed, 10/31/2007 - 2:01pm

To followup on the backups thread, against the Redskins, on offense (76 snaps), the Patriots only had four skill players with more than 33 snaps (Brady (67), Moss (64), Stallworth (61), Welker (60)). They had 6 players with 17-33 snaps, and 7 players with fewer than ten snaps. That's seventeen players in the 6 skill slots throughout the game.

On defense, Mike Reiss hasn't yet posted data :)

by Peter (not verified) :: Wed, 10/31/2007 - 2:02pm


The question is not regarding the personnel decisions at the end of a blow out, but the play calling philosophy. You can have Peyton Manning in there the whole game, but what does it matter if he's just turning around and handing it off to Joe Addai three straight times?

Again, it's not a matter of whether it's right or wrong to take one route versus the other, but rather an inquiry into how a tactical difference, rather than a difference in ability, affects one team's DVOA against another's.

by Purds (not verified) :: Wed, 10/31/2007 - 2:08pm

Now that we're discussing gut feelings on Halloween, mine hurts!

No, more seriously, myy gut feeling (even as a Colts fan) is that the Colts lose this one by falling behind and not catching up. Not a bad loss, but a loss.

However, the bright side is that seeing the Pats in person will help the Colts much more than the Pats for their (hopefully) eventual playoff rematch. My thinking is this: the Colts aren't doing anything different this year than in years past, and with just about the same guys. The Pats, however, are a completely different offense, and it will help the Colts (who are a reactive team, not a dictating team) so see the Pats before the playoffs.

Finally, in my Colt optimism I would note two things:
1) NE, win or lose Sunday, is going to get home field advantage. No way Indy loses only 1 game this year -- too many decent/good teams on the schedule left. Even if NE loses Sunday, I doubt the lose to anyone else.
2) Not sure being in a snow game would hurt the Colts more than the Pats. Sure, NE has run the ball, but mostly after softening up the defense with passing. And, clearly, the passing game is more dangerous for NE than its running game. So, if the playoffs are in NE, then I think it's almost a wash.

by Fire Millen? (not verified) :: Wed, 10/31/2007 - 2:12pm

#22 - Mike, there is no Official DVOA Home Team Adjustment, however in the playoff odds page there is a mention of adjusting for home field. Can you talk a little about this or point to where on the site it is discusssed?

I am easily confused and find the presentation of estimated wins confusing. I think it would be clearer if estimated winning percent were used.

by Starshatterer (not verified) :: Wed, 10/31/2007 - 2:20pm

Again, it’s not a matter of whether it’s right or wrong to take one route versus the other, but rather an inquiry into how a tactical difference, rather than a difference in ability, affects one team’s DVOA against another’s.
This question has been answered. The Colts and Patriots used similar tactics on the last drives. Brady mixed runs and passes until he was pulled this past week, so did Manning. Cassel threw three passes at the end of the game; Sorgi threw two. The Patriots tried to convert two fourth downs (and succeeded), but the DVOA gain for that simply offsets the DVOA penalty for failing on the preceding third down.

You want to believe that they pursued different tactics, fine. Knock yourself out. The tactics they chose would tend to give similar DVOA results late in the game.

by Wanker79 (not verified) :: Wed, 10/31/2007 - 2:27pm

I am easily confused and find the presentation of estimated wins confusing. I think it would be clearer if estimated winning percent were used.

While it's only very simple math to convert Estimated Wins to Estimated Winning Percentage, I whole-heartedly agree that it's make things much less confusing and we'd finally be rid of the weekly questions that the byes cause.

by MRH (not verified) :: Wed, 10/31/2007 - 2:29pm

Help me out here - ARI's remaining shcedule, and those team's respective DVOAs:
TB 18.6%
Det -12.9%
Cin -2.1%
SF -50.0%
Cle 0.0%
Sea 12.4%
NO -14.0%
Atl -25.6%
StL -40.4%

That averages to -12.7% DVOA. But the chart says ARI's future schedule is -14.2%. Am I doing something wrong? Or is it an error in the chart?

by Cyrus (not verified) :: Wed, 10/31/2007 - 2:32pm

RE: All of Peter's posts

I think you have a legitimate query, in that if the Colts were in the Patriots position, they would change their strategy while the Patriots have not. However, as Andy (I think) pointed out, the Colts have only pulled Manning in the Carolina game so far.

As far as the argument that they change philosophies while Manning is in, you would have to count how many times they had to punt because they couldn't convert the first down. Otherwise, whether they run it or pass it, they are doing the same thing the Patriots are doing, burning down the clock.

And finally, you pointed to the 35 points in 18 minutes by the Chargers. If you ignore the fumble recover (a horrible play by Houston, not caused at all by the Chargers), then they managed 28 points in the span of a quarter. 7 was from the defense, as you said.

Look to the Miami game-- the Patriots put up 28 points, all by their offense, in the second quarter.

by B (not verified) :: Wed, 10/31/2007 - 2:38pm

The thing about blowouts and their effect on DVOA is DVOA looks at each individual play and compares it to similar plays in similar game situations, so the Patriots passing plays in a blowout is compared to other teams passing plays in blowouts, and running plays are compared the same way. Then each play is given a score based on So the Colts running the ball more in blowouts isn't hurting their DVOA unless they are doing it badly. Likewise, the Patriots DVOA only gets a boost if they pass while ahead by a large margin if they're successful at it. Of course, it's easier to pass than run while leading by a large score, as defense tend to assume the leading team will just run the ball, but that just changes the success points for the plays, and it's reflected in the DVOA calculation.

by nat (not verified) :: Wed, 10/31/2007 - 2:42pm

Why don't you take a look at the play-by-play? (No cherry-picking, now!)

I'd suggest examining the fourth quarters for:

1) first down success rate (what portion of first downs lead to a score or another first down) compared to the team's average or average for that game.
2) length (time) of drive vs. starting line of scrimmage, compared to earlier in the game.

Consider only drives that start in the fourth quarter with a lead of 17 or more points, so at least three scores would be needed to tie the game. Ignore drives where the backup QB is in.

If your theory is right, you will see a drop in success rate, and more time consuming drives for the Colts. For the Patriots you will see no change in success rate, and drives that consume far less time, as the Patriots hurry to run up the score.

I suspect you will see neither. Instead, you'll see a slight improvement in success rate (as first downs become more valuable than extra yards or quick scores), and more time consuming drives. The teams' actual success at this strategy may vary a bit, and the actual plays called will differ (possession passing versus running, depending on the personnel) but it may still be clear.

Go do the research and let us know what you find out.

by MP (not verified) :: Wed, 10/31/2007 - 2:46pm

Kaveman --

The scripts you posted in #66 and #67 appear on my screen as identical. Just to satisy my awk curiosity (it's like being bi-curious, only less hip!) can you repost it with the words ANGLE_BRACKET where the angle brackets should be? That should avoid HTML rendering problems.


by mawbrew (not verified) :: Wed, 10/31/2007 - 2:57pm

Re: 145

I appreciate the info but I'm not sure what (if anything) you believe it demonstrates. Nobody ever plays the starting 11 for every snap of the game due to situational substitutions and normal rotations. Having Gaffney be part of the 4 wide package, or Vrable part of the goal-line package doesn't really qualify as playing your 'back-ups' at least in the context of this discussion. Perhaps I should have been clearer earlier, and said 'playing backups and pulling starters (for the balance of the game)', as that's really what's at issue here.

by Peter (not verified) :: Wed, 10/31/2007 - 3:02pm

Re 155:

I would love nothing more than to do the research. Unfortunately, I'm a structural engineering major (and you can probably guess how much time that consumes). I'm posting here during breaks between a lab report that's a spreadsheet nightmare. I was kinda hoping someone like Aaron or Mike might find credence in the conversation and do a similar investigation; it'd be an interesting thing to look at.

Re 153:

I would say the Chargers' first half performance against the Texans is about as good as the Patriots' against the Dolphins. Those two are the most impressive single half displays I've seen this year--maybe in several years.

by AndyE (not verified) :: Wed, 10/31/2007 - 3:09pm

157: Taking a look at the Patriots offensive depth chart, in the game against Cleveland, the only skill players not to get rotated in were Gutierrez, Cassel, and Rivers.

If you want to argue that any method of playing out the game except the one way you would do it is bad, well, I'mglad you aren't the coach of the team I root for.

And why are you focused on Cleveland? That's a game with 14 points scored by each team in the fourth quarter, after a scoreless third quarter, with a net of 17 points at the end of the game. I didn't think the most rabid PDS sufferer would call that running up the score. Since one of the Pats last scores came on an INT, and the other two non-3-and-outs in the fourth quarter were the second and third longest Patriots possessions, I think it is reasonable to argue that the Pats were running out the clock.

by AndyE (not verified) :: Wed, 10/31/2007 - 3:27pm

Patriots drive chart, shortened:
vs WAS: 7:16 TD, PUNT, 5:09 TD, 0:57 FG, Fumble, 1:32 TD (halftime) 7:46 TD, 7:56 TD, 2:37 TD (started WAS 45), PUNT, kneeldown.

I'm leaving Miami out, only 1 post-halftime TD.

vs DAL: 5:29 TD, PUNT, 2:11 TD, Fumble, 6:51 TD, kneeldown for halftime, punt, 5:24 TD, 1:32 FG, 2:13 TD, 6:04 FG, 3:26 TD.

vs CLE: 6:59 TD, Punt, 0:11 TD, punt, 4:17 FG, 0:46 TD (halftime) punt, punt, punt, 4:30 TD, punt, 5:04 downs.

vs CIN: 4:26 FG, punt, 2:09 TD, INT, 6:48 TD, halftime, 4:07 TD, 6:07 FG, 5:46 TD, 2:38 gameend.

vs BUF: 4:22 FG, 3:35 downs, fumble, 2:35 TD, 2:44 TD, halftime, punt, 5:08 TD, 1:37 TD, 5:06 TD, punt, 2:07 downs.

vs SD: 2:46 TD, 0:50 missed FG (damn you Gostkowski!), 5:58 TD, 5:35 TD, punt, INT, halftime, 3:47 TD, fumble, 10:07 TD, 2:47 gameends.

vs NYJ: 5:39 TD, punt, 4:23 downs, 3:08 TD, 0:14 TD (oh, wait, that's Ellis Hobbs!), 2:48 TD, 10:28 FG, 4:04 TD, 0:17 gameends.

Wow. First off, the Patriots have scored 11 times of the 14 times a drive started inside their own 20. (Just ponder that stat for a while. Most teams don't have a red zone percentage that good). Second, how anyone can consider taking 5,6,7,8, and 10 minute scoring drives in the second half to be evidence of running of the score or playing aggressively boggles my mind. If anything, you could argue that the Pats are toying with their opponents, keeping the opponents' offenses off the field for huge stretches of the game, refusing to score quickly. Most of the sub-2-minute scoring drives started within opposition territory, or were pre-halftime two minute drills.

by Al 45 (not verified) :: Wed, 10/31/2007 - 3:35pm

Re: AndyE

Wow. First off, the Patriots have scored 11 times of the 14 times a drive started inside their own 20. (Just ponder that stat for a while. Most teams don’t have a red zone percentage that good).

This was actually written about on NFL.com the other day. Check out the link in my name.

Here's a brief excerpt:

What makes the Patriots' ability to go long distance (11 of 14 attempts) even more impressive is how they stack up to the division leaders around the league. The other three AFC leaders are the Colts, Steelers, and Chiefs. Combined, they have gone the long field for a TD in 7 of 54 opportunities. The Patriots have them beat by four touchdowns in 40 less attempts. In the NFC, it is even more impressive. The Cowboys, Packers, Panthers and Seahawks sit atop their respective divisions and they have combined for just 11 touchdowns in 75 trips. New England matched their scores in 61 fewer chances.

That's amazing, regardless of which teams you're playing (as they're all NFL teams).

by Jens (not verified) :: Wed, 10/31/2007 - 3:37pm

Isn't the DVOA supposed to tell us how the teams rank, when the strenght of the opposition have been taken into account? If so shouldnt the rank of "past schedule" be spread fairly evenly? It isn't!

by M (not verified) :: Wed, 10/31/2007 - 3:41pm

Is it possible that the Patriots weren't actually trying to "run up the score" until this week, whereas this week they did because of all the criticism they've gottent this year? In a few games, they definitely seemed to be less crisp in the 4th quarter than in earlier quarters. If they were "running up the score", how in the hell would Miami make the game CLOSER in the 4th quarter? Plus, the Cleveland game was only decided by 10 points unil the last interception return. The only games I think that they could be accused of running up the score are Dallas & Washington. Dallas - I think that might have been in response to Wade Phillips' comments. Regarding Washington, the only chance they have of meeting Washington again while Joe Gibbs is coach is in a Super Bowl - something looking more unlikely each week. I guess it is possible there is some bad blood there - Belichek used to be the DC for NYG while Gibbs coached the Redskins.

My personal belief is that once or twice during a "magical" season, a great team will have a game or two where they do the "How much CAN we beat them by?" routine. Perhaps Belichek felt the Redskins offered the lowest risk possibility of doing such, as there isn't as much retaliation potential as other teams. On the other hand, even with the retaliation risk, I do think they will try to put up 60 in the Jets rematch. Perhaps FO should have a "guess the score" contest for Pats/Jets in week 15.

by DGL (not verified) :: Wed, 10/31/2007 - 3:43pm

#141: Maybe Scott's onto something to explain Belechick's not letting up on the gas. In the last two games NE played IND in IND, the Pats were winning early, but the game still came down to the last play. And if you get into the habit of easing up (either in playcalling, in personnel, or in execution) when you're up big against the likes of Miami, that habit could come out subconsciously when you're up against the likes of the Colts. So you don't let yourself get into that habit.

by Kal (not verified) :: Wed, 10/31/2007 - 4:13pm

Not that 'future schedule' is meant to be a predictive quality, but perhaps instead of using DVOA as a value it should be used as a rank (1st to 32nd) and then average the ranking of opponents. That would decrease the outliers significantly. It still wouldn't be perfect; facing 7 average teams would look the same as facing 1 awesome team, one horrible team, and 5 average teams. However, the predictive value of either case should be about the same.

by Cyrus (not verified) :: Wed, 10/31/2007 - 4:16pm

Re 162:

There was an article somewhere talking about how Gibbs and the Redskins beat Belichick when he coached Cleveland, by a score of something like 44-17. Not sure if that had anything to do with this massacre, but it was interesting.

And the article went on (or maybe another one, go to Patriots.com and look at the "News blitz") to describe how Bill Walsh ran up the score, so Ditka retaliated... so basically, they made the claim that every single coach has run up the score at least a few times in their life.

Belichick is "evil" because he is doing it for the season. I would laugh if he sets the points record, DVOA record, TD passes record and undefeated season, all after being accused of cheating his way to 3 Super Bowl wins. Thats a giant F U to the world :p

by mawbrew (not verified) :: Wed, 10/31/2007 - 4:41pm

Re: 159

You have somehow confused me with someone else. I've never complained (in this thread or any other) about the way the Pats have closed out their games. I questioned a factual claim about whether the Pats had pulled their starters/played their reserves in every game this year. I don't believe they did in the Cleveland game and as I stated, the data you provided doesn't address the claim.

Now, if you had bothered to ask, I would have said they never pulled their starters against Cleveland because they were never comfortably ahead in that game (other than the final minute after the defensive TD and then why bother). Which (going way back to post 132) was kinda the point - there was at least one Pats game where the score was close enough that they didn't pull their starters.

Hard to imagine why folks get frustrated with Pats fans here. Sheesh.

by Tom (not verified) :: Wed, 10/31/2007 - 4:41pm

I find it fascinating that people are talking about the Chargers scoring 35 points in 18 minutes yet somehow DVOA should penalize the Pats for trying after they were up 38-0 (with 20:47 left to play).

Am I reading something wrong? Doesn't the fact that one team was able to score that many points DURING THE SAME WEEK, kind of imply that a team with a lead should continue trying regardless of the current score?

I realize the Chargers' offense is more potent that that of the Redskins but I find this discussion regarding what should count and what shouldn't very interesting in that context. Just my 2 cents.

PS I haven't read all the posts so if someone made this point before me, apologies and props.

by AndyE (not verified) :: Wed, 10/31/2007 - 4:46pm

166 mawbrew: Apologies. Your question, without intrinsic context, didn't feel like a clarification question.

by Will Allen (not verified) :: Wed, 10/31/2007 - 5:03pm

Hell, M, just to drive some people nuts, I hope the Pats lead the Jets 70-0 going into the last quarter, and in their three 4th quarter possessions, the Pats drive down to a first and goal situation, kneel three times, and then punt the ball out of the stadium, while Belichik bows in deference in the direction of Mangenius, across the field.

by AndyE (not verified) :: Wed, 10/31/2007 - 5:14pm

169 Will Allen:

Actual, I only want that for one of them. Another, I want them to drive to the one inchyardline and turn it over on downs. On another, I want to see them kick a field goal. And on the last, I want to see them get to the one, then sit on the field, take a few delays of game to send the ball back to a reasonable distance, and then punt it out.

by shannon (not verified) :: Wed, 10/31/2007 - 5:18pm

re: future schedule - maybe what people are really looking for is 'expected future wins'?

by Will Allen (not verified) :: Wed, 10/31/2007 - 5:22pm

Maybe they could toss in a 99 yard safety in the final minute, to make the final score 70-2. Moss, split wide, spins 180 degrees at the snap, and sprints towards the other end zone. Brady, in shotgun, pump fakes to Vrabel in the end zone the Jets are defending, then spins and fires a perfect strike 60 yards in the opposite direction to Moss, who runs thorugh the end zone and spikes the ball in celebration.

by Purds (not verified) :: Wed, 10/31/2007 - 5:25pm


Read the posts. You're way off. No one thinks DVOA should criticize the Pats for anything. Some Pats fans overracted into a "circle the wagons" mode because someone asked a procedural question about how DVOA is calculated, wondering if a team with a big lead that continues to use its full offensive might in the 4th quarter would be seen as better (by DVOA, not the public or anaylsis) than a team that just ran the ball into the pile.

I think B's response in post #154 sheds good light on the question and the answer (basically: no, DVOA does not favor anything like that). Give it a read.

by Wanker79 (not verified) :: Wed, 10/31/2007 - 5:32pm

Re: 172

I've been one of the people taking shots at NE for they way they've been closing out games...but I'd laugh myself into a comma if that actually happened.

by Will Allen (not verified) :: Wed, 10/31/2007 - 5:39pm

Hey, it's football, not combat, right, Wanker? A little extreme absurdity injected into the proceedings can only help!

by slo-mo-joe (not verified) :: Wed, 10/31/2007 - 5:43pm

I'd love to see the Pats run their defense on offense and their offense of defense for the last two drives. Vince Wilfork said a while back that he is pretty good at throwing the ball, and Samuel can catch, so they may even score that way (Vince would be great on the QB sneak too, I bet, but that would be unsportsmanlike).

by Cyrus (not verified) :: Wed, 10/31/2007 - 5:44pm

RE: 174

I'd laugh myself into a period!

by B (not verified) :: Wed, 10/31/2007 - 6:19pm

I'm pretty sure the fact that the Patriots have been throwing the ball lately a lot in the 2nd half is directly related to the fact that Maroney isn't getting a lot of carries because Bellichick is being careful with him as he recovers from his various injuries. With Faulk and Heath Evans getting most of the rb carries, the passing game would give them their best chance to convert first downs maintain possession. Also, if you're completing 70% of your passes, why bother running at all?

by Frankie (not verified) :: Wed, 10/31/2007 - 7:34pm

Hmmm...lots of calls for the Pats to beat the Jets 79 - 0. I don't get it. Why is Mangini the bad guy? Is there any actual proof that he was personally responsible for "turning-in" Belicheat? And even if he did tip-off the NFL as to Belicheat's tactics, is that really something the Pats have a right to be upset about?

by Mike Tanier :: Wed, 10/31/2007 - 7:45pm

Pats run up the score talk. Very good. Very Nice. I think there's much more to be said on this subject. But in a different part of the site.

I would once again like to invite all "Patriots run up the score" posters to check out a special message board devoted to just such talk.

Just click the Open Game Threads button on the main page and look for Armeggedon, which is short for "Armageddon sick of all this Patriots talk."

by Don Booza (not verified) :: Wed, 10/31/2007 - 7:54pm

I've got news for you Mike. "Running up the score" is not a topic of conversation on the Armagedden board. In fact, we've specifically banned that topic of conversation because its way too boring. Instead, the Armagedden board is actually talking about the game this weekend, as we should. So I would direct anybody who wants to talk about "running up the score" to the Pats fan website.

by slo-mo-joe (not verified) :: Wed, 10/31/2007 - 8:51pm

re 180, I may also add that I registered for the board over one week ago, and my account has yet to be activated, despite a reminder e-mail to the administrator.

Not that I am in a hurry or anything, but if you want people to discuss the Pats there, you should actually let them. ;)

by PatsFan (not verified) :: Wed, 10/31/2007 - 9:17pm

Re: #180, #181

Ironically, the Irrational Armageddon Thread is the most rational discussion thread going.

by PatsFan (not verified) :: Wed, 10/31/2007 - 9:24pm

Re: #179

It's the whole backstory. Belichick reportedly plead with Mangini to take any job but the NYJ one. Then when Mangini took it, on the team's flight back to NE after the Denver playoff loss, he was trying to talk players into following him to NYJ. That lead to Mangini being locked out of his office at Gillette along with allegations that he stole a laptop full of proprietary information. There are also rumors that when Mangini was in NE, he (at least at one point) headed up the videotaping operation, so it's a bit hypocritical for him to go running off to Papa Goodell. Finally, if some articles (like the "Mob mentality" one by Peter King) can be believed, trying to recruit players and going to Papa Goodell over something like the taping are both no-nos in the thin blue line of coaching. It's also supposedly believed by many that if it wasn't for Belichick, Mangini wouldn't be in the NFL at all, let along being the HC of the NYJ.

So there's a lot of bad blood.

by Nat (not verified) :: Wed, 10/31/2007 - 9:59pm

180: Mike Tanier (if you stop by again)
Not to quibble too much, but this is the discussion for DVOA.

While a lot of the posting has been (or seemed to be) about trying to justify a lower or higher esteem for the Patriots or Colts (which will all shake out one way or the other on the field this weekend), people are at least pretending to be curious about DVOA and how it deals with situational adjustments.

In particular, are the situational adjustments for being up by three scores with 20 minutes to 5 minutes left in the game large enough and of the right type that you could "pad" your DVOA by playing a normal offense?

A simple statement from you or Aaron or whoever, such as "situational adjustments never change the value of a play by more than X %" or "adjustments treat all leads of more than Y as equivalent" would be helpful.

I would never ask you to "rule" on whether the Patriots were running up the score. That topic is bogus.

BTW, this is related to a topic that was a bit in vogue several years back. Is it possible to "deflate" your DVOA by switching to a clock-killing offense several drives earlier than most teams. (Pittsburgh was known for this tactic, and some of us wondered if DVOA undervalued them because of it.)

It's actually good stuff, and only degenerates because of the lack of solid information about DVOA's situational adjustments.

So enlighten us.

by Brian (not verified) :: Wed, 10/31/2007 - 10:29pm

185: Concur.
Re 22 (and others asking about using DVOA to predict game outcomes):

"Just so you guys know, there is no “official� home field advantage modifier for DVOA. In fact, we like to point out that DVOA isn’t a blunt instument for use when handicapping games."

Doesn't FO do exactly that every week (10,000 times actually)? When FO posts its playoff odds, the caption says, "The playoff odds report plays out the season 10,000 times. A random draw assigns each team a win or loss for each game. The probability that a team will be given a win is based on an equation which considers the current DVOA ratings of the two teams as well as home-field advantage."

So what is the HFA equivalent in DVOA?

One of the best ways to measure if DVOA (or any other theory or model) is meaningful is to test how well it predicts future observations.

by Kaveman (not verified) :: Wed, 10/31/2007 - 11:17pm

#156: Sure, below. Its just a quick massage of the input line to handle posters who have links on their names. Rare to see anyone interested in the elegantly minimal awk in these days of perl/python/ruby but if you are, the gawk manual is an adequate tutorial. :-)

wget -O - | awk '$1 ~ /::/ { if ($2 == "LEFT_ANGLEa") { split($5, t, "[LEFT_ANGLERIGHT_ANGLE]"); $2 = t[2]; } posts[$2]++ } END { for (n in posts) { if (posts[n] > 10) { print n, "--", posts[n]; } } }'

by Purds (not verified) :: Wed, 10/31/2007 - 11:19pm

PatsFan #184:

Some good extra gossip/dirt there that would definately explain BB's hatred of Mangenius. Much more than I had heard.

What I don't get is the NE fans/media being swayed by the part about Mangenius "betraying" BB by taking the NYJ job because BB didn't want him to do that. Isn't this the same BB who spit in the face of his former mentor Parcels in going to NE just days after Parcels orchestrated the NYJ job for BB?

by AndyE (not verified) :: Wed, 10/31/2007 - 11:24pm

Purds 188:
Yeah, but he was a Jet then. Totally different situation.

by Don M (not verified) :: Wed, 10/31/2007 - 11:39pm

Having looked at DVOA, given this some thought. Looked at the recent history of the NFL, I'm struck by the break with reality we seem to see from the performance of the Colts and Pats so far. While very impressive by DVOA period standards, and even before. Everyone is talking about how the Pats are going to go undefeated, or how even if the Colts beat them, the Pats will likely win out.. Yada Yada Yada. One NFL team has gone undefeated ever. There have only been a handful of one loss teams. My gut (Hardly a serious statistical analysis tool) tells me that we'll be looking back at this season after the Chargers/Cowboys/Unlikely teamX, have one the superbowl after the Mighty Colts & Pats finished 12-4/13-3 or so each, wondering what all this hype was about. Why? Besides history and reality, injuries are so often the deciding factor in NFL season.

by Josh (not verified) :: Wed, 10/31/2007 - 11:46pm

I'm a browns fan. BB wanted to beat the Skins for the first time at home in 35 years.

In 1991, Belichick was a 1st year HC of the Browns. Gibbs' Skins were 14-2 that year. The Browns were 6-10. The Browns kept the game close through the third, being down 21-17, on the road at RFK, with players like Ed King on the OL. The Skins had Jacoby, Grimm, etc. We had Stephen Braggs at DB. They had Darrell Green.

The Browns had a TD on a fake FG from Brian Hansen to Webster Slaughter! and were only down by 4 before Gibbs put them away. That was the Browns' only blowout loss that year, 42-17.

Got this off cleveland.com:

Coming into the game, the Pats were 1-6 all-time against the Redskins, including 1-4 at home, and had not beaten them since 1972 at the former Foxboro Stadium.

by nat (not verified) :: Thu, 11/01/2007 - 8:12am

Well, the playoff odds page has the expected number of wins for the Patriots at 14.2, and for the Colts at 13.5. The vaunted Patriots are given a little over a 12% chance of going undefeated for the regular season. And there is a 46% chance that someone other than the Colts or Patriots will win the Superbowl.

So I wouldn't accuse the DVOA analysis of going overboard, at least on this topic.

Both teams are having remarkable seasons, and the combination is unprecedented. Enjoy the high quality of play. After this season, you may never see its like again.

by Waverly (not verified) :: Thu, 11/01/2007 - 10:16am

Actually, what's the likelihood that the Colts and the Patriots just get better next year? Over the years it seems to me they have both been getting better and better, with some setbacks along the way.

Which "mean" is it that teams are supposed to regress to? I never really understood the proposed theory. (Yes, I'm familiar with statistical regression analysis, just not how it applies specifically to DVOA and the NFL.) After all, there is some skill and intent involved in improving performance -- it's not necessarily random according to any standard kind of distribution.

In any case we might very well see this quality of play next season or the one thereafter.

by Rich Conley (not verified) :: Thu, 11/01/2007 - 11:11am


"Thanks Gnat. That cleared it up. The Pats don’t prepare for heat and exhaustion as well. That must be why they do play as well against all dome / warm weather teams."

Its a comment about how the Patriots had the FLU last year, and the colts didn't (pats had starters vomiting on the sidelines). The colts 'allegedly' turned up the heat. That would most certainly affect a sick team more than a healthy one.

by Rich Conley (not verified) :: Thu, 11/01/2007 - 11:34am

157: Taking a look at the Patriots offensive depth chart, in the game against Cleveland, the only skill players not to get rotated in were Gutierrez, Cassel, and Rivers."

Thats got to be wrong. I most distinctly remember Rivers dropping a pass.

by nat (not verified) :: Thu, 11/01/2007 - 11:36am

I suppose I could have been commenting on the flu, but I really wasn't. I ended my original comment with
"Actually, it will be hot enough and loud enough without any artificial help."

I was commenting on homefield advantage and the ability to practice in game-like conditions.

But thanks for covering my back, anyway.

Don't let this guy get you started.

by AndyE (not verified) :: Thu, 11/01/2007 - 12:01pm

Rich 195 -
Not against Cleveland (play-by-play in my URL).

by Greg (not verified) :: Thu, 11/01/2007 - 3:30pm

Wow, am I late to the party.

As a PIT fan, I'm wondering how much the variance is affected by the PIT/ARI game - since PIT is #1 and ARI is #2? Where would those teams fall without that game? (Not that I think PIT wouldn't still have a high variance - just how high?)

by AndyE (not verified) :: Thu, 11/01/2007 - 4:48pm

195 Rich: Nope, that was against Washington in the third quarter.

by Nat (not verified) :: Fri, 11/02/2007 - 12:18am

Well, you could be right. But teams have a certain level of skill they acquire and maintain through effort and training, and a certain level of skill that they acquire in ways they may not be able to reproduce consistently -- like having a sixth round draft pick develop very quickly into a top-tier QB, or having players mesh extremely well in a short time, or simply not losing skill through injuries or age.

It's possible that both the Colts and the Pats will be better next year - it happened this year. But even if there is no systematic regression to the mean, you would still have to have both teams maintain a high level of play, despite the aging of their veterans, the usual injuries, etc.

"May never see its like again." isn't a statement of certainty.

But I'm with you. I'd prefer "The best is yet to come!"