Writers of Pro Football Prospectus 2008

Most Recent FO Features


» 2018 Free Agency Cost-Benefit Analysis

Is Kirk Cousins the best free-agent quarterback in recent memory? Should Trumaine Johnson or Malcolm Butler have gotten the larger contract? And what makes a free-agent contract good or bad, anyway?

13 Dec 2005

Week 15 DVOA Ratings

by Aaron Schatz

Lots of good stuff in this week's commentary for the math-o-phobic found here on FOXSports.com:

  • The best and worst DVOA games of the year
  • Why the Giants have the advantage in this weekend's Battle of the All-Star Tight Ends
  • Why the Colts are so far ahead of everyone in estimated wins
  • That Washington stuff from last week's mailbag, this time for the general public
  • Cleveland's secret star revealed
  • Brian Billick, NFL clotheshorse
  • Guess the obscure singer whose lyrics will completely piss off the Atlanta fans, although this was really just a space-filler since I couldn't think of anything new and interesting to say about the Falcons

I know I mentioned in last week's mailbag that I had done some numbers on home field advantage and was going to discuss those in today's power rankings. Instead, I got caught up in some of these other things and ran out of both time and space. I was going to just do them here for FO readers but got called for extra parenting duty. So we'll try to get to the home field advantage stuff next week.

I've made two changes to the tables. In an effort to reduce trollery, the first of the two tables is actually in order of WEIGHTED DVOA, i.e. the FOXSports.com Power Rankings. The offense, defense, and special teams ratings on that table, however, are total-season numbers, not weighted numbers. The other change is a minor one to the ESTIMATED WINS formula. In general, a higher variance means fewer estimated wins, but a higher variance now means more estimated wins for all teams with a total DVOA below replacement level (-26.6%). This has been recommended in the past, and I needed a way to stop San Francisco from constantly coming out with -0.4 estimated wins.

To save people some time, there is no need to post comments of this nature:

<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>

And please, save your lame accusations of bias. I rip on my own favorite team in this week's comments and I wrote a fawning article about a team in the Pacific Time Zone for today's New York Sun. It's one of my five or six favorite time zones.

Offense, defense, special teams pages to be updated shortly.

* * * * *

These are the Football Outsiders team efficiency ratings through 14 weeks of 2005, 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 based on strength of opponent as well as to consider all fumbles, kept or lost, as equal value. SPECIAL TEAMS DVOA is adjusted for type of stadium (warm, cold, dome, Denver/Mexico City) and week of season.

WEIGHTED DVOA represents an attempt to figure out how a team is playing right now, as opposed to over the season as a whole, by making recent games more important than earlier games. This is the statistic used for the FOXSports.com Power Rankings. WEIGHTED DVOA for offense, defense, and special teams is available on those separate pages.

IMPORTANT: Beginning with Week 12 of 2005, DVOA is based on second-order opponent adjustments. Until other years are updated with this new system, only first-order DVOA (second table) can be compared to previous seasons.

As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.

1 IND 40.8% 2 42.7% 1 13-0 34.2% 1 -11.6% 8 -5.0% 31
3 DEN 32.8% 3 41.1% 2 10-3 25.2% 4 -7.8% 12 -0.2% 21
2 CIN 37.6% 1 30.6% 3 10-3 28.4% 3 -7.1% 13 2.1% 10
4 SEA 28.1% 6 30.0% 4 11-2 22.3% 6 -5.5% 16 0.3% 16
5 NYG 27.7% 5 26.1% 5 9-4 6.7% 9 -13.7% 5 7.3% 3
8 KC 22.1% 8 25.0% 6 8-5 23.3% 5 -0.9% 19 -2.0% 27
6 SD 26.7% 4 24.7% 7 8-5 31.1% 2 5.2% 23 0.7% 14
7 JAC 23.0% 7 17.9% 8 9-4 4.9% 11 -14.5% 2 3.5% 8
11 WAS 12.9% 9 14.5% 9 7-6 1.0% 14 -12.0% 6 -0.1% 19
13 CAR 10.5% 13 13.2% 10 9-4 -6.0% 18 -14.5% 3 2.0% 11
17 MIN 0.7% 18 11.5% 11 8-5 -8.6% 22 -9.2% 9 0.0% 18
9 PIT 15.6% 10 8.7% 12 8-5 6.6% 10 -11.7% 7 -2.7% 28
12 TB 11.2% 14 8.4% 13 9-4 -2.0% 16 -14.2% 4 -1.0% 23
10 CHI 14.5% 11 7.5% 14 9-4 -18.0% 28 -31.0% 1 1.4% 12
14 DAL 10.3% 12 5.3% 15 8-5 2.7% 13 -6.2% 14 1.4% 13
15 PHI 6.6% 15 3.2% 16 5-8 -1.2% 15 -9.0% 10 -1.3% 24
18 ATL 0.5% 17 0.5% 17 8-5 10.1% 8 9.4% 26 -0.1% 20
19 NE -1.9% 20 -2.0% 18 8-5 14.8% 7 17.2% 29 0.5% 15
20 OAK -3.5% 16 -2.5% 19 4-9 3.8% 12 4.5% 22 -2.8% 29
16 MIA 0.7% 19 -4.4% 20 6-7 -13.6% 25 -8.6% 11 5.8% 4
22 GB -12.1% 22 -9.1% 21 3-10 -9.1% 23 -4.8% 17 -7.8% 32
21 CLE -7.8% 21 -9.7% 22 4-9 -5.9% 17 4.0% 21 2.1% 9
23 DET -15.9% 25 -10.8% 23 4-9 -14.1% 26 0.5% 20 -1.3% 25
24 BAL -17.6% 24 -13.2% 24 4-9 -27.4% 31 -5.9% 15 3.9% 7
26 ARI -18.5% 27 -15.7% 25 4-9 -10.5% 24 8.2% 24 0.2% 17
25 TEN -17.9% 23 -16.7% 26 4-9 -7.5% 20 15.7% 28 5.3% 5
29 NYJ -25.7% 30 -24.5% 27 3-10 -25.9% 30 -1.6% 18 -1.5% 26
27 NO -24.3% 28 -26.5% 28 3-10 -8.1% 21 12.1% 27 -4.1% 30
31 HOU -39.4% 31 -30.5% 29 1-12 -17.4% 27 30.2% 32 8.2% 1
30 STL -32.3% 29 -31.6% 30 5-8 -6.3% 19 25.4% 31 -0.6% 22
28 BUF -25.1% 26 -32.0% 31 4-9 -24.5% 29 8.2% 25 7.6% 2
32 SF -75.5% 32 -74.6% 32 2-11 -54.7% 32 25.0% 30 4.2% 6

  • 1st ORDER DVOA is the "older style" DVOA from before Week 12 2005 which only runs opponent adjustments once. This stat should be used to compare with previous seasons. (Will be added later this week.)
  • NON-ADJ VOA is total VOA without any adjustment for opponent strength, luck in recovering fumbles, or the effects of weather and altitude on special teams.
  • 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 WEIGHTED 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.
  • VARIANCE (VAR.) 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 IND 40.8% 13-0 51.0% 11.6 1 -11.6% 31 13.0% 6 6.4% 32
2 CIN 37.6% 10-3 37.9% 10.2 2 -0.8% 19 -6.0% 22 20.7% 11
3 DEN 32.8% 10-3 28.4% 9.8 3 8.0% 3 -3.3% 21 18.9% 16
4 SEA 28.1% 11-2 42.1% 9.8 4 -16.2% 32 5.6% 11 16.0% 24
5 NYG 27.7% 9-4 32.8% 9.3 6 -1.2% 20 12.3% 7 19.5% 13
6 SD 26.7% 8-5 23.3% 9.5 5 5.3% 8 36.2% 1 10.1% 31
7 JAC 23.0% 9-4 19.6% 9.0 7 2.8% 14 -40.6% 32 19.3% 15
8 KC 22.1% 8-5 16.5% 8.9 8 1.8% 16 27.1% 2 13.8% 26
9 PIT 15.6% 8-5 19.1% 8.2 9 5.1% 10 -3.0% 20 16.7% 22
10 CHI 14.5% 9-4 9.3% 8.0 10 -6.2% 27 1.0% 16 27.7% 4
11 WAS 12.9% 7-6 1.1% 7.7 13 3.9% 12 11.5% 8 21.5% 9
12 TB 11.2% 9-4 17.8% 7.3 15 -9.9% 30 -9.3% 25 22.3% 6
13 CAR 10.5% 9-4 22.2% 7.7 12 -6.5% 28 -6.9% 24 19.4% 14
14 DAL 10.3% 8-5 8.4% 7.5 14 6.0% 7 -1.3% 17 18.9% 17
15 PHI 6.6% 5-8 -3.5% 7.2 17 8.3% 2 -10.9% 26 22.2% 7
16 MIA 0.7% 6-7 -3.1% 7.0 18 -0.7% 18 -14.4% 28 18.7% 18
17 MIN 0.7% 8-5 -3.8% 7.8 11 -1.4% 21 1.0% 15 21.7% 8
18 ATL 0.5% 8-5 11.8% 6.6 19 -5.5% 25 9.7% 9 15.1% 25
19 NE -1.9% 8-5 -3.5% 7.3 16 3.5% 13 -6.8% 23 11.9% 27
20 OAK -3.5% 4-9 -1.0% 6.1 20 6.9% 5 19.2% 3 11.2% 29
21 CLE -7.8% 4-9 -17.3% 6.1 21 5.2% 9 -2.3% 18 18.1% 20
22 GB -12.1% 3-10 -11.1% 4.4 26 2.6% 15 8.1% 10 16.1% 23
23 DET -15.9% 4-9 -17.9% 5.3 22 -0.4% 17 4.3% 13 32.6% 3
24 BAL -17.6% 4-9 -23.9% 4.5 25 8.5% 1 -2.4% 19 17.7% 21
25 TEN -17.9% 4-9 -6.4% 4.6 24 -5.9% 26 14.5% 5 18.1% 19
26 ARI -18.5% 4-9 -11.2% 4.6 23 -8.4% 29 5.1% 12 10.9% 30
27 NO -24.3% 3-10 -23.3% 4.0 27 -2.4% 23 3.6% 14 24.0% 5
28 BUF -25.1% 4-9 -19.8% 3.9 28 -1.9% 22 15.7% 4 34.8% 2
29 NYJ -25.7% 3-10 -32.1% 3.6 30 4.2% 11 -12.8% 27 20.0% 12
30 STL -32.3% 5-8 -26.3% 3.7 29 -2.5% 24 -22.0% 30 11.8% 28
31 HOU -39.4% 1-12 -44.3% 2.9 31 6.9% 6 -24.1% 31 20.8% 10
32 SF -75.5% 2-11 -77.1% 0.6 32 7.2% 4 -14.7% 29 43.0% 1

Posted by: Aaron Schatz on 13 Dec 2005

181 comments, Last at 19 Dec 2005, 3:01pm by DGL


by DGL (not verified) :: Tue, 12/13/2005 - 9:32pm

Personally, I'm particular to the Hawaii-Aleutian Time Zone.

by charles (not verified) :: Tue, 12/13/2005 - 9:34pm

Good god, buffalo it's gotten that bad.

by Malene, cph, dk (not verified) :: Tue, 12/13/2005 - 9:36pm

wow... look at Jax' remaining schedule... I'm guessing whoever picked him up in fantasy's gotta like that...
He could end up being the best QB in the final 3 games...

by OWL JOLSON (not verified) :: Tue, 12/13/2005 - 9:37pm

I really thought Carolina would drop after there inability to convert 3rd downs.

by Malene, cph, dk (not verified) :: Tue, 12/13/2005 - 9:38pm

also, you gotta like a team that's 2-11 yet MASSIVELY overachieving...

by Kal (not verified) :: Tue, 12/13/2005 - 9:44pm

Weird - I would have thought that Indy's schedule strength after playing Jacksonville would have went up a bit, but it's still really high. Is that because other teams they played went down as well?

Variance for the Colts is kind of ridiculous. They're the only team that has

by bobman (not verified) :: Tue, 12/13/2005 - 9:46pm

Ya see Colts fans... no need to have panicked a few weeks ago. Sure Aaron's always interested in reading the sophomoric rantings of hyperventilating mouth-breathers, but the Colts are fine today and they were fine back then. They just needed to play a couple decent teams, slap 'em around, and now they are ranked where everybody assumes they should be.

Now, if the modified swinging gate play had just worked, I'm SURE their ST rank would have climbed into the high 20's! Curse you, Justin Snow! (Actually, Justin, you are a priceless long-snapper, I was just kidding, buddy, I love you. It's the pressure, man....)

One wonders what might happen in a Colts-Bengals AFCC game... no-huddle all the time. The Colts will be expecting it this time, so won't be quite as bad on D, but the D's will both get a bit tired... the first 100-point combined championship game? Probably not, since they both tend to use the NH offense not to score fast, but to control the game, letting the clock tick down mercilessly. I WISH they'd throw in a fast snap once in a while to keep the D honest....

by Catfish (not verified) :: Tue, 12/13/2005 - 9:47pm

Jax's last 3 games: SF, Hou, Tenn
SD's last 3 games: Indy, KC, Denver


by Catfish (not verified) :: Tue, 12/13/2005 - 9:52pm

Re: 6

It looks like your less than sign ate the rest of the comment (it starts an html tag). I think you meant that Indy is the only team with a variance below 10%.

by bobman (not verified) :: Tue, 12/13/2005 - 9:54pm

Here's the sad and crushing thing (if you look at Ariz) in terms of strength of sked (SOS). The teams with the easiest past SOS are in the playoffs or close--in fact the top two teams record-wise have the two easiest past SOS. Okay, you say, who's next? Tampa Bay at 9-4. Okay, you say, I see a pattern, who is next? Denver, Chicago, NYG?

Uh, no, Arizona at 4-9. Wow, they really ARE bad. At least SF and Houston with fewer wins have had pretty tough SOS.

As a Colt fan I had been keeping an eye on the 72 'fins saying "please keep giving us the easy SOS and I won't complain." But looking at the remaining SOS, it looks like when the year is over, they'll end up with an SOS ranked in the 26 range. Certainly "easy" but historically speaking, not one that will stand out in 20 years as laughable.

by Tecmo Bo (not verified) :: Tue, 12/13/2005 - 9:57pm

Houston's special teams still #1 after Brown missed that 30 yarder? I wonder if it's more accurate to weight close and late field goals more. A tenant of baseball analysis is that there's no such thing as a clutch hitter, but I believe that in football there are clutch kickers. And I say this knowing that my Giants will prob be relying on Feely to win another game this season. Of course, the Texans high ST ranking is prob due mostly to their kick reurns, but still, it bugs me a little to see them at #1 after last week.

by Tecmo Bo (not verified) :: Tue, 12/13/2005 - 9:57pm

Houston's special teams still #1 after Brown missed that 30 yarder? I wonder if it's more accurate to weight close and late field goals more. A tenant of baseball analysis is that there's no such thing as a clutch hitter, but I believe that in football there are clutch kickers. And I say this knowing that my Giants will prob be relying on Feely to win another game this season. Of course, the Texans high ST ranking is prob due mostly to their kick reurns, but still, it bugs me a little to see them at #1 after last week.

by Kal (not verified) :: Tue, 12/13/2005 - 9:59pm

Awww, crap. Yeah, that's what I meant - Indy is the only team with single digit variance, and the average variance looks to be 20 or so. Pretty big jump. It's not just that the Colts are beating people left and right - they're doing it consistently.

Anyway, anyone know why playing Jacksonville didn't raise their DVOA enough? Is it because SF really is THAT bad?

by Tecmo Bo (not verified) :: Tue, 12/13/2005 - 10:06pm

sorry about the double post

by PatsDanFan (not verified) :: Tue, 12/13/2005 - 10:09pm

I notice that the top 7 DVOA spots are taken by 7 of the top 9 Offensive DVOA teams. The next 7 spots are taken by 7 of the top 9 Defensive DVOA teams (the other 2 are also in the top 9 ODVOA). 7 of the top 9 STDVOA teams are ranked 20th or lower overall. Since the DVOA rankings tie relatively well to W-L records, this suggests offense is more important to winning games than defense.

I would be curious to see how the offense/defense "equation" impacts games between relatively closely ranked teams, because generally, as the adage says, offense wins games,defense wins championships.

by Will Allen (not verified) :: Tue, 12/13/2005 - 10:14pm

omigod, if the Bear-Trolls hear of this Aaron, you're going to need pepper-spray!

by Kevo (not verified) :: Tue, 12/13/2005 - 10:22pm

Bear-troll...is that like a hobbit?

by spenceKarl (not verified) :: Tue, 12/13/2005 - 10:25pm

RE: 15 (PatsDanFan)

"this suggests offense is more important to winning games than defense."

I'd be carefull how quickly you suggest a correlation from such a small sample.

RE: 12 (Techmo Bo)

" it bugs me a little to see them at #1 after last week."

It shouldn't. Yes, they are still number one, but notice that their special teams did drop significantly. 10.9% as of last week and now 8.2% today. The Bills didn't do much to improve their special teams this week and are therefore still in second with 7.6%

by emcee fleshy (not verified) :: Tue, 12/13/2005 - 11:03pm

Odd. The Falcons and Chargers, who a casual observer may reasonably think have been all over the map this year, have been two of the more consistent teams.

As a close follower of ATL, I'm not really shocked, but it is interesting to see it quantified. As not-a-close-follower of SD, I'll admit, I am a little shocked.

by Richie (not verified) :: Tue, 12/13/2005 - 11:09pm

Not sure if anybody has done this yet, but I went in and looked at how DVOA has done since 1998 in predicting the NFL playoffs.

I compared DVOA to just picking the home team in each playoff game, and to picking the team with the better record in each playoff game.

DVOA went 54-23 (.701) in the playoffs.
Best Record went 53-21-3 (.716). (If two teams had the same record, I used the home team. If the two Super Bowl teams had the same record, I called it a tie.)
Home Team went 50-20 (.714)

Out of the 7 seasons we can look at, there were 3 in which DVOA really had a problem picking winners. In the 2000 playoffs DVOA went 5-6, while the home teams were 8-2. There is no one explanation. DVOA said the Rams would win at New Orleans, but they did not. DVOA said Tampa Bay would win AT Philly, they lost. DVOA said New Orleans would have beaten Minnesota, wrong again. Basically, DVOA kept losing the games in which it said a road team would win, except it did predict that Baltimore would win at Oakland.

2001 was another rough playoffs for DVOA, going 6-5. But this was the New England year. If you remove all New England games from the study, all three measures went 6-2 picking the games.

2003 was the third tough year for DVOA, going 6-5. This was the Carolina year. Again, DVOA just couldn't do anything against an inexplicably hot playoff team. Throwing out the Carolina games, and DVOA went 5-2, Home teams were 5-2 and best record was 6-1.

So, in the remaining 4 seasons where DVOA did well in the playoffs, it did very well, picking games at an .841 winning clip, compared to .725 for home teams, and .714 for teams with the better record.

What does it all mean? I don't know. I just wanted to see if DVOA could help me pick players for my playoff fantasy league. And I guess if there is no Carolina in the field, then DVOA might do alright.

by Richie (not verified) :: Tue, 12/13/2005 - 11:14pm

That was my Carl moment for the day.

by admin :: Tue, 12/13/2005 - 11:15pm

Also, remember what happened on the play immediately prior to the missed field goal...

by Will Allen (not verified) :: Tue, 12/13/2005 - 11:16pm

Richie, did you make the home field adjustment in DVOA? I cannot remember what this has been pegged at, but who knows? This might worsen the results!

by Ted (not verified) :: Tue, 12/13/2005 - 11:23pm

Richie, did you adust for home field advantage? If I recall home field advantage is worth about 17% of DVOA. If you make an adjustment for that, Minnesota would have been expected to beat New Orleans and Philly would have been expected to beat Tampa. St Louis still would have been expected to win in New Orleans but it would have been very close. On the other hand, Oakland would have been expected to beat Baltimore.

by VarlosZ (not verified) :: Tue, 12/13/2005 - 11:24pm

Aaron has said that home field advantage seems to be worth about 17%.

by Richie (not verified) :: Tue, 12/13/2005 - 11:27pm

Will, I did not use a home field adjustment. I forget what the number is supposed to be. It would only have affected a few games I think. It would have improved two games in 2000 (Miami vs Oakland, Phi vs NYG) and it would have meant losing the NE vs Pit game last year.

DVOA picked NYJ to win at SD last year, but if HFA is 4.6 or more, then it gets that game wrong.

DVOA picked Min to win at GB last year, buf if HFA is 6.5 or more, then it gets that game wrong.

In 2003, DVOA picked Ten to win at Bal, buf if HFA is 3.9 or more, then it gets that game wrong.

In 2002, DVOA picked Atl to win at GB, buf if HFA is 6.5 or more, then it gets that game wrong.

by David Haines (not verified) :: Tue, 12/13/2005 - 11:29pm

I guess you won't give the Seahawks any respect until they destroy the Colts in two weeks.

by Richie (not verified) :: Tue, 12/13/2005 - 11:29pm

Overall, it looks like using a HFA against DVOA for predicting playoff games hurts the predictablity a bit.

Not sure what that says. (Small sample size?)

by admin :: Tue, 12/13/2005 - 11:30pm

Okey dokey, unit pages updated.

by Ted (not verified) :: Tue, 12/13/2005 - 11:32pm

Adding to my previous post which applied to 2000, I will now look at 2001 with DVOA home field adjustments of 17%. This method went 8-3. The snow bowl game was basically line ball but New England was a slight favourite, it correctly predicted Philly winning in Chicago, and only missed on the last two New England games and the first round game between Baltimore and Miami.

by IzzionSona (not verified) :: Tue, 12/13/2005 - 11:38pm

Also, are y'all using weighted DVOA or just full-season DVOA? I'm fairly sure the predictive correlation on WDVOA is higher.

by Richie (not verified) :: Tue, 12/13/2005 - 11:42pm

When you say 17%, I assume that means that a team with a DVOA of 5% will increase to 5.85%, while a 27% DVOA will increase to 31.6%.

Based on that, 1998 has no effect. 1999 is one game worse (Mia vs Sea), 2000 is one game better. 2001 is 3 games better. 2002 is 1 game worse. 2003 is the same. 2004 is the same.

by Richie (not verified) :: Tue, 12/13/2005 - 11:42pm

Total DVOA. I don't know if there is a weighted DVOA for past seasons.

by Comrade Jason (not verified) :: Tue, 12/13/2005 - 11:57pm

#27--I'm not sure if that was a joke or not, but the fawning article Aaron mentioned at the start is about the Seahawks. You can't read it without being a paid subscriber, but he does say: "The best team in the NFC, it turns out, was hiding in the Pacific Northwest the whole time."

by Vash (not verified) :: Wed, 12/14/2005 - 12:13am

I understand Chicago's offense is terrible, but how did Pittsburgh's defense lose about 5% DVOA when they allowed only about 150 yards until garbage time, 3-of-13 third downs, etc?
I can understand how that would be negative in DVOA, considering no turnovers, but how is it that ridiculously powerful?

by DGL (not verified) :: Wed, 12/14/2005 - 12:20am


No, the 17% DVOA adjustment means that you add 8.5 points to the home team's DVOA and subtract 8.5 points to the road team's DVOA. So a home team with a DVOA of 3 will go to 11.5, and a road team with a DVOA of 12.7 will go to 4.2. (I remember reading it in a post or comment from Aaron, but don't recall where exactly.)

by admin :: Wed, 12/14/2005 - 12:25am

Yep. That's how it works. Maybe. Based on research so far, anyway.

by Catholic Samurai (not verified) :: Wed, 12/14/2005 - 12:27am

Ok, Fox needs to get the commentary up there stat. I need something to occupy my mind while my roomate willfully watches "The Sisterhood of the Travelling Pants".

by admin :: Wed, 12/14/2005 - 12:37am

Commentary now up, click link on my name.

by Catholic Samurai (not verified) :: Wed, 12/14/2005 - 12:43am

Aaron, nice commentary, as uaual.

A question, however: In next year's book (Which I assume will have Kevin Jones predicted to have another league leading rushing season on the cover) are you going to cover the greatest rushing or recieving season of all time, or some of the worst so and so seasons of all time? I believe many a people learned a lot from the top 10 QB listing last year. Hell, I thought Bert Jones was Mike Jones' brother before I read the book.

by Chance (not verified) :: Wed, 12/14/2005 - 12:45am

FYI, it looks like there is something left incomplete in the KC entry on the Fox site, there's a comment that says 'Make table here'

by Vash (not verified) :: Wed, 12/14/2005 - 12:45am

And with the 2005 PFP down to 8 dollars on Amazon, it's finally time for me to pick up a copy.

by Fnor (not verified) :: Wed, 12/14/2005 - 12:46am

I hate to snark, but...



by Catholic Samurai (not verified) :: Wed, 12/14/2005 - 12:50am

RE #42:

I don't know how glowing of an endorsement this is, but PFP 2005 hasn't left my bathroom since it came out...

...If that ends up on the cover of next year's book, it's spelled C-a-t-h-o-l-i-c S-a-m-u-r-a-i.

by dedkrikit (not verified) :: Wed, 12/14/2005 - 12:50am

Yeah. Is that "make table here" on purpose or a note that was overlooked by the webguy? (KC comments)

by admin :: Wed, 12/14/2005 - 12:52am

Fixin' it now. And the plan, with so many RB having big years, is to try to do best RB seasons in PFP 2006.

by Catholic Samurai (not verified) :: Wed, 12/14/2005 - 12:55am

RE #46:

Cool, will that being using the "Z-type" or whatever the measurement was a few months back or is it some new forumla which requires me to learn more about math than I ever wanted to know even thought I went to an engineering school?

by Chance (not verified) :: Wed, 12/14/2005 - 12:56am

That Ashley Lelie comment confirms what I've been informally thinking over the past few weeks, I was just too lazy to go and look up the stats. If I recall correctly, he really poured it on in the last half or so of last season as well. Maybe he's just a slow starter.

by JonL (not verified) :: Wed, 12/14/2005 - 1:06am

Cooley as well seems to run hot and cold. I think he could be Jason Witten-good, but some weeks he just seems lazy.

Of course Witten is having an up-and-down year as well.

by Led (not verified) :: Wed, 12/14/2005 - 1:54am

In the Curtis Martin/Lamont Jordan Extra Points thread I argued that FO predicted Martin's raw stats right but for the wrong reasons. My bad.

by jimmo (not verified) :: Wed, 12/14/2005 - 1:57am

forget the $8 PFP, there's an autographed copy going for $83.38!

by CaffeineMan (not verified) :: Wed, 12/14/2005 - 2:25am

Rip? You call that a rip? That's not a rip. I want my rip! The Patriots have the lowest weighted and total DVOA of any 8-5 team! Their weighted DVOA is below a 7-6 team and a 5-8 team! Their total DVOA is below the same two teams, plus a 6-7 team! Their defensive DVOA is 19th and their pass defense DVOA is 30th! I was looking forward to a really hearty rip and now I'm disappointed. Hmph.

by CaffeineMan (not verified) :: Wed, 12/14/2005 - 2:28am

Drat. Meant to say that their defensive DVOA is 29th. Still missing my rip. Just like Teddy KGB, I feel so unsyateesfied.

by Fnor (not verified) :: Wed, 12/14/2005 - 2:28am

I'm mostly just depressed that the stillers are now below *shudder* the Vikings.

This has to be a nightmare.

by Becephalus (not verified) :: Wed, 12/14/2005 - 2:52am

Yeah I hate to say it Fnor, but I think the Vikings are going to win, I think they would get throttled at Heinz, but the game isn't at Heinz...

by emcee fleshy (not verified) :: Wed, 12/14/2005 - 3:34am

Speaking as an ATLien, (and Greenwich Mean fan)-

Pissed off about Tom Waits? In a world where where sportscasters routinely act surprised to hear any word with more than three syllables***, we should just be grateful.

I can't say I really get it. But Waits, like Paul Westerberg and a couple of others, is one of those songwriters that you can still enjoy even if you suspect that you don't really understand exactly what he's getting at.

***except "Turducken"

by NF (not verified) :: Wed, 12/14/2005 - 3:47am

Becephalus: As Aaron says, DVOA rankings are a statistical guide, not a substitute for common sense.

In the Vikings, I see a team with no obvious starting running back, an average QB at best, no clear #1 receiver, and a defense that has "discovered" itself in the last month. This team has feasted on the bottom half of the league, and beat its one quality opponent through a record-breaking amount of good fortune. The Steelers meanwhile, finally have a reasonably healthy Bus, a QB who is back from injury who rates second in value per passing play in the league, and a tough veteran defense. And they scored some touchdowns against the leagues #1 defense this weekend. I say the Steelers are favorites to beat the Vikings, and I'm not a Steelers fan.

by Becephalus (not verified) :: Wed, 12/14/2005 - 4:00am

And PIT nearly lost to BAL once and lost to BAL another time. I think they are more inconsistent then you make them out to be. And I am a PIT fan (at leats more so than I am a MIN fan). Not that MIN is that consistent either.

by pcs (not verified) :: Wed, 12/14/2005 - 4:14am

#44 Catholic Sam:
Can't believe you said that, because that's exactly where my copy's been since September.

by Jeff F (not verified) :: Wed, 12/14/2005 - 5:20am

Re:58 - Remove the game starring Charlie Batch and Tommy Maddox, and look at their numbers after that. They won't look so bad, the worst team to beat the Big Ben powered Steelers are the average Patriots, who were a much better team when the game started than when it ended.

by Becephalus (not verified) :: Wed, 12/14/2005 - 5:57am

Well MIN is a 3 point underdog right now, so it looks like common sense is going with PIT. The first BAL game Ben was in the whole game and they won 20-19 (hardly convincing against a team that has been very poor all year).

In the last 7 weeks Ben has played in 4 games vs BAL, @ IND, vs CIN, and vs CHI. Now granted CIN and IND are very good, but PIT didn't look good in either of those games (they looked real bad against IND). Now maybe my image of PIT is colored by the fact that the games I have watched of their's are vs NE, @ SD, vs, JAC, @ GB, @ IND, vs CIN, but I think PIT has had maybe 2 1/2 good games in the last six (if you count the CIN game as half a game).

On the other hand MIN is playing well and has done decently except for the atrocious offense in the NYG game (the DEF played ok in that game too so it wasn't just special teams). While I agree they do not have a RBBC so does PIT and until last week Cowher has seemed reticent about riding the Bus (as I think is important for PIT success). Moore and Bennett have both been contributing and servicable if not stellar.

Johnson has been playing efficiently, and the reciving corps is deep if not star studded (unlike PIT). Wiggins has been a great TE. All that said I think the entire game comes down to how well the MIN OL plays, which has played well since there was some suffling of players but who knows if that will hold up. I think the MIN DEF will be able to hold PIT to the low 20s and maybe get a score or set one up itself.

Anyway, like I said I like MIN this week and would put money on it, and even if I can see why PIT would be favored I hardly think it will be the pushover you guys seem to think.

Of course as the MIA/SD game reminded us anything can and will happen.

by mawbrew (not verified) :: Wed, 12/14/2005 - 10:06am

I think Pittsburg will attempt to follow the same script against the Vikes that they used vs. the Bears - get ahead early and force the opponent to throw much more than they would like. I'm guessing it will work again, but if the Vikes can get a few big plays early (on offense, defense or special; teams) the Steelers could be in trouble.

by Vash (not verified) :: Wed, 12/14/2005 - 10:24am

The Steelers' only real anomaly in the games with Roethisberger has been the Baltimore games.

Anyone else feel like Pittsburgh-Baltimore is mostly a product of a division rivalry? (Also note: The loss came under the great Thomas Maddox)

by djcolts (not verified) :: Wed, 12/14/2005 - 11:00am

Re #27: Even if you assume the Colts and Seahawks play their starters so this will be a real matchup - the Seahawks have not "killed" any team with a winning record all season. They are 3-2 vs. teams with winning records, and each of the 3 wins were by 3 points at home.

by michael (not verified) :: Wed, 12/14/2005 - 11:34am

#59 Can’t believe you said that, because that’s exactly where my copy’s been since September

Third that. it was on my desk for my fantasy draft, and now I get to read it daily for the pithy comments. (like Shaud Williams as a sleeper in Buffalo.)

It has replaced my Onion Ad Nauseum, which used to be a yearly fixture.

by NoJo (not verified) :: Wed, 12/14/2005 - 12:36pm


That's not really a game plan ("try to get ahead early so we're comfortable"), but maybe a goal. Really, do teams ever go into a game without the goal of trying to get ahead early? A gameplan would be a plan of how to get ahead early. (I know that you didn't use the words "game plan", but that seems to be your point).

by Will Allen (not verified) :: Wed, 12/14/2005 - 12:45pm

Well, I don't think Vikings/Steelers is a good investment opportunity, no matter which way one leans, but if a gun was pressed to my temple, I guess I'd lay the points.

That said, it is important that Kevin Williams is back for this game. If he and Pat Williams can win their battle in the interior, which certainly is not an outlandish prospect, then the Vikings secondary, which is playing pretty well, may be able to sandbag Roethlisberger into a couple of turnovers. The Vikings offensive line and running backs, however, need to improve and do an adequate job of protecting Johnson's blind side, or the Vikings are likely going to have a fumble or two well behind the line of scrimmage.

If Koren Robinson can make a big play or two on a kick-off return or long pass, the Vikings will have an excellent chance to win, if the defense plays as outlined above.

by spenceKarl (not verified) :: Wed, 12/14/2005 - 12:55pm

Interesting to note that the Giants are the only team, according to weighted DVOA, ranked in the top ten for offense (#9, 6.0%), defense(#2, -16.6%), and special teams (#6, 3.5%)

by spenceKarl (not verified) :: Wed, 12/14/2005 - 1:07pm

My bad...Jaguars too (#10, #8, #7)

by Israel (not verified) :: Wed, 12/14/2005 - 1:32pm

#54 & 55 (Becephalus and Fnor) - I have been worried that they would beat the Bears and the Vikings and then lose to the Browns. Still am.

by Cabbage (not verified) :: Wed, 12/14/2005 - 2:36pm

How have all of Bobby Wade's fumbled and the wind game affected the Bears special teams DVOA? I would have expected something in the bottom third instead of average.

Are the kickoff returns and coverage teams doing such an outstanding job that they've neutralized the overall rating?

by putnamp (not verified) :: Wed, 12/14/2005 - 2:55pm

Seattle's season is over. They were put on the cover of SI. Sonnova%@^#$!!!

by Pat (not verified) :: Wed, 12/14/2005 - 2:59pm

CHI special teams:

It's the excellent punting/punt coverage and kickoff coverage. Gould's kickoffs are short, but the returns are extremely short - about 62 yard kickoffs, but only 18 yard returns. That's a net of 44 yards, which is pretty good.

by Starshatterer (not verified) :: Wed, 12/14/2005 - 4:08pm

Putnamp (#72 )--

Only pretenders fear the SI cover jinx. How many times has Indianapolis been on it, this year? It seems that they've done alright afterward. So far, anyay.

by TomC (not verified) :: Wed, 12/14/2005 - 4:10pm

...and the long Bobby Wade nightmare is finally over.

by Aaron Boden (not verified) :: Wed, 12/14/2005 - 4:28pm

CHI special teams will only get better as the Bears just axed Wade.

Well, at lease they won't muff as many punts.

by DJECATEPEC (not verified) :: Wed, 12/14/2005 - 4:47pm

Wow, I can't believe there are so many nincompoops who actually take this DVOA crap seriously! Listen its a sham, a put on! It can't be real! To make up all this crap about all these different stats like how many pounds overwieght the inside linebacker is etc. etc. and then be so wildly off the mark each and every week is the dead givaway that this thing is a JOKE. There are a dozen other good ranking systems out there and every single one of them has been way choser to reality week in and week out than this nonsense! This is SO BAD, that like the Texans, IT CAN'T BE FOR REAL! You have all been PUNKED! One week he Colts at number 6 next week at number 2! The underachieving Chiefs and Chargers several positions higher than the gutsy Steelers and the surging Pats. Three or four weeks ago the 3-7 Dolphins were ranked AHEAD of the 7-3 Bucs. Don't listen to the bullshit explanations, that's like listening to a smooth tongued 75 year old syphalitic leper talking himself into the sack with a super model. (Even if he makes it, its still all bullshit!) Just take a look at what is happening on the football field week in and week out and then read this DVOA bullshit. (If they simply added one more record to their wall of stats like "Coach's history of success in playoffs" then San Diego would automatically fall to 15th, where they belong insead of 5th, where they don't! The FoxSports DVOA ranking system is final proof that you need to do proper rankings is ACTUAL, REAL FOOTBALL instead of 5000 meaningless statistics and a crack pipe!

by FizzMan (not verified) :: Wed, 12/14/2005 - 5:14pm

Re: #77
Oh, thank goodness, a voice of clarity and intelligence in this morass of analytical mumbo-jumbo...

by james (not verified) :: Wed, 12/14/2005 - 5:15pm

Lets see if I can put 77 into proper form letter form so the rest of us FOers can read it more easily.


by spenceKarl (not verified) :: Wed, 12/14/2005 - 5:18pm


Very entertaining. Thanks for nonsensical rant.

by Larry Richards (not verified) :: Wed, 12/14/2005 - 5:28pm

This DVOA and Aaron Schatz are the biggest piece of garbage there is! The ONLY stat that counts is the Win/Lost stat..... Period! Aaron Schatz should write about table tennis.......

by Ricky (not verified) :: Wed, 12/14/2005 - 5:28pm

Schatz - Sharon isn't "local" as you use it even though it is geographically. Sharon sports are terrible and maybe only 1 out of every 10 people from Sharon are fans of anything outside of a good game of BINGO. Pats are going bowling again. Remember that.

by Falco (not verified) :: Wed, 12/14/2005 - 5:28pm

SpenceKarl's comments above got me thinking (always a dangerous proposition) about the way Total DVOA is finalized. Currently, it is the sum of 3 figures, offense, defense, special teams. What if the best measure is not addition of the 3 components, but the multiplication of these same components?

Intuition tells me that teams with a "fatal flaw" may measure better than they really are. SD and KC pass defense, for example, are masked by the overall numbers, but in playing games in real life, opponents will abuse those components. A more balanced team with no major weakness, with the same additive DVOA numbers, may in fact be better.

A simple geometric example: A rectangle with width-2, length-6 will have the same perimeter as a rectangle with equal sides with length of 4. However, the area of the more balanced rectangle is substantially greater (16 vs 12). It is easier to cut through the rectangle than the square, because you can focus on the path of least resistance.

I just ran the overall offense, defense, and special teams DVOA numbers, (not weighted) and converted each to a positive number using a baseline of -80.0 = 0 for each category (to make sure even SF's pathetic offense showed a small positive number). I then multiplied the three numbers, rather than added.

As you may expect, the results are pretty similar. Here are the differences between this multiplied DVOA and the added DVOA:

-NY Giants, slightly behind Seattle in Total DVOA, move slightly in front of them for 4th. Jacksonville, trailing SD by over 3 points in Total DVOA, jumps around San Diego to 6th because of better balance compared to SD's weak defense value.

-Chicago drops from 10 to 14 because of poor offense, while Tampa Bay moves up to 11, and Dallas improves 2 spots to 12 because of better balance.

-Minnesota (16th) and Atlanta (17th) both move up a spot at the expense of Miami, which drops to 18th due to poor offense.

-Baltimore drops significantly from 24th to 27th due to anemic offense, while Arizona jumps to 24th, Tennessee stays and is now in front of BAL, and NO moves around BAL to 26th.

I am curious as to whether this method was considered and tested/rejected or not.

by DGL (not verified) :: Wed, 12/14/2005 - 5:41pm


Given that the only stat that matters is W/L, I suggest that all sports commentators - ESPN, SI, Fox, and so on - hereby eliminate their power rankings in favor of the "NFL Tiebreaker Power Ranking".

It's a very simple power ranking, actually. You rank the teams in order of W/L, and where two teams have the same W/L record, you use the official NFL tiebreaking procedure to separate them.

After all, why should the ESPN crew, Dr. Z., or anyone else spend any time trying to rank NFL teams, when all that matters is their W/L record?

by Randy Hatfield (not verified) :: Wed, 12/14/2005 - 5:50pm

can't get over the Cowboy loathing that goes on among journalists. The Broncos beat the Cowboys in OT because a muffed field goal by Dallas in the 4th kept them from winning it and they go up in the rankings. The Cowboys beat the Chiefs because a muffed field goal by KC kept them from tying it in the 4th and they go down - 10 places below the Chiefs. There's no telling how far the Cowboys would have fallen had they defeated KC by 20 pts. Sheesh!

by Falco (not verified) :: Wed, 12/14/2005 - 5:51pm

Re: 77, Hey, EJECTED CPA, I believe these so-called surging Pats were recently handled by the underachieving Chiefs very recently. Are you really surging if you just got punked three weeks ago?

Dear Santa,

all I want for Christmas is for you to permit the underachieving Chiefs to make the playoffs as the 5th seed, so they can play the surging Patriots.

by Ricky (not verified) :: Wed, 12/14/2005 - 6:06pm

dear santa,

all i want for xmas is for that idiot to get his wish so the chiefs can come to gillete and try and dethrone the champs.


all of new england

by admin :: Wed, 12/14/2005 - 6:13pm

Hmm. I missed the part where I went around the country, put a revolver to people's heads, and forced them to read Football Outsiders. Did we get a link posted somewhere mean again today?

Also, I hope everybody enjoys the veiled antisemitism of comment #82.

And now, two apologies. First, the site's been slow today. Second, I forgot to stick up defense vs. receivers. That's up now.

There's a new little element on the sheet that I was playing around with marked "PASS RUSH"? That column represents the difference between each team's pass defense DVOA and the combination of its "defense vs. receivers" DVOA for all five types of receivers. Theoretically, this should be a measure of a team's pass rush because it includes sacks and passes with no listed intended receiver. It doesn't quite work -- I don't think the Jets and Bucs, to give two examples, should be ranked so low -- but I thought it was interesting to look at.

(Later edit: Also, you know, last week's power rankings commentary had a whole thing about how the Eagles ranking still reflected the games where McNabb was healthy, and that the Eagles had been much worse since McNabb was injured. Do I have to start repeating this stuff every single week just in case the commentary gets posted on some message board that hasn't seen the previous weeks? For crying out loud...)

by Starshatterer (not verified) :: Wed, 12/14/2005 - 6:30pm

Aaron (#88 )--

Veiled antisemitism? I don't get it. Around where I live, Bingo is played in the Catholic church basements.

Or was it the comment about bowling?

by Pat (not verified) :: Wed, 12/14/2005 - 6:33pm

Intuition tells me that teams with a “fatal flaw� may measure better than they really are. SD and KC pass defense, for example, are masked by the overall numbers, but in playing games in real life, opponents will abuse those components. A more balanced team with no major weakness, with the same additive DVOA numbers, may in fact be better.

You're presuming that there exists one, uniform number which will allow you to compare any two teams and give you a measure of the relative strength of the two of them.

There doesn't need to be.

A team with one "fatal flaw" - let's take last year's Chiefs, for example - might match up horribly with one team, but terrifically with another team. And there's nothing preventing a 'circle of death' from forming - Team A can beat Team B which can beat Team C which can beat Team A.

Any linear ranking is going to suffer from this problem.

by Catfish (not verified) :: Wed, 12/14/2005 - 6:36pm

The trolls have slowed down. Last week they were the first to post.

by mtn man (not verified) :: Wed, 12/14/2005 - 6:42pm

hey "local boy", let me tell you somethin. you don't deserve to talk about sports after putting the miserable philadelphia eagles above the pats. are you serious? eagles are 5-8, half their team is out, what the hell is wrong with you?

you want to talk defence, lets talk about how the pats shut down martin and mcgahee and they will do the same this week and put the ball in the hands of simms.

wake up and remove your head from you know where.

by Pat (not verified) :: Wed, 12/14/2005 - 6:48pm

the pats shut down martin

Uh... is there a team that hasn't shut down Martin this season?

by JAT (not verified) :: Wed, 12/14/2005 - 6:55pm

Aaron, I don't know how you do it. I don't think I could continue to put up with an inbox full of comments like those in #92, which I'm sure is just the tip of the iceberg.

by putnamp (not verified) :: Wed, 12/14/2005 - 6:55pm

#83, I'm not a mathematical expert, but I think you're over-complicating it a tad. A team with a "fatal flaw" may get that flaw picked on, but they will just as often use their other strengths to adjust. A team with a strong pass defense and a weak run defense may stick another guy in the box frequently. Their offense may play more conservatively in order to run down more time on the clock.

Additionally, since we're talking about a team with a glaring weakness as compared to a more normalized peer, that team would also have some sort of matching strength that they could take frequent advantage of - this presents an inverse to the "fatal flaw", and I think they would cancel out.

by Starshatterer (not verified) :: Wed, 12/14/2005 - 6:58pm

Pat (#93 )--

Buffalo, week 6. Over 150 all-purpose yards and a touchdown for Martin.

The Jets lost and all, but Martin still had a pretty good game.

by Dan (not verified) :: Wed, 12/14/2005 - 7:01pm

You all should link to Aikman's ratings in Extra Points when they come out every week.

DVOA, DPAR and AER go hand in hand.

by Sean (not verified) :: Wed, 12/14/2005 - 7:01pm

Re 92:

Are you sure you aren't from Atlanta?

by B (not verified) :: Wed, 12/14/2005 - 7:08pm

Re 89: I beleive it was the idea that "people from Sharon" (i.e. jews) can't play sports.

by Ricky (not verified) :: Wed, 12/14/2005 - 7:11pm

Listen guy,

No veiled antisemitism. It's a fact, that Sharon is not a sports town and you can't argue against that, plain and simple. Now, while Sharon has an extremely high Jewish population, I'm sure there is no correlation between the two; however, if it is a stereotype that Jewish people love BINGO for some reason, then that was not the barb I was getting at and I apologize.

As for bowling, it was meant as in, "super-bowling". Just like when college football teams are going to a big time bowl, they are sometimes said to be "going bowling".

When's the last time Chris Simms has a won a big game? Never. What makes anyone think he's going to beat Tom Brady and the Patriots? No cannons on Saturday, only muskets.

by Starshatterer (not verified) :: Wed, 12/14/2005 - 7:12pm

Sean (#98 )--

Poor spelling, poor sentence structure, poor logic, raving about DVOA without understanding it, pseudonym of "mtn man" -- too obvious. It's some kind of plant.

My sincere hope is that these Pat-fan trolls are really KC plants, hoping to ruin New England's mojo. Or swagger, or respect, or whatever it is that wins games this week.

by Fnor (not verified) :: Wed, 12/14/2005 - 7:40pm

#100: By your logic, no one would ever win a big game because they had never won a big game before.

I see a not-terrible offence with perhaps a pretty good QB against a secondary that's been pathetic this year.

On the other side, I see a very good defence going against a great QB behind a suspect line with question marks at WR and RB.

Your team has issues. Deal with it. Being a fan doesn't give you license to be completely blind to the realities of what's going on in the league, regardless of how much of a "football town" you come from.

Then again, unless you're somehow all hooked together into a collective football-brain, it doesn't mean one bit what the people around you know, so I suppose that's a irrelevant.

by DGL (not verified) :: Wed, 12/14/2005 - 7:48pm

Let me back up James, for #92:


Hey, this is fun. Maybe someone could come up with a perl script that would automatically post these every week, then the trolls wouldn't have to waste their time.

by DGL (not verified) :: Wed, 12/14/2005 - 7:55pm

I wonder if Dr. Z. gets this kind of response. After all, he put the 4-9 Lions below the 1-12 Texans, the 6-7 Dolphins above both the 8-5 Chargers and the 7-6 Redskins, and the 8-5 Steelers above four count 'em four different 9-4 teams. Obviously it's because he's biased against Detroit and is a Miami and Pittsburgh homer who doesn't know anything about football...

by Fnor (not verified) :: Wed, 12/14/2005 - 7:57pm

DGL: He doesn't have to worry about it, because SI hires someone to filter his mail! I bet Aaron wishes he had that....

by Mikey (not verified) :: Wed, 12/14/2005 - 7:59pm

Is DVOA mostly intended to be an assesment of the entire season to date, or is it mostly intended to be a ranking of which teams are best *right now*??

If it's a gauge of the entire season, then I don't understand why more recent games are weighted more heavily.

If it's more of a ranking of where teams are at present, then I wonder if adjustments can be made for personnel changes due to injury, suspensions, etc. Is such an adjustment even possible?

I guess it just seems to me that if Roethlisberger, for example, is going to play this weekend, then the Steeler DVOA should be weighted away from games started by Batch and Maddox.

But maybe I don't understand the intent of the system. Any explanation appreciated.

by Drew (not verified) :: Wed, 12/14/2005 - 8:05pm

Re 99

Ah, a pseudo-inside reference. I don't feel too bad about not getting it then.

by Paul (not verified) :: Wed, 12/14/2005 - 8:10pm

NFL teams average about 20 points a game. 17% of 20 would be 3.4. points. Home field advantage is right around here, yes? Is this meaningful or coincidental?

by Drew (not verified) :: Wed, 12/14/2005 - 8:10pm

Re 101

They're probably your trolls. Respectable Colts fans had to bear the shame of our trolls a few weeks ago, so the Pats folks might as well get in on it too. Such is life.

by thad (not verified) :: Wed, 12/14/2005 - 8:16pm

re 100
well my feeling is
1. the pats pass defense is really bad.
2. Galloway is really good.
3. Brady is playing very well but...
4 The bucs secondary is one of the best.
Now i don't want to imply absolute certainty here but hey, Galloway could go buck wild and the Pats recievers could...not.

by Larry Richards (not verified) :: Wed, 12/14/2005 - 8:39pm

Accually if you go and look at all the other Power rankings you will see Colts #1 and Seahawks #2 (not #4)

by Starshatterer (not verified) :: Wed, 12/14/2005 - 8:48pm

Drew (#109 )--

Shh. In denial here.

Thad (#110 )--

1. The Pats' coverage is bad. They can possibly make up for this, with pressure from the front seven.
2. Galloway will probably get his catches. If that's the extent of what the Patriots give up, they should do okay.
3. Brady's ben downgraded to questionable for the game. *Gulp*
4. A good secondary is not as scary without a good pass rush, and McFarland is doubtful for Saturday.

Basicaly, the Patriots, on defense, have to stop Cadillac from running them over, get to Simms, and limit the damage downfield from what the Bucs' receivers get. On offense, they need to be patient, run the ball and block sufficiently to keep that good secondary from blanketing the Patriots' waterbug receivers. These tasks are certainly within the capacity of even this year's Patriots team.

FWIW, the Patriots have a slight edge in DVOA over the Buccaneers, when you add the 17% for home field.

by Falco (not verified) :: Wed, 12/14/2005 - 9:00pm

Re: 90, 95. I agree with what you say. Matchups matter. "Fatal flaw" was probably poor terminology on my part. It is probably better to say "unbalanced" teams-where one unit is very good, while the other is below average, and "balanced"- where units are across the board similar in strength.

**Additionally, since we’re talking about a team with a glaring weakness as compared to a more normalized peer, that team would also have some sort of matching strength that they could take frequent advantage of - this presents an inverse to the “fatal flaw�, and I think they would cancel out.**

And that's the question, does it really cancel out? Or do we just assume it does? I don't know, but its worth asking. If there are two good teams measuring similiar total DVOA, is it better to above average across the board (NY Giants, Jacksonville, Dallas), or to be extremely good on one side of the ball, and not nearly as good on the other(KC, SD, CHI, CAR), or does it not matter cuz it cancels out either way?

If the latter is true, then adding the three separate units gets to the best possible answer. If all other things being equal, balance is better than imbalance, then deriving a formula to multiply the relative strength of each unit may enhance the answer.

by TomC (not verified) :: Wed, 12/14/2005 - 9:09pm

16. omigod, if the Bear-Trolls hear of this Aaron, you’re going to need pepper-spray!

I would like to point out that FO currently remains 100% Bear-Troll-free.

- Smug Bear Fan

by bobman (not verified) :: Wed, 12/14/2005 - 9:15pm

Regarding #77, is there any way Aaron can comp this guy a free membership at FO for a year? No? Okay, then, we discount his annual membership to $50, that should make him happy. Dare I step into a political quagmire by suggesting that he is a typical Fox audience member--he probably got to the Fox Sports URL after his daily visit to the O'Reilly Factor website.... Okay, I'll stop that.

Next, he makes a great point about how the dolphins were stupidly ranked too high; he just neglected to mention last week's win over "the best non-playoff team in football" Chargers. (Some folks, like SI's beloved Z, had them in the top 3 or 4 recently!) I'm not saying the Chargers didn't take the day off, but Saban also has his guys not giving up.

And these rankings ARE ludicrous; if you look all around, you'll see all the other non-scientific rankings with the Colts and Hawks and Broncos and Bengals at the bottom, where they should be, those pathetic losers. Texans rule!

Oh boy, did I pick the wrong day to stop sniffing glue.

by DGL (not verified) :: Wed, 12/14/2005 - 10:28pm

#111: So? ESPN, SI, and DVOA/Fox are all saying that SEA is the best team in the NFC. DVOA/Fox are just saying that there are two teams besides Indy, namely Cinci and Denver, that are better than Seattle. Seattle has one more win than Cinci or Denver, but they also have had the pleasure of beating SF (2-11) twice, Arizona (4-9) twice, and St Louis (5-8) twice.

My point is, if all you want to go by is straight W/L, you could write a very simple program that would sort the teams based on W/L, H2H, division W/L, conference W/L, common opponents, etc., and everyone could save their energy. The idea of Power Rankings, though, is to give some insight that perhaps doesn't come out of straight W/L records. ESPN does it by taking the collective opinions of six journalists. SI does it by the opinion of Paul Zimmerman. Fox/DVOA does it by "break(ing) down every single play and compar(ing) a team’s performance to the league average based on situation in order to determine value over average."

ESPN puts the 8-5 Patriots ahead of the 9-4 Jaguars, the 4-9 Browns ahead of the 5-8 Rams, and all three 3-10 teams ahead of the 4-9 Lions, and SI "goes against" W-L in a number of cases I pointed out in comment #104. Does that make those rankings more accurate, less accurate, or simply providing different insights?

Personally, I think the value of "power rankings" is to provide insight beyond what can be gained from simply looking at W/L. DVOA says there's reason to believe that Seattle isn't as good as Cinci or Denver, based on looking at how they performed in specific situations. That's interesting. Dr. Z says that Detroit isn't as good as Houston, regardless of the fact that they've won three more games, because they look like they've given up. That's interesting. I don't happen to agree with him -- I actually (shudder) watched the Detroit-GB game Sunday night, and they looked to me like they made a lot of stupid mistakes and were badly coached, but it didn't look to me like they'd given up -- but I'm not going to say his Power Rankings are a piece of garbage and he should be writing about table tennis.

by Tim Gerheim :: Wed, 12/14/2005 - 10:30pm

Re 106

There are both DVOA and Weighted DVOA. The rankings are in terms of weighted DVOA, which is meant to indicate more how a team is doing currently. I'm pretty sure that non-weighted DVOA counts all games equally.

Changes for personnel would be great, but it's not as easy as it seems. If a guy started but got hurt after the first play, the play-by-play data doesn't indicate that; it just says he started. And players sub in and out all the time on both sides of the ball, and you can only be sure from the PBP that they were on the field if they actually participated in the play (threw, carried, or caught the ball, made a tackle, etc.). DVOA is calculated on a play-by-play basis, not game-by-game, so in order rightly to include missing players in DVOA you would really want to know the 22 guys in on each play. That would take a LOT more tape review, and among other things would mean that DVOA rankings don't come out by Wednesday.

by Fizzman (not verified) :: Wed, 12/14/2005 - 10:35pm

Re: 106 (Mikey)

Total DVOA is for the whole season.
Weighted DVOA is weighted to focus on where teams are at present.

Adjusting for injuries has come up and been discussed many times. However, how much (numerically) does it affect your team if your quarterback can't play? What about your #1 receiver? What about your left guard, or nickelback? Too many players get injured to too many degrees, with too much variability in their backups, be able to quantify any of this in a way that would be justifiable. So it's not done, and instead we are reminded weekly to remember, "that any statistical formula is not a replacement for your own judgment, just a tool to use in analyzing performance."

by Fizzman (not verified) :: Wed, 12/14/2005 - 10:38pm

Beaten to the punch...

by emcee fleshy (not verified) :: Wed, 12/14/2005 - 10:54pm

Several weeks ago, I watched in horror as some of my neighbors from the University of Georgia remedial algebra class s-trolled through here with similar comments. What happened?

Instaschradenfraude! Loss at home to Green Bay and losses to both major division competitors immediately follow. A fall thusfar broken only by teams from cities ravaged by both episodic and systemic disaster (Katrina & Corruption, Matt Millen & American cars).

After today's display, I might recommend that decent Bostonians focus on the Celtics for a few weeks. (a solace unavailable to the good people of The Phoenix City. Aquarium anyone?)

by Will Allen (not verified) :: Wed, 12/14/2005 - 11:34pm

TomC., I must tip my hat to Bears fans; I thought for sure that at least a few would be raising pointless hell about this week's power rankings. Instead, it appears to be the Pats fans who have decided to beat their chests and bare their canines.

by admin :: Wed, 12/14/2005 - 11:34pm

The Instaschradenfraude curse will not work on Patriots fans because *I'm* a Patriots fan. I'm just a realistic Patriots fan.

Mikey, the comments above make the point pretty well about weighted DVOA. I'll add that deciding which injuries are or are not important is subjective. Creating a system of weights based on trends from 2000-2004 is objective. Therefore, I do the latter, and apply that to all teams equally, and when I personally think the injury thing does matter, I mention it in the comments (as I did last week in regards to Philadelphia).

Unfortunately, by creating an objective rule that works best for all 32 teams over multiple years, you will end up occasionally with some really strange results. The current Washington ranking is a very good example of this. But an occasional odd ranking does not mean the system on the whole is broken. Reading an article on Sam Alito this week I was reminded of an old lawyer adage, "Hard cases make bad law."

by DGL (not verified) :: Wed, 12/14/2005 - 11:37pm

Emcee - it's the FO Comment Board Curse.

Falcons fans express their outrage in hundreds of posts to the DVOA message board, and the Falcs promptly lose three of their next four.

DVOA shows Indy at number 7, Indy fans calmly note on the FO comment boards that they really have played an easy schedule, and over the next few games it'll all work out, and the Colts promptly reel off four straight wins, including three against strong AFC opponents who many commentators judged had the best shots of stopping the Colts' winning streak.

Madden and SI got nothin' on FO.

by Catfish (not verified) :: Wed, 12/14/2005 - 11:56pm

Re: 123

Don't forget what happened to Washington and Denver:
Wash starts 3-0 but is ranked 26; people (including the since repentant james) complain; Wash loses two straight.
Denver is 5-1 but ranked 13; people complain; Denver loses on a last minute drive to the Giants.

by Chris (not verified) :: Thu, 12/15/2005 - 12:41am

The real amazing thing about the Falcons thing was the way everything came to a head at once. Atlanta was rated as an average team with an easy schedule, no run defense, inconsistancy throwing the ball, and an extremely lucky rate of fumble recovery.

After barely beating Miami, they barely recovered any fumbles against the Packers and lost, while getting run over. In the rest of the games they lost, Vick threw relatively well, but the Falcons D just got ran over and they weren't recovering 90% of their fumbles anymore.

What back then was considered ridiculous by normal standards (a 6-2 team being only average) turned into a team thats now 3rd in their division without any sort of crippling injury or even signifigant change in the quality of their play.

by Ted (not verified) :: Thu, 12/15/2005 - 9:42am

I must say that I really don't get the high ranking of the Vikings. Yes, I read and understood the explanation on FOX but even if you apply less weight to their crap early season form I can't agree with a #11 ranking in weighted DVOA. I realise they have won 6 straight, but who have they beaten? The Packers, the Browns, the Lions twice and St Louis. None of these wins have really been convincing. Throw in that complete fluke win over the Giants and that is their recent form. They are also on pace to be one of the ten luckiest teams in NFL history according to ye olde Pythagorean Theorem (from last week's DVOA mailbag). I just don't understand how with this form they can be ranked where they are. What's the deal? Are they dominating these teams but not having it really show up on the scoreboard? I'm not trying to troll here, I just would like an explanation from someone who maybe understands this system better than I do. Maybe one of those game by game graphs would help.

by Ricky (not verified) :: Thu, 12/15/2005 - 10:43am

I'll get back to everyone after Saturday's game. If Brady plays, the Pats will win. I'm not going to give some monday morning quarterback breakdown of the secondary vs. galloway (ridiculous) or the relationship between pass rush and coverage (pretty obvious). Just a guarantee.

by Tony (not verified) :: Thu, 12/15/2005 - 11:49am

You dont want to hear that the Pats are back! Ofcourse I agree, they have not played a winning team in their last 4 games, well, the Colts played nobodies until they beat the Pats, so give the Pats credit anyway. If the Colts had the Pat's schedule, they would NOT be 13-0!

by B (not verified) :: Thu, 12/15/2005 - 12:04pm

Yea, cause if the Colts had the Pats schedule, they would have lost to the Colts! Or something. Seriously, do other Pats fans really feel comfortable about saturday's game? The Patriots pass defense (Although thier rush defense is greatly improved over the last month or so) hasn't been able to stop anybody this year, and with Brady hobbled their offense might have to rely more on Dillon, and I don't really believe he'll be as good as he was last December. I haven't been this nervous since, well, since they had to play KC. On the other hand, it's not like this game is that important, they can clinch the division next week against the Jets, and the #3 seed is already out of reach.

by admin :: Thu, 12/15/2005 - 12:41pm

Actually, I was already thinking it would be interesting to look at a week-to-week graph of the Vikings so I'll see about putting one together...

by Drew (not verified) :: Thu, 12/15/2005 - 12:46pm

I had a thought last night as I was mopping my kitchen, and it doesn't really pertain to the DVOA rankings so much as the threads they tend to spawn lately. I never thought this site would be accused of bias against the Patriots. So I think it might be neat if Aaron created a "greatest hits" archive of some of the better "no respect" hate mails. Over time, it would eventually contain emails accusing him of being biased against every team in the league, and probably just about every region of the country. It could become a running feature, since I haven't noticed a "hate mail of the week" in a while.

by Andrew (not verified) :: Thu, 12/15/2005 - 1:01pm

On Colts vs. Pats Schedules.

Pats have played 6 "good" teams:
@ Panthers (how exactly would they beat the Colts? - by Jake Delhomme throwing picks to Cato June?)
@ Steelers (thrashed by Colts)
Chargers (to be beaten by Colts this week)
@ Falcons (please!)
@ Broncos (wholely owned subsidary of Colts)
@ Chiefs (perhaps the only competitive game if the Colts played them)

And 6 atrocious teams:
Raiders (Colts would kill them)
Bills (Colts would crush them)
@ Dolphins (Colts would win)
Saints (Colts would annhilate them)
Jets (Colts would win soundly)
@ Bills (another horsewhipping)

OTOH, would this years Pats beat Jacksonville twice? What about Cinci? St. Louis pre-injury to Bulger/Holt/Bruce? They did lose to the Colts of course.

Maybe the Pats would be 9-4 if they got past the Rams with the Colts schedule, while the Colts might be as low as 12-1 with the Pats schedule if the Chiefs surprised them like last year.

Too big a deal is being made of this. The Colts schedule looks weak because it has the Niners, the Texans and Titans twice and the Rams. But New England has the Bills and Jets each twice, the Saints, and the Raiders.

by B (not verified) :: Thu, 12/15/2005 - 1:16pm

The most difficult game for the Colts if they played the Pats schedule would have been at Denver. The Colts do not own Denver, they are 2-1 in meaningful games played between the teams in recent years, with all the games happening in Indy. The 2005 Colts are better than the 2004 version, but so is Denver.

by spenceKarl (not verified) :: Thu, 12/15/2005 - 1:19pm

RE: 129 (B)

"their offense might have to rely more on Dillon"

With Tampa's defense 2nd in the league at stopping the run, creating an efficient run game definately won't be as easy as what the Bills gave them; But maybe we'll find out that Brady is not as questionable as the report states.

RE: 131 (Drew)

I like that idea. Figure eventually Aaron and FO will be accused of bias against everyone. Would be entertaining to see.

by Andrew (not verified) :: Thu, 12/15/2005 - 2:33pm

B #133:

Denver has been drooling over that game in late December of 2003 for 3 seasons now. Its time to get over it and look in the mirror. You can't keep dreaming of Peyton Manning again having one of the worst game of his career (right there with Dolphins in 2001, the 2002 Jets playoff game, and the Jaguars this year) since his rookie season, as if it is going to happen again.

BTW, don't forget Manning also beat the Shanahan Broncos in 2001 and 2002, with the 2001 game another smack-down. That would make it a 4-1 series.

by Ricky (not verified) :: Thu, 12/15/2005 - 3:33pm


How many injuries did the Patriots have this year? How many times did Marvin, Reggie, Payton, and the gang thank God that Rodney went down for the year? Only now the Pats are getting healthy (though they won't be fully healthy at all considering some are done for the year). I think you give the Patriots their projected roster from the pre-season and they are 14-2 going into the play-offs AGAIN, maybe with one of those losses coming aganist the Colts. However, like the Steelers the year before, the Pats would win when it realy matters. As is now, the Pats are getting healthier and I'm expecting the same thing. I guess we'll see.

By the way, what's so different about this year's Colts than last years? All those good offenses that the Colts have played this year have been held in check by Indy's "D", right? Give me a break.

by Pat (not verified) :: Thu, 12/15/2005 - 3:53pm

Hey, instead of arguing Colts-Pats, I'm surprised no one is talking about the historically bad season Alex Smith is putting in. Right now he's at -54.9 DPAR, but he's done that in only 115 passes! Right now he's causing San Francisco to lose about one point every two passes!

Man, talk about a screwup. Why, exactly, does anyone think this guy's gonna improve? Aaron should run similarity scores on him. We're watching the next Akili Smith/Ryan Leaf, baby!

by Bencoder (not verified) :: Thu, 12/15/2005 - 4:21pm


I read your post about multiplying DVOA figures to achieve more realistic results. In an email to Aaron, he replied that overall DVOA is factored 3 parts offense, 3 parts defense and about 1 parts special teams. Are you weighting all three data points equally?

by Drew (not verified) :: Thu, 12/15/2005 - 4:30pm

By the way, what’s so different about this year’s Colts than last years? All those good offenses that the Colts have played this year have been held in check by Indy’s "D", right? Give me a break.

Well, you answered your own question there, right before you dismissed said answer. The defense is visibly better this year. DVOA agrees, rating them 8th, up from 18th in 2004. I'm not sure why that gets a "give me a break."

Now to the more important topic at hand, Alex Smith. He sucks. He powerful sucks. His -128.6% DVOA blows away the -82.6% of Craig Krenzel from last year. And Krenzel was so bad that it gave rise to the term "Krenzelian." Thus we can safely say that Alex Smith is sub-Krenzelian. Luminaries like Akili Smith and Ryan Leaf have never entered that territory.

I suppose the biggest reasons to think he will improve are 1) he's a rookie on a horrible team, and 2) surely he can't get any worse. I suspect he'll improve, but will not be a big-time star. His ceiling would seem to be "servicable", somewhere in the Brain Griese range. Of course, that's just my opinion. I'm not an expert, but I did stay at a Holiday Inn Express once.

by Pat (not verified) :: Thu, 12/15/2005 - 4:38pm

I read your post about multiplying DVOA figures to achieve more realistic results. In an email to Aaron, he replied that overall DVOA is factored 3 parts offense, 3 parts defense and about 1 parts special teams. Are you weighting all three data points equally?

I think that's embedded in the conversion from special teams to DVOA. Overall DVOA on the chart above is just offense-defense+special teams. You too can confirm this yourself easily. :)

by Andrew (not verified) :: Thu, 12/15/2005 - 5:40pm

Ricky #136:

Patriots at 14-2 if healthy, huh?

Well, they lost to Carolina healthy. A healthy Rodney Harrison and Patriots secondary had didly squat to do with Tom Brady throwing 4 picks at Kansas City. I'm unclear how being fully healthy would help Antwaan Randle El throw an unexpected lateral to Hines Ward. Richard Seymour was in action when the Chargers came to town and ran the Patriots over.

Maybe the Patriots would have beaten Denver if fully healthy. I can't see them having won any of the other games they lost though.

by Ferg (not verified) :: Thu, 12/15/2005 - 5:42pm

Re 139: Perhaps we need a new term-- "Alexandrian."

by Pat (not verified) :: Thu, 12/15/2005 - 6:19pm

"Alexandrian"? Nah, that'll never work. Geh. It's not fair that the worst NFL quarterback of all time has to have such a completely boring name.

C'mon, didn't the guy have any nicknames? Bad middle name? Anything? Jeez.

I just want to see Aaron Rodgers play. I swear, I will just laugh, and laugh, and laugh if I see Alex Smith continue playing like this, and Rodgers come in and play half-decent. Already, Frye's playing really well, so it's obvious that the 49ers at least made one bad decision. We don't know anything about Jason Campbell yet (allow me to continue laughing at the Redskins expense), but it'd be hilarious if Smith was the only bust of the top 4 QBs in the 2004 draft, and the 49ers specifically chose him.

Man. It just keeps getting worse for the 49ers. What happened to them? It's far past the salary cap issues now.

by Will Allen (not verified) :: Thu, 12/15/2005 - 6:24pm

Ted, having watched the Vikings games this year, I can say that the games against the Lions, Browns, and Rams never seemed competitive, no matter that the score tightened at the end of a couple of them. The Packers played them close, but the Packers have played everybody close. Yes, the Vikings offense was terrible against the Giants, but the defense played well, so when the Giants decided to not cover kicks that day (although, to be fair, Koren Robinson is a very good kick returner), the Vikings won. That said, I was very surprised to see the Vikings rise so suddenly.

I think it must be mostly related to the fact that this does not seem to be anything close to the same defense that gave up scads of points in the first half to the Bengals, Falcons, and Panthers. Injuries and personnel playing together for the first time seems to have the source of a lot of their problems on defense, situations which have mitigated quite a bit in the last six weeks.

The bizarre thing is that absent a phantom offensive pass interference call against Jermaine Wiggins in the season opener against the Bucs, they'd likely be 9-4 right now. Really strange.

by Ferg (not verified) :: Thu, 12/15/2005 - 6:43pm

Re 143: Don't forget that (allegedly) Paraag Marathe (sp?) liked Stefan LeFors better than Smith. So the ultimate comedy would be LeFors seeing some action and looking competent.

by Ricky (not verified) :: Thu, 12/15/2005 - 6:48pm

Like I said, we'll see, right AJ?

by Pat (not verified) :: Thu, 12/15/2005 - 6:54pm

Competent? Man, 49ers fans would be happy with Krenzelian.

by J (not verified) :: Thu, 12/15/2005 - 7:46pm

Quick question/comment.

I notice KC's RUN DEF is #1; however, KC's opponents have only combined for 307 rushing attempts, 1230 yards, 4.0 yards/attempt.

The 307 attempts against is 4th lowest in the league. 1230 yards ranks 6th. 4.0 yards/attempt ranks tied for 16th.

Teams run less against KC, and this, IMO, is more due to their high powered offense (4952 overall yards, rank 1st in NFL, 25.3 points/game ranks 6th) than their "great" run defense.

Does RUSH DEF consider the lack of rushing plays against...like PAR? From what I understand it does not....anyone know? If not, I'll send Aaron an email.

by Catfish (not verified) :: Thu, 12/15/2005 - 7:57pm

Re: 148

DVOA is a rate stat, thus it does not consider # of plays. It's interesting that you bring it up, as a similar argument was made regarding the Tennessee run defense several years ago. They had far and away the fewest attempts against, but ranked in the middle of the pack by DVOA. Scroll down in the link for more info.

by Falco (not verified) :: Thu, 12/15/2005 - 8:05pm

This is what happens when a team feels compelled to draft a quarterback at #1 overall even if there is no franchise quarterback available. Alex Smith would be a lot better off learning on the bench as a mid-first round pick, than having his confidence shattered and getting labeled a bust less than a year in. But before we get ahead of ourselves, I would like to some day see Terry Bradshaw's DVOA for his rookie year. It was a different era for QB's, but here were TB's raw passing numbers:
13 Games, 83 comp,218 att,38.1% comp, 1410 passing yards, 6.5 yds/att, 6 TD, 24 INT. So, there is hope, even if it is a small ray of light, 30 years ago, if a team is patient and willing to wait 6 years for a season with more TD's than INT's.

Re: 138, 140. As pointed out, that conversion has already been made. What I did was convert all numbers using a baseline of -80.0% = 0 for each unit (arbitrarily chosen so even SF off would be slightly positive). So, for example, Colts were Off=114.2, Def=91.6, ST=75.0, with 80.0 = a DVOA of 0.0%. I multiplied Off x Def x ST, then calculated the cube root of that product (Overall = 92.2), and converted the numbers back to the DVOA format (36.6% DVOA, #1 overall).

I ran the numbers for 2004 Total DVOA, most teams were about the same again. The biggest changes were 9-7 Buffalo (dropped from 5 to 9), and then the rearranging of the bunched middle of the pack teams, which went from

13) MIN; 14) JAX; 15) HOU; 16)TB; 17) WAS; 18) DET; 19) ATL; 20) SEA,


13) JAX; 14)DET; 15)TB; 16) SEA; 17) ATL; 18) WAS; 19) HOU; 20) MIN.

Of course, this doesn't establish anything, very limited sample size and all. Subjectively, I agree with the move of NYJ to 5, and BUF dropping well below them to 9, as I dont believe that BUF was overall the 5th best team last year considering the entire season, though they had a great defense. -

by Falco (not verified) :: Thu, 12/15/2005 - 8:15pm

Re: 148

It's more because teams pass on the Chiefs by choice because of the tendency to give up big plays in the passing game, and the inability to generate a pass rush without heavy blitzing, other than individual efforts by Jared Allen, and DC Gunther Cunningham is aggressive with blitzes and wants to take away the run.

by Pat (not verified) :: Thu, 12/15/2005 - 8:31pm

Does RUSH DEF consider the lack of rushing plays against…like PAR? From what I understand it does not….anyone know? If not, I’ll send Aaron an email.

Defensive DVOA is the average DVOA of all plays involving Kansas City's defense. If Kansas City has faced less rushing attempts than normal, you could compensate for this by making their defensive DVOA the average of the rushing DVOA and passing DVOA. However, it's important to ask whether opponents might be avoiding rushing against Kansas City because of their lack of success at doing it, or because they think that they can pass against them (which they can).

For the record, this would change Kansas City's defense from -0.9% to -3.85%.

by sully (not verified) :: Thu, 12/15/2005 - 8:32pm

I have been remiss in showing up this week. Especially after I was so obtuse last week.

As many of you have perhaps figured out by now, I am not the sharpest knife in the drawer. But, a random thought went through my mind when looking at this weeks DVOA.

Perhaps a negative correlation exists between quality of special teams play and projected wins / actual wins or even DVOA, and may be, depending on their numbers, statisically insignificant to the model.

Just a thought. I do not want to fry my laptop. Maybe the mainframe can figure it out. You guys have all the numbers.

by Pat (not verified) :: Thu, 12/15/2005 - 8:33pm

Buffalo, week 6. Over 150 all-purpose yards and a touchdown for Martin.


Yah, but I think I could run for over 100 yards against Buffalo.

by J (not verified) :: Thu, 12/15/2005 - 8:35pm

149, 151

Thanks for the input.

I do like these stats, and have been following them for a 1.5 yrs. They certainly do correctly show many elements of why teams win/lose. However, they certainly are not perfect.

I think this is a one of those times.

IMO, KC's RUSH DEF is not the best in the league. The 4.0 yrds/att suggests they are not the best.

I take it their #1 RUSH DEF rank means teams try to run on 3rd and 6, and fail to get the first down...maybe get only 4 yards. OR, they normally stuff the run, but give up more than the average amount of long runs.

Anyone know?

by admin :: Thu, 12/15/2005 - 8:42pm

The ST thing was the other thing that got cut from this week's commentary, along with the HFA discussion. This year is abnormal. Last year 5 of the top 7 ST were on winning teams although two of those were non-playoff teams (BAL, BUF). In 2003, the top three were all division champions (BAL, KC, PHI). In 2002, seven of the top ten had winning records incl. TB. In 2001, six of the top ten had winning records (incl. NE) plus there was 8-8 Denver. In 2000, the top two ST were MIA (11-5) and BAL (12-4, SB champion).

by Bret (not verified) :: Thu, 12/15/2005 - 9:12pm

148, 155

I don't know the reason(s) KC run def DVOA is #1, but part of it is that they are getting the largest opponent adjustment in that category (DVOA = -19.3%, VOA = -10.6%). They rank 10th in unadjusted run defense (tied w/ PHI) behind PIT, TB, CAR, CHI, SD, NYG, SEA, WAS, and IND, in order.

Presumably that means KC has faced some strong running teams who didn't run very successfully that day.

by thad (not verified) :: Thu, 12/15/2005 - 9:19pm

re 127
Why is that rediculous?
First of all I was not saying that would happen, only that it could happen.
Did you see the pass defense against the
It was very poor.
In 2004 the Pats defense against WR's gave up 16 td's(11th) and an average of 13.6 yprec(7th)
In 2005 the Pats defense has already given up 17 td's(30th) and 16.6 yprec(32nd) which of course helps explain why their passer rating on defense has fallen from 75.3 to 87.8.
Meamwhile Galloway is averaging 16.6 yards per catch, second in the nfl for all WR's with at least 40 catches.
And has allready scored 8 td's, tied for 6th.
I think my analysis was spot on.

by Mark (not verified) :: Fri, 12/16/2005 - 11:44am

Let me get this traight here, the NYG future schedule is +meaning harder, and they play 3 teams ranked lower than them? And the Cboys have -meaning easier and 2 of the 3 are ranked higher than them? Oh, lets not forget they got lucky last week, and I guess NYG didn't.Not to mention the big gap in DVOA from KC to PHI.
Heres my prediction for this week, NYG can't handle KC...DAL beats WAS. Resulting in NYG bening ranked #2 in the DVOA because they were unlucky and then Aaron quits his day job to go fultime as NYG chearleader.

by DGL (not verified) :: Fri, 12/16/2005 - 12:00pm


What's so hard to understand? The teams that DAL is playing are ranked 11 (WAS), 13 (CAR) and 30 (STL). The teams that NYG is playing are ranked 8 (KC), 11 (WAS), and 20 (OAK). Playing the teams ranked 8th and 20th is a harder schedule than playing the teams ranked 13th and 30th.

Future strength of schedule is an absolute measure, not a relative measure. Otherwise, the 49ers would have the hardest future strength of schedule of all time.

Oh, and I don't think the Giants have cheerleaders.

by Mark (not verified) :: Fri, 12/16/2005 - 12:07pm


I know they have at least 1. And NYG is ranked #5, all 3 games left are below them. I think the Aikman's pose a more realistic pic.

by Pat (not verified) :: Fri, 12/16/2005 - 12:12pm

There's several ways to consider strength of schedule.

First, consider that an average team, facing an average schedule, should win (on average) 8 games. So one measure of strength-of-schedule is "how many games, on average, would an average team win facing my schedule?" The difference between 8 and that number is one measure of the strength of schedule. What FO uses is the averaged DVOA of all oppponents - this is actually pretty much the same thing - it just has to do with what the chance that a team wins is.

The next question to ask is "how many games would a specific team win had they faced an average schedule?" That's a different measure, though it is interesting. It makes Indianapolis's schedule look ridiculously hard. Hard? Yah, because what would it matter to the Colts if you replaced HOU/HOU/TEN with ATL/OAK/MIA? But Indianapolis faces 4 teams of the ~11 teams that measurably reduce their chances of winning at all (and they faced JAX twice).

Incidentally, everyone who computes strength of schedule pretty much does it the same way FO does. But when you look at it the second way, it's surprising how difficult Indianapolis's schedule is.

by Pat (not verified) :: Fri, 12/16/2005 - 12:13pm

I think the Aikman’s pose a more realistic pic.

The Aikmans don't list strength of schedule at all. Was there supposed to be a point to that, or was it just a random troll?

by Mark (not verified) :: Fri, 12/16/2005 - 12:31pm

Pat, excuse my humor its a little dry...lol. My point is this, granted the Giants have a good solid club, and probably do have the edge over the Cowboys, no complaints there. But in all the power rankings out there, nobody ranks them up there in the top 5 like this one.

by spenceKarl (not verified) :: Fri, 12/16/2005 - 12:48pm

RE: 164 (Mark)

"But in all the power rankings out there, nobody ranks them up there in the top 5 like this one"

No one else analyses football the way Football Outsiders does either. It makes sense that their ratings look different. But is different neccessarily wrong? This sort of statistically progressive football analysis is groundbreaking.

by spenceKarl (not verified) :: Fri, 12/16/2005 - 12:50pm

Pardon me, Aikman does analyse football in a similar manner; But the differences in his methods have been covered here before. Football Outsiders methods are still more complete.

by Mark (not verified) :: Fri, 12/16/2005 - 12:58pm

Look, don't get all riled up. I am just havin a little fun. All the statistical analysis being done is great and groundbreaking. But be honest! The Giants are not 10 teams better than the Cowboys. NO! stats will show you that either.

by spenceKarl (not verified) :: Fri, 12/16/2005 - 1:13pm

Mark, have you watched and analyzed every single play of the Giants season so far? Have you watched and analyzed every single play of the Cowboys season as well? How bout the 10 teams seperating New York and Dallas? Well, DVOA has. In essence, DVOA has watched more football than you have.

PS: Not riled up; Just arguing a point.

by DGL (not verified) :: Fri, 12/16/2005 - 1:27pm


I don't think it makes sense to look at strength of schedule in the relative sense you're referring to (i.e., all the teams the Giants are playing are ranked below them, so they have an easier remaining schedule than do the Cowboys, two of whose remaining opponents are ranked above them).

If, say, Indy's next two opponents were Cincinnati and Denver, and the Giants' next two opponents were Denver and Cinci, does it make sense to say that Seattle has a harder future schedule strength than the Giants? I would say that their future schedule strength would be identical. Indy would be more likely to win the next two games, because all factors point to them being a better team - but the opposition they face, i.e., their future schedule strength, would be identical.

According to DVOA, the Giants face tougher opposition the last three weeks of the season; however, according to DVOA, the Giants are better than the three teams they face. According to DVOA, the Cowboys aren't as good as the Giants, but the opposition they face over the last three weeks isn't as strong.

Using WDVOA and the -8.5/+8.5 home field advantage factor, WDVOA predicts right now that Washington will beat both Dallas and NY, Carolina will beat Dallas, Dallas will beat St. Louis, and the Giants will beat KC and Oakland, with the WAS/NYG game being the closest to call. (It also predicts, as of today, that Washington will lose at Philly in week 17.) That makes the Giants the Division champs at 11-5 and Dallas and Washington tied at 9-7, which leaves the Cowboys out of the WC hunt due to the H2H tiebreaker with Washington.

I haven't done a week-by-week analysis of WDVOA with HFA adjustments to calculate its ability to pick results (there were a few comments a while back from someone who did an analysis, but it didn't include HFA). Your mileage may vary. But (as is often said around here), DVOA is an algorithm, and if your favorite team is lower than you think it should be (or your least-favorite team is higher), it's not because Aaron's a cheerleader, it's because that's what the algorithm calculates. Far better to call Dr. Z a Steelers cheerleader because he puts them above four 9-4 teams, or to call ESPN a bunch of Lions haters because they put them below three 3-10 teams -- there, it's purely subjective opinion.

by Mark (not verified) :: Fri, 12/16/2005 - 1:39pm

OK, DGL you make some good points. I'll crawl back in my hole, Have a good day. By the way what is that algorithm formula?

by admin :: Fri, 12/16/2005 - 1:46pm

As the guy who invented the formula, I would like to say that it isn't always right. That's what happens when you try to paint a broader picture that will be more accurate over multiple years -- it will occasionally screw up on some little quirky bits here and there. I think the current Washington ranking is one of those quirky bits, and I said that IN THE COMMENTARY this week. (Many people apparently don't read the commentary, because I got a LOT of hate mail about Washington, and I'm not talking about the "Mark" fellow posting here.) I personally believe that Dallas is the better team, no matter what my numbers say. I do very much appreciate the defense by readers, but I would like to make sure everyone understands that the system is certainly not infallible. This is one of the reasons I ask people not to judge DVOA solely as a formula for picking games against the spread -- there are always other factors involved.

And to Mark: I think most of the FO readers like to discuss the pros and cons of the stats here and whether they could be improved. But when you end your post by accusing me (or any other FO writer) of being biased in favor of, or against, any specific team, it stops you from being taken seriously. As noted here multiple times, I'm very honest about which team I root for, and I have that team rated lower than ANYWHERE ELSE on the Web. If that's not proof of the objective nature of the system, I don't know what is.

by Pat (not verified) :: Fri, 12/16/2005 - 1:48pm

I don’t think it makes sense to look at strength of schedule in the relative sense you’re referring to

Well, it can. If you want to know why a certain team has won more than you'd expect, then it's useful. "Absolute" schedule strength is less useful in that respect. Houston and San Francisco's future schedule are two of the easiest ones in the NFL. Think that means the two of them will win out? Not bloody likely.

But it's not useful in comparing team's schedules, because it's circular reasoning. For that, you need absolute schedule strength.

by spenceKarl (not verified) :: Fri, 12/16/2005 - 1:56pm

RE: 170 (Mark)

By the way what is that algorithm formula

It's a long read, but the DVOA methodologies are all explained here:

by james (not verified) :: Fri, 12/16/2005 - 5:59pm

concerning message board curse,
I think its fairly simple. Fan attitudes correlate almost directly with player attitudes. When fans get happy they get angry when the journalists bash their team. Its not much of a stretch to deduce that the players of whatever team are also feeling as though they are better than they actually are.

Colts fan/players had it right. They felt they were pretty good but nearly as good as everyone was saying.

I see this happen on gambling boards all the time. The best plays come when the fans come on the gambling boards and start waxing poetic. You can fade these teams at will.

This week we may see it with Patriots. However, with all of their injury problems and the way they have stomped their last two opponents I'm not entirely convinced that the Patriots aren't the second best team in the AFC. Clearly the AFC is blessed with 5-7 pretty good teams that haven't had to play each other much and therefore everyone in the conference has an awesome record.

by DGL (not verified) :: Fri, 12/16/2005 - 11:09pm

Aaron and Mark,

After I had posted (and been away from the computer for the rest of the day) it occurred to me that I had left out the usual "Your mileage may vary" disclaimer. I'm not sure that I agree with what I said in post 169 about the last three weeks of the season in the NFC East, so it's perfectly reasonable if someone were to say, "Hey, there are factors that WDVOA doesn't consider," as Aaron is always careful to disclaim.

If I have time, I want to run the numbers on WDVOA's predictive value taking into consideration HFA -- not against the spread, but as an absolute indicator of W/L probability. It'll be interesting to see how that compares to a simple "pick the team with more wins (or whichever is higher in the NFL tiebreaker heirarchy)" control strategy, to TMQ's simple-minded prediction strategy (pick the home team; or, as recently modified, pick the home team unless the visitors have two wins more than the home team), and to picks based on the various other "power rankings" on the net. But that requires a fair amount of web scraping, and given my level of ability in finding and scraping HTML content into Excel, it's going to be a largely manual (read: time-consuming) effort.

by R.J. (not verified) :: Sat, 12/17/2005 - 4:16am

I'm not sure what you meant when you suggested that the good teams in the AFC "haven't had to play each other much". My impression was the exact opposite (considering there are only 5-7 of them), i.e., that if they hadn't been playing each other so much we would really see a bunch of great records. Just look: NE has played 5 such games (Indy, Denver, SD, KC, and Pitt). Denver has also played 5 (NE, SD, KC twice and Jax). Jax has played 5 (Indy twice, Den, cincy, Pitt). Pitsburgh has actually played 6 (NE, SD, Jax, indy, Cincy twice). And KC, Cincy and Jax have all played 4 against one of the good AFC teams (with more to come for all of them). Again, considering that there are only 7 good teams we are talking about, it seems surprising to me that they have actually played each other so much.

by R.J. (not verified) :: Sat, 12/17/2005 - 4:18am

woops, not more to come for Jax. Okay, I'll agree Jax hasn't and won't have played many games against the good AFC teams.

by R.J. (not verified) :: Sat, 12/17/2005 - 4:21am

woops again, I take that back. Jax played five, like I said the first time.

by Ricky (not verified) :: Mon, 12/19/2005 - 10:28am

any of the many doubters see the Pats Saturday? weird.

by Mark (not verified) :: Mon, 12/19/2005 - 10:39am

Well, I guess I owe Aaron a big apology. My pride was telling me the Cowboys were a much better team, and his analysis was right on the money.

by DGL (not verified) :: Mon, 12/19/2005 - 3:01pm


While WDVOA may still indicate that Dallas will split its last two games, your comment has (at least for now) negated the dreaded FO Message Board Curse (FOMBC), thereby increasing the probability of the Cowboys knocking off the Panthers next week.