Week 3 DVOA Ratings
by Aaron Schatz
We have a new number one in the Football Outsiders VOA ratings this week: Buffalo. All of those interceptions were not very kind to the Patriots, and according to VOA the Bills were the much better team despite not winning until the final play. (Buffalo has 58.9% VOA for the game, while New England has -35.8%.) That kicks the Patriots down to fifth place and puts the Bills in first. The Patriots remain on top of DAVE, although the Bills have climbed up to fifth. The AFC East now has three of the top seven teams in VOA and three of the top five teams in DAVE. When you consider how easy the schedule is for Pittsburgh and Baltimore, it's easy to imagine at least one of these three teams missing the playoffs.
This is especially true because there seems to be another surprise contender in the AFC: Tennessee. The Titans rank second in VOA this week, a much bigger surprise than Buffalo being on top. It doesn't seem like the Titans played all that great in Weeks 1 and 3, but VOA gives them at least 25.0% for all three of their games this year. We'll have to see what happens when we start to incorporate the opponent adjustments -- the Titans will take a hit for playing Jacksonville and Denver but get a bonus for playing Baltimore (which now ranks third in VOA). I don't have a lot of expanded thoughts on the Titans right now but do want to try to watch some of their games to see what's going on there. Their defense, which doesn't really have any well-known star players, is currently on top of the league.
Another team with interesting numbers right now is San Francisco. The 49ers rank 12th in VOA; obviously that will fall when we start incorporating the opponent adjustments. However, being 2-1 in the NFC West right now is a huge advantage. When you combine that with the fact that our preseason projections had the 49ers as the best team in the division anyway, and the fact that there are so many teams fighting over the six spots in the AFC, our playoff odds report now lists San Francisco as the third most likely team to make the playoffs! Ahead of them are the two 3-0 NFC teams, Green Bay and Detroit.
As long as we're on the subject of the NFC North, let's talk about the Minnesota Vikings. The Vikings rank 14th in VOA, much higher than the other 0-3 teams. Of course, while they've been an overall average team, they haven't really been an average team. They've been a good team in the first half of games, and merde-tastic in the second half of games. That's led to three losses, all by a single score, thanks to blowing double-digit halftime leads in three straight games.
Of course, this had me asking how the Vikings' record of halftime collapses compared with teams in the recent past. Is this a sign that the Vikings are going to be blowing leads all season long? And so I went off into the DVOA database. I looked at every team from 2002-2010 to see which teams had the biggest difference between performance before halftime and performance after halftime in the first three games of the season. To keep things consistent with the numbers we have for the Vikings right now, I used VOA rather than DVOA. To save myself a lot of time and energy, I looked at offense and defense only, rather than trying to split up special teams as well.
The Vikings are not actually the most extreme team since 2002. That prize goes to the 2006 Seahawks. The difference is that the 2006 Seahawks won their first three games, rather than losing them. The Seahawks saw their offensive VOA decline by -48.3% in the second half of Weeks 1-3, while their defensive DVOA jumped by 92.0%. Yet they were still able to hold on to beat Detroit, Arizona, and the New York Giants. The Seahawks outscored opponents 55-6 in the first half of those games, and were outscored 40-17 in the second half. They were beating the Giants 42-3 going into the fourth quarter in Week 3 and gave up 27 unanswered points in the fourth quarter, but held on to win.
For the rest of the season, however, there was virtually no difference between the Seahawks' performance in the first half of games and their performance in the second half.
And here's a surprise: There's another 2011 team that is declining almost as much as Minnesota in the second half: Houston. The difference is that Houston is playing much better in the first half, so when they decline in the second half, they just stop scoring rather than actively blowing leads. Still, through three games the Texans have outscored opponents 66-13 in the first half and been outscored 47-20 in the second half. San Francisco has also declined significantly in the second half of games this year. (All this first half-second half change goes along with Jason Lisk's theory that the crazy passing numbers of the 2011 season are going to make late-game comebacks much more common than in the past. You can read about that here.)
Here's a look at the teams that had the worst second-half decline in VOA during the first three games of a season (2002-2010), and what happened to them the rest of the year:
| Biggest Halftime Collapses in Weeks 1-3 based on VOA, 2002-2010 | |||||||||||
| Weeks 1-3 | Weeks 4-16 | ||||||||||
| Year | Team | 1H Off | 2H Off | Dif Off | 1H Def | 2H Def | Dif Def | Dif Both | Dif Off | Dif Def | Dif Both |
| 2006 | SEA | 24.2% | -24.2% | -48.3% | -62.8% | 29.2% | 92.0% | -140.3% | -1.8% | -3.5% | 1.7% |
| 2011 | MIN | 25.8% | -23.4% | -49.2% | -32.6% | 34.0% | 66.6% | -115.8% | |||
| 2003 | MIA | 36.2% | -7.2% | -43.3% | -57.6% | 14.2% | 71.8% | -115.2% | -30.8% | -0.6% | -30.2% |
| 2003 | DAL* | 10.9% | 2.2% | -8.7% | -55.5% | 39.0% | 94.5% | -103.2% | -28.0% | -11.3% | -16.7% |
| 2011 | HOU | 47.3% | -4.3% | -51.6% | -15.3% | 29.3% | 44.6% | -96.2% | |||
| 2002 | MIN | -3.4% | -31.2% | -27.8% | -15.5% | 44.2% | 59.7% | -87.5% | 12.6% | -3.5% | 16.1% |
| 2007 | CHI | -28.5% | -54.8% | -26.4% | -32.3% | 25.9% | 58.2% | -84.6% | -10.8% | 0.7% | -11.5% |
| 2010 | NE | 64.2% | 24.4% | -39.8% | -1.1% | 41.4% | 42.5% | -82.3% | 8.9% | 4.0% | 4.8% |
| 2004 | TEN | -9.3% | 1.6% | 10.9% | -55.7% | 32.6% | 88.2% | -77.4% | -47.8% | -0.4% | -47.4% |
| 2005 | TEN | -12.6% | -49.5% | -36.9% | -12.6% | 27.7% | 40.3% | -77.2% | 9.4% | 8.6% | 0.8% |
| 2011 | SF | -3.5% | -47.4% | -43.9% | -34.3% | -1.7% | 32.6% | -76.5% | |||
| 2007 | BAL | 23.0% | -40.8% | -63.8% | -21.2% | -8.5% | 12.6% | -76.4% | 22.5% | 2.9% | 19.6% |
| 2008 | DEN | 51.9% | 30.1% | -21.8% | -4.0% | 49.1% | 53.1% | -75.0% | 13.6% | 14.7% | -1.1% |
| *Weeks 1-4, as the Cowboys had a bye in Week 3. | |||||||||||
For some of these teams, it looks like the second-half struggles continued for the rest of the season. Well, yes and no. Some of the teams in the table above continued to collapse at halftime, but for completely different reasons. The 2003 Dolphins saw both offense and defense divebomb at halftime of Weeks 1-3. The rest of the season, only offense was a problem. In Weeks 1-4, the 2003 Cowboys had the biggest second-half dropoff on defense of any team I looked at. The rest of the season, while their offense continued to have big problems after halftime, their defense actually got better at the half. The 2004 Titans present the opposite problem -- in Weeks 1-3, the problem is their second-half defense but the offense is fine, while the rest of the season, the problem is their second-half offense and their defense is fine.
Overall, looking at the entire database of teams from 2002-2010, there is no correlation between second-half decline in Weeks 1-3 and second-half decline in Weeks 4-17. (If you want to get technical, the correlation coefficient is .05 on offense, -.07 on defense, and .02 overall.) If we look only at the 40 most extreme teams, the correlation is slightly higher (.13). Still, most of this evidence suggests that the Vikings aren't going to have these kinds of second-half collapses all season. If they continue to play at the level they are playing at overall, they should be a .500 team the rest of the way. Unfortunately, they're starting in an 0-3 hole, but they're not going to end up in the Suck for Luck sweepstakes.
By the way, perhaps you are wondering about the flipside, the teams with the biggest halftime turnarounds in Weeks 1-3. The biggest by far would be the 2006 Giants, of the aforementioned game with Seattle. The Giants were outscored 72-17 in the first half of their first three games, but outscored opponents 64-24 in the second half. That included a memorable overtime victory over Philadelphia where the Giants overcame a 24-7 fourth-quarter lead and won on an overtime touchdown pass to Plaxico Burress. After halftime in these games, the Giants' saw their offensive VOA go up by 74.9% and their defensive VOA go down by -63.9%.
And what about the rest of the season? From Week 4 onwards, the Giants offense didn't improve at all after halftime, while their defense got significantly worse. In fact, eight of the ten teams with the biggest halftime jumps in Weeks 1-3 were actually worse after halftime in Weeks 4-17. These teams averaged a change of 87.5% in Weeks 1-3 (combining offense and defense) but averaged a change of -16.3% at halftime in Weeks 4-17. By the way, the 2011 Bills fall in this top ten area. So far this year, their offense has improved by 53.8% after halftime and their defense has dropped by -30.6%.
* * * * *
All stat pages are now updated, and the FO Premium database of DVOA splits should now be updated through Week 3. Next week we will begin applying opponent adjustments to DVOA, and also we will begin posting the full double-table that includes estimated wins and schedule strength.
Please note that, as announced on the site a few days ago, the KUBIAK midseason update will be available for download after Week 4 this year rather than this week, due to the Jewish holidays.
* * * * *
These are the Football Outsiders team efficiency ratings through three weeks of 2011, 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 VOA are adjusted to consider all fumbles, kept or lost, as equal value. SPECIAL TEAMS VOA is adjusted for type of stadium (warm, cold, dome, Denver) and week of season.
There are no opponent adjustments in VOA until the fourth week of the season, which is why it is VOA right now rather than DVOA. As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.
DAVE is a formula which combines our preseason projection with current VOA to get a more accurate forecast of how a team will play the rest of the season. Right now, the preseason projection makes up 55 percent of DAVE.
To save people some time, please use the following format for all complaints:
<team> is clearly ranked <too high/too low> because <reason unrelated to DVOA>. <subjective ranking system> is way better than this. <unrelated team-supporting or -denigrating comment, preferably with poor spelling and/or chat-acceptable spelling>
| TEAM | TOTAL VOA |
LAST WEEK |
TOTAL DAVE |
RANK | W-L | OFFENSE VOA |
OFF. RANK |
DEFENSE VOA |
DEF. RANK |
S.T. DVOA |
S.T. RANK |
|
| 1 | BUF | 44.4% | 6 | 18.1% | 5 | 3-0 | 51.5% | 1 | 9.5% | 19 | 2.3% | 12 |
| 2 | TEN | 38.4% | 5 | 6.5% | 14 | 2-1 | 17.1% | 8 | -26.2% | 1 | -4.9% | 24 |
| 3 | BAL | 30.1% | 8 | 21.4% | 2 | 2-1 | 7.6% | 12 | -21.4% | 3 | 1.2% | 13 |
| 4 | DET | 28.4% | 2 | 11.4% | 9 | 3-0 | 7.6% | 13 | -23.0% | 2 | -2.2% | 20 |
| 5 | NE | 27.5% | 1 | 28.5% | 1 | 2-1 | 46.4% | 2 | 22.2% | 30 | 3.2% | 10 |
| 6 | NYG | 26.8% | 14 | 15.8% | 7 | 2-1 | 22.1% | 7 | -5.0% | 9 | -0.3% | 16 |
| 7 | NYJ | 25.0% | 3 | 20.9% | 3 | 2-1 | 5.4% | 16 | -10.7% | 6 | 9.0% | 5 |
| 8 | GB | 24.7% | 7 | 17.4% | 6 | 3-0 | 34.7% | 4 | 8.5% | 17 | -1.5% | 18 |
| 9 | HOU | 24.3% | 4 | 14.8% | 8 | 2-1 | 24.3% | 6 | 10.4% | 22 | 10.4% | 3 |
| 10 | OAK | 18.8% | 15 | 0.5% | 17 | 2-1 | 36.2% | 3 | 11.3% | 23 | -6.2% | 27 |
| 11 | CIN | 15.0% | 9 | 0.3% | 18 | 1-2 | 2.5% | 18 | -14.6% | 5 | -2.2% | 19 |
| 12 | SF | 14.4% | 11 | 2.8% | 15 | 2-1 | -22.7% | 29 | -15.6% | 4 | 21.5% | 1 |
| 13 | NO | 11.2% | 10 | 10.4% | 10 | 2-1 | 26.1% | 5 | 19.6% | 29 | 4.7% | 9 |
| 14 | MIN | 8.1% | 16 | 2.3% | 16 | 0-3 | 6.2% | 15 | 7.1% | 16 | 9.0% | 4 |
| 15 | WAS | 7.2% | 13 | -2.5% | 19 | 2-1 | 7.3% | 14 | -3.1% | 12 | -3.2% | 22 |
| 16 | ATL | 6.6% | 18 | 7.8% | 12 | 1-2 | 2.2% | 20 | -4.8% | 10 | -0.5% | 17 |
| TEAM | TOTAL VOA |
LAST WEEK |
TOTAL DAVE |
RANK | W-L | OFFENSE VOA |
OFF. RANK |
DEFENSE VOA |
DEF. RANK |
S.T. DVOA |
S.T. RANK |
|
| 17 | PIT | 2.7% | 21 | 18.9% | 4 | 2-1 | -6.7% | 22 | -6.3% | 8 | 3.1% | 11 |
| 18 | CLE | -3.9% | 20 | -2.8% | 20 | 2-1 | -11.5% | 24 | -2.5% | 13 | 5.1% | 7 |
| 19 | TB | -4.2% | 22 | -6.0% | 22 | 2-1 | 1.7% | 21 | 17.4% | 26 | 11.6% | 2 |
| 20 | ARI | -5.4% | 17 | -11.3% | 25 | 1-2 | 2.3% | 19 | 8.8% | 18 | 1.2% | 14 |
| 21 | DAL | -5.8% | 19 | -6.4% | 23 | 2-1 | -6.9% | 23 | -8.1% | 7 | -6.9% | 28 |
| 22 | PHI | -8.3% | 12 | 9.6% | 11 | 1-2 | 3.6% | 17 | 11.8% | 24 | -0.1% | 15 |
| 23 | SD | -11.9% | 27 | 6.9% | 13 | 2-1 | 11.4% | 10 | 15.2% | 25 | -8.0% | 30 |
| 24 | CAR | -13.6% | 26 | -14.0% | 26 | 1-2 | 10.6% | 11 | 18.6% | 28 | -5.6% | 25 |
| 25 | DEN | -19.3% | 23 | -16.8% | 28 | 1-2 | -14.6% | 25 | 9.8% | 20 | 5.1% | 8 |
| 26 | MIA | -21.2% | 28 | -4.8% | 21 | 0-3 | 12.3% | 9 | 27.8% | 31 | -5.7% | 26 |
| 27 | IND | -23.4% | 25 | -17.3% | 29 | 0-3 | -15.0% | 26 | -3.7% | 11 | -12.1% | 32 |
| 28 | CHI | -27.1% | 24 | -8.4% | 24 | 1-2 | -22.4% | 28 | 10.0% | 21 | 5.4% | 6 |
| 29 | SEA | -33.7% | 31 | -27.3% | 30 | 1-2 | -23.9% | 30 | -1.3% | 15 | -11.2% | 31 |
| 30 | JAC | -36.8% | 29 | -15.5% | 27 | 1-2 | -34.8% | 31 | -1.4% | 14 | -3.4% | 23 |
| 31 | STL | -49.6% | 30 | -29.9% | 31 | 0-3 | -18.3% | 27 | 28.9% | 32 | -2.4% | 21 |
| 32 | KC | -63.1% | 32 | -35.0% | 32 | 0-3 | -37.5% | 32 | 18.5% | 27 | -7.2% | 29 |
Comments
128 comments, Last at 30 Sep 2011, 11:55pm
#1 by AT (not verified) // Sep 27, 2011 - 2:24pm
Let's get this out of the way:
Dallas is clearly ranked too low because Tony Romo is a hero and they beat Washington last night and Washington is ranked higher. By the way, what is a sintient computer program?
#4 by drobviousso // Sep 27, 2011 - 2:35pm
By the way, what is a sintient computer program?
Hal 9000
Data
Kit
Paul Krugeman
#33 by Kevin from Philly // Sep 27, 2011 - 4:01pm
Your examples are sentient. A sintient computer program would be RaiderJoe.
#38 by 0tarin // Sep 27, 2011 - 4:44pm
Nice.
#109 by Frank (not verified) // Sep 28, 2011 - 5:09pm
Krugman really needs to be removed from that list. None of the other three ever had a blog created solely to track their false predictions and math errors. Nor were any of the others former Enron advisors.
#110 by Aaron Brooks G… // Sep 28, 2011 - 5:21pm
I'm sorry Frank, I can't do that.
#13 by Karl Cuba // Sep 27, 2011 - 2:57pm
A program obsessed with Clint Sintim?
#2 by Bay Area Bengal (not verified) // Sep 27, 2011 - 2:25pm
After watching the Oakland/Buffalo match-up last week, I had a subjective feeling that both teams were for real ... and with Buffalo's schedule, I had them pinged to make the playoffs this year.
Looks like DVOA agrees on Buffalo, but Oakland still has a lot to prove (especially on defense!).
#7 by ChicagoRaider // Sep 27, 2011 - 2:38pm
Given that the cornerbacks are Stanford Routt and experiment-of-the-week, it could take some time the Raider defense to stabilize. That is assuming that they can control teams from breaking off tons of monster runs with not-MVP-candidate running backs. That said, if I was Tom Brady I would keep a special eye on Richard Seymour next game.
#3 by andrew // Sep 27, 2011 - 2:34pm
I still want to see a Minnesota-Buffalo game.
#5 by drobviousso // Sep 27, 2011 - 2:36pm
Can't. That would actually flip the stadium over at half time.
#6 by Raiderjoe // Sep 27, 2011 - 2:36pm
Raiders 10th and Jets 7th? Are Dvoa and DAVe drunk? Raiders lose oonly game by 3 pts. Beat Brincos. Beay Jeta by 10. All qualty opponests excelt for Bromcos.
#20 by AJD (not verified) // Sep 27, 2011 - 3:36pm
Pssst, there's a template for your comment, it's right before the comment section and it's in boldface type. Copying and pasting will save you the vexing problem of typing properly.
#26 by CraigoMc (not verified) // Sep 27, 2011 - 3:47pm
Not familiar with FO's patron jester?
#30 by Sablesma // Sep 27, 2011 - 3:50pm
That template applies to everyone but raiderjoe, he has his own special template and he is following it to a T.
#34 by Sidewards // Sep 27, 2011 - 4:05pm
More accurately, to a Y
#46 by Eddo // Sep 27, 2011 - 5:39pm
Awesome.
#8 by Will Allen // Sep 27, 2011 - 2:40pm
If the Vikings are still a top-20 offense by week 10, to say nothing of by season's end, I'll be astounded.
Vikings are clearly rated way too high, because the players are too busy during film sessions searching for foreclosure bargains in Ranchos Palos Verdes on their I-phones. Checking Trulia to see how many players are selling their homes is much better than DVOA! Jason Allen was beter with a mulet!
#40 by JS // Sep 27, 2011 - 5:01pm
Whoa, Will, you kind of went RJ on us there at the end.
#9 by Shawn // Sep 27, 2011 - 2:41pm
As a Bills fan who got way too excited about their start in 2008, I'm still not ready to put my guard down. But it is encouraging that they've improved their VOA ranking and DAVE ranking each week. In 2008, week 2 was actually their top spot, at #3 in VOA. The best the Bills did in DAVE was getting to #14 in week 3. So those are positives. Negatives include this first half/second half DVOA improvement as discussed and, more subjectively, the defense still looks terrible to me. Oh, and that this is the Buffalo Bills we're talking about.
#12 by Timmah! (not verified) // Sep 27, 2011 - 2:55pm
The Bills run defense looks really good to me actually. The pass defense, though- yeesh. No pressure whatsoever and there seems to be one player a week that they just can't cover at all (Pope, Moore, Welker respectively).
#57 by dmstorm22 // Sep 27, 2011 - 8:49pm
Just looking at their schedule in 2008, their 4-0 start was them beating a 4-12 Seattle team by 24, a 5-11 Jags team by 4, a 5-11 Raiders team by 1 and a 2-14 Rams team by 17. They then lost to eventual NFC Champion Arizona 41-17, and then beat the 8-8 Chargers by 9 (in part of the Chargers 4-8 start).
Obviously, we have no idea where the Chiefs, Raiders and Patriots will be, but other than the Chiefs who probably will be awful, the Raiders are mediocre-to-good, and the Patriots will probably be around top-5 all year long. Just the teams they beat seem to spell that this 3-0 is a lot more indicative of a good season than the '08 Bills 4-0 and 5-1 start.
#10 by Joe Mama (not verified) // Sep 27, 2011 - 2:42pm
As a Bills fan and statistics nerd, this is the exact opposite of 1,000,000 kicks to the crotch for me.
#11 by Burbman (not verified) // Sep 27, 2011 - 2:46pm
However, being 2-1 in the NFC West right now is a huge advantage. When you combine that with the fact that our preseason projections had the 49ers as the best team in the division anyway, and the fact that there are so many teams fighting over the six spots in the AFC, our playoff odds report now lists San Francisco as the third most likely team to make the playoffs!
Oh Fine Sir, you fail to underestimate the complete suckitude of the NFC West. I fully expect that all four teams will fail in their quest to Suck for Luck, and will end up with a four way late season battle to see who can be the first 6-10 playoff team in history.
#31 by bravehoptoad // Sep 27, 2011 - 3:53pm
Philadelphia, Tampa, and Detroit all coming up. Hard to see how the 49ers aren't going to be 2-4 to start week 7. The o-line is just so horrible.
Still, god love Brad Seely.
#67 by jebmak // Sep 27, 2011 - 11:52pm
So they'll only be tied for first then.
#72 by Mr Shush // Sep 28, 2011 - 6:37am
I wonder what odds one could get for the Rams starting 0-7 and making the playoffs . . .
#61 by Rhombus (not verified) // Sep 27, 2011 - 9:56pm
I think you overestimate the suckitude of the NFC West. Let's go over the actual games.
Games played outside the NFC West:
SF: 1-1; AZ 1-1; SEA 0-1; STL 0-3
Outside the division, their overall record is 2-6. This includes losses to Philadelphia, Pittsburgh, Baltimore, and the NY Giants, all pretty good teams, and all four of these losses were against the bottom two teams in the NFC West (Rams and Seahawks). The other two losses were a 24-27 OT loss to Dallas by the 49ers, and a 22-21 loss @ Washington by Arizona. This strikes me as a bad division, but not as bad as last year. They could easily be 4-4 outside the division, including 4 tough opponents.
I expect the NFC South or the AFC West to give the NFC West a run for their money as worst division in the league this year, especially after SD's poor performance against KC this week and against the Vikings last week.
#71 by Mr Shush // Sep 28, 2011 - 6:36am
I think you are very, very wrong. The Saints, Falcons, Buccaneers, Chargers and probably Panthers and Raiders would all almost certainly win the NFC West. The Chargers are just doing what the Chargers do. The Redskins and Cowboys are mediocre at best, and both prone to shooting themselves in the ass. I'm not convinced that any of the Eagles, Steelers and Ravens are as good as advertised, and the Giants have no healthy DBs (though they do appear to be better than I expected. 8-8 still probably wins this division.
#83 by justanothersteve // Sep 28, 2011 - 9:16am
Huh? The NFC South? Seriously? Other than maybe Carolina, every NFC South team would probably be the conference favorite if placed in the NFC West. And I think every NFC North and East team except maybe Minnesota (and that's only a maybe) would also be the favorite in the NFC Worst. It's that bad. There may only be 5 AFC teams that wouldn't also be the favorite (Chiefs, Broncs, Colts, Browns, Bungles).
#14 by Arkaein // Sep 27, 2011 - 2:58pm
Hmmm. No way GB is only the 8th best team. In fact, VOA seems down on quite a few NFC teams. NO at 13th? Bears at 28th? Sure, they're 1-2, but that's against GB, NO, and ATL, a pretty brutal opening schedule.
I suppose you can argue that neither GB or NO is really elite, and some none of these matchups have been major tests, but I'm not really buying it. More likely, it seems that these have been fairly tough matchups (throw in NO playing Houston, and they've had a tough schedule of their own), currently not accounted for due to lack of defensive adjustments but preventing any of these teams from registerin gmajor blowouts. Meanwhile, teams like BAL, NYJ, and DET are all good teams riding bit too high on blowouts of scrubs like the Rams and Chiefs.
#16 by andrew // Sep 27, 2011 - 3:01pm
Remember we don't really have defense adjustments in yet. If those teams are all really good then they'll get a boost when those adjustments kick in.
#69 by Fion // Sep 28, 2011 - 4:36am
I agree completely. I remain convinced that DVOA hates Green Bay for some reason.
#89 by bravehoptoad // Sep 28, 2011 - 10:26am
Great offense, middle of the pack special teams and defense. What's the problem?
#15 by RickD // Sep 27, 2011 - 2:59pm
"Buffalo has 58.9% VOA for the game, while New England has -35.8%."
Don't you think that's a bit weird for a game that was decided by 3 points?
#35 by nat // Sep 27, 2011 - 4:25pm
That is truly weird. Weird enough that I suspect something's broken this week.
All the usual suspects - fumble-luck, third and fourth down efficiency, returns, extremely long plays, etc. - fail to explain why the score would diverge so far from the VOA. This did not have the feel of a game in which New England was lucky to keep it close. Instead it felt like a close game that the more deserving team won.
Go figure.
#17 by Temo // Sep 27, 2011 - 3:03pm
I'm not surprised at the Dallas offensive ranking or that the o line has been primarily responsible. I am however baffled that the weakest members of the group thus far seem to be the two holdovers, doug free and Kyle kosier. Then again Kosier has only had 1.5 bad games and the other guard spot has probably been worse in total, but Free sucking was very unforeseen.
#27 by AnonymousD (not verified) // Sep 27, 2011 - 3:47pm
By VOA Dallas would have one of the best teams in the league if their running game came close to competent. Agreed on Free. His last two games have been really disappointing, and I don't have a suitable explanation for them.
#28 by Keith(1) (not verified) // Sep 27, 2011 - 3:48pm
Lol Costa...
#79 by jedmarshall // Sep 28, 2011 - 8:49am
Yeah the other guys may have sucked, but hiking it to the quarterback 4! times before he was ready, once with the game on the line wins the booby prize.
#82 by Temo // Sep 28, 2011 - 9:05am
Well theoretically that's not repeatable. And if it is, he gets replaced with Kowalski, who is pretty much the same quality center (acceptable, a touch above replacement level).
#88 by jedmarshall // Sep 28, 2011 - 10:26am
Yeah I've never seen snaps mangled so horribly before. I'm just saying losing 10 yards a pop because of random snaps in one game doesn't help the offensive VOA.
#18 by Thomas_beardown // Sep 27, 2011 - 3:05pm
So Buffalo and Detroit are both top 5 teams according to VOA, never thought I'd see the day. Coincidentally, if FO continues to build their database backwards, the next year is 1991 which might also have them both in top 5.
#37 by Aaron Brooks G… // Sep 27, 2011 - 4:39pm
You have seen that day, but that day was 1991.
#41 by Thomas_beardown // Sep 27, 2011 - 5:01pm
Well I was 4 at the time, so I don't remember seeing it.
#43 by Eric G (not verified) // Sep 27, 2011 - 5:16pm
Maybe the Bills were not the top 5 because Bruce Smith was injured most of the season. The Lions were an average team that won a lot of close games. That was probably the worst 12-4 team in NFL history.
#77 by Aaron Brooks G… // Sep 28, 2011 - 8:11am
The 1991 Lions had middling pythagorean results because when they lost, holy hell did they lose.
They lost 5 games on the season (13-5), and three of those losses were 45-0 to the Redskins, 35-3 the 49ers, and 41-10 to the Redskins. They were +151 in their other 15 games (13-2). The other losses were 20-10 (Bears) and 30-21 (Bucs). They only played 4 one-score games all year.
#102 by t.d. // Sep 28, 2011 - 1:58pm
they also destroyed the jimmy johnson cowboys in the playoffs, preventing them from taking out the redskins and starting the dynasty a year earlier (don't laugh-those cowboys matched up well against washington)
#84 by Mike B. In Va // Sep 28, 2011 - 9:21am
I dunno, man. That Bills team was ridiculous for the first 8 games - they never really seemed to completely recover from mailing it in for the Monday night week 9 game vs. KC, though.
#19 by srsbzns (not verified) // Sep 27, 2011 - 3:12pm
These rankings and common sense don't really match up. Jets above the Packers? The freakin Giants? Does anyone actually believe that the Giants would beat the Packers on an average Sunday? Don't get me wrong, I know how VOA works. But man, you get some weird looking rankings early in the season.
#22 by Alexander // Sep 27, 2011 - 3:38pm
Such is the way of unadjusted numbers.
GB has had a decent schedule so far (excepting my woeful bears, who are sure to be less woeful when they start requiring their receivers to tar their hands).
#36 by Turin // Sep 27, 2011 - 4:38pm
GB had a huge decline in offensive VOA between weeks 2 and 3, which I don't quite understand. Subjectively, their offensive performance didn't look that much worse than weeks 1 & 2 (the big difference was the 2 turnovers), but obviously the numbers think differently. We'll see what happens when opponent adjustments are added in.
#47 by Duke // Sep 27, 2011 - 5:49pm
As I recall--and the ESPN drive chart ( http://scores.espn.go.com/nfl/drivechart?gameId=310925003 ) seems to back it up--Green Bay's offense was a few really good long drives that resulted in scores, and the rest quick drives. I thought they had a worse game, offensively, than the Saints did against the Bears the previous week.
Still surprised it had a large effect, though.
#56 by Turin // Sep 27, 2011 - 7:52pm
That's what every game for every offense looks like (make sure you're ignoring kneel-down end-of-half/game drives). NO had 5 3-and-out drives against CHI. GB had 4. NO was 2-for-4 in the red zone against CHI, GB was 3-for-5. Both had similar times of possesion.
I can only assume all the false-start penalties, combined with Stark's truly awful performace, killed the VOA numbers.
#64 by DisplacedPackerFan // Sep 27, 2011 - 10:44pm
If you look at the splits, GB's pass VOA went up by a couple percentage points. Their running VOA went from 30 something % (they were top in the league) down to 16.5%. I can buy this as any run, be it Grant or Starks, to the outside was a loss of yardage. The play calls simply had Starks trying to run outside the tackles almost every time. You add in the fumble and it makes sense.
The effect was entirely a very good run offense falling down pretty hard which passes the eyeball test.
#21 by Danish Denver-Fan // Sep 27, 2011 - 3:37pm
You have Indy and Buffalo switched.
#39 by Theo // Sep 27, 2011 - 4:48pm
You have 2010 and 2011 switched.
#23 by Shattenjager // Sep 27, 2011 - 3:40pm
It amazes me how bad the Rams have looked so far after all of the coming-into-the-season optimism. I didn't buy into the idea that it was some sort of average or above-average team or even a likely division winner, but I still thought the 5.5 wins in FOA was likely to be low. Maybe not.
#73 by Mr Shush // Sep 28, 2011 - 6:43am
I think there was some optimism that Kendricks and Sims-Walker would represent enough of an upgrade over the guys catching passes last year to spark a significant offensive jump. That does not appear to be the case. That is an awful group of receivers.
Also, the Rams have a brutally front-loaded schedule. I can still see them winning the division at 8-8.
#80 by t.d. // Sep 28, 2011 - 9:00am
They really do have a better chance of starting 0-7 than of winning the division.
#92 by Mr Shush // Sep 28, 2011 - 11:10am
As I posted earlier in the thread, the two are not mutually exclusive . . .
#98 by greybeard // Sep 28, 2011 - 1:03pm
They did not win 8 games last year. They are worse than they were last year and the competition is better. What makes you think that they can win 8 games? They can easily go 2-4 in NFC West itself IMO.
#100 by Rocco // Sep 28, 2011 - 1:36pm
Are we sure the NFC West is better than last year?
#101 by t.d. // Sep 28, 2011 - 1:49pm
I think San Francisco (head coach) and Arizona (quarterback) are better, but the Seahawks and the Rams are worse. Also, either PFR or Jason Lisk wrote a pretty convincing argument this preseason that Bradford was pretty awful last season (did a side-by-side comparison with Rick Mirer's rookie season, and they were pretty similar), but received undue credit from a vast improvement in St Louis's defense, which was largely due to playing terrible opponents
#105 by Thomas_beardown // Sep 28, 2011 - 3:27pm
Wasn't Rick Mirrer's rookie season considered pretty good? Until everyone learned he couldn't throw to his left?
#106 by Mr Shush // Sep 28, 2011 - 3:29pm
I'm not at all convinced the Rams are worse. I think they've played some tough opponents and had a long cross-country road trip for a 1pm Eastern start, were the best team in the division last year and are probably the best this. Losing to or narrowly beating Seattle is far more unimpressive in my mind than any of the Rams' results to date. And yes, I know the 49ers won by 16, but I'm not willing to take those two fourth quarter TD returns as predictive of much.
Also, Bradford had bad rate stats /= Bradford was awful.
The article you're talking about is this piece on the PFR blog. I think Stuart is far too pat in dismissing the effect of Bradford's supporting cast, and its terribleness. 1993 Brian Blades was vastly better than anything Bradford had to work with, and Kelvin Martin may well have been too. A far more appropriate comparison to my mind is Donovan McNabb, the only quarterback I can think of who was asked to start as a rookie with a similarly terrible group. His rookie rate stats make Bradford look like Joetonny Mantanitas.
Maybe the brutal early schedule results in the Rams quitting on the season, and they collapse. But if they keep their collective heads together, they'll end up with between 6 and 9 wins, and even 6 might win that horrifying suckfest of a division.
#107 by dmstorm22 // Sep 28, 2011 - 4:21pm
What exactly is this long cross-country road trip for a 1 pm Easter start business about?
#111 by greybeard // Sep 28, 2011 - 7:43pm
Rams played 3 games and lost each on of them by at least 12 points differential. Last year they lost only three games with more than 12 points differential. That is 3 of 16 vs 3 out of 3.
I am not sure which Rams game would be any close to loosing to Dallas by 3 at OT and winning 2 other games or losing to Washington by 1 and winning another. Rams lost by 12, 18 and 30 points margins.
Anyway, we will find out about this in a few months.
#112 by BJR // Sep 29, 2011 - 5:26am
As your last sentence implies, it is too early to be leaping to conclusions such as "the Rams are worse than last year". Remember, by no stretch of the imagination were they good last year. So far they have been beaten handsomely by Philly and Baltimore, but I suspect that (barring injury) those two will be top 5 DVOA teams in the long run. And they actually outplayed the Giants for long periods of that game, but failed at key moments. They still have to play the Packers, Cowboys and Saints in the next 4 weeks so there is a real chance they could be 0-7 before their schedule then becomes easy.
Personnel wise, objectively they've probably had a minor upgrade on defence, losing Atogwe but signing Quentin Mikell and adding some experienced depth at linebacker. The problems lie on offence where they have already been hit badly by injuries (losing Amendola for the season and Steven Jackson for the first three games) and failed to address their biggest area of need: receiver. As pointed out above, Bradford's supporting cast, particularly without these two, is painfully bad and anybody who argues that the fault lies with Bradford is simply courting controversy.
#114 by t.d. // Sep 29, 2011 - 10:05am
Alright, I'll make a more nuanced assertion: the Rams are likely to regress this year because defensive improvement is more variable and less sustainable than offensive improvement, and their improved record was the effect of defensive improvement (as well as a historically easy schedule). Furthermore, Sam Bradford, while impressive for a rookie in an abject situation, was probably not one of the 20 best quarterbacks in the league last year. He may have a bright future, but he's already been passed by both Stafford and Newton in the 'quarterback on the rise' conversation and the jury is still out on him.
#24 by tonic889 (not verified) // Sep 27, 2011 - 3:42pm
re: BUF/NE -- Does DVOA do anything to adjust for "tip luck"? Weren't 3 of the 4 INTs based on tipped balls? It would seem to me that a teams luck in intercepting tipped passes should be counted the same way fumbles are counted. (Interceptions of non-tipped passes would of course be a different story)
#29 by CraigoMc (not verified) // Sep 27, 2011 - 3:50pm
To my eye it was 2 tipped passes and 3 bad throws that became 4 interceptions.
What's the definition of a tipped pass? One thrown by a quarterback who is above criticism? how would you account for the other side - bad passes that should have been intercepted and weren't?
#49 by Nathan // Sep 27, 2011 - 5:59pm
What's the definition of a tipped pass? One thrown by a quarterback who is above criticism?
One that is touched by a player causing it's trajectory to be significantly altered? There are plenty of things that are difficult to determine in charting football. If a pass was tipped is not one of them.
#51 by CraigoMc (not verified) // Sep 27, 2011 - 6:30pm
"One that is touched by a player causing it's trajectory to be significantly altered?"
And of course there's no reason why a QB should be penalized in any way for a pass that hits a player other than the intended receiver.
"There are plenty of things that are difficult to determine in charting football. If a pass was tipped is not one of them."
It is, apparently, given the number of people on this site who simultaneously imagined a third tipped interception by Brady.
And even so, who decides whether it was a "fluky" tip or not? Does it depend on what player tipped it? The intent of the player that tipped it? The distance the tipped ball travels, divided by the time it spent in the air? Becuase judging on the comments in Audibles and elsewhere, it seems to depend on how one feels about the QB in question.
If you throw the ball and it ends up in the hands of a defender, you threw an interception. There, that's easier.
#52 by ASmitty // Sep 27, 2011 - 7:12pm
I'm still trying to wrap my mind around how it wasn't Brady's fault that he threw a pass directly off of a defensive lineman's head.
In any event, it wouldn't matter if you treated that "tip" as a half of an interception. Earlier in the game Brady had another ball tipped at the line that he was lucky wasn't intercepted. Treat both of those as a half of an interception and, voila, you have four interceptions overall.
#55 by Nathan // Sep 27, 2011 - 7:28pm
And of course there's no reason why a QB should be penalized in any way for a pass that hits a player other than the intended receiver.
I wrote nothing that even remotely resembles this. I'm talking about what constitutes a tipped pass, not about blaming anyone.
#115 by RickD // Sep 29, 2011 - 12:33pm
Straw men are fun, aren't they.
Which passes were tipped should be easy to establish objectively. Assigning blame is something fans can do but needn't be done by DVOA. After all, it doesn't do so for fumbles, does it?
The original point was that tipped passes lead to a random situation similar to what happens with fumbles. While it's certainly Brady's fault if he throws the ball into a D-lineman's helmet, the question of who snags the ball when it veers 25 feet off course seems random.
#117 by BaronFoobarstein // Sep 29, 2011 - 1:41pm
That sounds right to me. But remember, DVOA says that fumbles are predictive, it's just the recovery of fumbles that aren't. I would suppose that likewise tips are predictive but intercepting them is not.
#62 by MJK // Sep 27, 2011 - 10:25pm
Almost right, by my memory. There were 3 tipped passes and 2 bad throws that became 4 interceptions (the third tipped pass should have been interception #5, but was dropped).
The tips weren't flukes, and all three of them were arguably on Brady or on good play by the Buffalo defense, not due to random luck. Two of them were off of Buffalo players, and hence were either a bad throw by Brady or a Buffalo player getting his body into the throwing lane and getting his arms up. The third was actually tipped by Woodhead, but it was a high and poorly timed throw by Brady.
Thing was, the Bills were getting no pressure to speak of on Brady. This was just his annual "makes horrible decisions and throws 4 interceptions" game. I hope it's his only one this year (it seems like every year he throws one or two of these in an otherwise stellar season).
If you want to bring the fumble analogy further, causing fumbles is a skill, but fumble recovery is luck. Tipping a pass is skill, recovering the tip as an INT is luck. So we shouldn't discount tip drill INTs...rather we should count ALL tipped balls, whether completed, intercepted, or incomplete, as something good done by the defense (or bad by the offense). But, as someone else alluded, that will never be in VOA because "tips" aren't listed in the play-by-play.
#85 by Mike B. In Va // Sep 28, 2011 - 9:26am
The tips weren't flukes, and all three of them were arguably on Brady or on good play by the Buffalo defense, not due to random luck. Two of them were off of Buffalo players, and hence were either a bad throw by Brady or a Buffalo player getting his body into the throwing lane and getting his arms up.
The Bills' D-line players have repeatedly stated that clogging the throwing lanes was part of the defensive game plan in interviews. I wonder how many more tipped Brady passes we'll see this year, given that Buffalo was so successful with it.
It's a scary way to play defense, though, since Welker ate them alive.
#108 by Anonymous1 (not verified) // Sep 28, 2011 - 4:25pm
They weren't flukes, but they were easily avoidable. Barring the NE coaches becoming complete idiots, if the Pats see a defense like that again they should eat it alive.
#116 by RickD // Sep 29, 2011 - 12:39pm
This was just his annual "makes horrible decisions and throws 4 interceptions" game.
He only had 4 picks in all of 2010. This game was very unusual. Actually, he's done so only five times before Sunday, and the most recent was in 2006.
See below (and ignore the fact that WEEI cannot count to six.)
http://itiswhatitis.weei.com/sports/newengland/football/patriots/2011/09/28/looking-at-tom-bradys-history-after-four-interception-games/
#32 by Keith(1) (not verified) // Sep 27, 2011 - 3:56pm
Only two were tipped, and only one of those was "fluky." They are not predictive, but they are not "not interceptions."
#44 by Anonymous1 (not verified) // Sep 27, 2011 - 5:28pm
Regardless how much blame you want to place on Brady, the team and offense should be penalized for some questionable play calling that led to those interceptions. When you are spotted to a three TD lead, run the damn ball. When you can kill the clock and eliminate the chance of the opposing team scoring just before halftime, run the damn ball. When you have a 7 point and are already in FG position with 7 minutes left in the game, run the damn ball.
#78 by Aaron Brooks G… // Sep 28, 2011 - 8:29am
It was in character, though. The Patriots were passing constantly on the Dolphins despite having no defense and having a large lead at almost all times.
#104 by Anonymous1 (not verified) // Sep 28, 2011 - 3:15pm
Actually, the Patriots didn't lead Miami by more than 14 points until the 4th quarter. In the middle of the 3rd quarter, it was still a tie game.
Aside the score, it is just sound situational football. The Pats didn't have nearly the amount of obvious run situations against Miami as they did Buffalo.
#118 by commissionerleaf // Sep 29, 2011 - 5:41pm
Brady's new contract stipulates that he must be permitted to pass often enough to break any and all offensive passing records during the course of the contract, except for career counting statistics. Belichick is not allowed to run the ball unless there have been three consecutive passes for at least 20 total yards, and Brady doesn't decide to check to a pass at the line.
#42 by JasonK // Sep 27, 2011 - 5:09pm
To answer your question, no. The VOA formula is derived almost purely from information available in the official play-by-play, and thus has no basis upon which to apply an adjustment for "tip luck."
#25 by greybeard // Sep 27, 2011 - 3:43pm
1st half SF is #2 in VAO. The glass is half full.
#45 by N8 (not verified) // Sep 27, 2011 - 5:35pm
Let's just remember the 2010 dvoa ratings after week 3. KC was #2.
http://footballoutsiders.com/dvoa-ratings/2010/week-3-dvoa-ratings
#48 by Johnny Socko (not verified) // Sep 27, 2011 - 5:53pm
I see your point, but let's also remember that Green Bay was #1 and Pittsburgh was #3.
#59 by Stuart (not verified) // Sep 27, 2011 - 9:20pm
It also correctly predicted 10 of the 12 play off teams, including KC.
#50 by The Powers That Be // Sep 27, 2011 - 6:00pm
That's nothing - five weeks later, they were #1.
#68 by jebmak // Sep 27, 2011 - 11:56pm
And it worked out pretty well. They won the division and made the playoffs (and made me a chunk of change).
#74 by Mr Shush // Sep 28, 2011 - 6:49am
From that perspective, sure. And (D)VOA was of course right that the Chiefs were much better than people had thought. But it was entirely wrong to think they were actually good, and the division win prediction turned out to be right essentially through fluke (the ungodly awfulness of the Chargers special teams, which neither the projection system nor anyone or anything else had forseen).
#53 by Jonadan // Sep 27, 2011 - 7:18pm
If #2 ranking after 3 weeks means you make the playoffs, Tennessee fans will be ecstatic.
---
"When you absolutely don't know what to do any more, then it's time to panic." - Johann van der Wiel
#54 by BaronFoobarstein // Sep 27, 2011 - 7:23pm
Are there any data released about the comparative predictive power of DAVE and "D"VOA after various weeks?
#58 by dmstorm22 // Sep 27, 2011 - 8:53pm
I looked at the team offense and was shocked that the Raiders have the #4 passing offense by VOA. I guess it probably has to do with Jason Campbell having thrown just 1 pick, and converting first downs. Shocking.
As for the Raiders overall, even in the years of suck (2003-2009), they would pull of random big wins at home, the apex of this being in 2009 when they beat Philly and Cincy at home. This time it felt different though. For once, they didn't have to play out of their heads to beat a good team, they just had to play their game. Contrast this to their unbelievable win over Pittsburgh in 2008 when Bruce Gradkowski decided to turn into Peyton Manning in the 4th quarter. That was obviously a mirage (not for hte Steelers, who were in the midst of five straight losses), but this win over the Jets seemed totally real and repeatable.
#75 by Mr Shush // Sep 28, 2011 - 6:54am
Hmm. I didn't see the game, but the Rex Ryan Jets have always struck me as a good team that can lose to anyone and the Raiders a bad team that can beat anyone, and it was a long cross country trip.
I buy the Raiders as the second best team in the AFC West, and they might go 8-8 or even 9-7. I'm not ready to buy them as actually good.
#93 by Sean McCormick // Sep 28, 2011 - 11:24am
I did see the game, and while I haven't seen the DVOA results for it, I would expect them to come out fairly evenly. Oakland hit a few big plays in the running game and generated a bunch of first downs through penalties, and they put the game away by scoring from a turnover that gave them a short field and then by stopping the Jets on the one foot line, which is basically a 14 point swing in the fourth quarter.
I think Oakland played well and that they would continue to match up well against the Jets O-line, particularly so long as Mangold was out, but I wouldn't favor them in a rematch. Agreed that Oakland clearly looks like the second best team in the AFC West.
#103 by dmstorm22 // Sep 28, 2011 - 2:43pm
I'm just saying that for once it wasn't back to back fumbles like against Cincinnati, or Donovan McNabb playing awful against Philly, or even a 57 yard field goal against the Jets in 2008. This seemed real. Yes, the Jets are probably better, but they looked like they belonged on the same field.
The Raiders didn't play perfectly. They were 0-8 on third downs, for one thing. They gave up 370 yards passing. But they won. They don't need to play perfectly or have their opponent play awful to win anymore, which is as much as Raider fans can ask for now. They aren't a "circled win" on other teams schedules anymore.
#60 by QQ (not verified) // Sep 27, 2011 - 9:48pm
I can't wait until after next week, when all the teams that have been slaughtering KC, STL, etc have their performances heavily downgraded due to opponent adjustments.
#76 by Mr Shush // Sep 28, 2011 - 6:55am
It doesn't work like that. The opponent adjustments get phased in very gradually, so you won't see radical swings.
#63 by MJK // Sep 27, 2011 - 10:30pm
Wow, I know VOA hates interceptions (or, conversely, loves them for the defense), but I didn't realize it was this strong. To my eye, the Bills were totally incapable of stopping the Patriots except by intercepting them, excepting one or two drives (of course, the reverse was also true). So the two teams played almost evenly, with NE having a slight edge (as reflected by the scores). The difference in the game was the NE turnovers, and the somewhat lackadaisical pace at which the Patriots scored their last TD, leaving them no time to counter if Buffalo scored on their final posession. But I have a hard time believing that four INTS account for almost a 90% difference in VOA, especially when Buffalo also had one (real) INT. I wonder what the VOA for this game would look like if you ignored all five real INTs (I say real, because Fitzpatrick also threw an INT on 4th-and-15 in the maroon zone, which is functionally equal to a punt and actually worse for the Patriots than an incomplete, so it shouldn't really count against Buffalo).
The only other thing I can think is that NE's day was "boom or bust"ier, with their running game basically shut down except for one drive, and their offense relying on picking up 12 yard passing chunks, whereas Buffalo was carving up the NE defense at will both on the ground and in the air. I know VOA perfers steady progress...
#65 by Purds // Sep 27, 2011 - 10:59pm
I watched the game, and then read the VOA numbers here and had the same reaction. I suppose my eyes could be a bit off, but I didn't think the game was as dramatically skewed toward Buffalo as the FO numbers suggest, even with the INT's.
My only guess: Perhaps my personal impression of the game was biased by how easily I could imagine NE and Brady taking control of the game again, and that I didn't really see what those 4 interceptions did to the game.
Still, 90%? Doesn't seem right. Maybe 10-20%.
I'd love to know how those numbers break down.
#66 by nat // Sep 27, 2011 - 11:42pm
For VOA in a single game, you should really look at half the difference rather than the difference. Because the same plays are used to judge both teams. You should say that VOA put the Bills between 36% and 58% better than the Patriots.
Still, to claim that the Bills were 47% more effective per play doesn't seem to match the game we saw. The Bills didn't actually achieve 47% more of anything except turnovers: first downs, yards, yards per play, points, return yards. Well, field goals. Even if you adjust yards with the usual 50 yards per interception, you still get less than a 15% advantage.
I wonder if there is a problem with the scaling for goal line plays. The Patriots were 1 for 5 from one yard out. Given that there is a bonus value for actually scoring, going one for five has got to be way below average. If there is an additional multiplier for the red zone, those plays could be ridiculously costly in VOA terms, even with both drives ending in TDs.
Like you, I'd love to see the breakdown. Because something seems weird with this one.
#125 by MJK // Sep 30, 2011 - 11:36am
Good point about looking at half. I had forgotten about that.
I went back and checked, and VOA MUST really love and hate INT's more than I thought. Quantitatively, using VOA's success definition, NE was "successful" on 53% of its plays. Buffalo was "successful" on only 48% of its plays. So my subjective impression that the two teams played fairly evenly, with NE having a slight edge (ignoring the INTs), seems correct. Of course, this is just a success percentage, ignoring the relative value of different successes and the average performance in a given situation (which VOA takes into account). Bug the big difference is that four of NE's failures were drive ending INT's (two near Buffalo's goal line, and one in their own end returned for a TD), and just two of Buffalo's were (one on 4th down, and the other near the middle of the field), so that's GOT to be the difference. However, I question how predictive one game can be.
Your goal line thought is a good one. I was thinking about it. In the middle of the field, on 1st and 10, getting 4 yards is a success. Getting 10 yards is a bigger success, and getting 20 is even bigger. So there's a steady increase in the value of the play with the yards you get. In order to beat the league average (which is probably somewhere in the 3-6 yard range), you need to get quite a few yards, and falling a little short of the league average just gives you slightly fewer "success points". But on the goal line, when you're one yard out, there is no steady increase. You either get the yard, or you don't, and the league average probably leans towards getting that yard at least half the time, so the league average value of that play has got to be pretty high. You can't do much better than the league average (there's no way to get 20 yards instead of 1), but if you don't get that 1, you've done markedly worse than league average. So any unsuccessful play from the 1 yard line is going to look really unsuccessful, and any successful play isn't that impressive. Basically, VOA thinks you really ought to score from the 1. So the fact that the Patriots went 1 for 5 from the 1 is probably depressing their VOA quite a bit.
#126 by Nathan // Sep 30, 2011 - 12:47pm
So the fact that the Patriots went 1 for 5 from the 1 is probably depressing their VOA quite a bit.
It's certainly depressing me.
#70 by Ephraim (not verified) // Sep 28, 2011 - 4:50am
And they're taking it again. All this computer takes into account is statistics. Try the only thing that matters to a TEAM -- the record. The fact that they play 19 (or 20) games and come out on top. That human aspect will never be taken out of the game, and this whole rating system is meaningless. I could care less who's on top because, as of right now, Titletown has won 9 straight. Not stopping anytime soon.
#81 by DGL // Sep 28, 2011 - 9:03am
The Packers are clearly ranked too low because all this computer takes into account is statistics, and the human aspect will never be taken out of the game. The only thing that matters to a TEAM - their record - is way better than this. Titletown has won nine straight. Not stopping anytime soon.
FTFY.
#86 by Keith(1) (not verified) // Sep 28, 2011 - 10:06am
You are looking at it all wrong. These are not rankings. They are simply how the team is playing compared to the rest of the league. The team with the 32nd ranked offense could be undefeated, but their offense still sucks. Right now, Green Bay is simply a top-10 team, and their record reflects a good team. You do not have to be ranked 1st in offense, defense, and special teams to win the Super Bowl. You just have to play the game correctly and beat the teams that need to be beat.
Stop taking it so personally. I am a Green Bay fan and these numbers look right to me. They are not dominating anything, but they are playing good football. Let them continue to play. The numbers will shake out. They will continue to be a top-10 team this year and they will continue to win games. But they will not win every game. Take your wins and take your losses, and understand that in certain situations, they are not the best team.
And, as it should be repeated, the best team does not always win the Super Bowl. Sometimes you get an 16-0 team against a 10-6 team, and the 10-6 team wins. Think about it.
#96 by justanothersteve // Sep 28, 2011 - 11:45am
As another Packers fan, I'll also say the numbers look about right. And they're just numbers. I'm far more concerned that my team is 3-0 than where they are ranked by VOA or DAVE.
#90 by bravehoptoad // Sep 28, 2011 - 10:29am
Try the only thing that matters to a TEAM -- the record.
Unless you're trying to improve that team, or figure out why they lose, or why they win, or where they're weak, or where they're strong, or how to beat them, or how to win with them.
Ignoring all that and a few other little things, then yes, all that matters is the record.
#87 by Will Allen // Sep 28, 2011 - 10:18am
O.K., I now have a non-mocking comment in regards to the Vikings, and the limits of statistical analysis. If I remember right, in 2009 the Vikings were similarly ranked after three games to where they are ranked now. Now, perhaps my eyes are lying to me right now, but you'd have to be drunk, R.J.-style, to think that this year's Vikings is anywhere close, even if they had managed to win one or two of their 3 games so far, to having the potential to be a couple of fumble bounces away from playing in the Super Bowl. The 2009 team, despite the similar ranking, obviously had personnel, mostly on offense, capable of putting pressure on the opposition that this year's team can't even get close to. That team was capable of getting a home playoff game, and crushing their opponent in that game, and then capable of going on the road to one of the more hostile environments, and outgaining the opposition in that environemt by 200 yards or more. This team may not win 6 games.
Weird, and deserving further exploration.
#91 by Jonadan // Sep 28, 2011 - 10:46am
I was going to make a careful reply, something based on the Vike's first-half performance this year and "if they can manage to start playing for 3 quarters instead of only two they'll start winning games", but I thought I'd go look up the 2009 Week 3 DVOA first.
Vikings were 5th, not 14th like this year, and that's a huge difference.
---
"When you absolutely don't know what to do any more, then it's time to panic." - Johann van der Wiel
#94 by DisplacedPackerFan // Sep 28, 2011 - 11:26am
Exactly, for those that don't want to follow the link they were 41.8% VOA (5.6% offense, -27.3% defense, 8.9% special teams) vs 8.1% VOA (6.2% offense, 7.1% defense, 9.0% special teams) is a big difference. My eyes tell me this Vikings team could still be an 8-8 team VOA tells me the same thing.
Perhaps Will was getting at the offensive rankings being similar and that team could crush opponents. Well by week 15 that offense was up to 14% DVOA and was at 23.4% VOA in the final rankings so it improved from what it was early in the year. Also when you have a defense that good, that gave shorter fields, you will score more points too. I don't even think the offensive ratings are that far off. Sure the defense in 09 slipped a bit but still finished at -3.4 % VOA, but early in the year it was very good.
I stuck with VOA from 09 since that is all we have now. They don't rank the same statistically.
#95 by Will Allen // Sep 28, 2011 - 11:39am
Yeah, that is the problem with not checking your memories. I now remember that the middle-aged self photographer was still playing fairly conservative in the first couple weeks of the season, and they had two cream puffs, the Browns and the Rams, early.
My eyes still tell me that this year's offense is worse than VOA indicates.
#97 by Sidewards // Sep 28, 2011 - 12:20pm
It might have to do with the sliding baseline that Aaron said he was going to work on. Changing offensive environment and all that.
#99 by BaronFoobarstein // Sep 28, 2011 - 1:22pm
Buffalo is clearly ranked too high because Ryan Fitzpatrick sucks. Ordering teams on their starting QB's college's acceptance rate is way better than this. Jerry Jones is America's nxt topp model!
#119 by commissionerleaf // Sep 29, 2011 - 5:44pm
Indianapolis is clearly ranked too low because Peyton Manning is injured and will be back. Taking into account the return of the best quarterback in history is way better than this. Go Coltses.
But seriously, folks. If Indy can hang with Pittsburgh without Peyton Manning, either Pittsburgh is hopeless or Indianapolis is much improved on last year (but for Manning).
#120 by Nathan // Sep 29, 2011 - 6:09pm
The former.
#121 by Thomas_beardown // Sep 29, 2011 - 6:40pm
I choose option c, it was a fluke game and the steelers will be fine.
#122 by BaronFoobarstein // Sep 29, 2011 - 7:24pm
I really want to think that, but then there's the Baltimore game.
#123 by Jerry // Sep 29, 2011 - 10:15pm
The correct answer is "we'll see". In the meantime, it looked like they took the Colts' best shot; unfortunately for Indianapolis, the team as currently constituted isn't quite good enough to beat the Steelers. It may be the equivalent of last year's game in Buffalo.
#124 by nat // Sep 30, 2011 - 7:58am
Or football is a team sport, and Manning wasn't the only reason for their success in the past. Not that you would ever consider giving credit to anyone else...
#127 by BigdaddyD (not verified) // Sep 30, 2011 - 9:29pm
I was looking at your numbers and wondering how Baltimore managed to get ahead of Detroit. It appears as though they are tied in Offensive VOA with 7.6, but Detroit has the better defensive VOA through three games, when I realized that Baltimore is only ahead due to a slightly better S.T. ranking, and now that I think about it.
All special teams have both offensive and defensive sides, so I'm now starting to question how accurate the current system is in the S.T. field. Unless your system was geared so that a positive number showed that a S.T. unit was more offensively minded (good return teams) and a negative number showed that the unit was more of a defensive mind(good coverage teams) but then that field would be more about balance than rank.
I think that O.S.T. and D.S.T. Fields need to be introduced in order to accurately rank special teams units, or the field should be abolished all together and O.S.T play should just fall under offense and D.S.T. play should fall under defense.
#128 by BaronFoobarstein // Sep 30, 2011 - 11:55pm
ST = |DST| + OST