by Aaron Schatz
As we enter the final week of the 2009 regular season, DVOA is a big mush of very good but not historically great teams. This week, the season-long ratings really start to look different from the weighted ratings that lower the importance of games more than two months old. Philadelphia remains in the top spot overall, but Green Bay is now the top team in weighted DVOA. The season-long ratings say that the Saints (fifth) are a far superior team to the Carolina Panthers (18th). The weighted ratings say that Carolina (eighth) is a slightly better team than New Orleans (ninth) at this point.
The top six teams in weighted DVOA are now separated by less than five percentage points: Green Bay, New England, Philadelphia, Dallas, Baltimore, and San Diego. Notice any teams missing from that list? Well, New Orleans and Minnesota, of course, but they've faded significantly in the last couple weeks. Indianapolis hasn't faded at all since the beginning of the season, but the Colts aren't in that pack of six either. The Colts are seventh in both total and weighted DVOA, which means that their decision to sit starters in the middle of the third quarter against the Jets hasn't just pissed off the entire state of Indiana. It has also cursed Football Outsiders to another year of dealing with that annoying problem: If DVOA is supposed to filter out luck and context to give the most accurate possible picture of team quality, how do we account for teams that sit their starters once they are locked into a playoff spot?
The answer is "we don't." In the past, our research has always shown that adjusting for "sit starters" games doesn't improve the ability of regular-season DVOA ratings to predict postseason outcomes. There's also a problem setting the boundaries for games should and should not count. Dwight Freeney and Robert Mathis only played third downs against Jacksonville in Week 15; if the game had mattered to the Colts, they probably would have played on first and second down as well. Should that game be adjusted? What about when a team sits the quarterback, but leaves in all the top receivers? Or what if a team doesn't play its best receivers, but does play the quarterback. And so forth, and so on...
Therefore, we don't adjust DVOA for all those plays where Curtis Painter was lousy. Sitting starters is a lot like injuries. We trust readers to know that you have to look at DVOA with common sense and a knowledge of context, just like any other stat. We know that the Colts are a better team than their DVOA rating shows. We know that the Jets are a worse team than their DVOA rating shows. We know these things, just like we know that the Baltimore Ravens team ranked second in DVOA has been decimated by injuries and isn't really the team out on the field right now, just like we know that the Pittsburgh defensive DVOA is much higher in games with Troy Polamalu healthy, and just like we know that the San Diego Chargers' rating for the whole season is still being dragged downward by the first five games of the season.
Out of curiosity, I did look to see what the Colts' DVOA would be if we took out every drive of Sunday's game after Peyton Manning came out. It makes a difference of about two percentage points on the season, which really isn't that much. But if you would like, feel free to view the Colts' ratings with a big fat mental asterisk. I do, which is why we actually ran our playoff odds simulation this week with a "no Painter" version of the Colts' rating, even though it doesn't make much of a difference. (The simulation isn't ready yet, but it also assumes that the Colts and Patriots will sit their starters this week but the Cardinals and Bengals will not. We had to make our best guess, and that's what we went with.)
I said a week ago that I would write a bit about Baltimore if they lost to Pittsburgh and were still high in the ratings, and there they are at number two overall, so let's look at the Ravens a little bit closer. They have been slowing down recently, although not by very much -- the ratings are so tight that a three percent drop in DVOA is enough to drop them from second to fifth. They don't stand out for the usual reasons that DVOA and conventional wisdom disagree about a team. Their schedule strength is average. Their fumble recovery luck has been average. They're unlucky in "hidden" special teams value, but seven points of bad luck doesn't explain an 8-7 record for a team that our estimated wins metric says should be 11-4. (Pythagorean projections agree, giving the Ravens 10.8 wins.) The Ravens have simply alternated close losses to good teams with huge blowouts over bad ones. The main exception to the pattern is their 30-7 Week 8 win over Denver. They've crushed Chicago, Detroit, and Cleveland twice, but they also have five losses by six points or fewer. The other problem for Baltimore is penalties. The Ravens lead the NFL with over 1,000 yards from penalties this year. All past research has shown that penalty totals (especially for defensive penalties) don't actually have a strong correlation with losing games, but the Ravens' extreme flag-happy ways have certainly cost them some games, particularly this last one. Then again, the penalty that has cost them the most is one of the penalties that actually is already incorporated into their high DVOA, defensive pass interference.
If the Ravens win on Sunday to earn a wild card, and can get some of their injured players (especially Ed Reed and Jared Gaither) back for the postseason, I think that they'll be dangerous. I also think the Steelers will be a dangerous playoff team if they manage to qualify and get Troy Polamalu back. With some of these other 8-7 teams, however, I'm worried about yet another year where a mediocre regular-season team suddenly turns it on in January and makes the regular season look meaningless. If the Bengals do sit starters on Sunday night, and the Jets back into the playoffs at 9-7 essentially thanks to two forfeits, and then they go on a run to the Super Bowl, I swear, I am going to hurt somebody. If Jacksonville somehow makes it in and goes on a run to the Super Bowl as an 8-8 team that was outscored by opponents by 80 points during the regular season, I am really going to hurt somebody.
In the NFC, the playoffs will be an interesting (albeit very small sample-sized) test of what momentum really means. If you look at how the NFC teams have played in recent weeks, it's pretty simple: Philadelphia, Dallas, and Green Bay are hot. New Orleans and Minnesota are not. Arizona isn't really trending one way or the other. (Their weighted DVOA is higher than overall, but they've had a couple of poor games recently.) However, as Bill Barnwell pointed out in Quick Reads [2], late-season momentum in the last couple years hasn't meant a tinker's damn. If the playoffs follow the trends of recent years, Anthony Herrera and Bryant McKinnie are going to become John Hannah and Anthony Munoz sometime around January 9. It actually makes for a really good NFC bracket. Based on total season DVOA, Arizona is the worst team of the six, but the Cardinals are far better than they were a year ago and certainly wouldn't stand out as a historically weak Super Bowl representative. Unlike, say, Jacksonville.
(bangs head on desk)
Anyway, individual stats are updated, team stats will be updated shortly, playoff odds simulation and premium update are coming later this evening. (UPDATE: Everything is now updated as of Wednesday morning.)
* * * * *
These are the Football Outsiders team efficiency ratings through 16 weeks of 2009, 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 [3].)
OFFENSE and DEFENSE DVOA are adjusted to consider all fumbles, kept or lost, as equal value. SPECIAL TEAMS DVOA is adjusted for type of stadium (warm, cold, dome, Denver) and week of season. WEIGHTED DVOA is adjusted so that earlier games in the season become gradually less important. It better reflects how well the team is playing right now.
As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.
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 DVOA |
LAST WEEK |
WEIGHTED DVOA |
RANK | W-L | OFFENSE DVOA |
OFF. RANK |
DEFENSE DVOA |
DEF. RANK |
S.T. DVOA |
S.T. RANK |
|
| 1 | PHI | 34.3% | 1 | 30.9% | 3 | 11-4 | 16.7% | 8 | -12.3% | 4 | 5.3% | 2 |
| 2 | BAL | 31.8% | 2 | 28.8% | 5 | 8-7 | 16.2% | 9 | -14.1% | 2 | 1.4% | 12 |
| 3 | NE | 31.3% | 5 | 31.9% | 2 | 10-5 | 31.0% | 1 | 1.8% | 16 | 2.1% | 11 |
| 4 | GB | 28.4% | 6 | 32.3% | 1 | 10-5 | 22.8% | 6 | -12.4% | 3 | -6.8% | 32 |
| 5 | NO | 26.6% | 3 | 18.4% | 9 | 13-2 | 30.3% | 2 | 0.9% | 13 | -2.9% | 28 |
| 6 | DAL | 24.8% | 7 | 29.8% | 4 | 10-5 | 25.0% | 4 | 1.5% | 15 | 1.3% | 13 |
| 7 | IND | 24.2% | 4 | 21.6% | 7 | 14-1 | 23.6% | 5 | -1.1% | 10 | -0.5% | 22 |
| 8 | SD | 16.8% | 13 | 27.4% | 6 | 12-3 | 25.1% | 3 | 8.8% | 23 | 0.6% | 18 |
| 9 | DEN | 16.7% | 10 | 11.1% | 13 | 8-7 | 4.3% | 19 | -12.0% | 5 | 0.5% | 19 |
| 10 | MIN | 16.1% | 8 | 10.7% | 14 | 11-4 | 15.8% | 10 | 4.8% | 19 | 5.1% | 3 |
| 11 | PIT | 15.2% | 12 | 16.4% | 12 | 8-7 | 18.5% | 7 | -1.1% | 9 | -4.4% | 30 |
| 12 | ARI | 14.3% | 11 | 17.5% | 10 | 10-5 | 11.2% | 13 | -0.5% | 11 | 2.6% | 7 |
| 13 | NYJ | 12.8% | 15 | 10.6% | 15 | 8-7 | -10.4% | 23 | -19.7% | 1 | 3.4% | 6 |
| 14 | HOU | 10.7% | 14 | 16.7% | 11 | 8-7 | 15.7% | 11 | 7.3% | 21 | 2.3% | 9 |
| 15 | NYG | 9.9% | 9 | 5.7% | 17 | 8-7 | 14.9% | 12 | 3.5% | 17 | -1.4% | 26 |
| 16 | MIA | 5.6% | 17 | 4.9% | 18 | 7-8 | 7.9% | 15 | 4.5% | 18 | 2.3% | 8 |
| TEAM | TOTAL DVOA |
LAST WEEK |
WEIGHTED DVOA |
RANK | W-L | OFFENSE DVOA |
OFF. RANK |
DEFENSE DVOA |
DEF. RANK |
S.T. DVOA |
S.T. RANK |
|
| 17 | CIN | 5.6% | 16 | 5.8% | 16 | 10-5 | 6.5% | 18 | -0.1% | 12 | -1.1% | 23 |
| 18 | CAR | 4.0% | 22 | 18.6% | 8 | 7-8 | 0.0% | 20 | -7.6% | 7 | -3.6% | 29 |
| 19 | ATL | 1.5% | 21 | -2.0% | 22 | 8-7 | 9.5% | 14 | 8.0% | 22 | 0.0% | 20 |
| 20 | SF | 1.0% | 20 | 3.9% | 19 | 7-8 | -10.0% | 22 | -10.4% | 6 | 0.6% | 17 |
| 21 | TEN | -5.7% | 18 | 1.4% | 20 | 7-8 | 7.7% | 16 | 12.4% | 27 | -1.1% | 24 |
| 22 | WAS | -6.1% | 23 | -1.8% | 21 | 4-11 | -4.0% | 21 | 1.0% | 14 | -1.1% | 25 |
| 23 | JAC | -6.7% | 19 | -13.8% | 23 | 7-8 | 7.7% | 17 | 14.0% | 28 | -0.4% | 21 |
| 24 | BUF | -15.3% | 24 | -15.2% | 24 | 5-10 | -20.7% | 29 | -3.3% | 8 | 2.2% | 10 |
| 25 | CHI | -20.2% | 25 | -23.8% | 27 | 6-9 | -19.8% | 28 | 5.4% | 20 | 5.0% | 4 |
| 26 | TB | -23.9% | 28 | -20.6% | 26 | 3-12 | -15.1% | 25 | 12.2% | 26 | 3.4% | 5 |
| 27 | CLE | -24.1% | 27 | -19.8% | 25 | 4-11 | -13.9% | 24 | 19.4% | 31 | 9.2% | 1 |
| 28 | SEA | -31.4% | 26 | -39.2% | 30 | 5-10 | -16.2% | 26 | 16.1% | 29 | 0.9% | 15 |
| 29 | KC | -32.4% | 30 | -34.9% | 29 | 3-12 | -18.9% | 27 | 11.4% | 25 | -2.1% | 27 |
| 30 | OAK | -33.6% | 29 | -30.6% | 28 | 5-10 | -24.2% | 30 | 10.1% | 24 | 0.7% | 16 |
| 31 | STL | -43.4% | 31 | -40.5% | 31 | 1-14 | -24.6% | 31 | 20.0% | 32 | 1.3% | 14 |
| 32 | DET | -51.8% | 32 | -50.9% | 32 | 2-13 | -28.2% | 32 | 18.8% | 30 | -4.8% | 31 |
| TEAM | TOTAL DVOA |
W-L | NON-ADJ TOT VOA |
ESTIM. WINS |
RANK | PAST SCHED |
RANK | VAR. | RANK | |
| 1 | PHI | 34.3% | 11-4 | 37.1% | 10.8 | 4 | -0.7% | 23 | 17.4% | 23 |
| 2 | BAL | 31.8% | 8-7 | 31.2% | 11.2 | 1 | 1.4% | 17 | 13.5% | 14 |
| 3 | NE | 31.3% | 10-5 | 26.2% | 10.6 | 5 | 4.7% | 9 | 21.1% | 29 |
| 4 | GB | 28.4% | 10-5 | 38.5% | 10.2 | 7 | -9.8% | 31 | 15.6% | 18 |
| 5 | NO | 26.6% | 13-2 | 30.5% | 11.1 | 2 | -2.4% | 25 | 15.0% | 16 |
| 6 | DAL | 24.8% | 10-5 | 23.6% | 10.6 | 6 | 0.9% | 18 | 7.1% | 2 |
| 7 | IND | 24.2% | 14-1 | 26.4% | 10.9 | 3 | 2.2% | 15 | 8.4% | 6 |
| 8 | SD | 16.8% | 12-3 | 18.5% | 9.8 | 8 | -0.1% | 21 | 7.5% | 3 |
| 9 | DEN | 16.7% | 8-7 | 14.9% | 9.5 | 10 | 5.1% | 8 | 16.1% | 19 |
| 10 | MIN | 16.1% | 11-4 | 21.6% | 9.3 | 12 | -7.1% | 30 | 10.7% | 10 |
| 11 | PIT | 15.2% | 8-7 | 20.5% | 9.4 | 11 | -2.5% | 26 | 11.9% | 12 |
| 12 | ARI | 14.3% | 10-5 | 21.1% | 9.7 | 9 | -10.4% | 32 | 19.4% | 25 |
| 13 | NYJ | 12.8% | 8-7 | 14.8% | 8.4 | 14 | 2.5% | 14 | 16.7% | 21 |
| 14 | HOU | 10.7% | 8-7 | 15.7% | 9.2 | 13 | -3.8% | 29 | 8.0% | 4 |
| 15 | NYG | 9.9% | 8-7 | 1.3% | 8.2 | 17 | 6.0% | 5 | 21.6% | 31 |
| 16 | MIA | 5.6% | 7-8 | 2.2% | 8.2 | 16 | 6.6% | 4 | 9.5% | 7 |
| TEAM | TOTAL DVOA |
W-L | NON-ADJ TOT VOA |
ESTIM. WINS |
RANK | PAST SCHED |
RANK | VAR. | RANK | |
| 17 | CIN | 5.6% | 10-5 | 8.8% | 8.3 | 15 | -0.2% | 22 | 17.6% | 24 |
| 18 | CAR | 4.0% | 7-8 | -1.9% | 7.9 | 19 | 6.8% | 3 | 21.2% | 30 |
| 19 | ATL | 1.5% | 8-7 | -1.4% | 8.1 | 18 | 7.2% | 2 | 10.6% | 8 |
| 20 | SF | 1.0% | 7-8 | 5.0% | 7.2 | 20 | -2.9% | 27 | 11.0% | 11 |
| 21 | TEN | -5.7% | 7-8 | -12.8% | 7.1 | 21 | 5.9% | 6 | 34.6% | 32 |
| 22 | WAS | -6.1% | 4-11 | -6.1% | 6.9 | 23 | 0.1% | 19 | 14.8% | 15 |
| 23 | JAC | -6.7% | 7-8 | -6.7% | 7.1 | 22 | 0.1% | 20 | 20.9% | 28 |
| 24 | BUF | -15.3% | 5-10 | -17.3% | 6.1 | 24 | 3.1% | 12 | 15.3% | 17 |
| 25 | CHI | -20.2% | 6-9 | -22.0% | 5.3 | 25 | 2.6% | 13 | 13.2% | 13 |
| 26 | TB | -23.9% | 3-12 | -29.0% | 4.7 | 27 | 9.8% | 1 | 16.8% | 22 |
| 27 | CLE | -24.1% | 4-11 | -27.6% | 4.9 | 26 | 1.9% | 16 | 16.3% | 20 |
| 28 | SEA | -31.4% | 5-10 | -27.5% | 3.3 | 30 | -3.4% | 28 | 20.1% | 27 |
| 29 | KC | -32.4% | 3-12 | -29.9% | 3.7 | 29 | 3.3% | 11 | 6.8% | 1 |
| 30 | OAK | -33.6% | 5-10 | -34.8% | 4.0 | 28 | 5.3% | 7 | 19.6% | 26 |
| 31 | STL | -43.4% | 1-14 | -40.7% | 2.7 | 31 | -1.1% | 24 | 10.6% | 9 |
| 32 | DET | -51.8% | 2-13 | -52.9% | 2.2 | 32 | 3.6% | 10 | 8.1% | 5 |
Links:
[1] http://ad.doubleclick.net/jump/footballoutsiders.fsv/ros;sect=ros;fantasy=yes;game=no;tile=3;sz=300x250;ord=' random_number '?
[2] http://www.footballoutsiders.com/quick-reads/2009/week-16-quick-reads
[3] http://www.footballoutsiders.com/methods#dvoa
[4] http://www.footballoutsiders.com/info/glossary#estimated_wins
[5] http://www.footballoutsiders.com/info/glossary#variance