Week 13 DVOA Ratings
by Aaron Schatz
With a dominant win over the Los Angeles Rams, the New England Patriots jumped four spots and now top the Football Outsiders DVOA ratings for the first time this season. The Patriots inched barely ahead of the Dallas Cowboys, who remain at No. 2 this week. However, the bigger story is not which teams are on top but how low their ratings are. The Year of No Great Teams has reached another pinnacle... or is it more appropriate to say nadir? New England has 19.9% DVOA and Dallas has 19.7% DVOA, making this the first time in the history of the DVOA ratings that the No. 1 team has been below 20%.
Four of last week's top five teams dropped in DVOA this week. No. 1 Atlanta lost a close game to Kansas City. Dallas dropped because their game with the Vikings was effectively a tie for DVOA purposes. The Philadelphia Eagles fade to No. 5 (No. 6 in the weighted rating) as DVOA's one-time favorite has put up its three worst games in the last three weeks. Even Seattle dropped a little bit this week; the Seahawks got a nice 45.4% rating for their big win over Carolina, but schedule adjustments drop them as a number of their past opponents (Rams, Dolphins, 49ers, Jets) got creamed this week.
Meanwhile, Oakland, Pittsburgh, and Kansas City all saw their DVOA ratings rise with good wins, and the Ravens jumped back into the top ten after destroying the Miami Dolphins. As a result, the top of our ratings are packed together extremely tight. The top five teams are separated by just 3.3% DVOA, and the top nine teams are separated by just 9.0% DVOA.
New England and Dallas do manage to clear the 20% DVOA bar in the weighted DVOA rating that drops the strength of earlier games, but the gap between the Patriots/Cowboys and previous No. 1 teams in DVOA is gigantic. Again: it's not just that there's never been a No. 1 team rated this low after Week 13. There's never been a No. 1 team rated this low, ever. Going all the way back to 1989, only one team ever was No. 1 with a DVOA rating under 25%, the 2011 Green Bay Packers with 24.7% DVOA after Week 15. If you want to revisit a similar past, you can find that article here, although the Packers were listed with a slightly higher rating at the time because we've changed the method a little bit since then. The 2011 Packers were actually sort of similar to the Dallas Cowboys of 2016. They went 15-1 but really weren't as good as that record. They won a lot of close games because their great offense overcame a poor defense.
(Another reason that conventional wisdom and DVOA differ on the strength of the New England Patriots: they've played the easiest schedule in the league so far based on average DVOA of opponent, although Tennessee or Indianapolis will probably still end up with the easiest schedule for the full season.)
How unique is The Year of No Great Teams? Not only has there never been a No. 1 team below 20%, but only twice before was the No. 1 team after Week 13 rated below 30%: the 2000 Titans (29.5%) and the 2011 Texans (29.9%). The Patriots are higher in weighted DVOA, at 25.0%, but that's also the lowest ever for the No. 1 team though Week 13. The previous low belonged to the 2006 Dallas Cowboys, who were No. 1 in weighted DVOA at 29.1% after Week 13. That was Tony Romo's first season as the starter, by the way, but the team peaked at midseason, finishing ninth in DVOA and backing into the playoffs with a wild card after a 1-3 record in the last four weeks.
But wait, there are even more tidbits that show how unique 2016 is as the Year of No Great Teams. Only twice since 1989 has the No. 3 team in DVOA after Week 13 rated below this year's No. 1 Patriots: the 1991 Saints at 17.1% and the 2003 Titans at 18.6%. In fact, in 16 of the 27 seasons since 1989, there were at least five teams with higher DVOA than any team has in 2016. In 2009, there were eight ddifferent teams with higher DVOA after Week 13 than any team has right now in 2016.
One big explanation for The Year of No Great Teams is that this is The Year of No Well-Balanced Teams. Right now, the top seven teams in offensive DVOA all have defenses ranked between 19th and 28th. Meanwhile, the top seven teams in defensive DVOA all have offenses ranked between 19th and 28th except for Seattle, which is now No. 13 on offense. Patriots fans, are you worried about your defense? Great, so are Cowboys fans, Falcons fans, and Raiders fans. Broncos fans, worried about your offense? So are Ravens fans and Vikings fans. The only teams that look at all balanced right now are Seattle, where the offense will likely rebound a bit from that horrible game against Tampa Bay, and -- surprisingly -- Pittsburgh, which now ranks No. 8 in both offense and defense.
The top teams are even more packed together when you consider that three of the top four teams in DVOA just lost great players to season-ending injuries. New England lost the best tight end in the NFL. Seattle lost the best free safety in the NFL. Desmond Trufant isn't the best cornerback in the NFL, but he was the best defensive player on the Atlanta Falcons. The playoff odds simulation has adjusted the ratings of these three teams based on a rough estimate of what these players might be worth. Remember that the drop with a replacement quarterback is a lot bigger than any other position, and losing a top offensive player tends to have a more predictable impact than losing a top defensive player. Our estimate of the value of Gronk is enough to move Dallas ahead of New England as our current favorite to win Super Bowl LI.
[ad placeholder 3]
You may be asking if The Year of No Great Teams is also The Year of No Horrible Teams. The answer is not quite. It's actually The Year of Not Too Many Horrible Teams. The Browns kept things close in some of their earlier losses, but they are scraping the bottom of the barrel now. Their DVOA of -38.1% is middle of the pack historically for the worst teams of each year, and the Browns are now ranked in the bottom five for all three units. The Jets are now No. 31 after the Colts just plain brutalized them this week. A 41-10 blowout loss to an Indianapolis team that still ranks only 25th in DVOA gives the Jets the worst game of the season by DVOA, with a dismal rating of -110.8%. The Jets rank in the bottom six for all three units. Also near the bottom of the league are Houston, San Francisco, Los Angeles -- although they are a Johnny Hekker-powered No. 2 on special teams, trailing only Philadelphia -- and Jacksonville.
However, you'll notice there's a pretty big gap between those six teams and everyone else in the league. Usually, there are more teams down below -10% DVOA. Every other season of DVOA going back to 1989 has had at least eight different teams below -10% DVOA after Week 13 except for three. In 1991, 22 of 28 teams were above -10% at this point of the season. In 1993, 21 of 28 teams were above -10% at this point, although that was the year of two bye weeks so teams had only played 11 games apiece. And in 2003, the first year of our website, 26 of 32 teams were above -10% after Week 13.
As strange as The Year of No Great Teams is, at least one general rule has reasserted itself over the course of the 2016 season. You may remember that 2015 was one of the rare years where the best defense was stronger than the best offense, and the worst defense was weaker than the worst offense. Early this season, it looked like that trend would continue in 2016. It hasn't. The top offenses, Dallas and Atlanta, are above 25%, while the best defense right now is Denver at -18.5%. The worst offense is now Los Angeles at -29.3%, and both the Rams and the Houston Texans offense are worse than the Cleveland defense at 17.3% DVOA.
* * * * *
 |
Once again this season, we have teamed up with EA Sports to bring Football Outsiders-branded player content to Madden 17 Ultimate Team. Each week, we'll be picking out a handful of players who starred in that week's games. Some of them will be well-known players who stood out in DVOA and DYAR. Others will be under-the-radar players who only stood out with advanced stats. We'll announce the players each Tuesday in the DVOA commentary article, and the players will be available in Madden Ultimate Team packs the following weekend, beginning at 11am Eastern on Friday. We will also tweet out images of these players from the a href="https://twitter.com/fboutsiders" target="_blank">@fboutsiders Twitter account on most Fridays. The best player of each week, the Football Outsiders Hero, will require you to collect a set of the other four Football Outsiders players that week, plus a certain number of Football Outsiders collectibles available in Madden Ultimate Team packs.
The Football Outsiders stars for Week 13 are:
- RB Devonta Freeman, ATL (FOOTBALL OUTSIDERS HERO): No. 2 RB with 56 DYAR in Week 13 (15 carries, 56 yards, 2 TD plus 49 yards and three first downs receiving).
- TE Dwayne Allen, IND: No. 1 TE with 61 DYAR in Week 13 (4 catches, 72 yards, 3 TD).
- WR Tyler Lockett, SEA: No. 4 WR with 51 DYAR in Week 13 (5 catches, 63 yards, plus 75-yard touchdown run and 46-yard kickoff return).
- LT Ronnie Stanley, BAL: No sacks allowed; Baltimore RB had 49 yards on 8 carries to the left.
- LE Derek Wolfe, DEN: Six run tackles for a combined 14 yards plus 2 QB knockdowns.
Jacksonville is now the only team that hasn't gotten a Football Outsiders player in Madden 17 Ultimate Team. We'll try to take care of that in the final four weeks.
* * * * *
All player/team DVOA stats pages are now updated through Week 13 of 2016. Playoff odds, snap counts, and the premium DVOA database are also fully updated. Drive stats and pace stats will be updated later this evening.
If you're looking for more of my thoughts on the Football Outsiders playoff odds and DVOA ratings, my playoff odds commentary at ESPN Insider will be running on Tuesday afternoons instead of Wednesday mornings for the remainder of the regular season.
* * * * *
[ad placeholder 4]
These are the Football Outsiders team efficiency ratings through 13 weeks of 2016, 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 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. As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE. As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.
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.
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>
(Ed. Note: My apologies. I had the second ratings table sorted wrong when I created today's HTML tables, which caused all kinds of errors with both weighted DVOA and non-adjusted VOA. Tables are now fixed below. Most of the questions in the comment thread prior to 1:00 am EST Wednesday morning are now moot. -- Aaron)
| TEAM | TOTAL DVOA |
LAST WEEK |
WEI. DVOA |
RANK | W-L | OFFENSE DVOA |
OFF. RANK |
DEFENSE DVOA |
DEF. RANK |
S.T. DVOA |
S.T. RANK |
|
| 1 | NE | 19.9% | 5 | 25.0% | 1 | 10-2 | 19.6% | 3 | 1.8% | 19 | 2.1% | 10 |
| 2 | DAL | 19.7% | 2 | 21.6% | 2 | 11-1 | 25.1% | 2 | 8.4% | 26 | 2.9% | 9 |
| 3 | SEA | 18.4% | 4 | 17.8% | 4 | 8-3-1 | 2.5% | 13 | -14.4% | 5 | 1.5% | 11 |
| 4 | ATL | 17.3% | 1 | 18.5% | 3 | 7-5 | 25.1% | 1 | 10.8% | 28 | 3.0% | 8 |
| 5 | PHI | 16.6% | 3 | 15.3% | 6 | 5-7 | -6.5% | 22 | -14.6% | 4 | 8.5% | 1 |
| 6 | OAK | 15.6% | 7 | 16.1% | 5 | 10-2 | 18.4% | 4 | 6.2% | 22 | 3.3% | 7 |
| 7 | PIT | 15.4% | 6 | 14.3% | 8 | 7-5 | 9.9% | 8 | -5.9% | 8 | -0.4% | 18 |
| 8 | KC | 12.2% | 8 | 14.8% | 7 | 9-3 | 2.8% | 12 | -3.0% | 11 | 6.5% | 4 |
| 9 | BAL | 10.9% | 13 | 13.3% | 9 | 7-5 | -12.3% | 27 | -18.4% | 2 | 4.8% | 6 |
| 10 | WAS | 7.8% | 10 | 9.8% | 10 | 6-5-1 | 14.3% | 6 | 7.4% | 23 | 1.0% | 13 |
| 11 | DEN | 5.4% | 15 | 3.9% | 13 | 8-4 | -10.6% | 24 | -18.5% | 1 | -2.4% | 23 |
| 12 | GB | 5.2% | 18 | 3.4% | 15 | 6-6 | 8.7% | 9 | 1.5% | 18 | -2.0% | 20 |
| 13 | NYG | 4.6% | 14 | 5.0% | 12 | 8-4 | -2.8% | 19 | -9.6% | 7 | -2.2% | 21 |
| 14 | MIN | 3.8% | 16 | 0.4% | 20 | 6-6 | -11.5% | 25 | -14.2% | 6 | 1.1% | 12 |
| 15 | NO | 3.7% | 11 | 5.4% | 11 | 5-7 | 15.8% | 5 | 8.2% | 25 | -3.9% | 26 |
| 16 | BUF | 3.3% | 12 | 3.1% | 16 | 6-6 | 8.4% | 10 | 5.1% | 21 | 0.0% | 16 |
| TEAM | TOTAL DVOA |
LAST WEEK |
WEI. DVOA |
RANK | W-L | OFFENSE DVOA |
OFF. RANK |
DEFENSE DVOA |
DEF. RANK |
S.T. DVOA |
S.T. RANK |
|
| 17 | SD | 3.0% | 17 | 1.1% | 18 | 5-7 | -0.3% | 15 | -5.1% | 9 | -1.8% | 19 |
| 18 | TEN | 0.7% | 20 | 2.3% | 17 | 6-6 | 11.8% | 7 | 7.9% | 24 | -3.2% | 24 |
| 19 | MIA | -0.4% | 9 | 3.8% | 14 | 7-5 | -0.7% | 18 | 0.0% | 17 | 0.3% | 15 |
| 20 | TB | -0.6% | 19 | 1.0% | 19 | 7-5 | -0.7% | 17 | -2.4% | 12 | -2.3% | 22 |
| 21 | CIN | -0.8% | 21 | 0.2% | 21 | 4-7-1 | 4.6% | 11 | 2.0% | 20 | -3.5% | 25 |
| 22 | ARI | -3.6% | 23 | -5.0% | 23 | 5-6-1 | -12.2% | 26 | -14.8% | 3 | -6.2% | 30 |
| 23 | CHI | -4.4% | 24 | -3.0% | 22 | 3-9 | -4.9% | 21 | -0.5% | 16 | 0.0% | 17 |
| 24 | CAR | -7.2% | 22 | -7.5% | 25 | 4-8 | -4.6% | 20 | -2.2% | 14 | -4.8% | 29 |
| 25 | IND | -8.4% | 28 | -6.3% | 24 | 6-6 | -0.4% | 16 | 14.9% | 30 | 6.8% | 3 |
| 26 | DET | -8.7% | 25 | -8.3% | 26 | 8-4 | 1.5% | 14 | 16.1% | 31 | 5.9% | 5 |
| 27 | JAC | -16.3% | 29 | -17.3% | 27 | 2-10 | -15.9% | 29 | -3.5% | 10 | -3.9% | 27 |
| 28 | LARM | -19.3% | 26 | -17.4% | 28 | 4-8 | -29.3% | 32 | -1.9% | 15 | 8.0% | 2 |
| 29 | SF | -22.3% | 27 | -25.6% | 30 | 1-11 | -10.3% | 23 | 12.5% | 29 | 0.5% | 14 |
| 30 | HOU | -24.5% | 30 | -25.3% | 29 | 6-6 | -19.8% | 31 | -2.2% | 13 | -6.9% | 31 |
| 31 | NYJ | -32.5% | 31 | -33.8% | 31 | 3-9 | -15.6% | 28 | 9.3% | 27 | -7.6% | 32 |
| 32 | CLE | -38.1% | 32 | -38.7% | 32 | 0-12 | -16.7% | 30 | 17.3% | 32 | -4.0% | 28 |
- NON-ADJUSTED TOTAL DVOA does not include the adjustments for opponent strength or the adjustments for weather and altitude in special teams, and only penalizes offenses for lost fumbles rather than all fumbles.
- ESTIMATED WINS uses a statistic known as "Forest Index" that emphasizes consistency as well as DVOA in the most important specific situations: red zone defense, first quarter offense, and performance in the second half when the score is close. It then projects a number of wins adjusted to a league-average schedule and a league-average rate of recovering fumbles. Teams that have had their bye week are projected as if they had played one game per week.
- PAST SCHEDULE lists average DVOA of opponents played this season, ranked from hardest schedule (#1, most positive) to easiest schedule (#32, most negative). It is not adjusted for which games are home or road.
- FUTURE SCHEDULE lists average DVOA of opponents still left to play this season, ranked from hardest schedule (#1, most positive) to easiest schedule (#32, most negative). It is not adjusted for which games are home or road.
- VARIANCE measures the statistical variance of the team's weekly DVOA performance. Teams are ranked from most consistent (#1, lowest variance) to least consistent (#32, highest variance).
| TEAM | TOTAL DVOA |
W-L | NON-ADJ TOT VOA |
ESTIM. WINS |
RANK | PAST SCHED |
RANK | FUTURE SCHED |
RANK | VAR. | RANK | |
| 1 | NE | 19.9% | 10-2 | 24.6% | 8.3 | 2 | -8.4% | 32 | -4.1% | 21 | 14.9% | 22 |
| 2 | DAL | 19.7% | 11-1 | 18.6% | 9.0 | 1 | 0.6% | 18 | 3.0% | 12 | 3.7% | 2 |
| 3 | SEA | 18.4% | 8-3-1 | 20.9% | 7.9 | 5 | -2.1% | 22 | -10.0% | 30 | 13.0% | 21 |
| 4 | ATL | 17.3% | 7-5 | 10.9% | 8.3 | 3 | 5.7% | 2 | -11.3% | 31 | 10.3% | 16 |
| 5 | PHI | 16.6% | 5-7 | 15.0% | 7.9 | 4 | 3.3% | 4 | 10.8% | 3 | 15.6% | 24 |
| 6 | OAK | 15.6% | 10-2 | 14.4% | 7.5 | 6 | 0.7% | 17 | 3.1% | 11 | 8.1% | 10 |
| 7 | PIT | 15.4% | 7-5 | 17.2% | 7.1 | 9 | 1.0% | 12 | -6.2% | 26 | 28.6% | 32 |
| 8 | KC | 12.2% | 9-3 | 12.1% | 7.1 | 8 | -2.4% | 24 | 6.2% | 7 | 17.6% | 29 |
| 9 | BAL | 10.9% | 7-5 | 16.2% | 7.0 | 10 | -5.0% | 30 | 12.8% | 1 | 15.7% | 25 |
| 10 | WAS | 7.8% | 6-5-1 | 4.1% | 7.2 | 7 | 3.7% | 3 | 2.4% | 14 | 9.6% | 14 |
| 11 | DEN | 5.4% | 8-4 | 12.0% | 6.4 | 16 | -0.2% | 20 | 12.1% | 2 | 11.4% | 18 |
| 12 | GB | 5.2% | 6-6 | 3.5% | 6.2 | 17 | 0.7% | 16 | 2.3% | 15 | 13.0% | 20 |
| 13 | NYG | 4.6% | 8-4 | 1.4% | 6.7 | 14 | 1.7% | 10 | 8.8% | 4 | 1.0% | 1 |
| 14 | MIN | 3.8% | 6-6 | 6.2% | 6.9 | 12 | -0.2% | 19 | -6.0% | 25 | 9.0% | 13 |
| 15 | NO | 3.7% | 5-7 | 8.5% | 7.0 | 11 | 1.0% | 11 | 3.1% | 10 | 4.1% | 3 |
| 16 | BUF | 3.3% | 6-6 | 6.1% | 6.2 | 18 | -0.9% | 21 | -13.9% | 32 | 11.6% | 19 |
| TEAM | TOTAL DVOA |
W-L | NON-ADJ TOT VOA |
ESTIM. WINS |
RANK | PAST SCHED |
RANK | FUTURE SCHED |
RANK | VAR. | RANK | |
| 17 | SD | 3.0% | 5-7 | 2.3% | 5.4 | 21 | 0.9% | 14 | -4.4% | 22 | 4.5% | 5 |
| 18 | TEN | 0.7% | 6-6 | 8.8% | 6.4 | 15 | -6.8% | 31 | -5.8% | 24 | 10.9% | 17 |
| 19 | MIA | -0.4% | 7-5 | 3.6% | 6.8 | 13 | -3.4% | 28 | -3.2% | 19 | 22.8% | 31 |
| 20 | TB | -0.6% | 7-5 | -3.7% | 5.8 | 20 | 2.7% | 7 | 5.0% | 9 | 19.6% | 30 |
| 21 | CIN | -0.8% | 4-7-1 | -2.6% | 5.9 | 19 | 2.7% | 6 | -9.0% | 29 | 7.1% | 9 |
| 22 | ARI | -3.6% | 5-6-1 | 2.6% | 5.3 | 22 | -2.8% | 27 | 0.6% | 17 | 8.8% | 11 |
| 23 | CHI | -4.4% | 3-9 | -3.0% | 4.9 | 23 | -2.5% | 25 | 2.0% | 16 | 16.9% | 26 |
| 24 | CAR | -7.2% | 4-8 | -12.5% | 4.7 | 26 | 2.9% | 5 | 6.9% | 5 | 6.0% | 7 |
| 25 | IND | -8.4% | 6-6 | -6.6% | 4.9 | 24 | -3.6% | 29 | -5.4% | 23 | 17.2% | 28 |
| 26 | DET | -8.7% | 8-4 | -5.1% | 4.9 | 25 | -2.6% | 26 | 6.3% | 6 | 8.9% | 12 |
| 27 | JAC | -16.3% | 2-10 | -13.5% | 3.5 | 28 | 0.9% | 13 | -7.1% | 28 | 4.3% | 4 |
| 28 | LARM | -19.3% | 4-8 | -15.2% | 4.7 | 27 | -2.1% | 23 | 2.5% | 13 | 15.1% | 23 |
| 29 | SF | -22.3% | 1-11 | -24.2% | 3.3 | 30 | 2.2% | 9 | -4.0% | 20 | 10.3% | 15 |
| 30 | HOU | -24.5% | 6-6 | -22.6% | 3.2 | 31 | 2.3% | 8 | -6.2% | 27 | 7.1% | 8 |
| 31 | NYJ | -32.5% | 3-9 | -32.6% | 3.4 | 29 | 0.8% | 15 | 0.1% | 18 | 17.2% | 27 |
| 32 | CLE | -38.1% | 0-12 | -42.4% | 0.0 | 32 | 6.1% | 1 | 5.2% | 8 | 4.8% | 6 |
Comments
76 comments, Last at 13 Dec 2016, 6:34pm
#2 by Perfundle // Dec 06, 2016 - 6:41pm
I guess we'll finally find out how important Thomas is to Seattle's defense, though they seem to have borne up well the last two games. It feels like Chancellor's absence was felt more than Thomas, and Wagner being gone in 2014 hit them pretty hard too. Of course, all this depends on how good Thomas' backup Terrell actually is.
#19 by theslothook // Dec 06, 2016 - 7:39pm
Defenses can adjust to players missing. In the case of Seattle, they won't let terrell to do Earl Thomas things. That might cause strain on the rest of the defense now shifting responsibilities, but defenses are rarely at the mercy of a single player ever and with Bennett back, the pass rush should improve.
#20 by Ian Chapman // Dec 06, 2016 - 7:48pm
Not only that, but we now have a pretty good starting data set of what the Seattle defense looks like without Thomas, and it's pretty danged good. Yes Seattle gave up two early touchdowns in the Tampa game, but those were the *only* points they gave up. Normally when you hold an opponent to 14 points, you'd expect to win the game (I make this point because we are talking about defenses). 14 points is right in line with what you expect from "Legion of Boom: Classic". Likewise, we have nearly three quarters of the Carolina game where Thomas didn't play, and Cam repeatedly tried to test the defense long (had to because Carolina was in such a hole) with very little effect sans one big play right after Thomas was injured.
I guess the early data suggests that while Terrell isn't Earl Thomas (and to be fair who is?), the Seattle defense is very nearly as effective with Terrell as with Thomas. Honestly the data suggests to me that Chancellor is more critical to Seattle's defense than Earl Thomas.
-Ian
#23 by dmstorm22 // Dec 06, 2016 - 8:29pm
The Bucs really should have scored more. I think they turned it over twice in Seattle's half of the field, then had Aguayo miss a kick and clock management issues stop them from attempting another kick at the end of the half.
#53 by MilkmanDanimal // Dec 07, 2016 - 1:20pm
Yep. The Bucs outplayed Seattle by way more than the 14-5 score would indicate, but just did lots of stupid things to not put points on the board. Seattle was patently awful in every sense of the word in that game.
#54 by theslothook // Dec 07, 2016 - 1:20pm
A thought experiment.
Defense A: Out of 10 drives, they allow three drives for tds but no points otherwise and 7 punts.
Defense B: Out of 10 drives, they allow 6 potential scoring drives - 2 tds and 4 field goals, but 2 are missed.
Which defense did a better job/ is more predictive of quality going forward?
#68 by Bright Blue Shorts // Dec 08, 2016 - 5:36am
If you're going to make it that basic then it requires a basic answer
Defense A concedes 21pts
Defense B concedes 20pts
(Assuming no 2-point conversions)
So defense B is better in this game.
But given that league avg. for FGs is 80% (roughly 3 out of 4) they could usually expect to concede 23pts and therefore going forward Defense A is more likely to perform better.
#75 by theslothook // Dec 11, 2016 - 4:07am
Oh I don't know. I'm sure there is one. I think I personally in game would hate the tds given up, but I suspect a defense that forces more punts is a better predictor of future performance than a good D that does well in the red zone but gives up a ton of drives.
#1 by RickD // Dec 06, 2016 - 6:37pm
If there are no great teams, we should keep an eye on the Weighted DVOA, to see who is "hot" coming into the playoffs. I could make a case for any of 4 or 5 teams in the AFC and at least 4 in the NFC. The Earl Thomas injury is huge, as is the Gronk injury, for bringing the Pats and Seahawks back to the pack.
#3 by ibrosey // Dec 06, 2016 - 6:45pm
This paragraph is not intuitive:
"Even Seattle dropped a little bit this week; the Seahawks got a nice 45.4% rating for their big win over Carolina, but schedule adjustments drop them as a number of their past opponents (Rams, Dolphins, 49ers, Jets) got creamed this week."
Seattle was dinged for the way past opponents have played *after* playing the Seahawks? I'm sure this also means ratings are upgraded if past opponents outplay the level they were at when they played the Seahawks? What's the benefit of altering value by shifting baselines? I don't get it.
#8 by RoninX // Dec 06, 2016 - 7:01pm
As the season goes on we have a better picture of the true talent level of a team (injuries cause issues with this theory - but thats the gist). Generally I agree with it, but presumably some of the Seahawks victories continue to gain/hold value as well NE, ATL (good play in a tight loss). DVOA continues to think highly of the eagles. Normally I can follow the logic, but that much negative movement after a decisive win is a little hard to reconcile with past opponent adjustments.
#33 by Bobman // Dec 06, 2016 - 11:09pm
he simplest example might be Team A crushes Team B in week 1. Well, when you look at a 48-0 it's "obvious" that Team B sucks, so Team A might be great or merely good. The opponent adjustments don't really enhance Team A's score, they might even bring it down for a weak opponent.
Five weeks later and Team B has roared to life, soundly beating quality opponents every time. Team A has lost one because of losing three fumbles in the red zone, but performed well statistically. They have the same record, but Team A's win over Team B in week 1 looks better and better in hindsight. And Team B doesn't really suck--yes they pooped the bed in week 1, but they lost to a very good opponent and managed to perform well against other good opponents.
As others have said, we get a better picture of a team's true quality as time goes on, so a win a few weeks ago might mean more after more data comes in. In a weird quirk, I think it's ten weeks and the data fades in importance because it's ancient history. Yes, we DO know more after week 10, but the recent stuff is much more relevant.
#55 by Richie // Dec 07, 2016 - 2:19pm
We don't want to judge the quality of an opponent purely by what they did BEFORE a team played them. We need the whole season of information to evaluate that opponent.
For instance, San Francisco clobbered the Rams in Week 1. They were the #1 team in DVOA after week 1.
Then, in week 2, Carolina walloped the 49ers. If we leave the 49ers strength as static at week 2, the Panthers win will continue to look like a great victory. But that's not accurate. After 11 more weeks of information, we can now see that the 49ers are a terrible team, so the opponent strength for Carolina's win over the 49ers has gone down and down all season long.
#66 by Mo S // Dec 07, 2016 - 8:37pm
The flip side of this is that teams often change. They find a player or a scheme that they weren't using much in the beginning that worked and kept using it. Or they have a young team, new members of the o-line, free agents that take time to gel. Alternatively, they have a scheme or a glaring weakness that wasn't identified early that gets exploited later, but wasn't obvious week 1. So Team B in week 1, while wearing the same laundry and with much of the same personnel, is very different from Team B in week 8. I mean how many times have you heard, "Oh their offensive line was terrible in the first half of the season, but they've gelled in the last half of the season and look really good"?
#67 by Richie // Dec 08, 2016 - 1:54am
True. But all you can do is try to rate teams by what they have actually done. If you just try to throw out everything they did early in the season, the ratings won't have much meaning.
This is why FO doesn't even try to exclude obvious things like "Tom Brady didn't play the first 4 weeks" because when they test these things, it doesn't improve the predictability of the ratings any more. Game 1 Patriots is still more relevant to Game 12 Patriots than any other objective analysis.
#31 by RickD // Dec 06, 2016 - 10:42pm
What do you mean by fixed? I'm seeing this phenomenon now. And I opened the page in multiple browsers to make sure it's not a cache issue.
In any case, Weighted DVOA should certainly correlate with full season DVOA, but a perfect ordinal correspondence seems very unlikely.
#5 by RoninX // Dec 06, 2016 - 6:54pm
Seattle's total DVOA was 19.2 after their debacle with the Bucs, and this week their total DVOA went *down* to 18.4 after murdering the Panthers? Seattle's offensive DVOA also dropped after hanging 40 on a mid tier (#14) defense. They were ineffective in the red zone (As they always are) but took chunks between the 20s including several runs of 10+ and a handful of explosive passing plays, and DVOA usually rewards that more than most metrics. This seems odd to me.
Is there something I'm missing? DVOA opponent adjustments are always shifting, but its rare to see a team's rating regress after a decisive win.
#9 by Perfundle // Dec 06, 2016 - 7:05pm
Like the article said, most of their past opponents were horrible last week. Ignoring the NE/LA game, because they played each other, here is how they each did:
Miami: Lost 38-6 to Baltimore
SF: Lost 26-6 to Chicago
NY: Lost 41-10 to Indianapolis
Atlanta: Lost 29-28 to KC at home
Arizona: Won 31-23 against Washington
New Orleans: Lost 28-13 to Detroit at home
Buffalo: Lost 38-24 to Oakland
Philadelphia: Lost 32-14 to Cincinnati
TB: Won 28-21 at San Diego
#12 by Ian Chapman // Dec 06, 2016 - 7:20pm
I understand the explaination, but it still doesn't add up. Could the editors of F.O run the numbers again. Specifically, TB won and won against a good team, so Seattle should be punished less for that loss. New England won, and Seattle beat New England very recently. Philly laid a stinker, it's true, but Buffallo played a respectable game (albeit a losing one) against the Raiders who are turning out to be a very good team.
Looking at the list above, it seems that about half won and half lost. Now add in that Seattle completely dominated Carolina (which while not a terrific team, should still be rating at least at middling), and it seems like Seattle is getting shafted this week. So I am wondering if there wasn't a mistake somewhere in the calculations.
-Ian
#15 by Perfundle // Dec 06, 2016 - 7:26pm
New England won, but they beat LA, who Seattle also played, so it's basically no change there.
Basically, Miami, SF, NY, NO and Philadelphia had very bad to catastrophic losses, and the other games probably didn't really move the needle.
#18 by Ian Chapman // Dec 06, 2016 - 7:30pm
Seattle played New England more recently than they played LA, however. Not only that, but I am not seeing how the stomps you mentioned (some significantly in the past) could possibly outweigh a complete drubbing by at least a respectable (if not great) Carolina team just completed.
It doesn't make sense which is why I am wondering if the calculations were done correctly. It doesn't smell right.
-Ian
#24 by dmstorm22 // Dec 06, 2016 - 8:31pm
What does the recency have to do with it. This is normal DVOA, not weighted.
It's very simple. A lot of their previous opponents look worse now than they were a week ago, so Seattle's DVOA in those games is worse now than it was a week ago.
#26 by Ian Chapman // Dec 06, 2016 - 8:40pm
First of all, Seattle also seems to be getting shafted on the weighted DVOA and recency DOES apply there. Secondly, that might be true, but that doesn't mean it's a good measure. In fact I find it very suspicious that Seattle can do so well (per Aaron) this last game and still go down. I find it very difficult to see how this is possible which is why I am wondering if this was done correctly.
-Ian
#13 by RoninX // Dec 06, 2016 - 7:22pm
Nine performances, and I'd call 3 good to fair: TB, AZ, and ATL (yes an L). One expected: BUF. But you're right the rest were just horrifically horrible. Also, SF and Rams probably have a net negative effect across all NCFW teams pulling down not just the value of games against them but the Cards vs. Seahawks game as well as a by product.
#17 by Jimmy Oz // Dec 06, 2016 - 7:28pm
It's a very good question.
I would have thought it to be opponent adjustments, as most of Seattle's prior opponents lost their games this week, and most of the teams that played Seattle's prior opponents racked up more points than when Seattle played them.
Seattle's Schedule - Opponent's week 13 game - *
Week 1 Dolphins - L 6-38 v Ravens *
Week 2 Rams - L 10-26 v Patriots
Week 3 49ers - L 6-26 v Bears
Week 4 Jets - L 10-41 v Colts *
Week 6 Falcons - L 28-29 v Chiefs *
Week 7 Cardinals - W 31-23 v Washington *
Week 8 Saints - L 13-28 v Lions *
Week 9 Bills - L 24-38 v Raiders *
Week 10 Patriots W 26-10 v Rams *
Week 11 Eagles L 14-32 v Bengals *
Week 12 Buccaneers W 28-21 v Chargers *
In games marked with *, Seattle scored less against these opponents than their week 13 opponents. Only 3 of the Seahawks' previous opponents won their week 13 game, and only 2 scored less points than when Seattle played them.
I would have thought that this would've shown in 'NON-ADJUSTED TOTAL DVOA', however that dropped from 18.2% in week 12 to 10.9% in week 13, despite the big win. So yeah, I've got no idea what's going on either.
#21 by rquamme // Dec 06, 2016 - 7:55pm
This is the unexplainable point from what I see. Aaron said that Seattle had a nice 45.4% rating for their game this week. This would indicate to me that the NON-ADJUSTED TOTAL VOA dropping from 18.2% to 10.9% would be impossible, unless there's something I'm missing about how raw VOA works.
#22 by rquamme // Dec 06, 2016 - 7:59pm
I think the saddest part of DVOA that I've been watching this year is that the Browns still sit at 0.0 estimated wins. One has to think that at some point they'll get that first tenth, but I'm slowly losing hope as the season goes along.
Even Detroit in 2008 had 2.1 estimated wins at the end of the year, but the Browns do not appear to be receiving any such signs of bad luck.
#28 by JoeyHarringtonsPiano // Dec 06, 2016 - 9:04pm
The other crazy thing is that Detroit has victories over the 5th, 10th, 14th, and 15th ranked teams. Yet they have losses to #23 Chicago and # 30 Houston that weren't really as close as the final score indicated.
#34 by herewegobrowni… // Dec 06, 2016 - 11:14pm
The 2012 Colts were right in that ballpark of 8-4ish/26th-ish in DVOA.
Regarding the Lions being below daaaa Bearz, I have to think Chicago's dominating win over SF this weekend (6 yards for Kaep!,) opponent adjustments or not, helped quite a lot.
#43 by MilkmanDanimal // Dec 07, 2016 - 9:54am
The Bears also beat the Vikings, who are clearly dropping like a stone but haven't hit bottom yet. That win will look worse as Minnesota continues to get Sam Bradford killed. The Bears also had reasonably close losses against a number of decent teams, and really aren't getting blown out at all.
They're bad, but they're not Cleveland bad. A few random bounces and the Bears easily have a couple more wins.
#49 by BJR // Dec 07, 2016 - 11:48am
The Bears fans round here may be able to confirm, but my impression is they aren't a 'bad' team. They may well end up in the very bottom reaches of the DVOA table, but that is only due to horrendous injury luck. Otherwise they'd just have been an average/slightly below average team.
In fact the last two Bears scorelines look highly creditable given the circumstances - playing the Titans all the way to the final play, and blowing out the (albeit terrible) 49ers. I suppose that reflects well on John Fox and his staff.
#52 by Thomas_beardown // Dec 07, 2016 - 1:15pm
The Bears are bad team, but they aren't awful. They're a 6-10 quality team that had bad luck.
Of course that isn't counting injury luck. With normal number of missed games, this team is probably 8-8. That's what I was expecting coming into this year and hoping for enough good luck to get 9-7 or 10-6. As it is, I'm now hoping my original assessment (8-8) is right and a juicy high draft pick pushes them into playoff contention.
I also really hope they don't stupidly fire John Fox.
#63 by TomC // Dec 07, 2016 - 5:45pm
Yeah, the Bears are a mediocre-talent-level team that's been devastated with injuries and hasn't won any close games. They don't deserve to be 3-9, but they don't deserve to be in the playoffs, either.
That said, I can't figure out why they're ahead of Detroit in DVOA. The Lions have played a very similar schedule to the Bears and have a +24 point differential, compared to the Bears' -66. This has to be one of those situations where DVOA "likes" the way the Bears are losing better than it "likes" the way the Lions are winning. And it's the Lions that are the real outlier. Usually point differential + strength of schedule is a decent proxy for DVOA, but by that metric, DET should be somewhere in the GB-MIN-BUF-SD-ARI pack, and definitely not below the Bears.
#64 by JoeyHarringtonsPiano // Dec 07, 2016 - 6:38pm
In one of his ESPN Playoff Odds articles, Aaron said that Detroit's rating is being dragged down primarily by three terrible games: Week 3, when they fell behind 31-3 in Green Bay, Week 4, when they played awful and lost in Chicago (Stafford's only really bad game of the season), and week 6, when they made Case Keenum look like Drew Brees and nearly lost at home to the Rams.
As those games fall further in the rearview mirror, maybe we'll see a weighted DVOA rating at least somewhat closer to the W-L record.
#70 by TomC // Dec 08, 2016 - 1:26pm
Thanks, JHP. But I think you can still ask *why* DVOA rates those games so terribly. Yes, they fell behind 31-3 to GB, but they came all the way back and only lost 34-27. A 7-point road loss to a good team shouldn't be a DVOA disaster. The Bears game was a 3-point road loss (though I know it wasn't as close as the score indicated). The Rams game was a win! How bad can a win be? If only barely beating the Rams gives you terrible DVOA, what about Seattle, TB, and Arizona, who actually lost to them? Finally, shouldn't making Keenum look like Brees at Ford Field be perfectly canceled by making Brees look like Keenum in the Superdome?
#27 by BlueStarDude // Dec 06, 2016 - 9:00pm
Would the Giants 1.0% variance be some kind of record?
How does that stack up against other teams from the past after 12 games?
I looked at a dozen or more years (final standings) and don't see anything that low.
In most years (all?) Dallas’s 3.7% would be number 1.
#29 by Raiderjoe // Dec 06, 2016 - 9:06pm
No way shodul Eags be ranked ahead of Raiders. Eags don't even have a winning record. If Raiders played Eags score wlidl be like 27-10
Falca ahead of Raiders. Fine. Even though Raiders are better, atleast Falcons bbeat Raiders.Pates, Cowbpys, seahawks ranked ahead
Fone
Can understabd even thougj Raiders really pepbably better than all three
#35 by DezBailey // Dec 06, 2016 - 11:20pm
Week 13 BES Rankings came out earlier today - http://besreport.com/week-13-bes-rankings-2016/
Oddly DVOA clings to the Eagles in the top-5 despite their 5-7 record. The BES has them ranked 26th after losing to the Bengals whom DVOA has ranked 21st. And where's the love for the Buccaneers?...three straight wins...beat KC, Seattle and San Diego. Perhaps the BES showed them too much love...ranking their defense No. 1 this week.
#76 by DezBailey // Dec 13, 2016 - 6:34pm
Week 14 BES Rankings were published this morning - http://besreport.com/week-14-bes-rankings-2016/
Patriots are now #1 in BES Overall with the Steelers making a big leap from #12 to #3. Bucs have also cracked the top-10, landing at #8.
#37 by Aaron Schatz // Dec 07, 2016 - 1:12am
My apologies to everyone. I had my spreadsheet sorted incorrectly when I created the tables for the DVOA article today, and then I went and disappeared for six hours.
The tables have now been fixed above. This will answer most questions posted in comments prior to Wednesday morning, including "Why are teams in the same order for both DVOA and weighted DVOA?" (They aren't); "Why doesn't weighted DVOA match what's listed on the playoff odds page?" (It should now); and "Why did Seattle's non-adjusted VOA go way down after the game with Carolina?" (It didn't, that was Atlanta's non-adjusted VOA listed as Seattle by mistake).
#46 by LyleNM // Dec 07, 2016 - 11:29am
Thanks for that, Aaron. I think the most puzzling question remaining (which I know you tried to answer in the text) is how Seattle went from 19.2% last week to 18.4% this week while putting up a 45.4% effort. The sniff test isn't sure what to think.
#47 by RoninX // Dec 07, 2016 - 11:38am
Yeah I agree that the numbers on that are still simply hard to reconcile with what we saw on Sunday - I guess I'm left with a new respect for the power that incremental opponent adjustments play in week to week DVOA ratings.
#48 by dmstorm22 // Dec 07, 2016 - 11:40am
Here's a thought experiment, simplifying the math far more than waht actually goes on admittedly.
We have 11 data points which tells us Seattle is 19.2% better than average. Assuming no changes, if we then get a 12th data point where they are 45.2% better than average, simply weighting that in, Seattle's overall rating would be 21.4% - again there is much more involved in what really happens.
Now, if we get many other factors showing that Seattle wasn't actually 19.2% better than average over those 11 data points, that they were worse, then even when factoring in the 45.4% performance in the last game, their overall rating goes down.
Backing out the simple calculation, if Seattle really was 15.9% better than average in their previous games based on the new information (their previous opponents were worse than we thought), than when you factor in the 45.4%, their overall rating still goes down.
Now, I imagine there is a lot more going on, and even then I would not have guessed one-weeks worth of new information on past opponents would have THAT much of impact, but maybe it can. Some of Seattle's old opponents did have really bad games (Jets, Dolphins particularly).
#38 by Bright Blue Shorts // Dec 07, 2016 - 2:57am
The lack of a great team reminds me that unheralded Leicester City won the English Premier League last season and have now headed back to obscurity. I'm wondering if there is a team waiting to get hot and take the Lombardi home.
#42 by MilkmanDanimal // Dec 07, 2016 - 9:47am
Allen Robinson is absolutely capable of taking over a game and putting up huge stats. Admittedly, it's only going to happen when the Jags are down 28 in the 4th quarter, but, let's face it, the odds of that happening seem pretty good over the next four weeks.
#40 by Cheesehead_Canuck // Dec 07, 2016 - 9:21am
Seattle and Pittsburgh seem to be the most balanced team. The Steelers D is ranked higher than I would have expected.
The best team the Patriots has beaten with their starting QB in the game is..... the 16th place Buffalo Bills. Ravens and Broncos are good teams with weak offenses.. they may not play a game against a team with a good offense until January now. Scheduling quirks!
#41 by dmstorm22 // Dec 07, 2016 - 9:40am
They had a similar stretch in 2011 when they played a string of crap QBs starting with a win over the Jets. I think they were 5-3 heading into that Jets game, coming off of back-to-back losses to the Steelers (Roethlisberger) and Giants (Manning).
Starting with the Jets game, they went 8-0, beating QBs like Sanchez, Tyler Palko, Tebow, Vince Young, etc.
#69 by Will Allen // Dec 08, 2016 - 12:26pm
More than usual, this year seems to me to be one where injury profile in January will have huge influence on who plays in February. When nobody is great, everybody is just a random event away from being mediocre. Even Seattle, which is defense dependent, and theoretically not quite as exposed to a single player injury disaster, would be likely wiped out if Wilson just had his mobility greatly hampered. I guess Dallas is the one team that might actually improve if the starting qb was hurt, so I'm inclined to view them as the favorite, but I have no strong confidence in any projections this year, compared to years past.