Week 9 DVOA Ratings
The New Orleans Saints' Sunday night slaughter of the Tampa Bay Buccaneers was big. Really big. Probably even bigger than you think it was. It was so big that the Saints jump up six spots to reign as the new No. 1 in Football Outsiders' DVOA ratings.
According to DVOA, the Saints' victory on Sunday night was the best single game played by any team since I launched Football Outsiders in 2003. After opponent adjustments, it's the best game of the century. It's the fifth-best single game ever measured by DVOA going back to 1985.
|Best Single-Game DVOA, 1985-2020|
Those ratings are based on end-of-season opponent adjustments, by the way, and eventually the Saints' rating for Sunday night will be based on end-of-season adjustments as well. That might knock it down a couple pegs if Tampa Bay doesn't play as well in the second half of the 2020 season, but the point stands: this was a massive, ultra-dominant victory. The Saints' total DVOA went up over 11 percentage points from 20.4% to 31.7%.
Opponent adjustments mean that the loss wasn't as bad for Tampa Bay as it was good for New Orleans, but it was really bad. Tampa Bay's total DVOA went down from 39.3% to 27.6% in one week. As noted in this week's Any Given Sunday column, this was the biggest drop ever for a No. 1 DVOA team in Week 8 or later. I searched through all the historical data, and Tampa Bay's single-game DVOA of -109.7% looks like it was the worst ever put up by any team coming off a week at No. 1. You can probably ignore the Week 17 games on this list, which featured teams that had clinched their playoff seeds and were resting starters.
|Worst Single-Game DVOA by No. 1 Team, 1985-2020|
You might be surprised to see that Tampa Bay, despite this horrendous performance on Sunday night, is still No. 2 in the DVOA ratings. The Buccaneers had a long, long way to drop from two weeks ago when they were No. 1 with 45.0% DVOA and over 12 percentage points separating them from the rest of the league. The DVOA system is still impressed by the rest of their performance this year, especially the big win over Green Bay. No matter how badly the Buccaneers got destroyed on Sunday night, that Green Bay win still happened. And yet, even if we take away both the Green Bay win and the New Orleans loss, the Bucs average 31.4% DVOA in their other seven games. Our system really likes this team this year. I don't know why we have Tampa Bay higher than other advanced metrics on the Internet, but last time I was worried about this same thing, Tampa went out and stomped all over Green Bay and made DVOA look pretty good. So, this is just how it is. Tampa Bay remains No. 1 in defense this week because Pittsburgh's defense declined a little bit as well. On offense, Tampa Bay drops from sixth to 13th.
After New Orleans and Tampa Bay comes Kansas City at No. 3, up one spot from last week. If you asked me subjectively which team I thought was the best in the NFL right now, I would say Kansas City, and I know that other advanced metrics do have them at No. 1. DVOA has them a bit lower because their defense and special teams come out close to average. Pittsburgh drops two spots to No. 4 with a close win over Dallas.
Then we have Baltimore, dropping from third to fifth, and Indianapolis, dropping from fifth to sixth, after a game between the two teams that was closer than the final score. Indianapolis gained 5.2 yards per play to only 4.2 yards per play for Baltimore, although some of that was on a final drive to nowhere where the Colts went 63 yards in the final two minutes but couldn't score and it wouldn't have mattered anyway since they were down by two touchdowns. Getting rid of drives like this or counting them differently is definitely on the to do list for the next iteration of DVOA.
Some of the big movement on the DVOA table was a lot bigger in terms of percentage points than it was in terms of ranking. Seattle drops from 22.8% to 16.3% after losing to Buffalo, but that's only two spots in the rankings to eighth. Buffalo moves up from 2.1% to 6.9%, but that only takes the Bills from 14th to 13th.
Another thing you'll notice if you look at the table below is that it's a bit different from years past because usually we introduce Weighted DVOA after Week 9. New in 2020, however, we're continuing our DAVE metric later into the year, incorporating our preseason projections all the way until 13 games have been played. Right now, you'll still find DAVE below instead of Weighted DVOA. However, if you want to see both DAVE and Weighted DVOA, you can check out the DVOA database in FO+.
There aren't a lot of teams with a big difference between full-season DVOA and weighted DVOA at this point, although a few stand out. Here are the seven teams where there's a difference of more than 2.0% between full-season DVOA and weighted DVOA after Week 9:
|Biggest Gaps Between Total and Weighted DVOA|
Where is the improvement coming from for Miami, and the decline for New England and Jacksonville? Since we've reached the halfway point of the season, I figured this was a good time to see how teams have changed over the year so far. For each team, I've run their offensive and defensive DVOA for the first and second quarters of the season. For most teams, that comparison is Weeks 1-4 vs. Weeks 5-9. For the two teams that had byes in Week 4, Pittsburgh and Tennessee, the split is Weeks 1-5 vs. Weeks 6-9.
It would be easier if I could just call this "September" and "October," but of course the first quarter of the season ended in the first week of October and the second quarter went two weekends into November, so never mind that.
Let's start with the team that stands out on both sides of the ball: Miami. On offense, the Dolphins have improved from 25th in the first four games to 10th in their last four games. However, that's not because it is Tua Time. The Weeks 5-9 offense for the Dolphins breaks down to 20.4% with Ryan Fitzpatrick at quarterback (Weeks 5-6) and -3.1% with Tua Tagovailoa at quarterback (Weeks 8-9). And Miami's defense has improved even more than Miami's offense. The Miami defense was dead last in DVOA for the first four weeks and ranks seventh in DVOA since.
The other team that stands out on both sides of the ball, and not in a good way, is New England. On offense, New England ranked 19th in the first four games but has dropped to 29th in the four games since. On defense, New England ranked 17th in the first four games but has dropped to dead last in the four games since.
The biggest offensive improvement between the first and second quarters of the season belongs to the Arizona Cardinals, who went from 24th on offense in Weeks 1-4 to third on offense in Weeks 5-9. Since Week 5, the only two offenses that have been better by DVOA are Tennessee and Kansas City. Other offenses that have jumped in the second quarter of the year are the Chargers, from 21st to eighth; the Giants, who have gone from 31st to 21st; and the Eagles, who have gone from dead last to 22nd.
One offense has gone very strongly in the other direction. Remember Week 1 when Jacksonville upset Indianapolis and Gardner Minshew was 19-of-20 with three touchdowns and no interceptions? That was clearly Jacksonville's best game of the year but the Jaguars still had positive offensive DVOA the next four weeks. Then things took a downturn, although the Jaguars bounced back a little bit above zero with Jake Luton on Sunday. So in total, the Jaguars went from fourth in offensive DVOA for Weeks 1-4 to 27th in Weeks 5-9. That's where the big Jacksonville drop comes from in the table above, since their defense has consistently been near the bottom of the league all year.
Other offenses that have fallen a bit in the last few weeks include Tampa Bay (from seventh to 18th), Cleveland (from 12th to 23rd), and of course Dallas without Dak Prescott (from 16th to dead last).
However, just to demonstrate how fickle these four-week splits are, I would have mentioned Buffalo's offense if I had written this last week. The Bills offense went from 19.5% DVOA in Weeks 1-4 (sixth) to -5.2% DVOA in Weeks 5-8 (21st). Then the Bills bounced back to 25.9% DVOA this week against Seattle, even after adjusting for the Seahawks' poor defense.
On to defense. I already mentioned the Miami Dolphins, who have had the most impressive defensive improvement in the last few weeks. Other defenses that have improved significantly are Cleveland (from 26th to 15th) and the Los Angeles Rams (from 19th to eighth).
Defenses that have gone the other way, besides New England, start with the injury-riddled San Francisco 49ers, who ranked fourth in defense in the first four games but have dropped to 23rd since Week 5. The other defenses that have gotten worse in the second part of the season belong to the two New York teams, with the Giants going from 12th to 28th and the Jets going from 11th to 30th.
Speaking of the New York Jets and their odds of going winless this season, here's an update on all the quirky odds I started tracking last week. However, with San Francisco, Arizona and Seattle all losing non-division games, I'm going to retire listing the NFC West props after this week.
- New York Jets go 0-16: 11.7% (up from 8.8%)
- Pittsburgh goes 16-0: 5.0% (barely up from 4.9%; their win over Dallas was highly probable according to last week's simulation)
- Entire NFC West makes the playoffs: 0.6% (down from 2.4%)
- Entire NFC West has a winning record: 6.3% (down from 17.2%)
- Entire NFC East has a losing record: 68.8% (up from 62.2%)
- Entire NFC East is 6-9-1 or worse: 31.5% (up from 26.2%)
- Entire NFC East is 5-10-1 or worse: 4.6% (up from 3.7%)
* * * * *
Football Outsiders playoff odds, snap counts, and the FO+ database are now all updated through Week 9.
* * * * *
These are the Football Outsiders team efficiency ratings through nine weeks of 2020, 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.
Because it is early in the season, opponent adjustments are only at 90% strength; they will finally reach full strength next week.
DAVE is a formula which combines our preseason forecast with current DVOA to get a more accurate projection of how a team will play the rest of the season. DAVE is currently 36% preseason forecast for teams with nine games played and 45% preseason forecast for teams with eight games played.
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>
- 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).
57 comments, Last at 13 Nov 2020, 12:59pm
#1 by Bobman // Nov 10, 2020 - 6:15pm
Kind of amazing Indy did not fall more offensively last week, but then again Balt's D is pretty good.
If you look at them, they're VERY similar and close by DVOA--offense is 0.3% apart and the Colts D is 3.1% ahead. But on ST... where Indy isn't half bad, Balt takes a substantial lead. I think in reality there is more space between them than this indicates.
#13 by jheidelberg // Nov 11, 2020 - 1:24am
Aaron, you mentioned that subjectively you feel that the Chiefs are the best team in the NFL. Would your own DVOA formula in some ways state this, in that the Chiefs have a low variance, and thus, are consistently playing at an excellent level/very good level being third in overall DVOA.
Are the following true?
1. In the playoffs, against the good teams, due to consistency of excellent/very good play, the Chiefs would be the most likely to run the table and win the Super Bowl. I would assume that the answer is yes because the top two teams are more likely to throw in a clunker game in the playoffs due to their high variance.
2 .In the regular season playing against the average teams, the Chiefs are very hard to beat because they are so likely to be above 20% DVOA in any particular game.
Does high variance predict future high variance? In other words, are inconsistent teams likely to continue to be inconsistent, while consistent teams are likely to stay consistent?
#24 by Joseph // Nov 11, 2020 - 9:42am
If you look at the DVOA table from last week, the Saints were at #2 in variance. Their change to #29 is ALL a result of the game Sunday night. Their variance went from "we play above average on offense and ST and slightly above average on D" to "when we take it up a notch, we play like the best team in the NFL like a lot of analysts predicted before the season started."
Now, Michael Thomas returning was always going to help the offense. Sanders off the COVID list after 2 weeks helped too. But somebody lit a fire under their defense this past week. The following weeks will tell us if the Saints flipped a switch, or if it was "getting up for a big game on SNF against a division rival that's trying to take our crown."
#3 by Raiderjoe // Nov 10, 2020 - 8:16pm
No way Raiders shoudl be 22nd. They just beat chargers. Rhat should coint for somethinf becasue eith any luck Chagers would be 6-2. So you cant really say they are bad team. Also,beat Chiefs.
Maybe Saints, zbuccaneers, Steelers, Chiefs, Nills, Seahawks, cardinLs, ravens, titans, and Raiders are top 10 teams. Thoigjt Clots coild play Raiders in AfC title game before seaeon started but can now see Totans as beinf Raiders' victims. Steelers could barely beat rhe cowboys. Ravens not good wjen trailng. Raiders get leaf on them and will win. Chifes tough game but Raiders have explosive pffense to match. Bills mice up and.comimger team but Raiders can beat them for sure. Nfc teams? Raiders would beat wjoever
#25 by DGL // Nov 11, 2020 - 9:43am
Raiders ar clearlt ranked too loe because just beat chargers and rhat should coint for somethinf. Pickinf top 10 teams by eyebal is way better than rhis. Thoigjt Clots coild play Raiders in AfC title game before seaeon started but can now see Totans as beinf Raiders' victims. Nfc teams? Raiders would beat wjoever
#31 by p_cj // Nov 11, 2020 - 11:47am
What is the point of advanced statistics if they aren't predictive? Tampa was the best team "by a mile", shouldn't that lead to some sort of reliable outcome?
I know single games are hard to predict, but it seems that the Tampa-Green Bay game was more of an anomaly than this one (since the outcome of the first matchup between the teams ended in a similar fashion).
#44 by Ryan // Nov 11, 2020 - 6:07pm
One tangible benefit could be team management using these metrics to review team/unit performance from the year before to identify specific strengths and weaknesses--personally groupings, routes run, etc.--and map that to player management.
One example: after Chris Ballard signed Devin Funchess, someone pointed out either on this website or twitter or something that his strengths in DVOA by route (I believe curls and outs in particular) matched the routes called most often by Frank Reich's outside receivers. Of course, Funchess broke his collarbone a few minutes into week 1, so we'll never know how it would have worked out, but have to believe advanced stats influenced the acquisition.
#5 by p_cj // Nov 10, 2020 - 8:43pm
I'm curious what the Pittsburgh-Baltimore game in 2007 looks like in DVOA.
With about 2 minutes left in the first half, Ben had thrown for 5TDs and the Steelers were up 35-0. Ben sat the rest of the game (IIRC) and Baltimore ended up with a total of about 100 yards of offense with 4 turnovers and 5 sacks given up.
Granted, as a Steelers fan, I watch mostly on Pittsburgh games and it wasn't a huge offensive output yardage-wise for Pittsburgh, but that was about as one-sided a game as I have ever seen. Of course, Balt wasn't very good that year.
#19 by p_cj // Nov 11, 2020 - 8:51am
"All thrown into the end zone" Yes, that is typically where they happen.
In any case, not really getting your point. Harrison had a memorable game, but football is a team sport. DVOA measure team efficiency, yes?
#32 by Israel P. - Ashkelon // Nov 11, 2020 - 12:34pm
All thrown into the endzone = no yards after catch. This wasn't receivers running with the ball. Just adds to Ben's accomplishment.
My point was that no matter how great a game the offense had, the game will always be remembered as the James Harrison Game.
Sorry if I was unclear.
#8 by dank067 // Nov 10, 2020 - 10:30pm
Well, Drew Brees is up to 3rd in DYAR and 4th in DVOA. Even if you leave out Weeks 1-2, his yards/attempt, air yards, and some other efficiency metrics do still appear to be down from where he's been the last several years, so he's not necessarily "back." But he and that offense are still highly effective and with a healthy Thomas they may even have room to improve as the season goes on.
#48 by jheidelberg // Nov 12, 2020 - 12:53am
The odds of making a 59 yard field goal are slim, so despite missing two extra points, the fact that he made this kick, in addition to another kick of over 40 yards without missing a FG in the game, would lead me to believe that his DVOA was positive and thus offset some of the disaster of which you are speaking.
#18 by Aaron Brooks G… // Nov 11, 2020 - 8:29am
But not the fewest wins!
(Washington and NYG played a tiebreaker game between 6-3-1 teams) in 1943.
I'd also like to tip my hat to the 1943 Lions, who played in two of the four lowest combined offense games in NFL history, in the same month, including their NFL record -53 rushing yards against the Cardinals on October 17, 1943. A game they won. They went 1-0-1 in those games, holding each of the Cardinals and the Giants to 84 total yards.
#22 by Aaron Brooks G… // Nov 11, 2020 - 9:27am
It was the Lions. That's usually explanation enough.
(They've been involved in three of the four lowest offense games in NFL history)
The game, oddly, was played in Buffalo. I guess those people hadn't suffered enough.
There's a Chicago Tribune article about the game that might shed some light, but I don't have access to it.
\Detroit's 0-0 tie with NYG in the *other* low-offense game was the NFL's last 0-0 tie.
#33 by ZosoLZ4 // Nov 11, 2020 - 1:03pm
In 1943 the NFL still lumped sack yardage in with rushing, like the NCAA currently does. That would be my guess for a good portion of the negative yardage. Anyway, here's the boxscore:
#23 by Aaron Brooks G… // Nov 11, 2020 - 9:33am
Tennessee is more efficient at both rushing and passing than Green Bay.
(hidden variable) Both Green Bay and Tennessee are more efficient at passing than running.
(hidden variable) Green Bay passes more frequently than Tennessee does.
Green Bay has higher net efficiency.
#28 by Aaron Brooks G… // Nov 11, 2020 - 10:32am
Educated guess, anyway.
Tennessee has a 255/246 pass/rush balance and GB has a 280/222 balance. Given how close their component DVOAs are, that's probably enough of a split to account for the difference. A straight naive multiple on that breakdown does yield GB as just slightly more efficient overall.
#29 by andrew // Nov 11, 2020 - 10:51am
What are we to make of the fact that the top 3 games all occurred in week 1?
The conclusion I would make is that the opponent improved, a lot, the rest of the year and by virtue of being week 1, they had a lot of time to do so.
I remember that '91 Lions team, they were absolutely steamrolled by Washington (all time dvoa champs iirc) week one, then won some close games vs very bad to mediocre teams, had a midseason swoon where they lost 3 of 4 then caught fire after the injury to Mike Utley. It looks like they only had 3 wins over teams with winning records in the regular season, Dallas, Chicago (they split) and Buffalo in overtime the finale where the Bills had clinched homefield in the AFC and rested their starters. THey then got Dallas minus Aikman for their last playoff win, and got the Redskins again. To their credit it was closer than the first matchup. They may not have been as good as their record made it seem, but they were a lot better than they looked week 1.
The '89 Steelers looked horrible in starting 0-2, being outscored 92-10. They won 5 of their last 6 to finish 9-7, made the playoffs, won their first game and only lost by one point in the second round to the AFC champs (the Broncos team that got used as a punching bag by the Niners).
This theory doesn't seem to hold up so much for the first game, though, those Browns (the inaugral season of the new browns) were really, really bad, 2-14, with one of their wins being the rematch wth Pittsburgh, in Pittsburgh (those Steelers were one of the worst of the Cowher era, 6-10). I think that is more just a case of that Browns team being really, really, really bad. They got 2 first downs for the game. They ran 9 times for 9 yards. Ty Detmer was 6 of 13 for 52 yards with a pick, and Tim Couch threw another 3 attempts with 0 completions and another pick. They were sacked for 21 yards, leaving them with 40 yards of offense for the game. They also lost two fumbles,while the Steelers ran 57 times for 217 yards between 4 running backs and Kordell Stewart. They also used a second qb, Mike Tomczak who was 8 for 8 for 78 yards and 2 TDS in mop up work. The Browns time of possession was 12 minutes. Their first drive was 4 plays and ended in a pick after 14 yards. The second drive was 5 plays, 29 yards followed by a punt. Their next 9 drives totaled 2 yards. The steelers meanwhile had their first drive end on a failed 4th down attempt. They scored their next 8 drives, 5 tds, 3 fgs. If we have any articles about this game I'd love to read more on it, but the stats are pretty horrific.
#30 by Aaron Brooks G… // Nov 11, 2020 - 11:44am
THey then got Dallas minus Aikman for their last playoff win, and got the Redskins again.
Aikman started that playoff game. He was pulled for Beuerlein, but he did play.
\Dallas was better with Beuerlein that year, anyway
#34 by Richie // Nov 11, 2020 - 1:23pm
"The Browns time of possession was 12 minutes. "
Ouch. That ends up being the lowest ToP since 1983. The 7 games with sub-15 ToP's since 1983:
1. 12:11, Cle-Pit 1999
2. 13:17, Phi-NYG 1983
3. 14:28, Mia-Chi 1988
4. 14:45, Oak-Atl 2008
5. 14:46, GB-Den 1999 (one month after the Cle-Pit game)
6. 14:53, Ind-Mia 2009 (and the Colts won, 27-23! Manning averaged 13 yards per attempt. Dallas Clark had 183 yards.)
7. 14:56, Det-SF 1991
#43 by NYChem // Nov 11, 2020 - 4:59pm
But what was the record pre-merger? I mean, I guess it's possible, but it seems almost impossible for a team to have the ball much less than 20% of the time, unless (or even if) its a Georgia Tech/Cumberland 222-0 type game.
#36 by MC2 // Nov 11, 2020 - 1:45pm
Please reduce the amount of clutter at the top of the screen, or at least make it possible to scroll past it.
As it stands, it's now impossible to view an entire DVOA table at once, i.e. without scrolling (unless you reduce the font to an eye-strainingly minuscule size).
#41 by Eddo // Nov 11, 2020 - 3:27pm
Co-sign. The header is huge and is anchored to the top of the page. Reducing its size as you scroll, or allowing it to be pushed off the screen, would be an improvement.
That said, even without it, the table doesn't fit onto one screen. I think that's just something you'll have to live with.
#37 by Steve in WI // Nov 11, 2020 - 1:58pm
Bears 28th in offensive DVOA, 4th in defensive DVOA. If there was a metric that measured number of years squandering a really good unit by not being able to field even a mediocre unit on the other side of the ball, the Bears would be the undisputed champions of that metric, right? I can't think of another team that has been so far from a championship for decades at a time while regularly having excellence on one side.
#38 by Aaron Brooks G… // Nov 11, 2020 - 2:31pm
Baltimore. Twice in 25 years have they had a good offense that was better than their defense.
Now, there's the Jekyll&Hyde version, the Saints, who are doomed to never being good at both simultaneously.
But there's really nothing like the Bears, who have basically played for a championship every season in the last 100 in which they had a top offense, except 1947, 1955, and the weird year of 1965 in which they won big but lost close and finished 3rd in their division.
#39 by LyleNM // Nov 11, 2020 - 2:49pm
Hoping that FO can get around to 1979 (specifically Nov. 4th) when the Seahawks were held by the Rams to one (1) first down and negative seven (-7) yards of total offense. Final score was only 24-0 so it may not end up as one of the best/worst total DVOA games but might be the worst offensive performance since the merger at least when FO gets there. Also the Seahawks were not really a terrible offense, just terrible for this one game.
#40 by Aaron Brooks G… // Nov 11, 2020 - 3:25pm
The Lions had 16 total yards, and 5 TOs.
However, that was still more yards than their opponents, who generated 14 total yards. Apparently the game was played in a monsoon and the resulting quagmire, and at one point eight successive plays were all punts.
#46 by jheidelberg // Nov 11, 2020 - 6:12pm
I looked up this game and the Rams had 475 yards for an incredible 475 vs -7 yards differential.
Although not one-sided another classic is the game between Buffalo and the NYJ on Sept 29, 1974. On an incredibly windy day in Buffalo, the Bills won 16-12 as Joe Ferguson went 0-2 passing and Joe Namath was 2-18, 33 yards and 3 INT's. Yes, there really were more passes caught by the defense than the offense in this game.
I remember watching this game at 15 years old as a Jets fan (I have recovered from this affliction, and have lived a healthy, happy adult life). Amazingly the Jets could not figure out that completing a pass was virtually impossible and threw 18 passes in this game.
Jets fans, its just another day in Jets history.
#51 by E // Nov 12, 2020 - 11:19am
The four worst teams by DVOA in the NFC are the 4 NFC East teams. The four worst teams by DAVE in the NFC are also the 4 NFC East teams. To say that differently - by both metrics, every other team in the NFC is better than every NFC East team. This is truly unbelievable. I can't imagine this has ever been true for any division in either conference in DVOA history. And we're halfway through the season so this isn't a weird data fluke or very likely to change. Wow.
#52 by Dan // Nov 12, 2020 - 4:08pm
The probability of this happening by chance with the current 32-team setup is 1/455 (16/16 * 3/15 * 2/14 * 1/13). (This is ignoring
The NFL only switched to 4-team divisions and 16-team conferences in 2002, so this is the 19th year. There are 2 conferences, which brings it up to 38 chances to have this pattern. Which is about a 8% chance of it ever happening.
But this is happening midseason, not at the end of the year. If you check every week then that's a lot more chances (38*17=646), but they're not statistically independent. So, going by intuition rather than calculation, that maybe triples the chance of it happening at some point during the season.
Pre-2002 there were fewer teams and more teams per division. With 5-teams per division, is the equivalent question whether the worst 4 teams by DVOA in a conference were all in the same 5-team division, or is it the worst 5 teams? For a 15-team conference with three 5-team divisions, the probability is 1/91 for the worst 4 (15/15 * 4/14 * 3/13 * 2/12) and 1/1001 for the worst 5 (15/15 * 4/14 * 3/13 * 2/12 * 1/11). So "worst 4 in a conference all in the same division" is much more likely to have happened than "a division is full of the worst teams in its conference" if we check all the way to 1983. If we check after every week since 1983, probably "worst 4 all in the same division" has happened.
#57 by Pat // Nov 13, 2020 - 12:59pm
It's also possible that it's partly influenced by schedule. Just turns out at this point in the season that the NFC East has played the fewest NFC teams of any division (not by a lot, just 2 games). Weirdly that doesn't even improve in the next few weeks (unless my math is off): it's not till like week 13/14 that they catch up, I think.
Obviously DVOA's opponent adjusted but fewer connections never helps for comparisons: fundamentally it means that a lot of the opponent adjustments for the NFC East are unique to them rather than the rest of the NFC: opponent adjustments for their games are mostly internal + AFC North. So you might expect more error in comparison when you group them like that.
#53 by theTDC // Nov 12, 2020 - 4:16pm
It's worth mentioning that the 2017 Rams-9ers game needs an asterisk. The rams sat all their starters because the game was meaningless. The gap to the next worst game is thus larger.
Also, why have my comments been getting deleted? There was nothing in them except football.
#54 by pchicola1 // Nov 12, 2020 - 4:57pm
Hi Aaron. Just wondering if you have more context on the reason of NE's D decline.
In FOA2020, some "regression towards the mean" was expected. This is waaaay worse than expected.
Is it a coverage issue? Lack of pass rush/pressure? All of the above.
#56 by Dan // Nov 12, 2020 - 11:40pm
They lost like half their players. The DVOA projections article had them as the 27th ranked defense. So not that much worse than expected.
The Almanac was published before a lot of the covid opt-outs, I think, so it would've been less pessimistic about them.
That still leaves open the question of what specifically they are doing badly.