Week 10 DVOA Ratings
by Aaron Schatz
Well, that was certainly an interesting week of games, wasn't it? Road teams went 11-3 in Week 10, while underdogs were somewhere between 9-3 and 11-3 depending on whether your gambling arena of choice listed a "pick 'em" line for the Dallas-Tampa Bay and Washington-New Orleans games. Seven of the top dozen teams in last week's DVOA ratings lost this week at home, and six of those losses were to teams that ranked lower in DVOA. As noted in the Any Given Sunday column earlier today, Andrew Healy's "Surprise Score" would rank Week 10 of 2015 as the second-most surprising group of upsets since 1979, trailing only Week 6 of 2001.
Despite all the upsets, the top four teams in DVOA remain the same and in the same order this week: Patriots, Cardinals, Bengals, and Panthers. However, three of the four teams saw their actual DVOA ratings drop even though they rank in the same places. Yes, I said three of the four. The Patriots at No. 1 fall from 40.0% DVOA to just 34.5% DVOA. The Bengals at No. 3 fall from 30.0% DVOA to 26.5% DVOA.
But here's the surprising one... Arizona at No. 2 plummets from 36.2% DVOA to 30.2% DVOA even though the Cardinals beat the Seahawks this week. Arizona's drop helps explain why Seattle shoots up from ninth (14.0%) to fifth (20.0%) despite losing. The DVOA ratings for this game were very, very different from the final score.
By DVOA, this actually came out as the best game Seattle has played all year, at 79.0%. And it came out as the worst game Arizona has played all year, with a single-game DVOA of -9.0% despite the boost that comes from opponent adjustments for playing Seattle. These numbers are even more surprising given that Arizona won despite Seattle recovering all three of the game's fumbles.
|DVOA (Opponent adjustments included)|
|VOA (No opponent adjustments)|
(Note that numbers in the premium database might be slightly higher for Seattle right now because after I had run the premium numbers, I discovered an error that wasn't penalizing Seattle for lost yardage on an intentional grounding call.)
Looking at yards per play averages for Sunday night's game, these DVOA results aren't actually that crazy. Seattle was actually the much more efficient offense in general, gaining 6.6 yards per play compared to 5.4 yards per play for Arizona. But these numbers expose two possible weaknesses in the DVOA system as it is currently built.
1) What happens when one team is more efficient per play, but the other team gets a lot more plays? If we remove plays cancelled by penalty and look at just runs and passes, Seattle ran 52 plays on Sunday night. Arizona, on the other hand, ran 84 plays. Arizona had three drives of 10 or more plays, while Seattle had none. Meanwhile, Seattle ended up with three drives that were quicker than three-and-out: the safety on first down, a third-quarter interception on second down, and then a 3-yard touchdown on the first play after Cliff Avril's sack-fumble-dance-completely-ignore-loose-ball-like-a-moron at the start of the fourth quarter.
2) DVOA measures passes and runs, but only a handful of penalties are currently included: intentional grounding, defensive pass interference, false start, and delay of game. In past years, when I tried to incorporate other penalties into DVOA, I couldn't find a way to incorporate them that improved the system's predictive accuracy. But every couple years, I take another go at the problem, and it might be time to look again, particularly at offensive holding. We know that offensive penalties tend to be more predictive than defensive penalties, and we know that offensive holding has been a particular problem for the Seattle Seahawks in recent years. Vincent Verhei brought this up in Audibles this week; Seattle gained 6.4 yards per carry on rushing plays against Arizona but that number doesn't account for three different offensive holding calls on runs that put Seattle into first-and-20 situations, plus a face mask called on Garry Gilliam on another running play that put Seattle into first-and-25.
Those first-down setbacks get into a further issue with DVOA and penalties. DVOA analyzes every play and compares it to a baseline determined by down and distance among other things. So the Seattle offense ends up looking better in part because it was efficient getting itself out of bad down-and-distance situations, but much of the time we're not penalizing it for getting itself into those bad down-and-distance situations. Seattle's average offensive play came with 11.4 yards to go; Cleveland was second this week at 9.6 and the NFL average was 8.7. (If you are curious, New England was last this week at 7.6 yards to go.)
Note that I introduced these two issues as possible weaknesses in the DVOA system, not definite weaknesses. These are both issues I've played around with in the past when I've worked on improvements to the system. The system we have right now is the one that is the most accurate when it comes to balancing two competing goals: correlation with wins and losses (without adjusting for schedule) and correlation with future performance (after adjusting for schedule). It's definitely possible that a system purely based on per-play efficiency is simply more predictive than one that accounts for Arizona getting to run so many more plays than Seattle. It's also possible that accounting for how offenses recover from bad penalties is more important than accounting for those penalties in the offensive ratings. Figuring this out requires more testing and adjustments to the system. The time for doing that is the offseason, if I even can find time among the 1,000 other things I want to do every offseason. So for now, the system stays as it is. It's the best system we have around here, at least right now, and it's definitely telling us that the Seattle Seahawks are by far the best 4-5 team in the NFC and still a serious playoff contender despite those five blown fourth-quarter leads.
(By the way, read this post by Trey Causey about what it means to write for public consumption. It's possible that I'm being a complete idiot by acknowledging that there are possible flaws in the stats we use around here. But hey, we've always tried to be honest in the dialogue we have with our readers. As I often say, Football Outsiders wants to lead the league in couching our opinions in caveats. But I digress.)
Is there any other team whose ranking right now looks more shocking than Seattle still sitting at No. 5 despite a 4-5 record? Why yes, there is. I've written about the Minnesota Vikings already this season but as they continue to win, their DVOA rating continues to look weird. The Vikings climbed seven spots with their big win over Oakland, but that still has them at No. 19 despite a 7-2 record. The new playoff odds report still has Green Bay as the favorite to win the NFC North even though the Vikings now have a one-game lead. Why does DVOA seem so wrong about the Vikings?
There are some general reasons why the Vikings rank so low in DVOA. First, they've had phenomenal luck recovering fumbles. The Vikings have recovered 7 of 10 fumbles on offense, and 4 of 5 fumbles on defense. (They've had no fumbles on special teams. They had a muff, but those are almost always recovered by the return team.) Minnesota's 73 percent fumble recovery rate is second in the league this year. Only San Francisco has had better luck, recovering 75 percent of fumbles. (That's one of the reasons the 49ers are so far behind the rest of the league in the DVOA cellar despite winning three games.)
Second, the Vikings have had a very easy schedule so far. Based on average DVOA of opponent, the Vikings' past schedule ranks 31st -- only Atlanta has had it easier -- while their upcoming schedule is the toughest in the league. Their schedule has been particularly easy when it comes to opposing offenses, which is part of why the Vikings still rank just 20th in defensive DVOA even though they are second in the league in fewer points allowed. Five of Minnesota's nine games have come against the five worst offenses by DVOA: No. 28 Detroit (twice), No. 30 San Francisco, No. 31 St. Louis, and No. 32 Denver. The best offense they've faced this year is Oakland, which again helps explain why this week they had their highest single-game DVOA of the year at 44.7%.
But it's single games that create the biggest difference between Minnesota's DVOA and conventional wisdom, as you'll see when you look at the week-to-week graph for the Vikings' season so far.
Minnesota has four games with a single-game DVOA rating below zero. One is the loss to Denver, as expected. Another is a close win over Detroit, where the negative rating is not a huge surprise given the heavy downward adjustments that come from playing Detroit this season. The other two games are the real issue here: a Week 6 victory over Kansas City and that 20-3 faceplant against San Francisco way back on the first Monday night of the season.
I wrote about the Week 6 Kansas City game back after Week 6, calling it the kookiest, craziest result in a week filled with kooky DVOA results. The Vikings won 16-10 even though Kansas City gained 5.8 yards per play with a 47 percent success rate while Minnesota gained only 4.7 yards per play with a 33 percent success rate. Those stats certainly back the idea that Kansas City was the better team that day. And while the ratings for that game are part of why we've been so wrong about Minnesota since midseason, they are also part of why we've been so right about Kansas City in the same time period. After that game, Kansas City ranked 18th in DVOA despite being 1-5. Since then, the Chiefs have won three straight and moved into the DVOA top ten. We really can't ignore what DVOA tells us about Week 6.
[ad placeholder 3]
On the other hand, it sure does look like we can ignore what DVOA tells us about Week 1, doesn't it? Minnesota's -92.7% DVOA for that loss to San Francisco is the second-lowest game by any team this year, trailing only Tampa Bay's Week 1 loss to Tennessee which now has a horrifying single-game DVOA of -121.3%. San Francisco may have three wins, but that's the only positive DVOA game the 49ers have had all season. There's extra reason to believe it might be a fluke: not only was it the first game of the season, but Minnesota was stuck playing the special late game of opening Monday night despite being a Central Time Zone team. Not that we have much history to look at a sample of Central and Eastern teams playing at 7:00 Pacific time, but it makes logical sense that like Pacific teams playing at 1:00 Eastern, this may have been a problem.
All the research we've done on the last 25 years of football tells us that the long-term view is usually better than the short-term, and we don't throw out games just because they look fluky. That game will count in Minnesota's DVOA all year. It will gradually fade out in the weighted DVOA formula, but it won't disappear entirely until Week 15. Nonetheless, take that game out and the Vikings look a lot different. This table shows how Minnesota's rating moves closer to conventional wisdom has you forget about Week 1, strength of schedule, and then fumble recovery luck.
|Minnesota Vikings DVOA, Removing Context Adjustments|
|Metric||Off DVOA||Rk||Def DVOA||Rk||ST DVOA||Rk||Total DVOA||Rk|
|Weeks 2-10 DVOA||-4.2%||22||-2.3%||13||5.7%||4||3.7%||15|
|Weeks 2-10 VOA (no opponent adjustments)||-5.1%||22||-5.2%||11||5.7%||4||5.7%||13|
|Weeks 2-10 VOAf (no opponent adjustments;
fumbles not all counted equal)
By the way, take that Week 1 game away from San Francisco and the 49ers drop from -36.5% DVOA to a frightening -44.4% DVOA on the season. Yikes.
The Vikings/Packers NFC North race is one of the main subjects of todays' commentary on our playoff odds simulation published at ESPN Insider. As we did the last couple years, we'll be writing this commentary each Tuesday afternoon until the end of the regular season. This week's article also discusses how Week 10 dramatically changed the race for the AFC wild cards. I'll be linking to that piece inside the DVOA commentary each week rather than giving it a separate discussion thread on our front page.
* * * * *
Once again in 2015, we have teamed up with EA Sports to bring Football Outsiders-branded player content to Madden 16 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. We will also tweet out images of these players from the @fboutsiders Twitter account on most Fridays. One player each week will only be available for 24 hours from the point these players enter packs on Friday.
The Football Outsiders stars for Week 10 are:
- MLB Lawrence Timmons, PIT (24-HOUR HERO): Led NFL with 6 defeats in Week 10, including 4 different tackles after third-down receptions that prevented conversions, plus a third-down sack and a run TFL. Also had two other tackles on runs for no gain.
- RB Andre Ellington, ARI: No. 5 RB of Week 10 with 38 DYAR; 5 carries for 61 yards and a touchdown, plus 3 receptions for 27 yards.
- K Stephen Gostkowski, NE: Game-winning 54-yard field goal, plus Giants didn't get any of his five kickoffs past the 20 (three touchbacks).
- RT Morgan Moses, WAS: Washington RB had 64 yards on 13 carries right side with no stuffs, plus no sacks allowed.
- CB Jimmy Smith, BAL: Allowed only 2 completions for 5 yards, plus a 27-yard DPI, on 7 targets.
* * * * *
All stats pages are now updated with Week 10 information (or will be in the next few minutes) including FO Premium, snap counts and playoff odds.
A quick note on the playoff odds simulation: We've made adjustments to some team's weighted DVOA being used in the simulation because of quarterback changes, in particular Pittsburgh and Dallas. We also have built in an estimate of what Andrew Luck's injury will mean for the Colts. However, we decided not to include any changes to adjust for the injuries on the New England offense. We could adjust for the loss of Julian Edelman through the playoffs, but the new preseason projection system would suggest that, based on last year's production, we also would need to adjust the Patriots upwards because they had to play the first few weeks without Brandon LaFell. The Patriots' offensive line makes things even more confusing -- the team got its starting center back, but he's now playing right tackle, but the regular right tackle should be back at some point, but the starting left tackle is done for the year, and on and on. In the end, we decided that the hit the Patriots took from their near-loss to the Giants was enough to suggest what the impact of life without Edelman and Dion Lewis will be like. We'll reconsider that if they struggle significantly on offense again this week.
* * * * *
[ad placeholder 4]
These are the Football Outsiders team efficiency ratings through 10 weeks of 2015, 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 for strength of schedule and 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. 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.
As of this week, opponent adjustments for all stats are at 100 percent strength and will be for the rest of the season.
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).
123 comments, Last at 24 Nov 2015, 11:11am
#1 by TimK // Nov 17, 2015 - 7:43pm
Interesting to see the Broncos with 32nd offence and 1st defence - an offence up to the standards of The Bucs would move the Broncos above the Seahawks so they don't have to aim too high to get the division. The old line about the defensive players asking the offence to just go out and hold the opposition sounds about right. Broncos special teams are about as good as I remember them being as well.
#3 by Danimal // Nov 17, 2015 - 8:01pm
there were years when broncos special teams couldn't get out of the high 20's in DVOA rankings, so I'm shocked that they've been flirting in the top 15 all year.
that offense.... it's offensive, what a waste of a year if B.Oswhatever can't do something close to replacement level and create a functioning system for the rest of the year.
betting on Manning not coming back, hoping he's healthy... just from the bench.
#34 by deus01 // Nov 17, 2015 - 11:52pm
I'm no fan of Fox but I think the Bronco's would have been better off keeping him and Gase and still hiring Wade for defense. Elway seems to have also really undervalued how important having a good Offensive Line is.
#42 by tunesmith // Nov 18, 2015 - 3:06am
I honestly thought Fox's handling of the offensive line was his weak point. Even in the Super Bowl year, the line wasn't anywhere near as good as Manning was making them look, and as soon as they faced a real pass rush, they were exposed. Then as the years go by you saw them make half-hearted attempts to bring more talent in, only to see them get waived or signed by other teams, or riding the bench. At least now Kubiak and Dennison are making a concerted effort to get the young players some solid playing time by rotating them in. Paradis, Garcia, and Schofield occasionally look good, and they're all very young. Beyond that, it really didn't help that they lost both Clady and his backup (a rookie) to injury. Ryan Harris is a solid right tackle but they have to rotate him and Polumbus at left.
Hopefully Denver will draft one more early-round tackle next year, and then that line might actually turn into a strength.
#95 by deus01 // Nov 18, 2015 - 3:27pm
The problem is the line now is completely incapable of run blocking and also bad at pass protection. They're too weak and uncoordinated, so they may get better with more practice but I don't think they'll improve enough to help the offense significantly. They should have realized this before the current season. If you look at the line as being exposed in the Super Bowl and how ineffective the Broncos offense was then it makes sense that even an average pass rush would hamper the offense with the terrible line play this season, especially with an injured Manning who's trying to process a new system.
Quarterbacks that aren't extremely mobile are always going to suffer behind bad line play. Previous season's Manning and even Brady at the start of last year play their worst when the line isn't holding up.
#2 by bluecjj // Nov 17, 2015 - 7:59pm
How many percentage points of DVOA do you think the Patriots lose now, now that both Edelman *and* Lewis are out?
Also, a sort of question on DYAR- how many "yards above replacement", about, would translate into one "point above replacement"? How many percentage points of DVOA would translate into being a one-point favorite over another team in a neutral site? Points are the more familiar language to me.
#6 by Aaron Schatz // Nov 17, 2015 - 8:21pm
The problem isn't that years ago -- and it was a LONG time ago at this point -- we switched from posting individual stats in terms of points to posting them in terms of yards. The problem is the difficulty of separating one player's contribution from the contributions of his teammates. How much of Edelman's value was actually Brady, or the system as a whole? How much of Dion Lewis' value was actually Brady, or the passing game opening things up for the run, or the blocking of the offensive line? Even if DYAR was still DPAR, it wouldn't change the fact that scheme and teammates are not filtered out of those numbers.
And if you're still that hung up on the change from DPAR to DYAR, read the Trey Causey link above.
#7 by Karl Cuba // Nov 17, 2015 - 8:27pm
And these are very complex problems that presumably have no ideal solutions. I hope I speak for the vast majority of your loyal fans when I say that we really appreciate your openness in acknowledging that these issues exist, it's the right way to do this. Unlike some other platforms that pretend they can rate individual players to three significant figures...
#64 by bravehoptoad // Nov 18, 2015 - 11:34am
Every July 7th, my FO cronies and I gather around and toast a sad toast to DPAR. There follows a silent moment, then a few maudlin anecdotes. Remember how DPAR roasted Michael Vick in 2004? 2005?
#80 by Raiderjoe // Nov 18, 2015 - 12:52pm
it is good you and your friends got togeher to drink. weird thing to drink over but not weird to drink so that is cool. would be reraly weird, but also cool. if you did you DPAR drinking in woods like bunch of high school kids who raided their dadyy's beer
#5 by Karl Cuba // Nov 17, 2015 - 8:13pm
The Vikings-Niners game has quite a lot of context that can be added to it:
- The niners had shifted to a zone run game after running power schemes for the previous four years.
- They'd also shifted from Fangio's approach that had two safeties deep more than pretty much any team and rarely blitzed to Mangini's 'I can outwit everyone' big blitzing, supposedly disguised system. The following few weeks showed how with regular season game tape to watch this new scheme delivers more matchup blessings for the opposing offense than it helps the defense.
So the Vikings had to play a team that they really couldn't have properly anticipated.
And the niners' -44.4% DVOA represents what happens when you couple the retirement and suspension strewn offseason from hell with one of the most outmatched coaching staffs I've ever had the misfortune to see. I really don't think it's too much of a stretch to suggest the top three coaches (HC, OC, DC) are at best the 32nd in the NFL; nobody else was going to offer Tomsula a head coaching job, Chryst was promoted from qb coach after they couldn't find anyone else (and the qb coach was working local radio the previous two years) and Mangini was the tight ends coach after working for ESPN for a few seasons. I did think the much more experienced coaches lower down the totem pole (Chris Foerster, Tom Rathman, Tony Sparano, Tim Lewis, Clancy Pendergast and Jason Tarver) might seady the ship but the cretinous trio at the top have held firm. This bunch might not be in the top fifty pro coaches in their respective positions, they might not be in the best one hundred and it shows.
It isn't like there's much hope because the interfering owner who likes to leak his bile to the press is going to put off any of the top coaching prospects. All I would say is that the players are less to blame that the idiotic system. Coaching matters, why else can you explain the Pats always doing so well ther than the greatness of their coach who took his team to the playoffs even when the had to replace Brady with Cassel.
#17 by Thomas_beardown // Nov 17, 2015 - 9:34pm
On the other hand you don't have to look at Belichick to see the effects of coaching. After his first year in Chicago, Lovie Smith was never worse than 7-9 in Chicago, and he's on track to be similar in Tampa Bay.
There is something to be said for a minimum level of competency.
#26 by Eleutheria // Nov 17, 2015 - 10:33pm
But that 11-5 came against the 28th hardest strength of schedule. It's not like teams were struggling against Cassel more then usual, and Cassel on the 2010 Cheifs (589 dyar vs 458) was statistically better then his 2008 campaign, and that was the one time Cassel did make the playoffs.
When you have Randy Moss and are facing defenses with nobody who can cover him, you're going to win games.
To me 2008 was more about how good Randy Moss is, and how bad the defenses the Patriots faced were, then how good of a coach Belichick is.
#30 by Karl Cuba // Nov 17, 2015 - 11:11pm
I'm regretting choosing this example and now finding the pedants less fun.
So Ele, disregarding my (obviously terrible) example that was the first that occurred in my head are you really disagreeing that coaching is vitally important?
I could have pointed to the Jim Johnson defenses that saw players leave, not be as good, return and succeed again or a myriad of other cases.
I think coaching is important. Am I wrong?
#43 by panthersnbraves // Nov 18, 2015 - 4:26am
I really like Norman. I think, though he has a little Ricky Williams vibe about him - not that he smokes weed, but rather he is on a different wavelength than many. I hope he stays. I think he's got a good setup and I think that he'd be unhappy in a lot of other places.
and could someone put my ID back in the captcha filter again so I don't have to do it every time?
#45 by herewegobrowni… // Nov 18, 2015 - 7:00am
Here's hoping Justin Gilbert can show the same kind of career turnaround (although I still don't think -last- year he was that bad, and PFF agrees with me with a yellow/average rating.)
There's no way from that OK State team all three first-rounders, Gilbert, Weeden and Blackmon can flop, right?
#53 by Dave Bernreuther // Nov 18, 2015 - 9:49am
That's a really odd thing to get annoyed about.
One could use the same Moss logic to the same effect to discredit Brady's 2007 season and award, no?
I tend to agree, of course, given that he was the key element of the top two offenses of all time (at that time) and a total freak... But Cassel is very, very bad, and plenty of very bad QBs have freak weapons and pile up good numbers but still lose frequently. Cassel's many weaknesses were avoided/hidden extremely well by Belichick and McDaniels, much as their current offense avoids their current players' (few) weaknesses to great effect. Credit is certainly deserved for that.
What always amazed me was that Scott Pioli of all people actually knew Belichick and McDaniels well, knew their strengths, and still decided that Cassel was worth not only acquiring but overpaying. I think very little of Pioli but that's a mistake that, given the information he had, surprises even me.
#62 by BJR // Nov 18, 2015 - 10:59am
The Chiefs 'only' gave up a second round pick for Cassel - not a huge price for a prospective starting QB. And he wasn't a total bust in KC - he had one solidly above average year (2010) in which the team won its division. He suffered injuries and was awful thereafter, but there have been worse trades in NFL history.
Why teams in 2015 are trading for Cassel, however, remains a mystery.
#88 by BJR // Nov 18, 2015 - 1:56pm
Yes they did.
Reading back about the trade I was reminded that Cassel had been franchised by NE after 2008 as his rookie contract was up (obviously with the intention of trading him), so the Chiefs were immediately forced to sign him to a long term contract. Obviously that explains why they 'only' gave up a 2nd rounder and got a veteran defender in return, and makes it look a lot less rosy for them.
#106 by Eleutheria // Nov 18, 2015 - 9:01pm
Though there is a massive difference between the slightly above average performance put up by Cassel in 2008 and the greatest season of the DVOA era and possibly all time.
Moss is in my opinion the greatest Wide Receiver of all time, so obviously when you pair him with a HOF QB like Brady, the QB has a good chance of having a career year (especially given the lack of talent the Patriots had on offense prior to 2007). And as BJR pointed out, the Casseltrade isn't as bad as people made it out to be.
#29 by herewegobrowni… // Nov 17, 2015 - 11:09pm
"I really don't think it's too much of a stretch to suggest the top three coaches (HC, OC, DC) are at best the 32nd in the NFL; nobody else was going to offer Tomsula a head coaching job, Chryst was promoted from qb coach after they couldn't find anyone else (and the qb coach was working local radio the previous two years) and Mangini was the tight ends coach after working for ESPN for a few seasons."
Mangini is worse than Rob "Lebowski" Ryan and Jim O'Neill? (Granted, the Pettine/O'Neill defensive combo looked strong for many years until now.)
I wonder how much of Mangini's recent resume can be attributed to tipping off Spygate. The NFL coaching fraternity might not like Belichick's videotaping, but they like a "rat" even less (especially since Mangini snitched on essentially his father figure.)
#32 by Karl Cuba // Nov 17, 2015 - 11:24pm
OK, I wasn't trying to get into a 'worst DC' fight. Mangini has been bad as a DC for the Pats, his defenses weren't good as the head coach of the Browns or Jets and after failing to get a job in either college or the pros and then spending a year as as niners' tight ends coach (?!?!) has been dire this year as the 49er DC.
Would you cover Larry Fitzgerald with a 265 lbs OLB? I guess not. Would you do it over and over again as you get whooped up and down the field? I hope not because that would mean that you're as foolish as Eric Mangini. That is just one of his moronic failings, I could give a list but I feel it distracts from the wider point I'm trying to make about how important coaching is and how weak the niners are across the board in that respect.
#54 by Dave Bernreuther // Nov 18, 2015 - 9:54am
I think has to do more with not being very smart than it does with being a rat, but the fact that he also didn't really have much of a resume on its own, as pointed out above, doesn't help his case either.
#13 by Hoodie_Sleeves // Nov 17, 2015 - 9:06pm
Because those stats measure a conglomeration of things outside the running game - like safeties having to play off the line because they have an elite QB and passing game. And the QBs ability to audible in and out of plays when running one type would be a bad idea.
Also, coaching and scheme matters. A ton.
#73 by Scott C // Nov 18, 2015 - 12:33pm
Yeah, and that one would suggest that the Chargers have a good running game because of it... Nope.
The overall point, that ALY is measuring the total effectiveness in a complicated context stands. Its part the line, part the scheme, part play-calling, part the opponents, part the RB.
#21 by Vincent Verhei // Nov 17, 2015 - 10:08pm
Because the rush offense DVOA listed here includes all running plays, not just what happened recently. They were crushing teams early in the year, but they've ranged from -3.6% to -30.8% in the last four weeks.
#22 by Vincent Verhei // Nov 17, 2015 - 10:13pm
Or, to put that another way:
From Weeks 1-6, NE's rush offense DVOA was a league-best 16.3%. much better than the second-best team (CAR, 9.6%)
Since Week 7, their rush offense DVOA is -15.4%, 21st.
#50 by RBroPF // Nov 18, 2015 - 9:04am
Cheer up, it may be again. Losing Solder hurts, but if/when they get Vollmer and Cannon back both tackle spots should be good to very good. Similarly, on the interior, they'll miss Wendell's veteran leadership and experience, but it looks like they've got 5 solid starters there.
Right now, it's the loss of Lewis and Edelman that really has me bummed. I think people underestimate Edelman's value to this offense and I expect them to really struggle without him. I really hope he's back and effective for the playoffs.
#48 by RBroPF // Nov 18, 2015 - 8:55am
I think part of the reason the run offense appeared so successful when the both line and skill positions were at full strength was because they were happy to throw the ball as much as they wanted. They ONLY ran the ball in situations where they were confident it would be successful.
With the line in the state it's currently in, they are running the ball much more to protect the line and to protect Brady, even though they know all those running plays won't be as successful.
#25 by Moridin // Nov 17, 2015 - 10:29pm
Not only that, but the Packer contribution to the Viking SoS (2 games) is predicated on GB having access to the team that was #4 going into Denver, because they've been on freefall since (DVOA 39%->15%). If GB was considered to be 0% (which would be uhm.. WAS or thereabouts) then the Vikings #1 future schedule likely gets bumped down as well. Not that playing Seattle and Arizona (thurs night on the road, bleh) aren't going to keep that schedule ranking up there, but 2 games against GB at #8 are certainly a strong contribution as well.
#23 by Eleutheria // Nov 17, 2015 - 10:19pm
Though if we're working under the assumption that the Vikings are underrated in DVOA, part of the Packers positive schedule is facing the Vikings twice. The other part of why the Packers future schedule looks so week is They'll be facing the Cowboys in week 14 with a healthy Romo. That's not something the strength of schedule model can anticipate.
Unless the Packers beat the Vikings twice, the Packers will probably lose the division, cause if they split the series, Green Bay will have to win two more games then Minnesota in their final five games.
#18 by owleabf // Nov 17, 2015 - 9:51pm
One thought on the Vikings fumble luck and why subjective measurements aren't reflecting that luck...
A significant number of the offensive fumbles for the Vikings have been unforced and without defenders nearby. Diggs put two on the ground vs Denver, in both cases there wasn't a defender nearby so he easily covered it up. AP and Teddy have had a couple botched handoffs that they just fell on. So those count as fumbles, but are situations where it was unlikely the D was going to grab them.
On the other side of the ball, as far as I can remember all of the Vikes defensive fumbles have been forced fumbles. So defenders nearby and aware that the ball is out, which you'll see a higher recovery percentage on.
It's a thought at least.
#24 by Eleutheria // Nov 17, 2015 - 10:23pm
Though DVOA counts fumbles based on the probability that it will be turned over based on the type of fumble.
And even then, aborted snaps or handoffs probably are recovered by defense 40-45% of the time.
I also don't think you can justify arguing that Diggs fumbling the ball with no defense around him as a legitimate reason to ignore those fumbles. If anything I'd be more worried cause this clearly means the fumbles are happening because of bad ball control by the vikings, and not because they've faced defenses that are good at causing fumbles.
#70 by owleabf // Nov 18, 2015 - 12:23pm
>Though DVOA counts fumbles based on the probability that it will be turned over based on the type of fumble
I didn't realize this, that pretty much negates my points.
>I also don't think you can justify arguing that Diggs fumbling the ball with no defense around him as a legitimate reason to ignore those fumbles
Not saying they should be negated, just that they weren't situations where he was going to lose the ball. As opposed to say a strip sack where players are all scrambling on the ground to cover the ball up and it's a little random who the ball bounces to.
Regardless, if DVOA adjusts for that then my original thought is wrong.
#27 by andrew // Nov 17, 2015 - 10:39pm
In the playoff odds you are now listing some scenarios involving the vikings (albeit remote chances).
In one you tag KC vs Minn as a "Superbowl III rematch". It would be a Superbowl IV rematch (III was Jets-Colts, which will never happen in the superbowl again barring realignment). I'd mention it there but there is no comments thread there (and it is referenced in this article).
#37 by DezBailey // Nov 18, 2015 - 12:12am
Week 10 BES Rankings - http://besreport.com/week-10-bes-rankings/
FAQ - http://besreport.com/about-the-bes/
BES and DVOA agree on the Pats, Bengals and Panthers. After that they differ widely. But the BES is more a gauge of momentum or heat. Vikes No. 4, Texans No. 7, Chiefs No. 13. The Seahawks seem hellbent on finding ways to dramatically lose games. The BES has them ranked 22nd. The BES is also higher on the Bears than most rankings as they ranked 11th this week.
DVOA's take on the Cards, Bills and Saints seems more accurate. The BES has the Bills 21st, Cards 17th and Saints 5th. Saints built up a nice head of steam but BES numbers suggest them capable taking down the Panthers. BES also has Indy at No. 14. They gave the Pats and Panthers hell before beating Denver. BES had the Colts as high as 10th back in Week 7 - http://besreport.com/week-7-bes-rankings/
Overall, it seem like Week 10 was a major shift in power. There will be more than a few teams making a Seahawks-like surge toward the playoffs. Could see a couple of unforeseen division Champs.
#61 by jmaron // Nov 18, 2015 - 10:43am
Minnesota dominated the game in the first half. KC managed 75 yards on 6 drives, resulting in zero pts and 6 punts. Meanwhile Minnesota managed 210 yds on 5 drives scoring 10pts.
The second half was roughly the reverse.
I didn't think that one team was clearly better than the other. Minnesota got a little lucky hanging on in the 2nd half, but KC was lucky they weren't down 20 in the first half.
#63 by jmaron // Nov 18, 2015 - 11:05am
I messed those stats up a little - KC's 1st 6 drives included a 34yd drive to start the 3rd quarter. Minnesota answered that drive with a 46 yd FG drive.
So it was 13-0 - with 23 minutes to go - KC had run 28 plays for 75 yds for a 2.7 avg. Minnesota had run 47 for 250 and a 5.3 yd.
The comment about KC averaging 5.8 to Minnesota 4.7 is very misleading for two reasons:
1) Minnesota ran 6 plays essentially killing the clock at the end of the half and the game. Those 6 play netted -1 yards...take those away and Minn averages 5.2 yds/play not 4.7.
2) KC piled up all their yards against a team protecting a lead. I'm skeptical of great yards gained when teams are behind.
I still think those teams are probably evenly matched, but there's no way in my mind KC was the dominant team in that game.
#44 by ammek // Nov 18, 2015 - 6:42am
It looks as though the Packers' adventure into above-average special teams is over. Thanks, guys, it was fun. Green Bay is heading for its ninth below-average DVOA in 11 years under Ted Thompson. That will, I think, be the most of any team in that timespan if Washington and Detroit continue to rank positively. Washington hasn't had above-average special teams since 2006. Cumulative ST DVOA, 2005-2014:
The Browns' ST are good again this season, but otherwise it seems there's a fairly significant connection between special teams DVOA and offensive/defensive DVOA. Only five teams with above average O/D DVOA have below-average ST DVOA, and only one of those (the Jets) ranks in the bottom quartile. Likewise, among the teams in the bottom half of O/D DVOA, only Minnesota and Cleveland are in the top 10 for special teams. Last year, of the top quartile for special teams, all but one had above average O/D DVOA. I wonder if there's any correlation here.
It's hilarious that the Lions had positive special teams last week despite missing two extra points and botching the recovery of an onside kick.
#120 by LionInAZ // Nov 19, 2015 - 11:28pm
It's not so hilarious given that the Lions had a 104-yd KO return that led to a TD, while Crosby laughably screwed up a game winning FG attempt. An attempt that wasn't even touched by a defender.
I can't even guess how Masthay's terrible punts figured into the results.
#49 by Since1stTMQExile // Nov 18, 2015 - 8:58am
The Washington chapter is always such a fun read in the annual FO book, but the Special Teams section is usually particularly entertaining. A constant churning of bad kickers and terrible coverage units. This year's was a bit different. Something like: "now that the Special Teams are no longer really, really old; let's see if they can actually be any good."
#58 by justanothersteve // Nov 18, 2015 - 10:13am
Green Bay is clearly ranked too high because they lost to freaking Detroit. At Home. Where they haven't lost to Detroit since Mike Tomczak was QB. Throwing darts on acid for rankings is way better than this. I think I'll find raiderjoe so we can get drunk on Sierra Nevada Pale Ale and I can be delusional about my team again.
#78 by Raiderjoe // Nov 18, 2015 - 12:49pm
well if GB was not so high and mighty, they wodul have traded for a receiver like nate Washington or steve smeith sr. before he got injured. but no, they are too high and mighty. "can't deal gold (draft pick) even 7th round gold for a vbet. becuaue it is like bad or something or the vet would give the packers cooties. so no, we will stick with our young guys and hope it works." *LOUD ANNOYING BUZZER NOISE* wrong! you lose to loisn, you dopes. N. Washington would've caught that 2-pinter at end of game . then packer sowuld have won in overtime
#103 by ChrisS // Nov 18, 2015 - 5:15pm
They still brew in Petaluma (one time home of the arm wrestling championships) as well as Chicago. Not sure where they brew for shipping to the midwest, have to check my next bottle. They did just sell 50% of themselves to Heineken.
#93 by mehllageman56 // Nov 18, 2015 - 3:15pm
The Jets should have tried that trade too. I'd rather throw a fourth down endzone desperation pass to Vernon instead of Kellen, even if it means losing a 6th round pick. Watch, they draft a hall of fame qb next year.
#59 by Joe Pancake // Nov 18, 2015 - 10:21am
Watching every Seahawks game of the season, it doesn't seem way out of line to me that they would be the fifth best team. Three (!) of their five losses have come to teams ranked ahead of them in DVOA, another is a road loss to the eight best team, and another a road loss to the 17th best team. And they have held a fourth quarter lead in all of them! You have to be very good to consistently be ahead of the best teams in the NFL late in the game.
Is there some hidden reason why the Seahawks can't hold leads? Maybe. But it seems more likely to me to be mostly chance. I think most people don't put enough stock into natural variance and schedule when determining who are the best teams.
As to the Cardinals game in particular, penalties were a HUGE factor in that game. They were calling the 'Hawks for something on seemingly every play (offensive and defensive). I can't imagine a crew calling more penalties. This doesn't mean the refs were "wrong," but it does mean a different crew on a different night probably would have been more favorable to the 'Hawks. It has just been that type of year so far.
#115 by Fierydemise // Nov 19, 2015 - 12:23am
We're clearly watching very different Seahawks games. Yes they've lost against 3 or the top 4 and 4 of the top 8 teams by DVOA but all of their wins come against well below average teams. They have wins against the number 31 and 32 teams in the league and against the number 31 team they needed a heroic goal line play (stipulating that while the refs clearly blew that call I think in an alternate universe where Wright tries to catch the ball they still win so I'm not going to focus too much on that). They barely squeaked out a win against the number 22 team fielding a backup QB. In fact Seattle has beaten backup QBs in 2/4 wins and 1 QB who has since been replaced by Blaine Gabbert, suffice it to say not exactly a great set of opposing quarterbacks.
Maybe the end of game stuff is just odd variance I doubt it. Seattle is playing immensely inconsistently, the offense will get a couple good drives a game and otherwise just a bunch of quick 3 and outs. The oline is playing a little better in that is no longer a garbage fire on 100% of plays, maybe more like 50% now. There is no consistency on defense either, Cary is just bad and it feels like trying to cover for him is causing more stress on scheme. Want to know how to beat the Seattle secondary, just TE seam routes all day, 30th in DVOA against TEs and 30th in defense against the deep middle. The only two teams worse against the deep middle? One team that fired it's DC and one team that fired it's coach, not great company. The most consistent part of the team seems to be the d line but where as in previous years where a pass rush felt like a bonus this year it feels mandatory. If the pass rush isn't getting home I don't trust the secondary to hold up.
Seattle is sticking around because they have some really talented players who can a couple times a game make totally game changing plays but it isn't sustainable and against decent teams the lack of sustainability becomes a major issue.
#60 by BretU // Nov 18, 2015 - 10:41am
Watching that Arizona - Seattle game, I have no difficulty thinking Seattle outplayed Arizona by a significant margin and that the referees closed the 88% DVOA gap. It was almost clockwork that Seattle would get a marginal holding penalty or non-existent face mask on offense or Arizona would get bailed out on a marginal defensive holding/pass interference call. It made for a terrible game to watch too because it broke up the game flow.
There is no reason to think it was anything other than just a horrendously called game where the referees had too big of an impact though. I have to imagine the league would have preferred Seattle to win to keep the division race closer.
#67 by Mugsy // Nov 18, 2015 - 12:07pm
Seattle is clearly ranked too high.
Not because they were beaten by the Cardinals-
(although the refs were unfair to them) ;)
But because their offense ranks 11th, and their defense ranks 7th in DVOA- putting them somewhere just below the Steelers, the Packers and yes, even the pathetic Jersey Jets.
#69 by nat // Nov 18, 2015 - 12:14pm
Regarding DVOA flaws:
What happens when one team is more efficient per play, but the other team gets a lot more plays?
That's not a flaw. No one in his right mind thinks that being more efficient guarantees a win. It doesn't even guarantee a more successful drive.
DVOA measures passes and runs, but only a handful of penalties are currently included.
That's only a flaw if those other penalties are predictive. It looks like they mostly aren't. I suppose you could give a secondary number like 'DVOA including non-predictive events'. That would let you say things like 'Team X looked like it would be the better team in most games, but was outplayed anyway'. There's not much value in that, except to people who look to DVOA for solace when their team loses or for validation when they win.
And if you're still that hung up on the change from DPAR to DYAR, read the Trey Causey link above.
Ah. A link to a mix of wisdom and self-pity. :-)
Don't fall into the trap of confusing criticism with trolling. (You mostly don't. Good for you!) Specifically, the switch from DPAR to DYAR came with a cost: the stats for different positions are no longer commensurate. TE yards and WR yards and RB yards are all different sizes. DPAR at least tried to equate everything to points that were all the same size. So it's easy for longtime FO fans to get wistful about DPAR, even if DYAR is overall a more useful stat.
You should consider writing a periodic column on DVOA and other FO stats from a more technical point of view. It would be too much to do weekly, but deserves more than being hidden in comments or a single off season article. It would also mean that constructive criticism wouldn't always feel so negative to you and the FO team.
You could have topics like "Can a Skill Be Non-Predictive, And Why Should We Care?" or "What Should 'Over Average' Mean When You're Down Twenty?" or "Could and Should Defensive DVOA Be Adjusted for Weather?" Potentially interesting topics, for the stats nerds anyway.
#79 by Scott C // Nov 18, 2015 - 12:50pm
> >What happens when one team is more efficient per play, but the other team gets a lot more plays?
> That's not a flaw. No one in his right mind thinks that being more efficient guarantees a win. It doesn't even guarantee a more successful drive.
Ah, but that is it --- DVOA likes the more efficient per play team better. That might indicate better future success, but actual success in one game is more of a drive success thing, which might like a team with a higher number of less efficient plays.
There isn't an easy answer. DVOA can favor the team that was more efficient but barely had the ball, which in an individual game might mean a loss to a less efficient team. But that per play success might be a better indicator of future wins.
Perhaps what is needed is a different stat that lives along-side of DVOA for teams that is more representative of past performance than a future predictor. For example, the percentage of total YAR that a team had (VOA * # of plays) in a game would probably correlate more with who won the game than VOA does. A cumulative 'value of play' stat essentially covers this case where one team was better per play but didn't have enough opportunity, but is likely less predictive.
#107 by Eleutheria // Nov 18, 2015 - 9:08pm
but in theory the more efficient a team is, in the long run, the more plays that team will run.
I suppose the exception to that (which may be a problem with this years Seahawks) is the consistency of that efficiency. If I get a VOA of 100% on one play, and -10% on next three. That's still a VOA of 17.5%.
So maybe there needs to be a stronger emphasis on consistency in DVOA. Though maybe putting too much focus on consistency would hurt DVOAs correlation coefficient with future wins.
#108 by Thomas_beardown // Nov 18, 2015 - 9:13pm
I'm not sure your theory is true. You would expect more points scored and more successful drives, but I don't know you would expect more plays (or drives). The most efficient offense imaginable would score on 1 plays drives, so nearly by default their opponents would run more plays.
#110 by Eleutheria // Nov 18, 2015 - 9:44pm
yeah that's true, but when you get that efficient, you're winning regardless of how many please your opponent runs.
And besides, the most efficient offenses do tend to run 5-6 plays per drive, 4 from the least efficient. Since 1-play TD drives are rare, but 3 and outs are common.
I should probably correct my statement and say that the more efficient you are, the less likely a team will beat you just through extra plays.
#111 by Thomas_beardown // Nov 18, 2015 - 10:12pm
I used an exaggerated example, but this isn't strictly academic. Arizona is 1st in yards per drive, 2nd in points, 2nd in drive success rate, and 3rd in offensive DVOA, but they are 15th in plays per drive. They like to throw deep.
#77 by Raiderjoe // Nov 18, 2015 - 12:44pm
No way should Raiders be bleow Bills, Jets or chiefs. Raiders beat Jets head to head. Bills and Chiefs not better than Raiders. No problem, though. Raiders will finish 11-5 or 10-6. bills or chiefs will lose in week 12, so that is at least lsos #6 for one of them righte there. Raiders will beat chuiefs in week 13.
#81 by alien1rock // Nov 18, 2015 - 12:54pm
For lack of a better place to ask this, is there a quick explanation for why Jamison Crowder's DYAR and DVOA are so low? He has been very effective for Washington with a very high catch rate and I believe a reasonably high first down rate. Two fumbles should hurt him, but not that much? Is it related to his very low yards per catch?
#86 by ChrisS // Nov 18, 2015 - 1:41pm
In light of the Trey Causey link. I like FO, a lot, I read it most days during the season, thanks for the product. DVOA is a model and as such it should not be expected to be perfect, it should be expected to give insights into reality and it does. I think pointing out potential weaknesses in a model is important and makes the results more useful.
#90 by Eddo // Nov 18, 2015 - 2:09pm
Am I wrong in relating the Arizona/Seattle and Kansas City/Minnesota games from this year to the infamous New York Jets/New England game from a few years ago?
The team that appeared better by DVOA concentrated their bad plays to several three-(or fewer)-and-out drives.
#94 by mehllageman56 // Nov 18, 2015 - 3:19pm
Seems reasonable. I did not get the impression that Seattle was better than Arizona watching that game; the Cardinals put up 39 points in Seattle, and could have scored more/allowed less with better ball security.
#97 by poplar cove // Nov 18, 2015 - 3:51pm
Totally agree........ I don't know how anyone could have watched that game and felt Seattle was better team. They benefitted in a big way from two defensive touchdowns to grab their only lead of the contest and then quickly gave that lead up on the very next drive by Arizona.
#96 by nat // Nov 18, 2015 - 3:48pm
Here's a link to the final part of that Jets/Patriots discussion from 2011: http://www.footballoutsiders.com/dvoa-ratings/2011/absolute-final-word-jets-patriots
This may be where Aaron learned (or relearned) that the best way to handle strange and perhaps broken results is to treat them as a learning and teaching opportunity. That seems on topic this week with Aaron's commentary.
If I recall and read correctly, there were multiple effects going on in that game to make its VOA a mismatch for what fans thought they saw. There was a small problem with how red zone bonuses were applied (do those still exist?) giving a small unearned bonus to the Jets. The Jets by pure luck avoided high expectation situations, and so got an easier average to compare to overall. The Jets got the ball last for a garbage time drive. The averages compared to in garbage time are tiny, giving a large unearned VOA bonus. The Jets' drives failed quickly when they failed, as you say about those other games. And, last but not least, the Jets had two truly awesome drives, turning a game's load of success into just fourteen points.
It was probably just bad luck that all these factors ended up skewing the VOA one direction, giving a seemingly wacky result for VOA.
Some of those issues were just timing ones. Football is like that: Four barely good-enough drives are better than two awesome ones. Some issues were perhaps weak points in DVOA - maybe unavoidable, but worth being aware of. For example, DVOA rewards teams that fall behind and almost catch up over teams that build leads and let them dwindle as they burn clock.
I have no idea if similar issues apply to the games you mention. Probably some of them do.
#102 by Eddo // Nov 18, 2015 - 5:07pm
The Cardinals/Seahawks game does seem to have some similarites. The Seahawks had some catastrophic failure drives (the safety, in particular) that were over quickly. And based off someone else's description of the Vikings/Chiefs game (I didn't see that one), it sounds like the Chiefs quickly went three-and-out a lot.
#123 by cstoos // Nov 24, 2015 - 11:11am
I am curious to see where KC ends up in the Week 11 rankings. My numbers have them as the #5 overall team (just behind Carolina) when looking at the whole season.
If you look at just the last 6 weeks though (including Chicago and MIN losses), they are #1.
Last 5 weeks, they are #1 by a LONG ways.
Not sure I have seen a turnaround quite this profound. Record wise, sure, but they are truly dominating on both sides of the ball.