Writers of Pro Football Prospectus 2008

Most Recent FO Features


» Defense and Rest Time

Do defenses really wear out over the course of a game? Do defenses benefit from long drives that give them more time to rest on the sideline? Guest columnist Ben Baldwin investigates.

18 Sep 2012

Week 2 DVOA Ratings

by Aaron Schatz

A lot of NFL power ratings around the Internet have the Houston Texans on top after two big wins to start the season. The Football Outsiders DVOA ratings are in agreement there: Houston ranks third in offense and first in defense, although the Texans are just 30th in special teams. Below the Texans, the DVOA ratings probably look a lot different from ratings elsewhere.

One of the big issues early is the issue of opponent adjustments. When you've got subjective ratings, you are always accounting for strength of schedule in the back of your mind, although very few subjective ratings give it the importance it deserves. With objective ratings, we're stuck with the system as we've worked it out for the last few years. After two games is too early to try to get a handle on strength of schedule. We introduce schedule adjustments in Week 4 and then gradually increase the strength until they're finally at full strength after Week 10.

The absence of opponent adjustments does a lot to explain why our ratings currently have New England second and, even stranger, Minnesota fourth. The Patriots blew out a Tennessee team that looks to be pretty poor, and then lost a close game to an Arizona team that has no offense. The Vikings had a close win over a bad Jacksonville team (a game that has a bigger VOA spread than the final score would otherwise indicate), then lost a close game to an Indianapolis team that has a nice quarterback but is otherwise rebuilding. In the middle, San Francisco has two convincing wins over two teams that made the playoffs last year. When the time comes to account for schedule strength, the 49ers will likely see their ratings for Weeks 1-2 move up, and the Patriots and Vikings will see their ratings move down. (No. 1 Houston has benefited from an easy schedule too, of course, whipping Miami and Jacksonville.)

The other reason why the Patriots and Vikings are ranked so high is that their losses are not seen as bad performances by the DVOA system. This week's Any Given Sunday column talks more about the Patriots' weird loss to Arizona, but Minnesota also had a game this week where it lost despite a higher VOA than its opponent, Indianapolis. The Vikings were at 28.9%, and the Colts at -3.7%. The Vikings lost even though they gained 0.6 yards more than the Colts per play. That doesn't include special teams, where the Vikings had 162 yards on returns and the Colts had just 30. The Vikings lost the turnover battle one to zero, but really that's not too bad. Put this game together with the Week 1 win over Jacksonville, and you get the Vikings in fourth place. Given our preseason projections, we see this as a bit of a fluke; the Vikings are still only 20th in DAVE, our rating that combines games so far with the data from the preseason forecast.

Two games in, you're just going to end up with some weird numbers, but this year seems particularly odd. For example: San Francisco being third is not a surprise, but the way they get there is. The 49ers currently rank first in the league in offensive VOA, but are just average on defense and special teams. There's that schedule strength issue again.

The 49ers are still only 13th in the DAVE ratings. Obviously, that seems too low to all of us. As I noted on Twitter earlier this week, "historical trends apply more often than not" still leaves a lot of room for "not." However, we don't know for sure that this is "not." A great example of a similar team would be the 2002 New England Patriots, the team that inspired the creation of Football Outsiders in the first place. This was a team coming off a year where they had gone from 5-11 to 11-5 and won the Super Bowl with a bunch of veteran free-agent defenders and a young quarterback who was not yet Tom Brady. If Football Outsiders existed in 2002, this is the kind of team where we would have been screaming "look out for regression!" before the season. And we would have looked like complete morons after the Patriots started the year by beating Pittsburgh at home 30-14 and then the Jets 44-7 on the road. Like with this year's 49ers, both opponents were coming off playoff appearances. After two weeks, the Patriots were second in the NFL with 65.5% VOA, behind only San Diego. They won their third game, 41-38 over Kansas City in overtime, and then the regression finally showed up. Starting guard Joe Andruzzi tore the meniscus in his knee. Daniel Graham dislocated his shoulder. Troy Brown had knee problems. The Patriots lost a big showdown game against the Chargers to start a four-game losing streak. By the end of the year, they were 9-7 and missed the playoffs on a tiebreaker. Weeks 1 and 2 were their best performances of the year.

As long as we're comparing current teams to teams from the past, how about those Arizona Cardinals? Look at the ratings, and you'll see them way down there at 24th despite being undefeated. The other five 2-0 teams all rank 11th or higher. The Cardinals are only the fifth team in DVOA history to rank 20th or worse after a 2-0 start. Looking back at teams with similar starts, it's surprising to see how many of them continued to win despite unimpressive DVOA ratings. Here is a look at the ten worst 2-0 teams by VOA ratings, and how they ended the season. The worst 2-0 team ever actually put things together, finished the season on a run, and went on to the Super Bowl.

Worst 2-0 Teams since 1991 by VOA after Week 2
    After Week 2 Season Final
Year Team W-L VOA Rank W-L VOA Rank
2003 CAR 2-0 -36.9% 28 11-5 0.6% 16
2004 JAC 2-0 -23.3% 28 9-7 1.7% 13
2012 ARI 2-0 -16.1% 24 -- -- --
2007 SF 2-0 -13.8% 22 5-11 -33.4% 31
2000 MIN 2-0 -13.7% 22 11-5 -6.3% 22
2006 MIN 2-0 -7.3% 19 6-10 -12.9% 23
2005 WAS 2-0 -7.0% 17 10-6 21.4% 7
1998 PIT 2-0 -5.8% 19 7-9 -1.2% 16
2001 IND 2-0 -3.5% 17 6-10 -11.3% 23
2009 SF 2-0 -2.9% 19 8-8 -1.2% 20
1996 BUF 2-0 -1.9% 16 10-6 2.4% 15

All stats pages are now updated with 2012 data, including the first tables for offensive line and defensive line. The last table to be added for 2012 will be "defense vs. receivers," which we will add sometime this week. The FO Premium splits database should be updated with 2012 data sometime this evening.

Make sure to also check out our brand new SNAP COUNTS page! The NFL is finally making snap counts publicly available, and we're counting them up for you and posting them free. The guys from Juice Analytics have been nice enough to design some excellent graphics for us which allow you to dynamically filter and sort through the snap data. Right now there is a table for each week, but we're working on a table that will combine all individual weeks and also show totals.

And now, I'm going to go start answering e-mails and working on a mailbag column. I mean it this time. Really.

* * * * *

These are the Football Outsiders team efficiency ratings through two weeks of 2012, measured by our proprietary Defense-adjusted Value Over Average (DVOA) system that breaks down every single play and compares a team's performance to the league average based on situation in order to determine value over average. (Explained further here.)

OFFENSE and DEFENSE VOA are adjusted to consider all fumbles, kept or lost, as equal value. SPECIAL TEAMS VOA is adjusted for type of stadium (warm, cold, dome, Denver) and week of season.

There are no opponent adjustments in VOA until the fourth week of the season, which is why it is listed as VOA right now rather than DVOA. As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.

DAVE is a formula which combines our preseason projection with current VOA to get a more accurate forecast of how a team will play the rest of the season. Right now, the preseason projection makes up 75 percent of DAVE.

To save people some time, please use the following format for all complaints:

<team> is clearly ranked <too high/too low> because <reason unrelated to DVOA>. <subjective ranking system> is way better than this. <unrelated team-supporting or -denigrating comment, preferably with poor spelling and/or chat-acceptable spelling>

1 HOU 62.2% 7 19.0% 4 2-0 20.8% 3 -60.1% 1 -18.7% 30
2 NE 39.1% 3 38.2% 1 1-1 22.3% 2 -24.0% 3 -7.1% 26
3 SF 36.4% 12 3.9% 13 2-0 37.6% 1 -1.2% 15 -2.5% 22
4 MIN 34.2% 8 -4.6% 20 1-1 17.8% 5 -4.8% 13 11.6% 4
5 BAL 34.0% 2 16.3% 5 1-1 12.6% 8 -9.6% 10 11.8% 3
6 ATL 33.8% 5 24.6% 3 2-0 15.9% 6 -8.5% 11 9.4% 7
7 GB 30.7% 14 25.6% 2 1-1 2.6% 16 -18.3% 5 9.8% 5
8 NYJ 24.4% 6 13.1% 6 1-1 13.7% 7 -1.9% 14 8.8% 9
9 DEN 24.3% 1 10.4% 8 1-1 6.1% 12 -17.4% 6 0.8% 17
10 SD 20.3% 15 2.2% 15 2-0 7.3% 11 -6.5% 12 6.4% 12
11 PHI 16.2% 11 7.7% 11 2-0 -20.4% 29 -50.4% 2 -13.8% 27
12 CAR 12.5% 22 9.1% 9 1-1 19.7% 4 6.0% 22 -1.2% 20
13 SEA 12.4% 16 -8.2% 21 1-1 -20.1% 28 -20.7% 4 11.9% 2
14 CHI 7.2% 4 12.8% 7 1-1 -18.0% 27 -15.8% 7 9.4% 6
15 TB 1.3% 10 -3.5% 18 1-1 1.1% 17 0.2% 17 0.4% 18
16 MIA -0.1% 29 -8.7% 23 1-1 -12.6% 24 2.4% 19 14.9% 1
17 DET -1.5% 13 -3.8% 19 1-1 -0.3% 18 9.0% 23 7.8% 10
18 CLE -5.6% 23 -18.4% 29 0-2 -17.9% 26 -14.0% 9 -1.7% 21
19 BUF -6.0% 30 0.5% 17 1-1 -1.6% 19 5.4% 21 1.0% 16
20 DAL -8.3% 9 0.7% 16 1-1 8.2% 10 1.8% 18 -14.7% 28
21 STL -10.1% 18 -37.7% 32 1-1 -12.4% 23 -1.0% 16 1.3% 15
22 NYG -11.4% 26 3.3% 14 1-1 5.8% 13 17.3% 26 0.1% 19
23 NO -15.1% 17 7.6% 12 0-2 12.5% 9 34.2% 30 6.7% 11
24 ARI -16.1% 21 -18.9% 30 2-0 -28.0% 32 -15.0% 8 -3.2% 23
25 WAS -20.4% 20 -8.3% 22 1-1 4.5% 14 4.8% 20 -20.2% 31
26 IND -25.2% 25 -16.5% 28 1-1 -10.5% 21 11.4% 24 -3.4% 24
27 PIT -27.5% 31 8.5% 10 1-1 -11.4% 22 19.8% 27 3.6% 14
28 OAK -33.1% 19 -13.5% 26 0-2 3.0% 15 14.4% 25 -21.7% 32
29 CIN -41.1% 32 -12.4% 25 1-1 -2.6% 20 47.3% 32 8.9% 8
30 TEN -41.9% 27 -11.4% 24 0-2 -24.8% 31 21.1% 28 4.0% 13
31 JAC -50.8% 24 -22.0% 31 0-2 -23.4% 30 21.6% 29 -5.8% 25
32 KC -70.2% 28 -15.0% 27 0-2 -14.3% 25 38.2% 31 -17.7% 29

Posted by: Aaron Schatz on 18 Sep 2012

159 comments, Last at 25 Mar 2013, 7:08pm by yoga studio chico


by TheSlinger :: Tue, 09/18/2012 - 6:51pm

The Vikings blew out a Jacksonville team that looks awful

26-23 in OT is a blowout?

by Aaron Schatz :: Tue, 09/18/2012 - 7:16pm

True. I forgot that VOA liked Minnesota's game more than the final score indicated. I'll go edit that.

by jimm (not verified) :: Tue, 09/18/2012 - 7:25pm

I suspect this will be the highlight of the season for me as a Viking fan. I did notice they were beating their opponents in yds/play by quite a margin and their special teams were amazing so I figured they would look good in DVOA. I suspect this week will bring them back to earth but would love to be wrong.

by ODBvernon :: Tue, 09/18/2012 - 8:03pm

I have to ask then, watching the Colts game did you feel like they were outplaying the Colts and just missing opportunities? As a Colts fan, aside from the rush offense and passive play-calling on defense in the 4th quarter, I thought the Colts prettily handily controlled the game - particuarly at the end of the first half/beginning of the second and on defensive 3rd downs (which I believe carry more weight in VOA).

I thought the VOA would be fairly close given the 4th quarter, but never, ever imagined seeing this big of a gap in favor of the Vikings. I'm curious if my "eye test" was completely broken in this game being a Colts fan and I overlooked the fact that Vikings greatly outplayed the Colts.

by Anonymous12313 (not verified) :: Tue, 09/18/2012 - 9:08pm

I'm a Colts fan and I thought the Vikings definitely played better.

by Revenge of the NURBS (not verified) :: Wed, 09/19/2012 - 8:18am

I'm in the same boat. I felt like the Colts were playing better for most of the game, up until the 4th quarter, at which point things were clearly going Minnesota's way.

I will say, though, that at least some of that was likely a result of penalties. The Vikings committed some bad penalties at some really bad times. For example, roughing the punter. In my mind, that's essentially a turnover. Same for the late hit on Luck when it would have been about 4th and 13. I wonder how DVOA treats these penalties.

I also wonder how DVOA treats a TD pass that gets batted in the air at the line and knocked around before finding its way into a friendly receiver's hands. I assume it treats it as a standard TD pass, which it was definitely not.

by jimm (not verified) :: Wed, 09/19/2012 - 10:01am

I thought the Vikings were playing a very good DVOA game...they were moving the ball very well in 5 yard chunks - where as the Colts were getting in third and long and converting. DVOA doesn't like such things. Also the Colts turtled in the second half. I didn't come away thinking the Vikings were a better team..I suspect they are both about the same - bottom 10 in the league.

The thing I took away most from that game was that Luck is going to be a great QB.

by Raiderjoe :: Tue, 09/18/2012 - 7:41pm

Yes Vikings crappish. Going to start losing sooon enough..

Cannot argue with Raiders rnalkingg. But tema clung to rebound . Will be 2-2 before you blink an eye. Head coach great up and cominger, C. Palmer still working out kinks with new rwcievers. Figure 11-5.

by DowntownChico :: Tue, 09/18/2012 - 8:10pm

Wait...what?! I can't understand from your "english", but I think you just stated that Raiders are gonna go from 0-2 to 2-2, and go 11-3 throughout the rest of the season to finish at 11-5.

The next three Raider games are vs Pittsburg, @ Denver, @ Atlanta. If they start the season 1-4 from here, that will be performing above expectation.

This is an analysis website, not a Raider Nation website. Let's back up bold statements such as this with rational reasoning.

by tally :: Tue, 09/18/2012 - 8:20pm

Somebody let himself get trolled.

by Paddy Pat :: Tue, 09/18/2012 - 8:53pm

RaiderJoe is no troll. His command of a keyboard is questionable, and he makes errant statements about the strength of his home team, but few people have such a strong working knowledge of players from the past. He's actually a well-informed commentator, and I would take him over Aikman or Simms any day!

by Anonymoussdf (not verified) :: Wed, 09/19/2012 - 12:11am

The fact that hes well informed doesn't keep him from being a troll

by tuluse :: Wed, 09/19/2012 - 12:22am

True, but I don't think he's doing anything to intentionally rile people up either. At worst he's doing a bit, but it's harmless.

I choose to believe RJ is exactly what he looks like. Just a guy who loves football, the Raiders, Sierra Nevada, but not spelling.

by Pied :: Wed, 09/19/2012 - 8:33am

You can always easily identify how long someone has been an FO reader by how they respond to Raiderjoe.

by bravehoptoad :: Wed, 09/19/2012 - 12:33pm

That's both a soothing and interesting way to interpret anti-Raiderjoe posts. Why do those always come across as so snotty?

by DowntownChico :: Wed, 09/19/2012 - 1:53pm

Ummm... this topic quickly got off on a tangent.

I made no inferences about RJ or his credibility. I was attacking his beliefs, not him as a person. (well, that doesn't sound much better...)

The topic is (was) about the Raiders rebounding from 0-2. I think he said they'd go 2-2 from here, and then 11-3 r.o.s. to end 11-5.

My point is their schdule does not dictate that. Even if they did play exceptionally better than they have been, (which I don't believe will happen), "The next three Raider games are vs Pittsburg, @ Denver, @ Atlanta". "If they start the season 1-4 from here, that will be performing above expectation."

I'm sure RJ is tickled by everyone coming to his defense. I know this is America, but let's not argue about the man behind the comments, let's debate the comments themselves.

by Marko :: Wed, 09/19/2012 - 5:06pm

"Ummm... this topic quickly got off on a tangent."

That's how we roll here. Wait until you read the comments on a MMQB column.

As for Raiderjoe, I echo the comments of the other longtime readers/commenters. His knowledge of football history is encyclopedic and astounding. His analysis of teams other than the Raiders and their division rivals is astute.

As for his analysis of current Raiders teams, it's obviously not meant to be taken seriously. Trying to argue with or debate him about the current state of the Raiders therefore is pointless. No one takes his comments about the Raiders seriously. As someone said below, he is like a performance artist, and his comments about the Raiders and their division rivals is for entertainment purposes only. As for his spelling and typing, it's also obvious that his "typos" are intentional and for comedic effect.

So just chuckle when you read his comments about the Raiders. Enjoy the misspellings when he talks about the Clots, Beras, Loins, Pates, etc. Have fun trying to interpret what he is saying when there are so many typos that it's indecipherable as written. And you might learn something along the way when he talks about NFL history or provides comments on other teams.

by Guest789 :: Wed, 09/19/2012 - 8:08pm

All of this. I've been reading for 3 years, commenting for 1, and Raiderjoe has always been my favourite.


“Treat a man as he is, and he will remain as he is. Treat a man as he could be, and he will become what he should be.”

by BigCheese :: Wed, 09/19/2012 - 10:31pm

I made no inferences about RJ or his credibility. I was attacking his beliefs, not him as a person. (well, that doesn't sound much better...)

Actually that sounds A LOT better. Attacking people's beliefs and opinions and not the people themselves is what we should all strive to. We're not all going to agree all the time (nor should we. How else could we learn something?), and in those times, attacking the other view-point, and defending ours form such attacks is what should happen and how a good debate that can grow knowledge happens.

The fact that it has somehow become taboo to attack others' beliefs (I think this might stem mainly from when they are religious, which some poeple think should get a pass), and that some people take these as personal attacks (and respond with actual personal attacks) is a hinderance to good discourse.

- Alvaro

Phil Simms is to analysts what Ryan Leaf is to NFL QBs

by Dan Slotman :: Wed, 09/19/2012 - 2:05pm

Raiderjoe is great. He is easily my favorite commentator. For a long time, people thought that Raiderjoe was a sockpuppet for Mike Tanier, though I think that's pretty well put to rest now.

by jds :: Wed, 09/19/2012 - 9:26pm

I haven't given up on that theory yet.

by JetFanInMD (not verified) :: Thu, 09/20/2012 - 9:07am

Really, when did Tanier die?

That said there was a Jets board where RaiderJeaux used to post until the locals got tired of his shtick and I don't imagine that Tanier would waste time trolling random internet message boards.

by Raiderjoe :: Thu, 09/20/2012 - 1:21pm

Iss different guy. Also saw Raiderjoe guy on abroncos sute

by Sophandros :: Wed, 09/19/2012 - 12:50pm

"He's actually a well-informed commentator, and I would take him over Aikman or Simms any day!"

I don't think that you realize just how awesome that would be.

Sports talk radio and sports message boards are the killing fields of intellectual discourse.

by Mr Shush :: Tue, 09/18/2012 - 8:50pm

Welcome to Raiderjoe.

You may be surprised to learn that 1. When it comes to teams other than the Raiders, RJ's assessments tend to be decidedly astute and 2. The guy's knowledge of football history is quite simply astounding. Just let the homerism and chat-unacceptable spelling wash over you.

by young curmudgeon :: Tue, 09/18/2012 - 10:49pm

It's time for my annual comment that Raiderjoe is not actually the alcohol-sodden oaf that he can appear on superficial reading, but is actually a brilliant performance artist.

by dmstorm22 :: Tue, 09/18/2012 - 10:52pm

He could still be an alcohol-ridden oaf. Just one with very high tolerance except for motor functions (hence the bad typing).

by DRohan :: Wed, 09/19/2012 - 8:23pm


by RickD :: Tue, 09/18/2012 - 11:35pm

For a while last year, after Al Davis died and raiderjoe was not heard from for a few days, there were some who thought raiderjoe was Al Davis's nom de plume.

by Pat (filler) (not verified) :: Wed, 09/19/2012 - 8:41am

Really, does anyone think dying would stop Al Davis?

by Hurt Bones :: Wed, 09/19/2012 - 9:07am

I am so ready for the irrational "Al didn't die, he's alive and living in Tracy/Al was already dead, his return again from the undead is imminent" thread.

by EricL (not verified) :: Thu, 09/20/2012 - 4:35pm

Santa Cruz. The undead are always in Santa Cruz.

by Independent George :: Wed, 09/19/2012 - 10:22am

I agree completely. And once you get past the silver & black tinted glasses, he's also a heck of a football commentator.

by FrontRunningPhinsFan :: Wed, 09/19/2012 - 12:07pm

Let's not forget that:

A) No matter how much he misspells words, he spells player names correctly. Every time. When very few knew who Asomugha was, he was letting everyone know how great he was, and spelling his name correctly all the while. Same goes for all the historical players he mentions.

B) Sometimes the spelling is humor in and of itself. I remember one time years ago he was saying the Raiders would be great on defense and started naming a few players. It went something along the lines of: "Raiders have [player x] [player y] nad Shields."

Nad shields! Genius! It still makes me laugh to this day.

It gets restated every season: RJ deserves his own column here.

Fire Jeff Ireland.

by bravehoptoad :: Wed, 09/19/2012 - 12:46pm

I can only think of them as the Clots now....

by Mike B. In Va :: Wed, 09/19/2012 - 1:28pm

Indeed. I always mentally re-arrange the letters no matter what I'm reading now.

by Marko :: Wed, 09/19/2012 - 4:48pm

"No matter how much he misspells words, he spells player names correctly. Every time."

That's not true at all. I know he usually spelled JaMarcus Russell's name incorrectly. And some others.

But I agree with the rest of your comment.

by DraftMan (not verified) :: Wed, 09/19/2012 - 6:31pm

He only spells _player_ names correctly. By that policy, does JaMarcus really count?

by dharrell :: Wed, 09/19/2012 - 3:04pm

Raiderjoe combines the Raider-fanaticism of Al Davis, the typing ability of a horse (maybe a Clot?), and the encyclopedic knowledge of Tanier into one glorious internet persona.

by McLuvin1983 (not verified) :: Wed, 09/19/2012 - 1:52pm

Another idiot homer blathering things that have no basis in reality.

Gotta love these ESPN forums-oh wait.

by McLuvin1983 (not verified) :: Wed, 09/19/2012 - 1:53pm

Your gimmick isn't funny and the raiders are trash. Wake up and smell the oakland.

by KB (not verified) :: Wed, 09/19/2012 - 5:24pm

I believe that is the weed you are smelling.

by Jerry :: Wed, 09/19/2012 - 5:52pm

Thank you for reminding us what a troll post really looks like.

by Insancipitory :: Tue, 09/18/2012 - 7:42pm

The Seattle Seahawks are clearly ranked too low because of their uniquely physical style of play in every phase of the game. A weekly census of opposing fans ranting on youtube is way better than this. U kno wit that electirc and SMART rook, str8 THUG defense, and BEASTMODE they are gonna bring it every week!!1eleventy!!

not as cathartic as i expected. :c

by Karl Cuba :: Tue, 09/18/2012 - 8:24pm

I dunno, my rant last week calmed me down and might have sprinkled magic 49er dust on Aaron's DVOA computer. 3rd without opponent adjustments, I'll take that.

by Insancipitory :: Tue, 09/18/2012 - 9:35pm

Well, if my Hawks do to the Packers what they did to the Cowboys, I probably won't be terribly invested in what DAVE does or doesn't of them that week. So, here's to cautious optimism!

by RickD :: Tue, 09/18/2012 - 8:04pm

The New England Patriots is clearly ranked (too high) because (Bill Belichick is intent on under-utilizing his best wide receiver). (Chimpanzee poo flinging) is way better than this. <:P>

by lefty :: Tue, 09/18/2012 - 8:07pm

Something is wrong with the DVOA; it's not putting the Eagles absurdly high yet!

by TimTheEnchanter (not verified) :: Tue, 09/18/2012 - 9:42pm

Although, that Defensive rating is well on its way! Before long Reid's DVOA-optimized offense should get rolling.... When do they play the Skins?

by Pied :: Wed, 09/19/2012 - 8:37am

Where was that like button for FO comments? I'm trying to push it here...

by bingo762 :: Wed, 09/19/2012 - 1:31pm

Philly beat Cleveland(ranked 18th). Baltimore beat Cincinatti(ranked 29th). Philly beats Baltimore. Philly ranked 11th, Baltimore ranked 5th.

by DowntownChico :: Wed, 09/19/2012 - 1:58pm

Please tell me if I'm wrong, but VOA measures outcomes of plays, not outcomes of games.

by tuluse :: Wed, 09/19/2012 - 2:06pm

It does.

by bingo762 :: Wed, 09/19/2012 - 2:38pm

I was joking. See the comments before mine

by Vicious Chicken of Bristol (not verified) :: Tue, 09/18/2012 - 8:20pm

NYJ = 8. PIT = 27.

PIT crushed NYJ.

DVOA = Junk math

by Mr. Obvious (not verified) :: Tue, 09/18/2012 - 8:38pm

Its currently VOA not DVOA. Read first, then comment.

by JimmyOz (not verified) :: Tue, 09/18/2012 - 9:11pm

I think this early it's pretty useless. It's interesting from a betting perspective to have performances objectively quantified cos i can't watch all the games each week, but i'm just as apt to take a subjective opinion at this point in the year.

is clearly ranked because . is clearly ranked because . is clearly ranked because .

is way better than this.

by JimmyOz (not verified) :: Tue, 09/18/2012 - 9:20pm

Hey JimmyDickhead - those chevron arrow pointy brackets greater than less than signs whatever the fuck they are... they fuck shit up. Anyways, as i was not saying....

DAVE ignores players changing teams & coming back from injury & coach changes. VOA doesn't adjust for opponent, VOA & DVOA has small sample size.

by jimbohead :: Tue, 09/18/2012 - 10:05pm

DAVE is a linear combination of current DVOA and pre-season projections, as the blurb says. Preseason projections definitely include variables for significant free agent signing, injury luck, and I believe they've done some research on coaching changes (though, my recollection is that coaching changes do very little that's statistically significant; bad teams fire coaches and improve, but its not clear whether that improvement is greater or less than what would be normal regression to the mean).

Check your facts bro, and clean up the language. There are children on this site. Maybe. ;D

by Jimmy Oz :: Wed, 09/19/2012 - 12:46am

Rubbish. DAVE has no concept of how to judge whether outlier performance is sustainable. DAVE mainly looks at DVOA & DVOA change and takes an uneducated guess based on what some dude programmed into an algorithm for what they think are matching scenarios. Like all the FO stats, they're just an averaged history of what's considered a similar circumstance.

by Boo-urns (not verified) :: Wed, 09/19/2012 - 9:31am

That's right. That's why they were so wrong on the 49ers last year, and again so wrong on them this year. Harbaugh is about as far from Singletary as possible, so incorporating Singletary-era stats is essentially meaningless. Common sense > DAVE here. Maybe the 49ers get hit with a huge rash of injuries (which was not the totality of the regression claimed by Schatz et al.), but it's already clear that DAVE was wrong.

by Karl Cuba :: Wed, 09/19/2012 - 9:41am

Shutup Shutup Shut up! FOMBC! Shhhhh!

by bravehoptoad :: Wed, 09/19/2012 - 12:34pm

Karl, don't you know it's bad luck to be superstitious?

by Karl Cuba :: Wed, 09/19/2012 - 2:22pm

Well that leaves me without an out doesn't it?

However, the existence (or non-existent supersitition) of the FOMBC does serve to reduce the number of 'DVOA is wrong. Aaron hates thur ninerz who R the gretest team evur!' style nonsense.

The prediction their numbers came up with were based on clear reasoning that stood up. If the 49ers do go on to exceed their projection then that would be a cause to look to improve the model but it is too soon for carping about FO's possible failure. Injuries could still hit the defense, turnovers haven't been occuring at last season't rate and Smith is going to start throwing interceptions again at some stage. If that doesn't happen then we can try and work out why not, which would be interesting.

So if I can use the FOMBC's voodoo to stop each and every message board from descending into '06 Falcons style dreckfests then I will.

by Boo-urns (not verified) :: Thu, 09/20/2012 - 12:34am

Straw man and a rather stupid one at that. I didn't say the 49ers were awesome, but rather made the statement that incorporating previous year stats into estimates of how the 49ers would do this year, without accounting for coaching changes (which I explicitly stated) and without fully accounting for personnel changes (which I did not), is stupid. Common sense could tell us that the 49ers were likely to be a lot better on offense this year (full season under the new scheme, lots of talent brought in) and that increases in turnovers were likely t be matched by increases in red zone efficiency. Quant models have value but only so much value as we construct them with. See, e.g., financial crisis. Whether or not the 49ers experience an abnormally high turnover rate and/or injury experience doesn't matter, Fo was wrong about its basic outlook. If you don't understand that you don't understand stats 101.

by Karl Cuba :: Thu, 09/20/2012 - 12:11pm

I really wasn't aiming my last post at you, the FOMBC has defeated all who took it on. I also appreciate that this will look silly after my last post.

As for the utility of FO's approach I think that the effects of the lockout will have screwed with any statistical projection derived from last year. As for the new talent, there's one new starting player who wan't on the team last year.

I do think two weeks is too soon to be trashing FO's model but it wasn't posts like yours I was referring to, it's the prospect of many, many ninertrolls showing up if they continue to win.

Edit: To be clear again, not aimed at you but to see what I was talking about then check out the highly cerebral sparklepony in post 128.

by Will Allen :: Thu, 09/20/2012 - 9:46am

You do understand that "The model does not account for all factors, epecially those which do not lend themselves to quantification", is not a terribly insightful statement, don't you? Nor does it mean that a model is "wrong about its basic outloook", unless what you mean is that every single model ever constructed by human beings, to give insight into how a complex system works, has been "wrong about its basic outlook", which, again is not a terribly insightful statement.

by Brice (not verified) :: Thu, 09/20/2012 - 10:43am

When you make bold predictions based on your model, predictions that go against common sense and don't take into account huge factors obvious to the casual observer, you can expect to get called out when those predictions turn out to be fantastically wrong. Maybe the model useful for some things. But it was plain to see that it's predictive usefulness was being overstated. I am sure many were waiting to jump on this. So lie back and take what's coming to you.

by Karl Cuba :: Thu, 09/20/2012 - 12:26pm

The day anyone comes up with a model that predicts the NFL with 100% accuracy is the day that it becomes pointless to watch the games. Does that mean that attempts to model and assess things that you are interested in are a waste of time? I don't think so.

Additionally, FO have been extremely up front in saying that they have some reservations about the projection, look up the preseason staff projections column. They are open and honest in saying that they have an objective method and they stick to it. To begin to make subjective tweaks would devalue the entire project.

Quite a lot of things have to happen for the 49ers to repeat last year's success that have been unsustainable historically. Alex Smith needed to continue his level of play that was an outlier when compared to the rest of his career, the defense would need to stay healthy and to force a lot of turnovers, the offense would need to continue to not give the ball away at a historically low rate and all of this against a tougher schedule. To do a rough calculation, take those five broad variables and give them a 50% chance of coming off and you get a 3% chance of getting all five. Give them a 70% chance and you get a 17% chance of them all occurring, which is quite close to the odds FO gave of the niners being an elite team.

It's also only been two games, a tiny sample size and I'm pretty sure that Aaron and his minions will look at the model if the niners become one of the top seeds. At this stage it is too soon to judge anything, I'd also add that there is no need to be so rude.

by Will Allen :: Thu, 09/20/2012 - 1:32pm

The biggest potential factor I see after two weeks, in terms of where my confidence that the Niners would outperform the FO projection might have been at least a little outsized, is that the 6 division games are looking more difficult than I thought they would be.

by Will Allen :: Thu, 09/20/2012 - 1:28pm

Uh, I have read Aaron on many occasions asserting that there are instances where he doesn't think his model is capturing reality. I am sure that you have created a concept of "bold predictions" in your mind that makes you feel good, for some strange reason.

As for what is "coming" to me, again, and I say this with the deepest respect, that's a really, really, weird thing to write. I'm just a guy who watches football, who thought FO's projection on the Niners was wrong.

On what planet do you live, where it is remarkable that a model, which attempts to predict that which cannot be predicted with anything resembling absolute accuracy, makes erroneous predictions?

by Brice (not verified) :: Thu, 09/20/2012 - 1:41pm

What made the prediction "Bold" in my opinion is that it flew in the face of conventional wisdom. That doesn't mean that it was necessarily wrong. But the justifications for this prediction were pretty flimsy. When you go against conventional wisdom and are very vocal about it, you tend to get more flak when you are wrong. If you go with conventional wisdom and are incorrect, then you don't get as much flak.

The get what is coming to you comment was kind of joking. But it stemmed from the fact that you and a few other posters seem to take criticism of FO pretty personally.

"On what planet do you live, where it remarkable that a model, which attempts to predict that which cannot be predicted with anything resembling absolute accuracy, makes erroneous predictions?"

What use is the model if the predictions are not accurate? Doesn't that make the whole exercise pointless?

by tuluse :: Thu, 09/20/2012 - 1:47pm

Define accurate, because missing on one team out of 32, even if missing badly doesn't seem like an inaccurate model to me. Note, I'm not saying that's how good this model is, or that it is accurate, but you have not provided any other evidence that the model is inaccurate.

by Will Allen :: Thu, 09/20/2012 - 1:53pm

If I am to understand your use of the English language, if a person develops a quantitative model that generates a prediction that is at odds with conventional wisdom, conventional wisdom which is largely not based on quantitative methods, such a prediction is "bold", no matter that the developer of the model has stated on several occasions that his model will at times fail to capture reality. I suspect that you do not employ the word "bold" in the way I do.

I said I thought the Niners FO projection was wrong. Why would I take it personally if someone agreed with me?

Your last question seems to imply that the inability of model to predict with accuracy 100% of the time renders the model useless. This is an extremely strange idea.

by Brice (not verified) :: Thu, 09/20/2012 - 2:08pm

Definition of Bold: "Showing an ability to take risks; confident and courageous." The prediction was bold and that is not a negative thing. The prediction was also wrong.

How accurate is the model? I read an article elsewhere which stated that the FO model is about as accurate as random chance? Can you point me to any FO articles which do post season analysis of preseason prediction accuracy? I am new here.

by Will Allen :: Thu, 09/20/2012 - 2:26pm

Please explain what is risky or courageous in saying "I have developed a quantitative model which I believe is useful, but is imperfect at capturing reality, which is making a prediction that differs with the conventional wisdom on the number of wins a football team will have." I don't mean to beat this into the ground, but I really think you have created something in your mind that does not exist in reality, ironically enough.

If you are saying that you suspect that the FO projections are about as accurate as random chance, why didn't you just come out and say that? I haven't checked on it in a while, but a few years back I remember reading that the FO projections of playoff teams were doing better than various widely published non-quantitatively based playoff team projections.

by Jerry :: Thu, 09/20/2012 - 5:47pm

If the model's predictions looked the same as conventional wisdom, they wouldn't be very interesting. There are typically a couple of "bold" predictions every year, and they turn out to be right often enough that they're worth paying attention to. And, to be fair, FO doesn't run away from the wrong calls.

by bravehoptoad :: Thu, 09/20/2012 - 12:56pm

Are there any intelligent straw men? Besides in Oz, of course.

The trouble with "common sense" is that it only works in retrospect. It's terrible at making predictions. The only real use for it is deflating predictions after the actual results are in. That's also not that difficult to do; there's always a long line of people anxious to throw around some "common sense" after the fact. They're often the kind of people who are happy to use adjectives like "stupid" a lot, too.

So tell us, oh great and powerful Boo-urns, how would you quantify coaching changes in predictive models? All ears.

by Will Allen :: Thu, 09/20/2012 - 1:26pm

First, you get some red shoes......

by Karl Cuba :: Thu, 09/20/2012 - 1:35pm

The 49ers do play in red shoes...

by Will Allen :: Thu, 09/20/2012 - 1:41pm

Well, then, unless I see flying monkeys at The Stick, I say the Niners win at least 12.

by tuluse :: Thu, 09/20/2012 - 1:48pm

Just wait until Harbaugh gets that Gatorade bath and starts melting.

by Brice (not verified) :: Thu, 09/20/2012 - 1:50pm

But most football commentators did predict the 49ers to be one of the top 5 teams in the league before the season started. So in this case, common sense was much better at making predictions than the model.

As far as quantifying coaching changes. I don't know how you would do it. But I do think that the NFL is filled with big, strong and fast men and there is very little difference in the physical abilities of different teams. The most important variable is coaching. That is the reason that New England wins year after year. They don't have better athletes than other teams, they have better coaches. Any model that doesn't quantify coaching changes is doomed to failure.

by Karl Cuba :: Thu, 09/20/2012 - 2:20pm

If you had read the 49ers chapter of FOA then you would have seen Danny Tuccito argue that the 49ers coaching staff could be a reason that they might beat the projection.

It is still too early to know if things like defensive injuries will regress to the mean. This debate would be better left until after the regular season. ack in 2006 there was a debate about how the 4-0 Broncos had beaten the projection, they ended up 8-8 and missed the playoffs.

by bravehoptoad :: Thu, 09/20/2012 - 7:14pm

...and there is very little difference in the physical abilities of different teams.

Now, there's a bold assessment.

I more often hear the opposite -- that the GM is a more important person than the coach because physical ability is harder to find than decent coaching. If there's anyone on the planet who would be skeptical of that theory, it's a 49er fan, but still I find yours equally dubious.

There's another theory making the rounds that contradicts yours, that the teams that win consistently are the ones with great quarterbacks. Recently the teams that buck this trend are rare, so I'm thinking it's a pretty big threat to your own theory.

What we'd need to test your assertion would be a team that routinely was in the playoffs with average players across the board and a great coaching staff. It's hard for me to imagine such a team doing well without any pro-bowlers. They'd have no reliable matchup advantages.

by Brice (not verified) :: Thu, 09/20/2012 - 8:22pm

What makes a great quarterback? Matt Flynn, Matt Cassell and Kevin Kolb all looked like great quarterbacks when they were in great systems with good coaching. When they left those systems they looked average or below average. That is the best illustration of how system and coaching matter above physical ability. So far, Alex Smith in a better system with better coaching is looking like a completely different quarterback.

Winning teams put players into the Pro Bowl. I don't think a player making the Pro Bowl is an objective measure of how good that player is. I am not saying that there is not a difference in ability between different players, only that when you average out the ability levels of all the players on a team, the difference between teams is not that great.

The 49ers are my team, so watching them has influenced my opinion. I have seen the difference between extremely bad coaching and extremely good coaching.

by bravehoptoad :: Fri, 09/21/2012 - 11:21am

Matt Flynn -- sample size too small to tell, and he hasn't played since he left, so wtf?
Matt Cassell -- looked like an okay quarterback on a good team, but never a great one
Kevin Kolb -- really really never looked like a great quarterback.

Counter examples, of obviously great quarterbacks, are numerous. Peyton Manning is a good place to start because of how the team fell apart without him.

I'm a 49ers fan, too, but that doesn't lead me to the conclusion that talent is equally distributed across all teams. Some have way more talent. The 49ers are one of those at the moment; in fact, they were so talented they could go .500 or close to it with the worst coaching in the league.

by Brice (not verified) :: Thu, 09/20/2012 - 9:05pm

I would like to add that the reason teams have pretty much equal talent is that the NFL is run like a socialist country. Good teams are penalized with low draft picks and all teams are held to the salary cap. A team cannot stockpile good players like in college football, where a few top schools get most of the top recruits.

by zenbitz :: Wed, 09/19/2012 - 10:45am

Interestingly, if you ignore opponent adjustments (which you kinda have to on week 2), The niners have regressed significantly on defense and special teams, right where DAVE expected them too. Turnovers, too (+1 in 2 games)

On the other hand, their offense (driven mostly by the OL gelling and RG upgrade) is the opposite of regressing and is actually good (probably not predictable with the kind of data DAVE gets fed).

by bravehoptoad :: Wed, 09/19/2012 - 12:38pm

Why do you have to ignore opponent strength? I understand why VOA has to, but why do you?

by zenbitz :: Wed, 09/19/2012 - 12:53pm

because it's badly underdetermined. "We all know" that Green Bay and Detroit have good offenses and bad defenses. Hence, we expect that the Niners offense VOA will adjust down towards the average and their defense VOA will adjust up to where we think it will be.

But "We all know" is wrong - it's just a guess based on the last couple years, heavily weighted to 2011. This is why FO drops DAVE after a few weeks...at some point, 2012 actual performance is more predictive than previous performance.

And actually all 4 statements (GB offense is good, DET offense is good, GB defense is bad, DET defense is bad) could be wrong (in fact, DAVE doesn't like the Lions offense either)

by bravehoptoad :: Wed, 09/19/2012 - 12:37pm

Wouldn't that be DAVE that was wrong about the 49ers for coaching reasons, and not DVOA?

It's interesting that advancednflstats was also dead wrong about the 49ers for a completely different reason; they don't account for special teams AT ALL.

by zenbitz :: Wed, 09/19/2012 - 1:06pm

You can't make that statement. DAVE could be wrong for any number of reasons -- including random variance. We can suspect it's simply the Harbaugh effect... but there's no way to test that.

For example - if DAVE proves to be right on turnovers/defense/st (as it has so far), but badly wrong on offense - this *could* be mostly due to personnel upgrades at WR and RG along with OL gelling. I think DAVE has a component for OL "history" or "cohesion" or whatever they call it... but it cannot predict very well (I don't think) when a mediocre line gels -- if it's more of a sudden phase transition than a gradual improvement.

To call it the difference "coaching" -- a point I do NOT disagree with, btw -- is just confirmation bias in action.

by jimbohead :: Wed, 09/19/2012 - 10:41am

Ok, so this is a different argument than before. Whereas earlier you were claiming that DAVE made no adjustments for x,y,z, now you're talking about it's lack of ability to effectively evaluate outlier performance, and predict it's sustainability.

Of course! It's a linear regression model! Your characterization of DVOA prediction, as an uneducated guess based on some dude's algorithm for what they think are matching scenarios, is of course way off-base; the approach is as sophisticated as a typical grad project in econometrics. But linear regression models like this have real value; they tell us what to expect on average. As fans, sometimes we need to hear that. And as people, we get to adjust for what we think are extraordinary circumstances. However, extraordinary circumstances do not justify artificially changing the model or how the model is reported, except to add an asterisk that they think this is probably wrong, which they've done pretty consistently this offseason.

Anyways, this 49er fan says, let's just enjoy the ride together.

by Will Allen :: Tue, 09/18/2012 - 8:34pm

The shame of the Vikings loss for Vikings fans is that they are going to find games where they are tied in the last minute increasingly hard to attain. Ponder looks like he might be a middle of the pack qb, which would be much better than many have predicted, but when I look at the schedule, I could still see them easily losing 8 straight to finish the season.

The Niners division schedule looks like it may be a bit harder than was expected, although I'd still anticipate that their defense will make their division opponents' qbs look overmatched.

The Broncos may have just played their most difficult matchup, once all factors are considered.

by Mr Shush :: Tue, 09/18/2012 - 8:55pm

Possible (the Broncos) but given that they still have to play Texans, @Patriots, @Chargers and @Ravens, I think it's a bit early to say.

But yes, while I'm happy to see my team at the top of the VOA rankings, my subjective rankings would have Atlanta first at this point, just ahead of the 49ers and Texans.

by dmstorm22 :: Tue, 09/18/2012 - 9:00pm

A road game in primetime is really hard to win, especially when teh home team is good (as in, not like Jacksonville in their one or two token home night games).

I think @Ravens is the only game left that will be tougher. Maybe @NE, but that place isn't really known for being loud and throwing off communication the way that a game in Baltimore might.

by Will Allen :: Tue, 09/18/2012 - 9:08pm

It's the road game against a very good East Coast (or close to it) opponent, in a very loud venue in front of the overserved Monday night crowd, which constrains a no huddle offense, early on in the Manning era, which makes me think this may have the most unfavorable situation for them. I think at altitude, Manning is going to just gas opposing defenses, especially as the season wears on.

by Mr Shush :: Tue, 09/18/2012 - 9:33pm

Fair comment. I thought before the start of the season that the Broncos had the third best chance of any team in the AFC to be a serious contender, and I don't think my view's changed (though the Chargers look better than I expected, and to a lesser extent the Ravens too).

Minor counter-arguments: the Falcons run a 4-3 (and Manning has historically been less dominant against 3-4s) and they can't run the ball worth a damn (I'm not yet sold on Denver's run defense).

by Will Allen :: Tue, 09/18/2012 - 9:43pm

Nolan was running a pretty nontraditional 4-3 last night, and he has had about as much success against the elder Manning as any d-coordinator. I was surprised to hear last week how much success Manning has had against LeBeau through the years. I guess it is the Pats and Chargers which produce those 3-4 numbers.

by Just Another Falcons Fan (not verified) :: Tue, 09/18/2012 - 10:32pm

Actually, the Falcons spent most if not all of the night in a 4-2-5 with WitherSPOON! (wonder if he would appreciate and/or get a Tick reference) and Nicholas as the linebackers. MLB Akeem Dent only hit the field for special teams and maybe short-yardage/goal line plays.

by LACharger (not verified) :: Tue, 09/18/2012 - 10:36pm

Chargers used to beat Manning-led Colts over and over, but a lot of those games were very close.

(Now that I think about, most of my favorite NFL memories are Chargers-Colts games. Spoiling their perfect season with a Michael Turner 80 yard TD run in the 4th quarter. 6 interceptions in a game including 3 acrobatic ones from a then-backup (albeit 1st round rookie) CB. The game where Sproles returned two kicks for TDs. The playoff game that the SD punter dropped 6 punts inside the 8 yard line or something ridiculous...I've never seen a punter have so much impact on a game. And even more amazing, the playoff game where San Diego lost its QB/RB/TE (two sure-fire HOFs and 1 all-pro, and the three most important players on the team at the time), a set of backups willed them to a game-winning 4th quarter TD drive. Leading to the great FO comment here "QB/RB/WRs are fungible". Almost makes up for having to watch PIT/NE/NYJ beat us every frigging time and knock us out of the playoffs every friggin year (every year we make it that is)).

Sorry for that walk down memory lane.

The point I was going to make is "Manning against a 3-4" is a bit silly. The Pats 3-4 is very unlike the Chargers 3-4. I don't know how other teams beat Manning, but for the Chargers it seemed to be luck (see above) and former ILB Stephen Cooper who just seemed to have the guy's number and would stand near the LOS barking out shit that would drive Manning nuts. I don't know if he was messing with Manning's cadence or translating the play calls, but it was a glorious thing to see.

I don't mind seeing Manning in Denver. Tebow drove us nuts.

by Scott C :: Wed, 09/19/2012 - 12:44am

Oh, Memories!

Thanks for that.

I agree that there was certainly luck involved in many of those wins (Vinateri missing a 25 ish yarder?) But the last time Manning had a good game against the Chargers Drew Brees was the QB and Indy won the shootout. Slowing down Peyton wasn't all luck, but winning some of those games definitely required more than that.

by dmstorm22 :: Wed, 09/19/2012 - 12:53am

So his 400 yard passing day wasn't a good game? Both of his interceptions touched the hands of Colts receivers first. The first one, granted, wasn't a great throw, but the second one on the screen to Kenton Keith was just a complete screw up by Keith.

Also, in 2008 he had 310 yards and no picks in the Wild Card game, facing the worst field position ever for a QB in a playoff game.

His 2008 regular season performance wasn't all that bad either (32/44 255 yds 2/1 td/int - a low ypc/ypa day, but good completion numbers) in a 23-20 win.

He's played bad against the Chargers in recent years twice. Even his game in the 2005 loss wasn't bad (one of his picks was on a drive when teh game was already over).

by Pied :: Wed, 09/19/2012 - 8:44am

God damned robo-punter.
I will never, ever forget that frustrating game.

by Revenge of the NURBS (not verified) :: Wed, 09/19/2012 - 10:22am

Agreed. Being removed from it by the years makes it a little less frustrating, but I'll never forget that game. Scifres was the undisputed MVP of that game. I've never seen anything like it.

by Mr Shush :: Wed, 09/19/2012 - 7:12am

It may be crude - and it may also be out of date - but FO definitely did a study (in a Pro Football Prospectus, maybe?) a few years back which found that Manning's career DVOA against 3-4 teams was substantially worse than against 4-3 teams. That still meant he was really good against 3-4 teams - this is Peyton Manning we're talking about - but there was a significant difference.

by Mr Shush :: Tue, 09/18/2012 - 9:37pm

The Texans defense is very, very good at annihilating bad opponents. It's not yet clear that it can cope with an elite spread offense (witness the Saints game a year ago). The Texans offense is both strong and balanced, with better receivers than a year ago, but worse blocking (though Ben Jones did a nice job rotating with Antoine Caldwell, which could go some way to redressing that imbalance if it continues). The Texans special teams . . . do indeed look pretty horrifying so far.

by RickD :: Tue, 09/18/2012 - 11:38pm

DVOA/VOA tends to favor the teams that can dismantle opponents. Whether that is desirable or not is unclear. In any case, if you're aware of this trend, you can make a mental adjustment as you desire.

by Scott C :: Wed, 09/19/2012 - 12:46am

Well, Dismantling inferior opponents correlates more highly with post-season success than barely beating quality opponents. (GUTS vs STOMPS)

Dismantling quality opponents of course correlates with postseason success the most.

by t.d. :: Wed, 09/19/2012 - 1:47am

i'd like to see that research reevaluated, given the recent wtf-ization of the playoffs. seems to me that any conclusions based on such a limited data set as 'nfl playoff outcomes' is bound to have problems, especially since the size and structure of the playoffs keeps changing, as does the way the game is played.

by Mr Shush :: Wed, 09/19/2012 - 7:44am

Sure, because narrow wins are just quite likely to be flukey. But we're talking about DVOA not margin of victory here. My case is that teams probably have different . . . I dunno, scaling factors maybe, for strength of opponent. I think against weaker offenses, the Texans defense is the best out there. Against stronger ones, there are plenty worse, but also several better. I don't think that's an idea scenario for post-season success.

by Kal :: Tue, 09/18/2012 - 11:00pm

I will have to say that without opponent adjustments YAR looks even crappier than usual.

by t.d. :: Wed, 09/19/2012 - 1:38am

not surprised the Cards grade out poorly. In addition to them being good at all the things dvoa regards as non-predictive events, that quarterback situation is still a big problem. they've got to be worse than the 2000 Ravens on offense (not comparing their defenses), so they have a built in low ceiling

by TomKelso :: Wed, 09/19/2012 - 11:26am

Most teams are worse than the 2000 Ravens on offense -- either by conventional stats or FO stats. Go look it up -- and next time, don't just swallow the usual media narrative.

by Ryan D. :: Wed, 09/19/2012 - 11:44am

The same Ravens offense that went 4 weeks without an offensive touchdown during the regular season? I don't care what the stats say, that's a pretty terrible offense, especially when you consider how many short fields their excellent D and return teams spotted them.

by tuluse :: Wed, 09/19/2012 - 12:50pm

-9.5%, ranked 23rd.

Not awful, but still pretty bad.

by Ryan D. :: Wed, 09/19/2012 - 1:44pm

Correction, FIVE weeks without an offensive touchdown.

by jebmak :: Wed, 09/19/2012 - 10:17pm

Yep. 0 for October. All five weeks.

by Aloysius Mephis... :: Thu, 09/20/2012 - 5:12pm

Their offense wasn't craptasmagoric but wasn't good or average either. Tony Banks was their QB for half the season and when Trent Dilfer stepped in he wasn't much (if at all) better. Running game was good but not dominant. And yet they won two of those five games in October and their worst defeat in that span was an 8-point loss to a Titans team that finished 13-3. I hear their defense was kinda OK.

by AndersJ (not verified) :: Wed, 09/19/2012 - 2:38am

Houston should end up with the best D before adjusting because they are playing 6 games against the 3 of the worst offenses in the NFL

by bubqr :: Wed, 09/19/2012 - 5:12am

he NFC West went pretty quickly from horrible division to home of 3 of the potentially 5 best defenses in football.

by bravehoptoad :: Wed, 09/19/2012 - 12:39pm

Also four of the more physical teams in the league. Once they're all done playing each other, I can imagine all of these teams are going to be beat up.

by Purds :: Wed, 09/19/2012 - 7:52am

Wow, looking on the quarterback stats page I noticed that DYAR and QBR rate some QB's very differently. DYAR loves Tom Brady but QBR hates him, while QBR loves Andrew Luck and DYAR hates him. As a Colt fan, I thought for a second this might just be evidence of FO's love of all things NE, but then I noticed that only one other QB is treated like Luck (high QBR, low DYAR): Matt Cassell. Yuck! I think I'd rather have my rookie QB lumped in the Brady group. Not good.

by dbostedo :: Wed, 09/19/2012 - 8:27am

QBR has some bonuses for things that fans care about, but that don't necessarily show how well a QB played or will play. (I.e. QBR gives extra weight to certain game situations, especially late game, trying to measure "clutch" and related to the awe fans seem to have for 4th quarter comebacks. But that is not necessarily descriptive or predictive of the quality of QB play.)

That's why, while it's better than the older QB rating - much better - it still has some oddities and isn't really done right in my opinion.

by Pied :: Wed, 09/19/2012 - 8:50am

It's only two games, it's only two games...
Manning was crap to start his rookie season too...
/don't hyperventilate don't hyperventilate calm down calm down

by Purds :: Wed, 09/19/2012 - 2:37pm

Yes, pass the paper bag over here, quickly!

by Revenge of the NURBS (not verified) :: Wed, 09/19/2012 - 3:48pm

Good Lord, people. Here's what you should look for from Andrew Luck during this season -- progress. That's it. We didn't draft Andrew Luck to win the Super Bowl this year.

So if you want good news, look at it this way -- he was better in Week 2 than he was in Week 1, by any metric. Not only that, the team as a whole was better in Week 2, again, by any metric. So hey, cheer up. He's already better than Curtis Painter...

by Purds :: Wed, 09/19/2012 - 10:30pm

I hope you realize I was kidding throughout this. We are not going to know much about any rookie through 2 games.

by Roy G (not verified) :: Wed, 09/19/2012 - 9:18am

These ratings look funny as well due to the connectedness of the schedule. Cincinnati, Cleveland, Baltimore, and Philadelphia have only played each other, so they are separated.
Wk 1: Cleveland played Philadelphia, Baltimore played Cincinnati
Wk 2: Cleveland played Cincinnati, Philadelphia played Baltimore

Another separate group is: Buffalo, NY Jets, Kansas City, Pittsburgh, Atlanta, and Denver
Wk 1: Buffalo played NY Jets, Pittsburgh played Denver, and Kansas City played Atlanta
Wk 2: Pittsburgh played NY Jets, Denver played Atlanta, and Kansas City played Buffalo

by A Nonymous (not verified) :: Wed, 09/19/2012 - 9:30am

the jets crushed buffalo. buffalo crushed kansas city. kansas city was also similarly crushed by atlanta, so by corollary, the jets crushed atlanta. atlanta handled denver rather comfortably. denver handled pittsburgh rather comfortably. thus, the pittsburgh/jets game was actually a fluke, and the sample of all other games indicates that the jets are much, much superior to the pittsburgh steelers

by Roy G (not verified) :: Wed, 09/19/2012 - 10:23am

Based on the results of those 6 games (Wins and Loss only, no homefield), the optimal ranking of the teams in a Linear Least Squares sense is:
1. Atlanta, 2. Denver, 3. Pittsburgh, 4. NY Jets, 5. Buffalo, 6. Kansas City
If you rank the teams this way, then there were no "upsets".

If you factor in the scores of the games and add 3 points for homefield, you get:
1. Denver, 2. Pittsburgh, 3. Atlanta, 4. NY Jets, 5. Buffalo, 6. Kansas City

by bravehoptoad :: Wed, 09/19/2012 - 12:41pm


Sadly it looks like the site has been abandoned sometime late last year, but it was super fun while it lasted.

by DowntownChico :: Wed, 09/19/2012 - 2:27pm

Thank you for allowing me to waste time at work looking into this process. It is incredibly interesting.

It looks like the site has move to www.beatgraphs.com and now incorporates more sports leagues.

by Brice (not verified) :: Wed, 09/19/2012 - 9:55am

The 2012 49ers are not the 2002 Patriots, they are the 2001 Patriots. Last year was the rebuilding year. They hired a new head coach who had to bring in his type of players and teach the team his system just like the 2000 Patriots. The 49ers of last year had the disadvantage of no off season due to the strike and were beset by injuries at one key position. They lost their most productive wide receiver, Josh Morgan early in the year and the guy they brought in to be a deep threat, Braylon Edwards, in the preseason. He wasn't they same player the rest of the year that he looked to be in the first preseason game. Then they lost Ted Ginn, who was also their starting kick returner. I seem to remember that having some effect on the NFC Championship game. They actually had to play a significant number of plays with Brett Swain at wide receiver and they almost pulled it off.

The fact that all of this happened in the rebuilding year and they still managed to go 13-3 and within a hair of the Super Bowl is amazing, but this is the year they should reach their potential.

by Coop16 (not verified) :: Wed, 09/19/2012 - 10:22am

Browns fan here (yes, I know, "how can you root for those bunch of ..."; I hear it all the time). Anyway, interesting to see them at #18. All of the "experts" rank the Browns at #32. I'm not going to defend the Browns too strenuously because (a) I'm sick of their constant losing/rebuilding, (b) they're 0-2 as always, and (c) I agree that they've got a way to go before becoming relevant. But still, it irks me to see guys like Pete Prisco ranking the Browns at #32 even though the Browns have played two close games while teams like Kansas City are getting blown out every week. Doesn't make a lot of sense. I put a lot more faith into FO, where objective analysis is done. Even so, #18 is higher than I would expect for the current Browns. I'm not sure whether to be gratified or mystified by the ranking. Anyway, against the Bengals, the Browns' offense showed an actual pulse for the first time since 2007, so I guess maybe things are looking up. Until end of year, of course, when Haslam decides to fire everybody, blow the whole team up and start over.

by bravehoptoad :: Wed, 09/19/2012 - 12:43pm

Yes, as another fan of a team perennially at the bottom for a decade, I noticed that teams at the bottom of power rankings always got a lot less attention than teams at the top. The closer to the bottom, the more the ranking was based on reputation.

by Coop16 (not verified) :: Wed, 09/19/2012 - 1:13pm

You root for a different team that's been at the bottom of the rankings for a decade? Who would that be? Who dares trespass on the clammy basement territory that the Browns have claimed as their own for the past 14 years? I would have guessed Detroit a couple of years ago, but they've crawled out of the basement now. And the Arizona Cardinals, who used to share a basement room with the Browns for years, actually went to a Super Bowl not long ago. (Yeah, although I root for the Browns, I've heard of the Super Bowl. The Steeler fans have described it to us.) So as far as I know, the Browns have the NFL basement all to themselves now.

by Revenge of the NURBS (not verified) :: Wed, 09/19/2012 - 2:15pm

While not quite as bad on the field, I'd say the Bills can give the Browns a real run for the title of most irrelevant team of the past decade.

by bravehoptoad :: Wed, 09/19/2012 - 2:28pm

Exaggerating slightly. 8 years, 2003-2010, no winning records.

by DavidL :: Wed, 09/19/2012 - 10:35am

VOA hates a 2-0 Eagles team?

Holy crap, the world really is ending!

by AndersJ (not verified) :: Wed, 09/19/2012 - 10:55am

It does just not like our ST and the insane amount of turnovers (VOA really hates turnovers). If the Eagles manage to get few turnovers per game there offensive VOA should rise alot.

by Ryan D. :: Wed, 09/19/2012 - 11:12am

"A lot of NFL power ratings around the Internet have the Houston Texans on top after two big wins to start the season."

Who? The power rankings provided at ESPN.com, Foxsports.com, NFL.com, and MMQB all have SF over Houston for the top spot.





by Mr Shush :: Wed, 09/19/2012 - 12:09pm

Texans are also first by combined EPA over at Advanced NFL Stats, for whatever that's worth.

by zenbitz :: Wed, 09/19/2012 - 1:12pm

The most hilarious "power rankings" are the objective SRS at pro-football-reference!

(really just illustrates the funny stuff that happens with Strength of Schedule / Opponents adjustment after 2 weeks)

by Aaron Brooks Go... :: Wed, 09/19/2012 - 5:39pm

Schatz will be happy to see that SRS considers the 49ers the worst team in the NFC West.

by jfsh :: Wed, 09/19/2012 - 11:30am

The #1 passing offense in the league according to VOA so far is... the New York Jets.

Wonder what odds you would have to have given someone before the season to get them to take that bet.

by Nick Bradley (not verified) :: Wed, 09/19/2012 - 11:36am

Actually, NOBODY has the Houston Texans as #1.

I know you feel bad for overemphasizing the plexiglass principle with the 49ers, but this is absurd.

by bravehoptoad :: Wed, 09/19/2012 - 2:17pm

Not nobody.


It's not like I'm the most widely-read guy on the internet, either.

by Ryan D. :: Wed, 09/19/2012 - 2:45pm

The combined overall ranking still has a tie at the top between SF and HOU. Of the five ballots scored, SF is ranked higher than Houston on 3 of the 5 ballots.

by bucko (not verified) :: Wed, 09/19/2012 - 7:46pm

Just glad the Green Bay defense is not at the bottom. If the defense can be decent and the special teams adequate I will keep the faith that the offense will get untracked.

by Valarius (not verified) :: Thu, 09/20/2012 - 10:50am

Aaron, do you know why you were wrong about San Francisco? It's because you failed to look beyond the numbers and to what caused them.

The 49ers tied for the lead league in turnovers because they are more physical than any other defense, and the opposing QBs are skiddish as a result, while skill players end up feeling like they have to make every play count to make up for the erratic quarterback play, and thus take chances they wouldn't normally, resulting in fumbles.

This was not luck, it was a result of the physical nature of the defense.

As for their relative lack of turnovers, this was again due to far more than statistical happenstance. Their offensive scheme is focused on ball control, which obviously limits the possibility of turnovers, and their quarterback for the last two years has been extremely cautious in when and where he throws the ball.

These things, while appearing as statistical anomalies in light of the history of the team, are easily explained in actual changes that have occurred in the roster and coaching staff. That is, the components that make up the team are the components that correlate greatly with high turnover differential. Thus a more informed analyst would predict the opposite of what you did regarding the return to the mean. That tendency assumes that other variables remain constant, and in this case the relevant variables have changed between 2008 and 2011, and they remained the same between 2011 and this season.

by Karl Cuba :: Thu, 09/20/2012 - 12:08pm

Two games is clearly enough to call it for the year, it's not too early at all. I agree Aaron, it's time to close the site down and go back to spinning Radiohead tracks the sheer weight of evidence is crushing you.

It doesn't even matter that if the 49ers were on pace to force as many turnovers as last year then they would currently have five and only have two, that's not any sort of 'evidence' of regression to the mean is it?

"Thus a more informed analyst would predict the opposite of what you did regarding the return to the mean." - That's just special.

And it's skittish.

by bravehoptoad :: Thu, 09/20/2012 - 1:19pm

How are you measuring physicality? If the 49ers were more physical than any other defense, and that was what lead to more turnovers, then why did they only tie for the turnover lead? Is Green Bay the second most physical defense, since they tied for first in turnovers? Weren't the 49ers just as physical in 2010, with almost all the same players?

If their more physical play was causing quarterbacks to be more skittish than any other team, why were they tied at 7th to 9th for number of sacks? Wouldn't teams with more sacks cause quarterbacks to be more skittish?

Why did other teams that focus on ball control have more turnovers?

Given how easily explained their turnover differential is, are you predicting +28 or better again this year?

by commissionerleaf :: Thu, 09/20/2012 - 3:35pm

If the 49ers were more physical than other teams, you'd expect them to suffer more injuries. However, their injury history is above average over the last few years, especially on defense. They'll be a good defense until Smith regresses or is injured. Virtually everything they do is based on the fact that he is a force on every play. It allows them to play seven and eight in coverage (meaning Willis and Bowman, good cover linebackers), and to leave seven or six defenders in the box without permitting efficient running. The only single defenders in the league close to as important for their teams are Freeney and Wilfork.

by bravehoptoad :: Thu, 09/20/2012 - 7:16pm

Agree...I think it's one reason the 49er defense wore down last week -- Justin Smith getting tired at the end of the 4th. I'm hoping they'll start giving him a blow more often.

by Mr Shush :: Fri, 09/21/2012 - 6:33am

Really? Not Revis? Or Peterson, for that matter? I dread to think what adjustments the Texans would have to make if Joseph went down, and while he's a fine player, he's not in the same category as those two.

by dmstorm22 :: Fri, 09/21/2012 - 10:10am

I'm actually interested to see in what happens to the Texans D if Joseph gets hurt.

To me, the Texans defensive improvement is as tied to the pass rush and guys like Barwin, Reed and especially Watt, as much as the secondary improvement.

by yoga studio chico (not verified) :: Mon, 03/25/2013 - 7:08pm

My spouse and I stumbled over here from a
different page and thought I might check things out.

I like what I see so now i'm following you. Look forward to looking at your web page again.