Writers of Pro Football Prospectus 2008

Most Recent FO Features


» 2018 Free Agency Cost-Benefit Analysis

Is Kirk Cousins the best free-agent quarterback in recent memory? Should Trumaine Johnson or Malcolm Butler have gotten the larger contract? And what makes a free-agent contract good or bad, anyway?

04 Feb 2013

Super Bowl Quick Reads

by Vince Verhei and Aaron Schatz

For the second straight year, the team that lost the Super Bowl came out better in our metrics.

DVOA (with opponent adjustments)
BAL 24% 27% 11% 8%
SF 26% 45% 3% -16%
VOA (no opponent adjustments)
BAL -7% 16% 31% 8%
SF 21% 44% 7% -16%

This is what's going to happen when you have great, close games. One team will get a few more of the breaks than the other team, or do things that are great but don't score as highly in our DVOA system because they aren't necessarily sustainable over time.

For example, Jacoby Jones' 56-yard touchdown doesn't score as high in DVOA as a 28-yard completion followed by a 28-yard touchdown, even though the result is the same on the field, because the latter set of plays is more indicative of repeatable success. And Jones gets penalized for his muffed punt; not much, because muffed punts aren't often recovered by the punting team, but there is a penalty there. In reality, of course, we know that the 49ers had no chance at all to recover that muff because nobody from the 49ers was anywhere near Jones when he dropped the ball. The Ravens' special teams also take a hit on punts, because Sam Koch put two punts in the end zone instead of landing them inside the 20 (one from his own 44, one from the San Francisco 42) and on the other punt, Ted Ginn had a 32-yard return.

Actually, San Francisco coming out with the higher VOA doesn't look quite as strange when you consider that the 49ers actually averaged 2.6 yards more per play than the Ravens, with more first downs and the same number of turnovers. That Jacoby Jones touchdown return was a huge play.

Other teams that lost the Super Bowl despite a higher single-game VOA rating included the 2011 Patriots, the 2008 Cardinals, the 2005 Seahawks, and the 2003 Panthers.

Joe Flacco BAL
Once again, Flacco was much better throwing to his right (11-of-18 for 139 yards with one touchdown and seven other first downs, plus a 14-yard DPI) or up the middle (4-of-4 for 80 yards and two touchdowns) than to his left (7-of-11 for 69 yards and three first downs). However, he was great on third downs (7-of-10 for 158 yards with two touchdowns and four other first downs, plus the DPI and two sacks), and he had a good day on deep throws (4-of-8 for 129 yards and a touchdown).
Colin Kaepernick SF
Kaepernick was fantastic as a rusher, with seven carries for 62 yards, one touchdown and three other first downs. His only failed carry was a 3-yard gain on first-and-10. His passing performance, though, was more erratic. He was even more accurate and productive on deep balls than Flacco (6-of-11 for 157 yards and a touchdown), but he did throw an interception, and played his worst at the most critical points of the game. On third and fourth downs, he went 2-of-5 for 15 yards with one first down and two sacks. In the red zone, he went 2-of-6 for 22 yards with one first down and two sacks. He also didn't throw a single pass to his running backs.

Running Backs
Frank Gore SF
Gore's longest runs went for 33 and 21 yards, but he was stuffed for no gain or a loss six times in 19 carries.
Bernard Pierce BAL
Pierce's longest run was just 8 yards and he was stuffed for no gain twice in 12 carries, but he gets extra DYAR for a trio of short-yardage conversions.
Ray Rice BAL
Rice was stuffed for no gain or a loss three times in 20 carries, and his longest run was only 12 yards. He had three first downs on the day, but converted only twice in five carries with 3 yards or fewer to go for a first down. He caught each of the four passes thrown his way for 19 yards, but one of those catches was a lost fumble and another was a 7-yard gain on third-and-8.
LaMichael James SF
-16 -16 0
His three plays: 9-yard gain on first-and-10; 1-yard loss and fumble, also on first-and-10; 2-yard gain on first-and-goal from the 7.

Wide Receivers/Tight Ends
Anquan Boldin BAL
Each of Boldin's catches went for a first down, including a 13-yard touchdown and four total third-down conversions.
Vernon Davis SF
Davis' shortest catch was an 8-yard gain on first-and-10. Each of his other receptions gained at least 11 yards and a first down, including gains of 24 and 29 yards.
Jacoby Jones BAL
Jones had one incompletion and a 56-yard touchdown on third-and-10. This does not include his 108-yard touchdown on a kickoff return. Remember how Jones' Houston career ended with a couple of fumbles in a playoff game? This was like the opposite of that.
Michael Crabtree SF
Crabtree had four catches of at least 19 yards. He was thrown three passes on the 49ers' final drive, all incomplete, for a total of -21 DYAR.
Ed Dickson BAL
23-yard gain on first-and-10, 14-yard gain on second-and-3.
Delanie Walker SF
Walker's three catches: 14-yard gain on second-and-4, 28-yard gain on second-and-10, 6-yard gain on third-and-13. This statline does not include his crushing block at the goal-line on Ed Reed.
Dennis Pitta BAL
Three catches on first-and-10 for 9, 9, and 7 yards, plus a 1-yard touchdown on second-and-goal. His incompletion came on third-and-2.
Torrey Smith BAL
Smith's first target resulted in a 20-yard catch on first-and-10. His next four passes were all incomplete. He finished with a 15-yard catch on first-and-10 and a 14-yard DPI on third-and-9.
Randy Moss SF
Moss' first three targets were all incomplete, but then he had a 9-yard gain on third-and-8 and a 32-yard gain on second-and-5.

Posted by: Vincent Verhei on 04 Feb 2013

95 comments, Last at 09 Feb 2013, 2:24am by Subrata Sircar


by Will Allen :: Mon, 02/04/2013 - 3:08pm

I suspected the Niners would have a substantially higher VOA; they really moved the ball well. The fumble recovered by the Ravens, and Jones' two long tds, really overcomes a large volume of moderately successful plays by the Niners.

by Yaguar :: Mon, 02/04/2013 - 3:13pm

Hi, I'm Randy Moss, I don't fight for the ball anymore.

by Aaron Brooks Go... :: Mon, 02/04/2013 - 3:16pm

It's not his fault Kaepernick threw one pass two seconds too late, and another two yards too high.

by Will Allen :: Mon, 02/04/2013 - 3:28pm

It doesn't matter if the ball is two yards too high. The receiver, if he intends to fufill his professional responsibility, has to jump for it, because, if for no other reason, he might decrease the chance, even by a very small amount, that the db will intercept it, by interfering with the db's field of vision for a moment.

by Guido Merkens :: Mon, 02/04/2013 - 4:12pm

I don't know if tipping a pass up in the air right in front of the safety would reduce the chance of an interception. And if the pass was too high to tip, then simple physics says that Moss wouldn't have obstructed Reed's line of sight toward the ball.

Appearing to give no effort toward a ball that was picked off is a bad visual, but I don't see any way in which it could have affected the play.

by Will Allen :: Mon, 02/04/2013 - 4:21pm

If you can tip it, then you can't be sure that you can't catch it; nobody's judgement is that fine. You also can't be sure what is taking place downfield, because your eye is on the ball, so you can't judge how your leap will impede the vision of someone behind you, and you don't have to actually block the view of the ball to make the catch harder for the db behind you. Throughout his career, Randy Moss has failed to appreciate that his professional duties, when the ball is in the air, extend beyond making a catch.

by Noah Arkadia :: Mon, 02/04/2013 - 11:56pm

You can't have players on the field, especially in a freaking Super Bowl, thinking about whether making an effort may or may not affect the outcome of the play. By the time they're done thinking, it's all over. You've got to make the effort! Find out later if it meant anything or not.

FO posters are a peacock. You got to let us fly!

by RickD :: Mon, 02/04/2013 - 4:49pm

"It doesn't matter if the ball is two yards too high. The receiver, if he intends to fufill his professional responsibility, has to jump for it, because, if for no other reason, he might decrease the chance, even by a very small amount, that the db will intercept it, by interfering with the db's field of vision for a moment."

I think it's far more likely that a receiver in that position will get clocked and injured than that he'll do anything to significantly hinder the vision of a d-back 10 yards behind him.

Blaming Moss for the pick takes Moss-hating to a new level.

by Will Allen :: Mon, 02/04/2013 - 5:00pm

I didn't blame Moss for the pick. I wrote that receivers need to jump for high passes. If your position is that receivers should not jump for high passes, because they might get hit, well, we'll just have to disagree.

by snoopy369 :: Mon, 02/04/2013 - 6:00pm

Agreed. Even if he's not going to make a difference, show the effort as a teammate. I'd feel a lot better as a QB if I saw my WR jump for balls I throw at him, even if they do get INTed, than if he just watches them go. Same for baseball - a pitcher is a lot happier to see his CF jump at the wall for home runs, even if they're a row back, or at least go up to the wall to see if he can catch them. Heck, that's probably a higher injury risk than WRs (see Rowand, Aaron).

by Aaron Brooks Go... :: Mon, 02/04/2013 - 6:22pm

Sort of like how Crabtree made a show of effort, and tipped away a pass that would have gone to Moss for a TD?

by RickD :: Mon, 02/04/2013 - 6:38pm

Yes, it's important for players to make futile gestures so fans can feel that they are "giving their all."

by Perfundle :: Mon, 02/04/2013 - 6:45pm

Except that the jump wasn't so futile that Crabtree had no chance at it. Are you seriously faulting a receiver for trying to catch a ball thrown in his general vicinity?

by RickD :: Tue, 02/05/2013 - 1:54am

I don't think that Crabtree had no chance at that ball. But he might have known that the ball was intended to another receiver. If the ball was, in fact, intended for Moss than Crabtree should not be "making the effort" if "the effort" results in an incompletion when no effort would have resulted in a completion.

At the time I didn't think the pass was intended for Moss, but I was only looking at the first view.

by Mike B. In Va :: Tue, 02/05/2013 - 9:27am

I really think you hang that on Kaepernick, not Crabtree. A more experienced QB throws that with touch so it arcs over Crabtree, a la what Flacco did on the first Baltimore TD. It was impossible for Crabtree to tell that wasn't just a high throw to him.

by James B (not verified) :: Tue, 02/05/2013 - 12:36pm

I can't recall the play you're talking about, but shouldn't Crabtree be aware of the play design and the fact that there's another WR behind him?

by Will Allen :: Mon, 02/04/2013 - 8:52pm

It is only in hindsight that you can have certainty about the futility of the effort. That's the point. Players should not pretend to be blessed with hindsight in real time.

by RickD :: Tue, 02/05/2013 - 1:55am

I disagree completely about the futility of making judgments about which efforts are likely to be successful and which are not.

Crabtree's effort might have been successful. The jump that everybody is demanding that Moss make had no chance at being successful.

by Will Allen :: Tue, 02/05/2013 - 9:16am

That is because you are defining success in a way that is too narrow.

by Aaron Brooks Go... :: Tue, 02/05/2013 - 10:13am

The team motto of the Vikings. =)

"Your definition of "success" is too narrow."

by Perfundle :: Mon, 02/04/2013 - 6:43pm

In fact, yes. Jumping to try to get that ball was the correct response there. The fault lies with the scheme that had two receivers stacked so close vertically (although that could've been because the protection broke down and everyone was scrambling, I can't remember) and with Kaepernick, for throwing a ball not short enough for Crabtree to catch or for not recognizing the stacked receivers and lobbing a throw to Moss.

by Scott C :: Tue, 02/05/2013 - 12:39am

Moss was not in position to catch that ball. In the replay, it was clear that if Crabtree didn't touch it, it would have gone straight out the back of the end-zone.

by moore2012 (not verified) :: Tue, 02/05/2013 - 4:22am

" I'd feel a lot better as a QB if I saw my WR jump for balls I throw at him"
I'd feel a lot better as a WR if my QB didnt try to get me killed when i go across the middle

by RickD :: Mon, 02/04/2013 - 6:26pm

Well, my initial exposure to NFL football was watching Darryl Stingley get paralyzed going for a pass over the middle. So I don't really see any need to jump for a ball that's "two yards" over the receivers head. I think you're kidding yourself if you think such a jump would have any effect on Ed Reed's vision or ability to make an interception.

My opinion is based on the "two yards" distance. If we were talking about two feet, my opinion would change. But trying to screen a safety's vision by jumping at a ball six feet above one's head would be like me trying to win a jump ball from Dwight Howard.

The highest vertical leap recorded at any NFL combine is 45 inches. A jump at a ball 72 inches above a receiver's head would provide Ed Reed an amusing show, but would in no appreciable way affect his ability to catch the ball.

by Will Allen :: Mon, 02/04/2013 - 9:09pm

Daryl Stingley's injury had nothing to do with leaping high for a pass, and in fact, his front foot was nearly on he ground, with his chin tucking into his chest, and had it been a very high pass that he was leaping for, he probably escapes permanent injury. I think you are kidding yourself if you think a receiver can have some sense of what is happening 180 degress from his field of vision, and can thus judge the futility, or lack thereof, in leaping for the ball.

by RickD :: Tue, 02/05/2013 - 1:59am

So, because Stingley was not hurt on a leaping attempt to catch a ball, such leaps are safe?

Surely that's not your argument.

I know that I wasn't arguing that Stingley had made a leaping attempt at a catch. I was making the argument that catching balls over the middle of the field is dangerous. We know from a lot of other evidence that leaping exposes the receiver to a "defenseless" position. Stingley was brought in to emphasize that a receiver in a defenseless position can be seriously injured. That fact is true regardless of whether a receiver is leaping at the time or not.

by Will Allen :: Tue, 02/05/2013 - 9:20am

Look, if you think receivers should stop jumping for balls, just say so. I disagree. If you think Moss could, at that moment, determine, with 100 percent accuracy, that there was absolutely zero to be gained by jumping, just say so. I disagree. He should have jumped.

by justanothersteve :: Mon, 02/04/2013 - 9:41pm

The vertical leap is measured by an athlete at a standstill. Randy Moss's combine vertical leap was 39". (He has not ever had a vertical leap of 51". It's one of those internet legends that won't die.) He also has an arm length of over 33". At 6'4" and on the run, it's not out of the question for Moss to be able to catch a ball 2 yards over his head. Plus, Moss has shown incredible hand/eye coordination several times.

I will give Moss the benefit of the doubt on this play. It's impossible to know all the circumstances on that INT. I think the catchability of that pass is a strawman that people who dislike Moss based on his past history are using. As a Packers fan, I have a long history of disliking a lot of Moss's actions. But I agree with you that fans have no way of knowing if Moss even had a chance at that ball.

by Aaron Brooks Go... :: Tue, 02/05/2013 - 1:25am

Let me put it this way -- in the history of the NBA, no man has been capable of plucking a coin from the top of the backboard.

Plucking a coin only requires a man to get a single finger to the top of the backboard; a mere 13 ft. Dwight Howard, who is 7" taller than Randy Moss, and can leap higher, cannot do it.

A football thrown that high requires at least one full hand at that height to catch it, and that's still a miraculous catch.

There was no point to jumping for it.

by Will Allen :: Tue, 02/05/2013 - 9:22am

Only if you conclude that there is only something to be gained if the receiver can touch it with a full hand. This would be a inaccurate conclusion.

by Subrata Sircar :: Sat, 02/09/2013 - 2:13am

As it happens, picking quarters off the top of the backboard is one of those things that people have done. Calvin Murphy is most commonly mentioned here - 5'10" with long arms and really, really serious hops - although that's likely apocryphal. (Cool story bro time - I used to play pickup with a guy who played major-college ball - 6'10'', bad knees and generally creaky guy when running, but give him the ball in the post and all those years of practice shone through; every move was well-oiled and graceful as could be. He actually went to tournaments with players like Elvin Hayes and Murphy and claimed to have lost a bet when Murphy picked a quarter off the top of the backboard.)

That being said, there exists sufficient evidence for me that I believe that it has been done. What this has to do with Moss and whether he should have jumped for something that he almost certainly had no chance at I leave to the reader.

by RickD :: Tue, 02/05/2013 - 2:05am

"it's not out of the question for Moss to be able to catch a ball 2 yards over his head."

Yes it is.

And it certainly was out of the question that Moss could have caught this ball.

The Randy Moss that you think can catch balls over 12' off the ground hasn't been seen for at least five years, if ever.

by Led :: Mon, 02/04/2013 - 8:33pm

That's my instinct as well. However, I don't think player safety should be exclusively the responsibility of the defense. Defensive players can be given more safety-friendly incentives via penalties and league discipline. The only way to give offensive players similarly safety-friendly incentives is through peer pressure (or the reduction thereof, when it comes to dangerous conduct). So if we're really interested in concussions and player safety generally, then we shouldn't ask receivers to expose themselves to severe injury in futile gestures. Similarly, we should be highly critical of QBs that throw passes that expose receivers to risk of severe injury.

by Yaguar :: Mon, 02/04/2013 - 3:45pm

I was talking more about the almost-interception to Cary Williams. Moss didn't come back to the ball and calmly let Williams step in front of him.

Then he made the same elementary mistake with Smith covering him over the middle, only a minute of game time later.

by sundown (not verified) :: Mon, 02/04/2013 - 7:42pm

Well, the one thing I think the SB finally put to rest is the conspiracy theory Mike Florio had been floating all year over at PFT that Moss was being misused and/or treated as a decoy and was going to break out for a huge game any week.

by Anonymous_ (not verified) :: Mon, 02/04/2013 - 3:17pm

Why are returns yards not counted in total yards? The Ravens had around 100 yard advantage there which makes the total yards virtually identical for each team.

by Guido Merkens :: Mon, 02/04/2013 - 4:15pm

Return yards aren't nearly as valuable as field position. If a player returns six kickoffs from five yards deep in the end zone out to the 20 yard line, he will have gained 150 return yards, but with exactly the same effect as taking a knee each time.

by justanothersteve :: Mon, 02/04/2013 - 4:33pm

But a 108 yard kickoff return for TD is huge compared to 4-27 yard returns.

by RickD :: Mon, 02/04/2013 - 4:51pm

Yes, there has to be a better solution than to simply ignore return yardage.

by RickD :: Mon, 02/04/2013 - 4:53pm

Return yardage for any kick into the end zone should be measured in terms of the yardage gained relative to a touchback. So a returner who gets back to the 20 gets credit for zero yards. And it shouldn't matter if the ball is taken at the -9 yard line or the -1 yard line.

Certainly this is how kickoffs should be dealt with. I'm tempted to extend this logic to all kickoffs. Punts are more problematic since it's much easier to bury a team inside its own 5 yard line with a punt than with a kickoff.

by snoopy369 :: Mon, 02/04/2013 - 6:07pm

I think you can only use 'all kickoffs', anything else would be uneven. I'd have to see first whether the distance mattered or not, as to whether this is useable. IE, if you field the ball at the 5, are you more likely to get to the 20+ than fielding it at the goal line? It seems logical, but at the same time a shorter kickoff might give the return team less time to set up (perhaps less relevant in the post-wedge era). If eventual return yard line does positively correlate with yard kick received, then you probably should not credit receptions past the goal line from the 20, as they are getting some advantage from a weak-footed kicker.

Then again, how do you correspond the two? Presumably the advantage gained from fielding at the 5 is not 5 yards, as the 'meet-up' point with the coverage team is less than 5 yards downfield. IE, catch at the 0, other team is at the ~40 by then, so meet-up point is at around the 20. Catch at the 5, other team at the ~40, meet-up point is at the 22.5 - so you really only gained 2.5 yards from the short kickoff.

I suspect this is one of those things that will just always have to be difficult to represent. DVOA and such accounts for it, presumably, by measuring the difference in scoring probability; but anything simple, like return yards, probably will be flawed no matter what you do. I suppose the best compromise might be considering any kickoff from the endzone to have begin at the goal line (so a maximum 100 yard return is possible).

by wmatt (not verified) :: Mon, 02/04/2013 - 3:30pm

Haven't read the previous posts that mention it, but I wonder if Flacco being throwing to his right is because Anquan Boldin usually lines up on that side?

by ChuckC (not verified) :: Mon, 02/04/2013 - 3:51pm

I this game, I think it was because Culliver was lined up to the right and Tarell Brown was on the left.

by navin :: Mon, 02/04/2013 - 4:03pm

Is this the largest "upset" in a Super Bowl according to DVOA history? I believe Seattle, Carolina, and New England only had small DVOA advantages.

However, 28 DVOA pp is a 5-6 point expected margin for the 49ers.

by DavidL :: Mon, 02/04/2013 - 4:25pm

Which is pretty much dead on, since there's seven non-predictive points of kickoff return touchdown that are only slightly reflected in DVOA.

by navin :: Mon, 02/04/2013 - 4:42pm

I'm pretty sure the TD return is reflected in the special teams DVOA. Things like FG blocks are non-predictive, not kickoff returns.

by DavidL :: Mon, 02/04/2013 - 5:21pm

Yes, but the longer a play is, the less weight each additional yard gets (see above about two 25-yard plays being "better" from a DVOA perspective than one 50-yard play), plus special teams gets far less weight than offense or defense in the overall score, so even a strongly predictive event will have less impact than if it happened on a play from scrimmage.

by navin :: Mon, 02/04/2013 - 8:21pm

Yardage is capped since it's hard to distinguish a 50 yard TD run from an 80 yard TD run. This is because the length is arbitrarily determined by the starting yardline and since both result in TDs. Kickoffs are different. My guess is a 108 yard TD return is much more valuable than a 50 yard return with no TD.

And of course special teams get less weight--they only happen on about 1/7 of all plays. Count how many plays in a game occur on special teams and I bet it is reflected almost exactly by the special teams weighting.

by Aaron Brooks Go... :: Tue, 02/05/2013 - 1:27am

That's bunk. There's a clear difference in expected point value as a basis of field position change.

by Bruce Lamon :: Mon, 02/04/2013 - 11:11pm

Patriots over Rams in '99 must have been a bigger DVOA upset.

by Will Allen :: Mon, 02/04/2013 - 11:55pm

Especially since they didn't play each other.

by RickD :: Tue, 02/05/2013 - 2:10am

I think the comment is about the DVOA recorded on game day, not the DVOA ratings coming into the game. And "upset" means "team won despite scoring a lower DVOA in the Super Bowl."

I think for the latter category, Giants over Pats (2007-8) holds the record for Super Bowl DVOA upset. (52.9 to 1.9 at the end of the regular season). Pats (7.9) over Rams (25.9) isn't close to that.

by Paul M (not verified) :: Mon, 02/04/2013 - 4:23pm

This is not an attack, but merely the cry of a skeptic. Barnwell maybe said it best today when he said that the success of GB, NYG and Baltimore may not be simply some random deviation or merely sample size completion but instead proof that we don't really know what makes a great team great anymore.

By anyone's definition, and particularly here, Baltimore was a barely good team 6 weeks ago. And Denver, New England and Seattle were three of the greatest teams to have walked the planet in the past 22 years, with SF not all that far behind. There was no real precedent for the Ravens to achieve what they achieved, unless the assessment of either them or their opposition was at least a bit faulty. Baltimore beat three of those historically or near-historically great teams in succession. And now we learn that the rating system of choice around here says SF was the better team last night.

Soccer and hockey involve relatively few scoring plays, so a lucky or fluky outcome is more likely (check out the last month or so of the EPL results if you don't believe that). Hockey and baseball-- more so than soccer where the goalkeeper more rarely plays a decisive role due to the size of the goal-- also involve single dominant players who can pretty much determine an outcome all by their lonesome. In the case of hockey, the hot goaltender can decide an entire best of 7 series, whereas in baseball the equivalent can only do so for two or perhaps three games. But all three of those sports are upset-friendly.

Basketball-- at least at the pro level-- is not-- too many outcomes within a given game mutliplied by 4 to 7-- and enough players so that just one cannot dominate but a select few can (thus most of the NBA champions are teams that feature more than one HOF-caliber player. And even many NCAA champs have two or three big-time NBA players of the future) There are very few best of 7 upsets, and not many examples where either the best or 2nd best team in the league going into the playoffs did not win the playoffs.

Where does football lie in this universe? A QB can have a big impact, but not as much as a goaltender or pitcher. A strong offensive or defensive line can too, but not like three superstars in a NBA lineup. There are 150 or so real plays a game-- fewer but not too dissimilar to the number of shots-- Field goal attempts plus free throws-- in a NBA game, but only a handful or so are scoring plays. A team can do everything right until they get to the 7 yard line and still end up with nothing.... And the postseason games are all one-and-done. Football falls in the middle, but more toward the predictable side of the team sport upset spectrum, IMHO. Which means that I believe the metrics can get better and the ability to forsee what teams like GB, NY and BAL have done the past three years (Or what Arizona nearly did two years before) can be improved.

by Will Allen :: Mon, 02/04/2013 - 4:56pm

Baltimore's offensive personnel had a 27% turnover from 6 weeks ago. It literally was not the same team in the playoffs. San Francisco's defense changed substantially for the worse when it's most important player tore his triceps in December.

Rules changes whch make the game more centered on qb play has made football more upset prone, but 22 years, from which the last few a supposed trend is established, just isn't enough data to really tell us much. I'm really looking forward to seeing the DVOA numbers back to 1978, because the biggest change since the rules that opened the passing game up has been the imposition of the salary cap, once the cheating on the cap got cleaned up. From 1978 into the mid 90s, teams were able to stockpile talent in a manner which is impossible today.

by Perfundle :: Mon, 02/04/2013 - 5:03pm

"Which means that I believe the metrics can get better and the ability to forsee what teams like GB, NY and BAL have done the past three years (Or what Arizona nearly did two years before) can be improved."

First of all, GB does not belong in the same group with the other three teams. They had a very good DVOA rating, and was second in the league in point differential.

Second, I don't think there's any possible way to predict the other three teams' success. I read Barnwell's article and I have no idea what he's arguing for in lieu of sample size randomness. There's no reason that sample size randomness can't explain these results.

The biggest difference between football against other sports like basketball, baseball and hockey is simply the number of regular-season games. I remember an article noting that the 2010-2011 Miami Heat, the runner-up that year, had a 9-7 stretch and an 8-8 stretch that year. They also had a 21-1 stretch that year. The point is that you can build any number of narratives about any of the successful NBA teams if you only look at 16-game stretches.

by RickD :: Mon, 02/04/2013 - 5:09pm

The problem is that Baltimore was weighed down by a handful of games that did not reflect its true quality level. The Ravens of early December were, in a very real sense, not the team that was playing the playoffs.

The cause of this string of lower seeds winning is simple: parity. There really was not a huge difference between the top four teams in the AFC, or between the top five or six in the NFC. It's very hard in the salary cap era to build a team that gets anywhere near the dominance level of the 49ers of the 80s or the Cowboys of the early 90s. No team can afford to stockpile that much talent.

As a result, the small number of franchises that are well-run cannot break away from the pack as was possible in the decades gone by. And then there's the problem of small sample size. DVOA tends to over-reward wins and over-punish losses. So if you have two evenly matched teams, random chance alone will lead one of the two to have a better game on a given day, and then DVOA will read too much into the result. That's part of why Denver was consistently rated as high as they were, in contrast to Baltimore.

DVOA also rewards consistency. I'm not sure that's the best way to make a judgment about who the best team is. Let's say you have two teams, and one gives you five DVOA results of (to make up numbers) +50, +45, +50, +45, +55, while the other team has +60, +65, +55, +60, -25, DVOA will prefer the former team. But a professional statistician would look at those numbers, recognize that the -25 if an aberration, an outlier, and that it should be discounted. If you wanted to make a 1-game prediction about the second team, you should make the guess that it will play a level that is probably higher than the level of the first team.

This leads us to the problem of the teams themselves refusing to play at their (recent) historical levels. The 49ers of the first 12 weeks simply were not there during the playoffs. Their defense gave up a lot of points to both the Falcons and the Ravens. History is, after all, nothing but prologue. Even the most accurate recording of past behavior has to accept that "things change."

by Will Allen :: Mon, 02/04/2013 - 5:34pm

I don't know how to statistically adjust for one of a team's most important players getting hurt, but remaining on the field, albeit at a markedly lesser level of effectiveness, in the playoffs, but it is a commone occurrence at the most elite level of athletic competition. I think the inability to develop and retain depth of talent, due to the cap, has made this phenomena even more important.

by Perfundle :: Mon, 02/04/2013 - 5:40pm

"DVOA also rewards consistency."

Well, DVOA rewards consistency within a single game. It doesn't make any kind of statement about consistency over an entire season (Their "ESTIMATED WINS" and "Forest Index" stuff does that.) In any case, the "+50, +45, +50, +45, +55" example is not particularly applicable, because NO NFL team is that consistent. Even Minnesota, the most consistent playoff team, had DVOA ratings of +60 and -40. Considering that the Giants had a fairly high variance last year and a very low variance in 2007, I don't think there's any kind of correlation between regular seson variance and playoff performance. Especially when you have stuff like Seattle's 58-0 win bringing their variance from 2nd to something like 28th.

by zlionsfan :: Mon, 02/04/2013 - 5:09pm

I'm not sure where you got the idea that there aren't many upsets in the NBA playoffs.

2011-12: #2 Heat defeat #2 Thunder
2010-11: #3 Mavericks defeat #2 Heat
2009-10: #1 Lakers defeat #4 Celtics
2008-09: #1 Lakers defeat #3 Magic
2007-08: #1 Celtics defeat #1 Lakers
2006-07: #3 Spurs defeat #2 Cavaliers
2005-06: #2 Heat defeat #4 Mavericks
2004-05: #2 Spurs defeat #2 Pistons
2003-04: #3 Pistons defeat #2 Lakers
2002-03: #1 Spurs defeat #2 Nets

Only once in the last 10 years did both top seeds make it to the NBA Finals; six times in 10 years, both top seeds were knocked out before the Finals.

And it's not like top seeds dominated in the decade before, either. Yes, Jordan's Bulls frequently faced the #1 seed in the West ... but that era also gave us #6 Houston as a champion and short-season #8 New York as a runner-up, and even Jordan's 1992-93 team failed to get the top seed in the East.

A best-of-anything series in any sport is upset-friendly. Seven games is a much smaller sample size than people want to believe.

by Jerry F. (not verified) :: Mon, 02/04/2013 - 5:30pm

It depends on what is meant by "upset." In the last decade, that looks like a lot of high seeds winning it all. Not once has a seed lower than three won. There are usually a handful of teams going into the playoffs that are seen as "legit" contenders, and though a given one might fall early, by the end one of those legit teams wins in the Finals. Yes, you can say there are upsets in that the top seeds don't always win, but lower seeds don't advance far, and, throwing out seeding, the teams that people tend to believe in make it far and those with good records that remain unconvincing as dominant forces do not. (I like this about the NBA, as opposed to the upsets of NCAA, as at the end of it all I usually really do feel like the best overall team has emerged.)

by snoopy369 :: Mon, 02/04/2013 - 6:15pm

I don't think I agree. The Bulls were the top seed the last two years, and they weren't favored by anyone vs the heat in 2011. In 2012, they lost their star player in the first round. The Thunder similarly last year were the 'best team' despite the #2 seed. Basketball more than anything has this weird differential between playoffs and regular season, where a good regular season team and a good post-season team are not entirely the same. To do well in the regular season, you need a deep bench and a coach who's always on. Come playoffs, that team probably doesn't win, between the fact that you have a lot more reset in the playoffs in the NBA, and the fact that playing all out for five months leads to mental exhaustion that isn't fixed with a few rest days.

Sure, NBA basketball has some upsets, but overall the winning team is usually one of the top teams in the league - and when it's not, it's often because the 'top team' was misunderstood (see Mavericks). The NBA is also a star-heavy league with little parity; the NFL has great parity, and is relatively star-ambivalent - you can win without a superstar, as long as your QB isn't a total idiot and you have depth and balance.

by RickD :: Mon, 02/04/2013 - 6:36pm

The lowest seed I see in the NBA Finals is #4. Compare that with a similar history in the NHL. Or the NFL, for that matter, which has fewer teams.

I think you also need to have a fuzzier notion of what "upset" means. I suppose that, technically speaking, it could be considered an upset if a 2 seed beats a 1 seed, but in popular usage we're looking for something more dramatic than "beating a higher seed."

Nobody really thought it was an upset when the 49ers beat the Falcons.

"A best-of-anything series in any sport is upset-friendly."

A best-of-seven series is less upset-friendly than a best-of-five series, which is less upset-friendly than a best-of-three, which is less upset-friendly than an elimination game. That's just math talking. Presuming the terms have been defined properly (an exercise for the reader).

by dejour (not verified) :: Mon, 02/04/2013 - 8:04pm

Certainly there is a greater deviation in regular season winning percentage among NBA teams than among NHL/MLB teams.

Logically, this means that a .500 team knocking off a #1 seed is less likely in the NBA than in the other sports.

by dmstorm22 :: Mon, 02/04/2013 - 5:16pm

I've changed my thinking into this:

In a sport with a 16-game schedule (by far the shortest of each of the major sports in the US, or even EPL or college BB), and a single-elimination tournament, it seems more statistically off that the high seeds dominated so much before. 16 games aren't really enough to tell is if one team is better than another, and putting in a one-game playoff makes it even more random. The results of the last five or so years isn't that shocking in that environment. Seeing 19 #1 seeds in the NFC win their opening game in a row (1988-2006), is shocking.

by snoopy369 :: Mon, 02/04/2013 - 6:20pm

100% agree. NCAA basketball is similar - even though there's a longer 'regular season', a lot of those games aren't useful for predictive value because they're games like Duke vs Virginia or whatever where clearly one team is superior, and is expected to win by 40. I doubt most teams play more than 20 'interesting' games in NCAA basketball, and then jumping into a single elimination tournament leads to very quirky results. You do end up with higher seeds winning much of the time (as there is much less parity) but it's very rare to have all higher seeds advancing in any round.

In the NFL pre-salary cap, you had drastic imbalances, leading to more 'obvious' results. Now that some of the salary cap-era imbalances have adjusted themselves out, you see so much parity that it is indeed hard to predict who's better in a short season; and add to that the significant schedule imbalances from the 4 team divisions and the intentional scheduling of more difficult teams for some teams (division winners in particular), and you have a very hard time predicting the effect a regular season win has on the likelihood of post-season success.

by RickD :: Tue, 02/05/2013 - 2:18am

I don't think you can say something like this:

" 16 games aren't really enough to tell is if one team is better than another"

In any kind of tournament setting, the number of comparisons needed is a function of how close together the objects are. You don't need 16 games to tell that the 49ers are better than the Jaguars.

In the past, there were several seasons where there were dominant teams such as the '85 Bears, '91 Redskins, '07 Patriots, etc. that were clearly beyond the competition. We didn't need 16 games to know that the Bears were the best team in 1985.

But if the teams are closely bunched together, then you would need more sampling (i.e., games/competitions) to determine the difference in quality level. But that's a problem in football since more games lead to more injuries, which mean that the quantity being tested (so to speak) is too variable.

by Aaron Brooks Go... :: Tue, 02/05/2013 - 10:22am

The other aspect that gets lost is playing style.

Some inferior teams just give otherwise superior teams fits, especially in a single-elimination format.

This is fairly common in single-person tournaments (fighting competitions, tennis, etc), where a lower ranking participant can have repeated, consistent success against a particular, objectively better player, just because something they do causes enormous problems for that better player. That doesn't mean the winner is better across the board, just that their style beats the better player's style.

An example of this would be the 2000 NCAA Tournament, where the 8th seed Wisconsin knocked off a #9, a #1, and a #6 into the Final Four, because they kept beating fast break teams who couldn't cope with Wisconsin's deliberately molasses style. They then lost (for the 4th time that season!) to MSU, who didn't care what style opponents played, and was happy to take a 19-17 halftime lead.

by Will Allen :: Tue, 02/05/2013 - 11:00am

The Rams/Pats Super Bowl may be a good example of this, albeit more coaching oriented. I remain convinced that if Mike Martz hadn't been so hell bent on appearing to be Mike Martz, Football Suuuuper Geeeenius, and thus willing to keep handing the ball off to Marshall Faulk, the Rams likely win the game somewhat comfortably.

by Ryan D. :: Tue, 02/05/2013 - 11:44am

Did Mike Martz just get criticized for RUNNING the ball too much?

by Will Allen :: Tue, 02/05/2013 - 12:01pm

I didn't write that clearly. He should have tried to run the ball down the Patriots' throats, because he had the offensive line and running back equipped to do so. It would not have been as prominent a way to display his high estimation of himself, however.

by Ryan D. :: Tue, 02/05/2013 - 3:57pm

I was sure that was what you meant, but my mind just kept interpreting it incorrectly.

by Jerry F. (not verified) :: Tue, 02/05/2013 - 12:03pm

The phrasing was weird there, but he definitely meant the opposite.

by DisplacedPackerFan :: Mon, 02/04/2013 - 5:57pm

I have no idea about soccer for this, but all the other major US sports play series not single games, the NCAA basketball tournament being a big exception and there are upsets in that all the time, and the champion is not always considered to have been the best team, but the team that played the best in the tournament.

Are we shocked that OKC beat Miami in a couple of games but lost the series in the NBA? Would you be shocked if SF beat Baltimore two out of three games? The better team can and does lose. Also GB in 2010 was not a slouch of a team by FO metrics. While they were a 6th seed (they had bad injury luck and they played an above average schedule) they were the 2nd best team in the NFC and 4th in the NFL, 0.2% DVOA behind Philly, who they barely beat. Weighted DVOA had them as an even stronger team (#1 in the NFC). So they are actually a counter example, the metrics on this site did say they were playing the best of any team in the NFC. Yes they were underdogs to Pitt in DVOA by 10 - 13% depending on weighted vs full season. That SB was a 3 point game with 7 minutes left much like this one was a 2 point game with 9 minutes left, the outcome could easily have been different.

Keep in mind what DVOA is saying too. For an oversimplified explanation. A team that has a 50% DVOA on average is 50% better than the baseline play in that situation vs that defense (and is now adjusted for the yearly environment). A 50% DVOA is big DVOA, but that means if the average team gets 5 yards the 50% DVOA team gets 7.5 yards. So with a 22% and 11% team they are getting 6.1 and 5.55 yards instead of 5. So in this case the 49'ers were the 6.1 team and Baltimore was the 5.55 yard team. On average San Fran could be expected to get about .6 yards more on a play than Baltimore would have on a play where an average team would have gotten 5 yards.

Do I think the better team won last night? No I don't. If they played again this Sunday in Indy and then a week after that in Arizona I would expect San Fran to win both of those games.

Do I think DVOA could be a better stat? Yes I do. I think the biggest improvement would be to be able to incorporate injury data. Something that may actually be possible with snap counts being tracked and published (though it wouldn't work for historical DVOA where we don't have that data). Though really the difference between regular season and post season Baltimore was the play of the offensively line after they changed 3 of the 5 starters (or at least changed the position they were playing) so that might still not be something easy to capture in a stat.

But really while I think you are correct that football is less likely to have an outcome dominated by a single player, not having a series of games makes it harder to say definitively if the best team won. It's not even the best team that day that always wins. Another factor is how much more important a gameplan is for football, how much harder it is to change what you are doing mid game if you never practiced for it during the week. Green Bay playing man coverage all game and trying to have the front 7 keep contain failed miserably when the front 7 was getting bullied all over the place (Raji getting blown 6 yards off the line, etc). Even if Capers adjusted better than he had, they still would have been playing a style they hadn't really practiced for. I hear this from players who are on the top levels all the time. Since you need so many people to be coordinated on a play, what you practiced the week before really does matter and it's not so easy to just change it if the other team has been practicing something that is consistently beating it.

Finally if football upsets were so predictable someone would be making a killing betting on it and while they might be able to keep it quiet for a while, I doubt that the majority of people would have such a hard time picking games.

I think what we are seeing recently is that the overall talent in the NFL has always increased since at least the 60's. Everyone is bigger, stronger, faster, has more exposure to the game, has better coaching (because there is more history of coaching to draw on), and quite likely is taking better drugs to mitigate other issues. Your average player today is simply better than the average player 20 years ago. The physical abilities being better now than in the past I think is fairly non controversial, much of that can be quantified. The mental side that I think it better as well can be argued but just the fact that things like instant sideline photos, higher quality film, etc exists and tends to be used well by teams lends me to think they are.

However I don't believe the superstar players have improved as rapidly as the average player has. Adrian Peterson is not that much (if any) better than Barry Sanders, Emmit Smith, OJ Simpson, or Jim Brown. Also how much better than his current peers is he? Is he that much better than Chris Johnson was? How much better was he than Alfred Morris this year? How much better than the 16th best back?

What I'm getting at is that the margins are just smaller. Salary caps and free agency changed the landscape in the mid 90's and we are seeing the effects of that settling in too. The average player now has a higher chance of making a single great play out of no where. The very best players are closer to average player than they used to be, and even closer to the "top 5" or whatever at their own position. I haven't run the numbers from available advanced metrics but I've got a feeling you would see this all bear out. The more evenly matched teams become the more likely it's a coin flip or one or two great plays completely change the outcome, or heaven forbid a blown call or two change the outcome. So if one of the 8 best teams in the league wins it all, I'm not all that surprised. So we've had 8, 12, 4, 6, 4, 14, 7, 4, 2 by DVOA in the last 9 years, with the Giants being the big outliers with the 12 and 14. Why did I go 9 years? I think it covers an "NFL Generation" fairly well. The average career of an NFL player who makes an opening day roster is 6 years, and the average career of an NFL player who makes the pro bowl is 11.7 years and the average career of a play who lasts 3 pension credited seasons is 7.1 years. 9 years might be too long, but it seems to cover the players that will make up the core of a team during a "generation" so the incremental changes that are always happening should be fairly well covered.

by abc123 (not verified) :: Tue, 02/05/2013 - 9:13am

This is a great response. I'd expound on it by encouraging anyone to read Richard Dawkins on Ted Williams hitting 400. The theory, simple really, is that there are real physical human limits to what athletes can do, humans cannot run a sub 4 40, they cannot bench 900 lbs, etc, so as sports become more professionalized athletes become more bunched up towards the limits of human possibility. Jim Brown was physically more superior to Nick Pietrosante and even Willie the Wisp Galimore than Adrian Peterson is to backs today, not because Peterson isn't great but because modern training and, frankly, pay scales ensure that the best alternative to Peterson is the next best RB. In this relative absense of separation, writ large, it is more difficult to find separation among teams.

Teams have responded by bringing intelligence to bear, complex game plans, film review, some even using advanced metrics themselves, figuring that if the players are more similar we'll find separation by using them more efficiently. But today even these things are less lasting. Nearly a decade after Paul Brown had his team doing film study and digesting thick playbooks Chuch Schmidt said the Lions were just running to the ballcarrier. Then, a decade after that they made that run to the ballcarrier LB their Head Coach and de facto head of the legal.ization while the Cowboys were making forays into computerization (no wonder they sucked for 20 years). Conversely today due to the pace of turnover both among players and assistant and head coaches as well as technology generally these ideas when they work permeate quickly. Paul Brown kept his original assistants together far longer than any similarly successful coach could today, and the ideas stayed in house. Ray Lewis disects plays at home on an ipad - a piece of technology my Fortune 500 employer just started supporting. Those advantages have been largely taken away.

Brave new world.

by Jester (not verified) :: Tue, 02/05/2013 - 2:22pm

I think that was Steven J. Gould, in "Full House" rather than Richard Dawkins.

by Paul M (not verified) :: Tue, 02/05/2013 - 3:41pm

In the big club soccer tournaments such as the Champions League, the late rounds are decided by a home-and-home series, at least until the final, which is a one game knockout. It can lead to some very interesting strategy since if the games are split, either goal differential or, if that is also tied, the team that scored more away goals determines the winner. Last year for example in the semi-finals of the Champions League I believe Chelsea first defeated Barca at home 1-0, meaning that when Chelsea (Ramires, assisted by Lampard) scored to tie the subsequent game 1-1 at Camp Nou, Barcelona had to score twice more to win and advance. Even a man down after Terry's red card, Chelsea weren't going to allow two more goals. Not sure it changes the "fluke" quotient that much-- it just adds a layer of tactics and strategy.

by dmstorm22 :: Tue, 02/05/2013 - 4:31pm

Your history is a little wrong. Barca scored two goals to go up 2-0 in the 2nd Leg before Ramires scored, so Barca only needed to score the one goal.

Personally, I don't like the away-goals rule. The only advantage I see is it limits Extra Time and more importantly PKs.

by johonny (not verified) :: Mon, 02/04/2013 - 7:26pm

The Raven's have been a solid franchise for years and the sites statistics seem to back that concept up. It isn't like this Ravens team came out of nowhere. More to the point while key players on the team are old, the Ravens have a core of young players which makes you think they will be in the AFC mix for years. We are talking about the team that was one play from being in the Super Bowl in 2011. That doesn't sound like a team "barely" good just 6 weeks ago. The short NFL schedule makes people overreacted to loses. Lose a few games in baseball or basketball and by the end of the season you've completely forgotten the "losing" steak. But if injuries make you run a dry patch in football and everyone loses the big picture. The big picture is the Raven Franchise is right there with Pittsburgh, Indy and New England in producing competitive rosters in the AFC on a yearly basis. I believe this site on a year in and out basis reflects that fact.

by dejour (not verified) :: Mon, 02/04/2013 - 7:56pm

I don't know. I think the QB is comparable in importance to a hockey goalie.

I think football should be theoretically more predictable than hockey since there is maybe 4-5 distinct scoring plays per team per game. Hockey has 2-3.

But the problem is that with the short NFL season, you only have so much data.

I remember reading some analytic website a few years ago. I believe they used a Sagarin or SRS metric for year 20xx. Then they went back in time and looked at the Vegas spreads of each game of that same season 20xx. They determined who was the favorite versus the spread. Anyways even with that end-of-season info, a time-traveling gambler would only get 60% of games right. It was argued that that was an upper limit of the accuracy of team rankings.

by Revenge of the NURBS (not verified) :: Tue, 02/05/2013 - 1:45pm

You lost me at the end. A time-travelling gambler just needs a list of what the scores actually were to be 100% accurate; he doesn't care about Sagarin ratings. That's how Biff Tannen made his first million betting on a horse race.

by cjfarls :: Wed, 02/06/2013 - 4:03pm

You're thinking too much.

Wierd happens... if Rahim Moore does inexplcably line up 2 yards shallower than any of his fellow safeties, and then take a horrible angle, Baltimore is done.

In the playoffs, most every team is typically good. Baltimore certainly wasn't a bad team, and they got healthier in the playoffs which should've improved them too. There are very few total eggs, as to get in, the teams have all typically shown at least the ability to play at very a high level (even if inconsistent in many of the lower-win teams). The variability between the best and the worst playoff team on a game-to-game basis is easily enough to lead to the inferior team winning, and once you add randomness like variable penalty calls (there is holding on every play, whether the refs call it or not is seemingly randomn), fumble recoveries, etc... in a lose and go home framework, we're gonna get odd winners.

While it would be great to figure out some new aspect of football greatness that no one currently recognizes and measures, its also completely not necessary to explain what has happened.

by Subrata Sircar :: Sat, 02/09/2013 - 2:24am

I think you have that exactly backwards for baseball and basketball.

In baseball, you can prevent a hitter from having nearly any impact, because he only comes up 3-4 times a game and you don't have to pitch to him when he does. A pitcher can affect more of the game, but only for ~7 innings usually and only to the extent that his defense lets him. Jack Morris doesn't get a sniff of the Hall of Fame if Kirby Puckett doesn't reel in a deep drive in Game 6 and thus allow him to take the mound in Game 7; even in Game 7 Morris loses if Lonnie Smith doesn't get faked out on the basepaths. Baseball is the only one of the four major sports where the defense puts the ball in play and thus has the advantage of initiation. The best players in the history of baseball have been worth about ~15 wins over a freely-available minor leaguer over the course of a 162-game season - you can do the math there.

Basketball, on the other hand, can be dominated by an individual player because you can't stop him from getting the ball on every offensive possession, if he can dribble up the court. Until a short while ago, zone defense was illegal in the NBA, so you could run a 1-4 clearout for your best guy every time if you wanted. It is no coincidence that the basketball teams with the best individual players win the most games. Hoops is dominated by the high-efficiency, high-usage scorers.

by Aaron Schatz :: Mon, 02/04/2013 - 5:19pm

I'll have more on this issue of trying to figure out randomness in a post tomorrow.

And yes, the TD return is reflected in the special teams DVOA.

by Purds :: Mon, 02/04/2013 - 6:21pm

One point I haven't seen anywhere else. Am I correct in what I thought I saw on Jones' fumbled return? I thought I saw a 49 on the bottom of the screen beat the blocker at the last second but hold up, making sure he didn't overshoot Jones if Jones caught it and ran by him. In other words, it looked to me like there were no 49ers close to snatch the fumble in part because Jones had burned them already for a TD return, and they were deciding to hem him in, not attack him, and thus no one was close enough to make a stab at the fumbled ball. Did anyone else see it that way?

by Are you Benjamin Button? (not verified) :: Mon, 02/04/2013 - 9:17pm

Given that Jones muffed the punt in the second quarter, and his touchdown return was at the start of the third, I'd say that no 49er was influenced by the touchdown during the muffed punt.

In fact, Jones didn't even have the touchdown reception yet - he caught that on the ensuing drive.

by Purds :: Tue, 02/05/2013 - 1:42am

Wow, I really wasn't paying attention, was I? I looked back at the play-by-play, and I guess I was thinking of the good punt return and then influenced by the announcers talking about Jones every time SF kicked or punted.

by theslothook :: Mon, 02/04/2013 - 6:40pm

Can we ever really say we knew what kind of team is at any time throughout the season? Teams change all the time, through injury, through tinkering, through play adjustments. Its highly unlikely the week 1 baltimore ravens were anything like the sb ravens. Ditto for the 49ers.

I want to add, we are essentially trying to measure a highly volatile system with really imperfect tools. DVOA play by play data is great, but what people seem to want- incorporating meta stats like scheme, pass rush, coverage concepts, special teams, injuries and then appropriately weighting them all somehow is exceedingly difficult to do, especially mid season. It would take an army of game charters along with plenty of highly complex math to actually break that down and even then, it would be pretty questionable how informative the whole exercise really is.

In the end- little details ultimately have a huge impact in determining the sb winner. The ravens, much like the giants of last year, had breaks go there way at very opportune times. The kind of things that are non-predictive and would've completely changed the outcome had they not.

by Vincent Verhei :: Mon, 02/04/2013 - 7:24pm

Is this the largest "upset" in a Super Bowl according to DVOA history? I believe Seattle, Carolina, and New England only had small DVOA advantages.

Not even close. San Francisco was 20.1% better than Baltimore in DVOA over the regular season, 15.8% better in weighted DVOA. In the 2007 season, the Patriots’ edge over the Giants in the same two categories were 51.0% and 39.9%. So no, this is nowhere near the biggest upset in DVOA history.

On the subject of Baltimore, their DVOA in each playoff game was at least 22%. They only had five games above 20% DVOA in the regular season.

by dmstorm22 :: Mon, 02/04/2013 - 7:32pm

I think he means 'upset' as the losing team having a higher DVOA in the Super Bowl game than the winning team.

by Nathan Forster :: Mon, 02/04/2013 - 9:50pm

I know this is pretty random, but did anybody else think there as an inordinate amount of advertisements promoting CBS programming during the Super Bowl? When watching a major network, I generally expect maybe one advertisement for network programming coming out of the commercial break, but during the Super Bowl I caught several instances of network advertisements doubling up.

This struck me as a poor choice by station execs. If you think of what the highest economic value of the spots are, it is probably showing a pricey advertisement for Pepsi, Doritos, or some car (that like all other car commercials, will bill itself as the "pinnacle of human engineering"). The highest economic value is probably not the twelfth attempt to cajole some poor sap into watching Two Broke Girls (or CSI: Buffalo). Not to mention that those station promos seem pretty lame when paired with the effort that's put into the "real" Super Bowl commercials.

Sorry JPP!

by RickD :: Tue, 02/05/2013 - 2:29am

In the past, I recall several years when the host network would bombard us with ads for a single show. ("His father's the district attorney!" - Oh wait, that was baseball. And I remember the Craig T Nelson show "Call to Glory" being advertised to death, but that was the Olympics, it seems. Oh well, so much for my examples.)

I think that if CBS is advertising their own shows, that's because they couldn't get enough other outsiders to pay the rates that they wanted.

by Jerry :: Tue, 02/05/2013 - 4:25am

I know that during the season, there are 21 breaks that are 1:50 long. Each includes 90 seconds of commercials, then 20 seconds for the network. Those 20 seconds can be a network promo, billboards ("Today's game is brought to you by...."), or sometimes just going back to the booth for extra commentary. The breaks may be longer in the Super Bowl, but it's probably the same structure. I remember a couple of extra network promos during the blackout, and one set of commercials that they repeated after play resumed, but nothing else out of the ordinary.

by Ryan D. :: Tue, 02/05/2013 - 11:41am

My (limited) understanding of the ad structure was that advertisers pay for a commercial at Time/Break X in Quarter Y. If they take a commercial break to start the power outage (say, break #2 in Quarter #3), CBS can't take another break for paid ads until the clock moves again. Someone specifically paid to have their ad air between the 3rd and 4th quarter, or at the two minute warning, and you can't move those commercials up. What you get during this extended break is filler, which is why we got a LOT of ads for How I Met Your Mother, Two Broke Girls, NCIS, etc. and a whole lot of Solomon Wilcots.

by Fo (not verified) :: Tue, 02/05/2013 - 3:54pm

If you break DVOA down by quarter, do early leads correlate with decreasing DVOA?

by MITCH (not verified) :: Fri, 02/08/2013 - 10:27am

Once again I'd have to disagree with this weak Dvoa method. Ravens out-played the 49ers in the areas that have a high correlation to winning games.