Sterling Moore and Split-Play WPA

Sterling Moore posted a stellar +0.47 +WPA in Sunday's AFC championship game. That's very good -- only Patrick Willis (+0.52) and teammate Dane Fletcher (+0.49) beat that mark for defenders on championship weekend. After all, Moore made arguably the defensive play of the weekend when he knocked what would have been the go-ahead touchdown pass out of Lee Evans's hands with mere seconds to go.

But most of Moore's +WPA actually comes from his contributions on the following play, the failed third down pass targeted for Dennis Pitta which set up the fateful fourth down on which Billy Cundiff kicked the Ravens out of the playoffs. The Ravens were still in excellent shape on that third down, and the failure to convert or score a touchdown took their win probability down from 83% to 43%, giving Moore a +0.40 WPA on the play. That leaves just +0.07 for his other successful play, the strip of Evans in the end zone.

That seems intuitively way too low, and that intuition is correct. Although technically the entire play from snap to throw to almost-catch to strip just cost the Ravens 7% of a win, if Evans holds on to the ball and Moore doesn't strip it, Baltimore's ticket to the Super Bowl is all but punched. But with the way the data is fed into our system, it's impossible to give out separate credit for different aspects of plays.

But let's say for a second it was possible. How would each aspect of that play have played out in the eyes of WPA?

On Opponent Strength and Team Strength Correlation

This post at Football Outsiders caught my eye today. The IgglesBlog noticed something odd with their team rankings. I’ve notice the same phenomenon in my own systems—that team ranking methods that adjust for opponent strength tend to produce rankings that correlate (inversely) with a team’s strength of schedule. In other words, top ranked teams appear to have weaker schedules and low ranked teams appear to have stronger schedules. The problem is, assuming that a ranking method properly adjusts for opponent strength, it ostensibly should produce no correlation between each team’s ranking and its opponents' average ranking. In fact, we might expect the opposite result because of the two “strength of schedule” games each season—Last year’s 1st place teams play other 1st place teams, and so on.

In 2011 FO’s “DVOA” method correlated with opponent strength at -0.66, which is considerable. Here at ANS, Generic Win Probability correlated with Average Opponent GWP at -0.60 this season. FO notes that in other years the correlation isn’t nearly as strong, but there is an apparent tendency for negative correlations for most seasons.

This phenomenon was first pointed out to me a couple years back by a reader, and I too thought it was either a) randomness, or b) a flaw with my methodology. But I soon realized this is exactly what we should expect given the NFL’s scheduling rules. It’s neither luck nor a flaw. In fact, it's a sign the method is doing something right.

Consider a fictional four-team football league. Presume we have a perfect team ranking system that can peer omnisciently into each team’s soul to know its True Winning Probability (TWP). The Sharks, Knights, River Dogs, and Jack Rabbits each have a TWP of 0.75, 0.60, 0.40, and 0.25. (Notice the TWPs average to 0.50, as they would have to.)

Should The Niners Have Kept The Punt?

Who would have thought that Ted Ginn Jr.'s absence might have made all the difference in the NFC championship game? Kyle Williams' two fumbles on punt returns kept the Giants in the game and all but won it for them in OT. During the course of this 22-punt game, Jim Harbaugh was forced with a few 4th-down decisions. Earlier this year, Brian wrote about Harbaugh's decision to keep the 3 points after David Akers made a 55-yard field goal and the Cowboys were called for a 15-yard penalty. In the third quarter, down 10-7, the Niners were faced with a similar conundrum, this time with a punt. On 4th-and-6 from midfield, Andy Lee hits a beautiful punt the to the Giants' 7-yard line. Justin Tuck is called for running into the kicker, but Harbaugh declines and takes the punt, pinning the Giants deep. But, was this the right decision?

After the punt, the 49ers win probability was 37% (and their expected points were +0.34, meaning they were actually expected to be the next team to score even though the Giants had the ball). So the question is as follows: does going for it on 4th-and-1 after the penalty increase the Niners' chance of winning? The estimated success rate on 4th-and-1 is 74%. If San Francisco succeeds, their win probability jumps to 47%; if they fail, it falls to 31%. So, if we let x be the chances of converting on 4th-and-1, we have the equation 0.31*(1 - x) + 0.47*x > 0.37. Thus, the 49ers should go for it if x > 37.5%. Since the estimated conversion rate is 74% (almost twice our break even point of 37.5%), this seems like a no-brainer: the correct decision would be to take the penalty and go for it.

Roundup 1/21/12

Using portfolio theory to analyze fantasy football strategies. I tinkered with portfolio theory a while back, but ultimately understood it's not appropriate for real football analysis. It is however, well-suited for fantasy analysis.

A commenter linked to this a couple weeks ago. Correlation != causation.

2011 Giants = 2010 Packers? I buy that.


A different kind of look at Flacco.

Is the new rookie wage scale the reason for the record number of underclassmen declaring for the draft?

This is a good analysis of when teams ahead should try to score rather than run out the clock. I agree and made the same observation at the time on the WP graph comments. Fans and analysts typically call for teams to 'run out the clock' far too early. The sport has changed over the years to where offenses only need 1 minute to drive the length of the field for a TD. The 2-minute drill is an antiquated term. Two minutes is an eternity. (See the NO-SF game: 4 TDs in the 4 final minutes.) Helmet-knock: FO.

And Then There Were Four

A lot of media attention naturally goes to the quarterbacks at this point in the season. Much of it this week has shone on Joe Flacco in particular. While it's true he didn't light up the scoreboard like the other three QBs did last week, he was the only one playing against one of the league's best defenses. The others faced the 17th, 22nd, and 29th ranked defenses in terms of overall efficiency. A lot of the debate on Flacco has revolved around the notion that he just wins. I hear a lot of people cite his overall win-loss record as a starter as well as mentioning he has a very good defense backing him up. On one hand he's been to the playoffs 4 out of his 4 seasons, but on the other he's not the reason the Ravens win.

Win Probability Added seems it was made to settle debates just like this. And WPA says that since his second season in the league, Flacco is one of the main reasons why the Ravens have been a winning team. In fact, over his entire career his WPA has averaged +0.10 per game.  And his last two seasons were +0.21 and +0.19 WPA, well above average. In other words, his performance would make a .500 team a .700 team, all other things being equal. But not everything is equal. In particular, quarterbacks naturally have positive WPA simply because passing is more lucrative than running, plus it's been getting easier over time.

Running in the Cold

A few posts ago, we looked at how temperature affected the passing game. This time, we’ll look at the running game. Often, analysts will discuss how winds and cold might affect passes, but unless the conditions are exceptionally snowy or muddy, rarely does anyone consider how cold weather affects running. And why would they? I’d agree there isn’t much reason to suspect that cold temperature alone would cause runs to be any longer or shorter than in moderate weather.

Before we look at the numbers, I should note that running and passing are connected in game theory terms. The better a team’s passing attack, the more an opposing defense needs to respect it, possibly allowing bigger running gains. And same goes for a great running attack. The better it is, the more the defense needs to be on guard near the line of scrimmage, lowering its guard against the pass.

Cold temperatures, or at least the kinds of conditions that go along with cold temperatures, appear to reduce the effectiveness of passing. With that in mind, defenses might be worried slightly less about deep passes and stack the box in cold temperatures. Thus, we might expect that cold temperatures could indirectly reduce the effectiveness of running.

This is where it gets really interesting, because that’s not what happens at all.

Conference Championship Game Probabilities

Weekly game probabilities are available now at the nytimes.com Fifth Down. This week, I discuss some considerations about how the four teams are perceived, including factors like recency bias and the randomness of turnovers.

Postseason Projections: Conference Round

We'd previously warned that the eventual winner of the NFC East was not to be underestimated in the postseason, and the Giants showed why in last Sunday's win over the Packers. By a slim margin, the model now sees the Giants as the strongest of the four remaining teams, though their advantage is lessened by the fact that they will have to meet the 49ers on the road. Overall, the model now gives New York about one chance in three to win the Super Bowl, odds on par with those of the Patriots.

None of the remaining teams is particularly dominant at this point and none is a complete long shot. San Francisco, with the lowest probability of a Super Bowl win, is still given a 15% chance. And now, submitted for your approval, the final postseason projections of the 2011 season, with the table below listing each team's percent probability—first of advancing to the Super Bowl and then of winning the whole thing. Enjoy.

Percent Probability to Advance
Team	Super Bowl	Sup Bowl Champion
NE	64	35
BAL	36	18
SF	41	15
NYG	59	32