When STATA meets Sunday

By
Friday, January 30, 2009

So there’s a football game being played this Sunday. You’ve probably heard of it. And you’re probably sick of all the media coverage that goes with it – except for the commercials. The talking heads discuss intangibles, “smash-mouth football” and let’s not forget the importance of establishing the run. But what you might not have heard is that the time-honored cliches and conventional wisdoms of the highest-rated professional sport in America are now being challenged by a new breed of sports analysts.

Enter footballoutsiders.com. Founded with the objective of disproving the importance of establishing the run, a group of fans and Zeta Delta Xi alums started doing to the NFL what sabermetricians had been doing for years in Major League Baseball: use statistics. Rather than rely solely on accounts of player toughness and conventional benchmarks of individual performance, they sought to create their own measurements.

Why is the importance of establishing the run overrated? Because correlation and causality frequently get screwed up. Most teams that win comfortably happen to run the ball successfully. Often, this is because they already have a lead, and they are trying to milk the clock. But there’s no point in a team repeatedly running Larry Johnson into the line for no gain. If you don’t think any head coach wouldn’t understand this, re-watch Herm Edwards “coach” the Chiefs against the Colts in the 2006-2007 playoffs.

Cliches aside, what’s the problem with conventional statistics? If I were to ask you which running back played better this year, Minnesota’s Adrian Peterson or Houston’s Steve Slaton, it probably wouldn’t be a very difficult choice. Sure, they both averaged 4.8 yards per carry, but Peterson had more touchdowns and almost 500 more rushing yards while leading the NFL in rushing and the Vikings to a division title.

Problem is, football isn’t tennis. More specifically, Peterson and Slaton don’t play on one-man teams. In fact, their performance is based largely on how well their teammates and their opposition play. More than any other game, the team dynamics of football are astronomically important. Which is why it seems so odd that individual stats for NFL players don’t even try to take these factors into account.

So the Football Outsiders developed stats that do more than shrug their shoulders at how team play affects individual performance. Two of these are Defense-adjusted Value Over Average (DVOA) and Defense-adjusted Yards Above Replacement player (DYAR). “DVOA breaks down every single play of the NFL season to see how much success offensive players achieved in each specific situation compared to the league average in that situation, adjusted for the strength of the opponent,” the site reads. DYAR estimates the amount of yards a player gained given his DVOA over an average player in his position.

Wouldn’t you know it? Peterson had a DVOA of -.2 percent, while Slaton had 5.8 percent and over 100 more DYAR. Or for those of you who prefer English, it turns out that having a monster offensive line like the Vikings and playing the Lions twice a year makes your running backs look better.

Obviously there’s no perfect way to control for all the other players on the field, and certainly Peterson still deserves credit for a great season. That doesn’t mean these new stats aren’t legit. Long gone are the days where scouting reports and gut feelings were all you needed to be an MLB general manager. They’ve been replaced by Billy Beane, Theo Epstein and a cohort of statisticians to supplement conventional analysis. Slowly, the same trend is starting to take over the NFL. Football Outsiders is becoming a popular stats source, notably cited on ESPN news blogs. And newly hired Detroit Lions head coach Jim Schwartz, a former Patriots intern and Tennessee Titans defensive coordinator, got his start with the statistical finding that fumbles are random and unrelated to team performance.

So what do these new stats say about Sunday’s game? Well, Pittsburgh’s weighted DVOA was first in the NFL at 30.8 percent, while Arizona was good for 21st at -7.9 percent. And as of last weekend, the Cardinals had a 13 percent chance of winning it all. Everybody already knew the Cards were underdogs.

But what’s probable doesn’t always happen. Just ask last year’s New York Giants. The Cardinals are still playing, even though they probably should have been slaughtered long ago. The fact that they’re still alive attests to one everlasting and often overlooked statistical truth – regardless of how small the probability, nothing is impossible as long as there’s a chance.