Sunday, January 14, 2007

10 yards vs. 15

The current hot-button issue for the results-driven analysis on ESPN is the Eagles' fourth down play calling in last night's game. With two minutes left, two timeouts, and the ball near midfield, trailing by three, Philadelphia elected to go for it on fourth and 10. After they completed a 16-yard pass, the play was called back on a penalty, and they elected to punt rather than attempt to convert fourth-and-15.

As usual, the "experts" focused on two points of view:

1. Philadelphia lost the game, so they must have made the wrong decision
2. Philadelphia gained more than 15 yards on their last attempt, so they made the wrong decision

Of course, neither of these things mattered. The only important factor is whether the Eagles have a better chance of winning the game by attempting to convert on fourth down or by punting. What's my take? I'm not certain. Win expectancy in football is tricky to calculate, and there's not a statistically significant amount of historical data to look at.

What I do know is that there must be a break-even distance, somewhere between one inch and 50 yards, where punting and going for it create an equal win expectancy. At this point, the Eagles should be indifferent between their options; it represents the line dividing the strategies of always going for it and always punting.

As I said, I'm not sure what this exact distance is, or even whether it's somewhere between 10 and 15 yards. I am sure that Andy Reid doesn't know it either. I also know that while Reid can be excused for not being able to calculate the precise figure in his head, the Eagles cannot be excused for not having someone on staff who knows how to make these calculations and apply them.

For $25,000/year, or perhaps even less, the Eagles could have hired an expert in analyzing football situations by win expectancy. I believe it's very conservative to say this expert could increase their chances of winning each game by 1% on average, for an additional .16 wins over the course of a regular season.

In an industry where teams spend over $10 million per win, $25,000 is a pittance. It can buy the Eagles .0025 wins, or .16. The Eagles, and every other NFL team, have made it clear where they stand on the issue. (There may be teams who hire these types of personnel, but they don't listen to them very often, so it doesn't really matter.)

Traditionally, games of skill and chance were dominated by the "old school" of thinkers, who based their conservative strategies on fundamentals passed down to them and honed them with years of experience. All these games are gradually being taken over by quantitative types who challenge every assumption and use math and computers to develop better approaches. Backgammon was permanently taken over by the "quants" in the early 1980s; later that decade, the math types hit Wall Street and made a fortune. More recently, sabermetricians have developed new statistics and analyses that are now used by all 30 MLB teams. Poker is experiencing the same revolution; the new ruling class is comprised not of road gamblers, but of quantitative geniuses like Bill Chen, author of the excellent Mathematics of Poker, who won two events at this year's World Series of Poker.

It's time for an NFL team to take the leap. Statistical analysis in football is nothing new; The Hidden Game of Football has been around for nearly 20 years, and Football Outsiders is doing good work and expanding their knowledge base every day. The first team in the pool usually reaps the most benefits, whether it's the Dodgers for signing Jackie Robinson and Roy Campanella; the NBA teams who drafted "raw" high schoolers Kevin Garnett, Kobe Bryant and Tracy McGrady; or the A's for handing the reins to Billy Beane.

Yes, it's hard for an NFL coach, who has been around the game for 30 years, to hand over the decision-making duties to a 25 year old nerd with a calculator and a spreadsheet. But if the coach cares more about helping the team win than his own ego, he'll do the right thing, and his team will be better off for it. The next time you turn on your TV, maybe your favorite team won't be punting on fourth-and-inches from the enemy's 35-yard line.

No comments: