Monday, December 18, 2006
Updated: April 15, 3:45 PM ET
If I Only Had The Nerve
By Michael Lewis
Three years ago I published a book about baseball, called Moneyball, that sought to explain how the A's won so many games with so little money. The short answer was that A's front-office folks had abandoned old-fashioned baseball management for a new complexity. They'd found better ways to use stats, which enabled them to value players and strategies more accurately than their competitors did. Before the book was released, I sent a copy to Oakland GM Billy Beane. He had no idea what I'd written, and I was worried what he'd make of it. But Beane didn't care that the rest of baseball might read the book, ape his methods and eliminate his advantage. "You don't really think anyone in baseball is going to read your book?" he almost snorted.
And he was right. Baseball people generally dismissed Moneyball, along with the possibility that they'd squandered a lot of owners' money. Ah well, I thought, that's just baseball. But I was wrong. There's another industry in which people have found it easier to resist a useful new idea than to rethink their old ones: professional football.
Exhibit A is a paper that was available on the Internet more than four years ago by David Romer, a UC Berkeley economics professor. The paper, solemnly titled "Do Firms Maximize? Evidence From Professional Football," which was published this past April in the Journal of Political Economy, tried to answer a simple question: What should an NFL coach do on fourth down?
Romer concluded that coaches punt or kick field goals on fourth down far too often. For example, when facing fourth and goal inside the opponent's 3-yard line early in the game, a coach nearly always opts for the field goal. (In Romer's sample, teams kicked 38 of 47 times.) The team gets three points, fans cheer politely and announcers agree that the coach made a sound decision.
But he didn't, at least not according to Romer. A team that goes for it in that situation scores 43% of the time, which means the expected points from going for it are the same as the expected points from kicking a virtually certain field goal: three.
But the team that goes for it and doesn't make it leaves the opponent, on average, in far worse field position than the team that kicks. Romer calculates that a team in this situation would need less than an 18% chance of scoring a TD when it goes for the first down to make a field goal attempt preferable. According to Romer, each time coaches sent out their field goal kickers, they reduced the chances they would win by 3%. Romer's research seemed a small but meaningful contribution to football strategy.
But it has had, if anything, the opposite of its intended effect. Back in 2001, before Romer circulated his research, the average team went for it on fourth down 15.1 times per season. During the 2005 season, the average NFL team went for it on fourth down 14.5 times. Which raises an interesting question: What are they thinking? A few possibilities suggest themselves:
Romer is wrong. It wouldn't be the first time an economist tried to screw up the real world with a crackpot idea. But it's been more than four years since Romer circulated his paper, and it has been picked apart by some very smart people—including academically minded NFL front-office types, who spoke with me but asked to remain nameless—and no one has been able to refute it. However they may quibble with this or that assumption, no one has found a serious flaw in Romer's math or his thinking or, for that matter, his intuition. Which brings us to a second possible explanation:
Romer is right, but no NFL coach is smart enough to understand him. Romer hints at this in his paper delicately, saying, "Many skills are more important to running a successful football team than a command of mathematical and statistical tools … Decisionmakers may want to maximize their teams' chances, but rely on experience and intuition rather than formal analysis."
Pro football has a false reputation for being the game of choice for rockheads. Given half a chance, coaches will encourage this view—and at least one enterprising journalist, Greg Garber of ESPN, gave them half a chance by thrusting Romer's paper in front of them and asking them to comment.
"This is a professor from Cal-Berzerkely?" asked then-Giants head coach Jim Fassel, a former Stanford assistant who couldn't resist a dig at his old rival. "It's easy to sit there and apply a formula," said Bill Cowher, "but it's not the easiest thing to do on Sunday. There's so much more involved with the game than just sitting there, looking at the numbers and saying, 'Okay, these are my percentages, then I'm going to do it this way,' because that one time it doesn't work could cost your team a football game, and that's the thing a head coach has to live with, not the professor." To which he added: "If we all listened to the professor, we may be all looking for professor jobs."
When an NFL coach pretends he has nothing in common with professors, he is merely guarding his reputation with fans and the media. But the complexity that a coach deals with week to week would cause the average professor's brains to melt and seep out his ears. More to the point, NFL coaches revel in the difficulty of their jobs and work very hard behind the scenes to make them more, not less, complicated. They spend countless hours analyzing film, creating schemes, managing salary caps and trying to outwit their opponents. They employ teams of assistants, some of them embarrassingly highly educated, to help them think through whatever problems must be thought through. If coaches didn't have some good reason for avoiding Romer's argument, they would embrace it, understand it and use it. Which suggests a third—and, to me, the most likely—possibility:
Romer is right, NFL coaches are perfectly capable of understanding that he is right, and they are not as perfectly interested in winning games as they pretend to be. Why, you surely must be wondering, would any coach ignore information that would help him, however slightly, win games? Romer has wondered about this too. He raises the possibility that fans or the owner might exert pressure on coaches to behave irrationally. But on short fourth downs, fans are usually yelling to go for it, so they would seem an unlikely source of pressure in the other direction. (This may be one of those cases in which mob sentiment is more intelligent than the individual who resists it.) And owners are typically agnostic on the nitty-gritty details of strategy; they just want to be loved and admired, and to attract fans. The best way to achieve their desires is to win games—a goal more likely to be reached if the coaches they pay so handsomely went for it more often on fourth down.
Another possibility—that's harder to dismiss—is the mind-numbing effect of the sports media. People whose job it is to voice their opinions have a special talent for second-guessing even the smartest decisions that don't work out. Any coach who breaks with conventional wisdom and goes for it on fourth down is taking the risk of provoking his loudest critics. Yes, he might improve his team's chance of winning, but at the price of being hounded by some guy whose idea of being clever is to call himself Mad Dog.
Still, if coaches care about the opinions of the media, they have a funny way of showing it. When coaches think about the media at all, they seem to think about them mainly with contempt. And there's a reason. The difference between what the media know and what coaches know is greater in football than in any other sport. Media types generally don't know the plays, for instance, and generally aren't invited to watch game tape. So they never truly know the why or wherefore of any matter on the field. This gap in knowledge between insider and outsider breeds condescension and disdain.
On the other hand, there is one group of people whose opinion coaches care intensely about: other coaches. Cast your eye upon the sideline and you will find evidence of their desire to get along with one another. Invariably you'll spot some older fellow, himself once a less-than-successful head man, now paid in the high six, or even seven, figures to coordinate the offense for his old pal, the current head coach. Further down you will see some young fellow in charge of the special teams who longs to be put in charge of the defense. None of these people could make anything like the living they do outside of football. And the main requirement for remaining inside football is that the other people inside football want you to be there.
Here may be the nub of the fourth-down quandary. It's not that the decision is so complex, but, rather, that it's so conspicuous—far more so than, say, a new rule or scheme. Go for it on fourth down more often than any other coach, and you not only set yourself apart from your peers, but you call into question their intelligence. If your decision doesn't pay off—if you go for it routinely and your team fails—you'll stand accused of malpractice. If your decision pays off—if you go for it routinely and succeed—you'll be branded a maverick or, worse, a gambler. (The same tendency, by the way, has been demonstrated on Wall Street, where many fund managers shy away from big investments that may distinguish them from their peers.) And God help you if you stop winning. Break with the pack on such a conspicuous decision and you don't merely improve your chances of winning. You insult all coaches who do things the old-fashioned way. It may be excellent football strategy, but it's a risky career move.