Scorecasting (7 page)

Still another abstract benefit of playing for Pulaski: The experience is so different from traditional high school
football that the Bruins’ players feel as though they’re part of something unique, an elite unit amid regular cadets. The team bonds have solidified; the offensive and defensive players consider themselves kindred spirits, bracketed together by their singular coach. And there are so many trick plays and intricate formations that players, by necessity, are alert at all times.

Happy as Kelley is to unleash his empirical evidence, these are the numbers that matter most to him: In the years since he took over as head coach, Pulaski is 77–17–1 through 2009, winning 82 percent of its games, and has been to the state championship three times, winning twice. All this despite drawing talent from only a small pool of private school adolescents. “I’m telling you,” says Kelley. “It works.”

It’s up for debate whether Kelley’s operating principles would work in all cases, for all teams, on all levels—for the record, he thinks they would—but his success at Pulaski is beyond dispute. With that record, you’d think other coaches would try to implement some form of Kelley-ball, but although he has become a minor celebrity in coaching circles and speaks at various banquets and conferences, he has not been flattered sincerely by imitation. Other coaches have cribbed the West Coast offense from
Bill Walsh, the former Stanford and San Francisco 49ers coach, or the spread formation from
Mike Leach, late of Texas Tech, but Kelley draws little more than a curious eye. “If there’s another team out there that don’t ever punt,” he says with a shrug, “I haven’t heard of ’em.”

Several years ago, a prominent college coach paid a visit to Kelley’s office at Pulaski, a nondescript box off to the side of the basketball court. The coach—Kelley doesn’t want to name him for fear it might hurt the future recruitment of Pulaski players—asked for a primer on “that no punting stuff.” Kelley happily obliged, explaining his philosophy and showing off his charts. “He wrote all sorts of stuff down in this big old binder and I’m thinking, ‘Finally someone else sees the light.’ ” But when Kelley watched the coach’s team play the next season, he saw no evidence that he had a disciple. Even armed with the knowledge that he was disadvantaging his team by his decision to punt, the coach routinely ordered the ball booted on fourth down.

That mirrors David Romer’s experience. In his paper, Romer, the Berkeley economist, argued that the play-calling of NFL teams shows “systematic and clear cut” departures from the decisions that would maximize their chances of winning. Based on data from more than 700 NFL games, Romer identified 1,068 fourth-down situations in which, statistically speaking, the right call would have been to go for it. The NFL teams punted 959 times. In other words, nearly
90
percent of the time, NFL coaches made the suboptimal choice.

Inasmuch as an academic paper can become a cult hit, Romer’s made the rounds in NFL executive offices, but most NFL coaches seemed to dismiss his findings as the handiwork of an egghead, polluting art with science. Plenty admit to being familiar with Romer’s work; few have put his discoveries into practice.

It all lays bare an abiding irony of football. Here are these modern-day gladiators, big, strong Leviathans. It’s a brutal, unforgiving game filled with testosterone and bravado. Players collide off each other so violently that there might as well be those cartoon bubbles “Pow” and “Bam.” The NFL touts itself as the baddest league of all. Yet when it comes to decision-making, it’s remarkably, well, wimpy.

There’s not just an aversion to risk and confrontation; coaches often make the
wrong
choice. In other words, they’re just like … the rest of us.

Time and again, we let the fear of loss overpower rational decision-making and often make ourselves worse off just to avoid a potential loss. Psychologists call this loss aversion, and it means we often tend to prefer avoiding losses at the expense of acquiring gains. The psychologists
Daniel Kahneman and
Amos Tversky are credited with discovering this phenomenon. (Kahneman won the Nobel Prize for this work in 2002; Tversky died in 1996 before being recognized.) As the late baseball manager
Sparky Anderson put it: “Losing hurts twice as bad as winning feels good.”

For most of us, the pain of losing a dollar is far more powerful than the pleasure of winning a dollar. In a frequently cited psychology experiment, subjects are offered two gambles that have identical payoffs but are framed differently. In the first gamble, a coin is flipped, and if it lands heads, you get $100; if tails, you get nothing. In the second gamble you are given $100 first and then flip the coin. If the coin lands heads, you owe nothing; if tails, you pay back the $100. Subjects dislike the second experiment much more than the first even though the actual gains and losses are identical.
^*

Marketing and advertising execs cater to this bias. Would you rather get a $5 discount or avoid a $5 surcharge? The same change in price framed differently has a significant effect on consumer behavior. A study of insurance policies, for instance, found that consumers switch companies twice as often when their carrier raises rates, as opposed to when the competition decreases its rates by the same amount. In everyday life, loss aversion causes people to make suboptimal choices. Many home owners looking to sell their houses right now would rather keep them on the market for an extra year than drop the price to $5,000 less than they paid, even though keeping the home for an extra year will surely cost them more than
$5,000. A study of home sales by two economics professors,
David Genesove and
Christopher Mayer, then at the University of Pennsylvania’s Wharton School of Business, showed this pattern. Home owners were reluctant to reduce the sale price below what they paid for the house even when continuing to own it meant incurring carrying costs—mortgage, utilities, maintenance—far exceeding the reduction in price needed to sell it. The idea of a loss was just too painful for them. In contrast, home owners facing a gain on a house often sold too early and for too little. The gain didn’t matter as much as long as there wasn’t a loss.

On Wall Street, fear of loss is often behind dubious investment strategies. Mutual fund managers, for example, will hold well-known or recognizable companies instead of obscure companies that are expected to deliver much better performance. The rationale: If you lose money by buying Walmart or Microsoft—recognizable blue chip companies—no one will blame you. You won’t get fired; they’ll chalk it up to “bad luck.” Even though a small, obscure company might be a better bet, on the off chance that it doesn’t pay off, you risk losing the client. So it is that many mutual fund managers will choose good companies over good investments.

On the television reality show
The Biggest Loser
, obese contestants compete to lose weight. The more they lose, the more they are rewarded. Two Yale professors,
Ian Ayres, an expert in contract law, and
Dean Karlan, a behavioral economist, were desperate to lose weight. Like the
Biggest Loser
contestants, they tried to find motivation in rewards. It didn’t work, and so they flipped the
Biggest Loser
concept around and tried to motivate themselves with
loss aversion. They entered a weight-loss bet with each other, and each one committed to pay the other $1,000 a week if he didn’t lose the required weight. In addition, once the weight was lost, it couldn’t be gained back without incurring the $1,000 penalty.

Two years later, neither professor has seen a dime of the other’s money—and they’ve lost almost 80 pounds between them. They launched a company,
stickK.com
, to help people facilitate personal commitment contracts for
weight loss and other personal goals by using loss aversion. If you don’t live up to your end of the
contract, they give your money to charity or a designated beneficiary. (In another variation, the losers have to donate the money to a cause that runs counter to their political sensibilities: gun haters contributing to the NRA, pro-lifers contributing to Planned Parenthood.)

This same
loss aversion affects coaches. They behave much like the shortsighted mutual fund manager who forgoes long-term gains to avoid short-term losses and the amply girthed professors who could lose weight only when faced with a loss rather than a reward. The coaches are motivated less by potential gain (a touchdown) than by fear of a concrete loss (the relative certainty of points from a field goal).

More broadly, many coaches ultimately are motivated less by the potential of a Super Bowl ring than by the potential loss of something valuable they possess: their job. And in sports, there are few faster ways to lose your employment than by bucking conventional thinking, by trying something radical, and failing. A coach ordering his team to punt on fourth and three—even when it’s statistically inadvisable—faces little backlash. He is the money manager who plays it safe and loses with Walmart. If he goes for it and is unsuccessful, there’s hell to pay. He is then the money manager who loses on that unknown tech stock and now risks losing the entire account.

It makes for an odd dynamic in which the incentives and objectives of coaches aren’t perfectly aligned with those of the owners or the fans. All want to win, but since the owners and fans can’t be fired, they want to win at all cost. Give a coach truth serum and then ask what he’d prefer: go 8–8 and keep your job or go 9–7 and, because of what’s perceived to be your reckless, unconventional play-calling, lose your job?

It’s not just football coaches who make the wrong choices rather than appear extreme. In
basketball, for instance, prevailing wisdom dictates that coaches remove a player with five fouls, particularly a star, rather than risk having him foul out of the game. But does this make sense?

We can start by measuring how long a player sits on the bench once he receives a fifth foul. We analyzed almost 5,000 NBA games from the 2006–2007 to 2009–2010 seasons and found that when a player receives his fifth foul, on average, there is 4:11 left to play in the game. He’s benched for about 3:05 of that remaining time, leaving only 1:06 of actual playing time with five fouls. Stars are treated a little differently.
^*
On average, they don’t receive their fifth foul until there is 3:44 left, and coaches bench them for a little more than two minutes.

The strategy of sitting a player down with five fouls and waiting until the end of the game to put him back in presumes that players, particularly stars, are more valuable at the end of the game than at other times. But this is seldom the case.

Statistical analysts in basketball have created “plus-minus,” or an “adjusted plus-minus,” a metric for determining a player’s worth when he is on the floor. Simply put, it measures what happens to the score when any particular player is on the court. When a player is plus five, that means his team scored five more points than the opponent when he was on the floor. Thus, this measure takes into account not only the individual’s direct influence on the game from his own actions but also the indirect influence he has on his teammates and his opponents. It measures his net impact on the game.

As often as we hear about “clutch players,” for the average NBA player, his contribution to the game, measured by plus-minus, is actually almost two points
lower
in the fourth quarter than in the first quarter. This is also true for star players and is even the case in the last five minutes of the game. Thus, the strategy of sitting a player down with five fouls to save him for the end of the game seems to be based on a faulty premise—he is no more valuable at the end of the game.

Now consider who replaces the player when he sits on the bench. The average substitute summoned in the fourth quarter to
replace the teammate in foul trouble, not surprisingly, has an even smaller impact. Replacing the star player in foul trouble with a sub has the net effect of reducing the team’s points by about 0.17 for every minute the star is on the bench. This is a heavy price to pay. (We considered that a star player in foul trouble might compete conservatively, so maybe the difference between a sub and a star who plays conservatively with five fouls isn’t all that great. But no, it turns out that’s not true. If anything, star players have an even
higher
plus-minus than normal when they are in foul trouble.)