The Baseball Economist: The Real Game Exposed (19 page)

Rickey saw an inefficiency in the market and exploited it at the expense of his competitors. And it didn’t take long for other teams to see what a bargain Branch Rickey had found. Eleven weeks after Robinson joined the Dodgers, the Cleveland Indians signed Larry Doby to be the first black player in the American League. In his first full season in the majors (1948) he helped the Indians win the World Series. In fact, the teams that were quick to integrate were reaping huge rewards in wins. By exploiting the inefficient use—or nonuse—of black talent, teams were able to win more games.

Eventually, black talent spread throughout the major leagues, with the best Negro League talent jumping to the AL and NL, until there were no more gains to be had by adding additional talent. This demonstrates one of the other characteristics of innovation: successful innovative practices will spread. Once the rest of baseball realized the success of black athletes, the inefficiency went away. Players who were once cheap and untouchable became caught in bidding wars between teams. The price of talent again rose to its old level, where black and white players of equal talent cost teams the same. The market for baseball talent was no longer inefficient when it came to race.

The great lesson in the reversion of rewards from innovation is that inefficiencies in markets don’t persist for long. To win by innovation one must be adept at finding inefficiencies and swift at acting on them once they’re found. No organization can build a baseball dynasty off one lucky strike. Only the continued pursuit of inefficiencies will lead to perpetual success. Which brings us back to the A’s in
Moneyball
.

As I mentioned above, in
Moneyball
, Michael Lewis recounts Billy Beane’s entrepreneurial skill. Billy Beane didn’t have a huge untapped talent pool to draw from. Many teams in the 1990s and 2000s were scouring the globe for talent. To find new ways to win on their small budget, the A’s needed to develop new ways for finding talent that was hiding just beneath the nose of every scout across the country. The answer lay in the complexity of the data. Baseball generates many quantifiable results with every play, so many that it’s hard for even a group of people to generate meaning from these numbers. Everyone can look at what Barry Bonds does and see that the things Bonds does well cause his teams to win. He hits for average and power, draws walks, and doesn’t strike out.

There’s no way the A’s could afford to pay Bonds to come across the Bay from San Francisco to play in the Oakland Coliseum, so the A’s needed to find something like Bonds to win. But one player who resembled Bonds would still be too expensive. Gary Sheffield could have been considered a Bonds-lite at the time, but he too commanded a huge salary in the market for baseball talent. Instead, the A’s found several players who did things important to Bonds’s success just enough better than the average player to be valuable, but not well enough for the rest of baseball to know it.

The first step in building a Frankenstein Bonds from several players was to find out what it was about Bonds that made his teams win. Then they had to figure out the going rate for those qualities in the open market and, finally, check out the options of baseball talent packaged in individual human beings. Once these steps were completed, the A’s could snatch up any mis-priced players who would yield the wins that everyone else was missing.

While Bonds has many noble qualities, I’ll focus on the offensive stats of on-base percentage and slugging average. Bonds is very good in both of these areas, but that does not mean they are equally important. Winning baseball games requires scoring more runs than the other teams you play. Therefore, to discover the importance of each of these metrics in scoring runs, Beane’s assistant, Paul DePodesta—an economics major from Harvard who would become the general manager of the Dodgers for a short time—used statistical methods to estimate how much each statistic impacted run scoring in recent history. To the surprise of many, DePodesta found something that everyone else was missing: each point of on-base percentage was worth about three times each slugging percentage point in producing runs. And this finding just happened to coincide some very favorable conditions in the market for baseball talent:

In major league baseball . . . Paul’s argument was practically heresy. . . . Heresy was good: heresy meant opportunity. A player’s ability to get on base—especially when he got on base in unspectacular ways—tended to be dramatically underpriced in relation to other abilities. . . . The one attribute most critical to the success of a baseball team was an attribute they could afford to buy.
⁵⁸

In this passage, Michael Lewis captures the essence of entrepreneurship. DePodesta had just stumbled upon a market inefficiency, a tool that would allow the A’s to purchase wins at a price less than what other teams had to pay for those wins.

Economists Jahn Hakes and Skip Sauer have verified the underpricing of on-base percentage and overpricing of slugging percentage at the time
Moneyball
was written—during the 2002 season. And this is just one innovation that would allow the A’s to purchase wins at a lower cost than the rest of the league. However, the success the A’s would have with on-base percentage is something that would not even last until the book hit the best-seller lists. Hakes and Sauer find that the inefficiency did not last for long:

This diffusion of statistical knowledge across a handful of decision-making units in baseball was apparently sufficient to correct the mispricing of skill. The underpayment of the ability to get on base was substantially if not completely eroded within a year of Moneyball’s publication.
⁵⁹

Innovation is a process, and though its returns are positive and real, they are short-lived. Just as the wins from integration evaporated, so did the wins from exploiting the underpricing of on-base percentage. To stay on top, the A’s would have to do new things. While the A’s had the third highest on-base percentage in the AL by 2000 and 2001, in 2002 they would fall to fifth and in 2003—the year
Moneyball
was released—the A’s were a woeful tenth in on-base percentage. How did the A’s manage to win ninety-six games in 2003 if on-base percentage is so important? The A’s found new inefficiencies to exploit. Having a good on-base percentage isn’t the only way to win. Beane didn’t reveal all his secrets in
Moneyball
. Maybe this was partly to hold on to some trade secrets to use against his rivals, but I think it’s more likely that those secrets just hadn’t been discovered and employed yet.

It was probably not a single magic bullet that propelled the A’s’ success; instead, it was an intellectual infrastructure put in place by Beane’s predecessor, Sandy Alderson. Beane, who worked closely with Alderson until he left the club, continued to modify the organization so that it could adapt to new information and the baseball market. By 2005, Beane had lost or traded away nearly all of the players that had been a part of the A’s recent success, yet the A’s kept on winning. An important lesson of this story is not just that an inefficiency can be exploited, but that these moments of inefficiencies are fleeting. The market acts fast to correct for its mistakes, so if you want to exploit them, you have to develop a method to always be looking for them and ready to act.

One of the misunderstandings of
Moneyball
is that Billy Beane had found a honey hole of success that twenty-nine other intelligent GMs chose to ignore. Nor was Beane the only GM stumbling on to new and innovative ideas that helped his team win. He is just one example of a successful entrepreneur. Economists refer to the pursuit of new innovative ways of doing things at the cost of discarding outdated methods as
creative destruction
—a term coined by economist Joseph Schumpeter. Creative destruction is a natural function of the competitive process, which is as old as humanity. Just as the wheel replaced dragging and the CD replaced the vinyl record, so too have many inferior practices been replaced through the process of creative destruction in baseball. Integration forced many marginal white players to leave the game, and the use of statistical tools may replace some of the work of traveling scouts. In the short run, this process seems unfriendly and mean; but the end result is unquestionably good: we get to consume more and better baseball for less.

Unlike creative destruction in the product market, new technologies do not lead to ever increasing wins for a team. For example, if someone invents a new method to produce corn, the price and quantity of corn will change permanently. The extra effort and money used to produce and purchase corn can now be devoted to producing and purchasing other goods and services. However, in baseball, wins and losses are a zero-sum game. One team’s wins necessitate other teams’ losses. As one innovation spreads from one team to the league, the innovative edge simply disappears. So how does creative destruction benefit baseball as a whole? If teams can identify and use all the best talent available, the outcome will be a more interesting game. Younger players may be identified earlier, struggling teams regenerate faster, plays become closer and more spectacular as everyone improves together, etc. Off the field, those resources that used to fund old methods can be devoted to making the ballpark experience better. Fans welcome cheaper tickets, more comfortable seats, and bigger, clearer scoreboards. These are just a sample of the benefits that result from managerial innovation. It is often overlooked that fans benefit substantially from such innovation.

11

Scouts vs. Stat-Heads

—MICHAEL LEWIS, MONEYBALL
⁶⁰

A GENERAL MANAGER must be efficient with his operation. In economic terms, being efficient means the GM attempts to put all of his resources to their most highly valued uses. If the market overvalues a particular baseball talent, such as saves, you liquidate your assets in this area. If the market undervalues a talent, such as on-base percentage, you acquire it while it’s cheap. It’s all very simple in theory, but difficult in practice. And to continue winning, a team must continue to innovate, as the profits returned from new inventions will disappear.

Finding new ways to win is a goal that all GMs aspire to. Whether it’s hitting, pitching, fielding, aging, speed, intelligence, or fan popularity, it’s the GM’s job to put a value on it. Once the values are in place, he can do the wheeling and dealing the fans love to analyze. All of this translates into winning. One of the newer ways for GMs to evaluate talent is to use statistical methods to scout players. Thanks to
Moneyball
, we know it’s something that the Oakland A’s use heavily, but they’re not the only team doing so.

The proper role for statistical methods is a sensitive issue in baseball these days.
Moneyball
made the debate public, but it was already happening. The backlash came from the traditionalists within the game, but not the same traditionalists who think the DH and lights in Wrigley are travesties. It came from those who believe in the traditional method of identifying talent by “scouting,” by putting a pair of human eyes on potential players. How can anyone learn more about what’s happening on the field by staring at numbers or charts on a computer screen?

Traditional scouting isn’t all that different in its approach from “performance scouting,” the subject of the opening quote, in terms of identifying future major-league ballplayers. The goal of both approaches is to predict the future production of major-league talent from a vast pool of prospects. It’s the method of prediction that divides these schools of thought. These two philosophies are often thought of as extreme stereotypes. Scouts are old men armed with radar guns and Skoal, who spend their time on the road scouring the high school and college ranks. Performance scouts, or “stat-heads” as they are more commonly known, are computer geeks who took
The Matrix
a little too seriously. It’s the numbers that matter to them, and not just any numbers. Stat-heads often use nontraditional stats, some developed and tested with sophisticated empirical techniques.

Stat-heads don’t just stare at numbers or combine numbers at random to judge players, and they certainly aren’t ignoring the human element of the game. The bias against statistics isn’t something unique to the baseball establishment; it’s something pervasive in society. Using statistics requires computers and numbers, very impersonal things, but they are tools that can help us see things that our eyes miss on the field. And in many cases, numbers are more honest than our eyes.

But the idea that you can evaluate players using statistics alone just cuts against everything the old scouts are about. There’s no gut instinct or personal experience from watching hundreds of games a year. Rather than viewing players in person, the performance scouts look for specific markers of future success as predicted in statistical models. These markers were chosen because many models involving hypothetical simulations, regression analysis, or just thought experiments said they were the best. While judging human beings solely on numbers is a little unsettling, think about all the transactions in life that are made without personal contact. Banks make loans based on credit scores, insurance agencies set premiums based on the number of speeding tickets you’ve had in the past five years, and eBay auction winners feel safe sending money to a stranger because he has a 97.4 percent positive feedback score.

This is not to say that the any team could succeed without putting some eyes on players, but in order to beat the competition they’re also looking for the things that the old methods might have missed.

It’s not that numbers don’t matter to scouts; they do. But over the years scouts have learned, correctly, that numbers can be deceiving. With the baseball talent spread so thin in amateur ball, it’s hard to know what those numbers mean. High batting averages, home runs, and strikeouts may be a product of poor competition, not major-league skills. These statistics can have very little to say as raw metrics, a problem that stat-heads are aware of. Scouting involves putting a pair of eyes on the players to try and pick out the guys who are going to play pro ball.