The Sabermetric Revolution: Assessing the Growth of Analytics in Baseball (2 page)

Moreover, in baseball, performance is much less interdependent than it is in other team sports. A batter gets a hit, or a pitcher records a strikeout, largely on his own. He does not need a teammate to throw a precise pass or make a decisive block. If a batter in baseball gets on base 40 percent of the time and hits 30 home runs, he is going to be one of the leading batters in the game. If a quarterback completes 55 percent of his passes, though, to assess his prowess we also to need to know something about his offensive line and his receivers.

So, while the measurement of a player’s performance is possible in all sports, its potential for more complete and accurate description is greater in baseball. It is, therefore, not surprising that since its early days, baseball has produced a copious quantitative record. Although one might not know it from either the book or the movie
Moneyball
, the keeping of complex records and the analytical processing of these records reaches back at least several decades prior to the machinations of Billy Beane and the Oakland A’s at the beginning of the twenty-first century.

Our book proceeds as follows. To clarify some matters of artistic license presented as fact,
Chapter 1
discusses the book and the movie
Moneyball
, what they get right, what they get wrong and various sins of omission.
Chapter 2
traces the growing presence of statistical analysis in baseball front offices.
Chapters 3
and
4
introduce and survey the current state of sabermetric knowledge for offense and defense, respectively.
Chapter 5
sketches the
Moneyball
diaspora, that is, the growing application of statistical analysis to understand performance and strategy in other sports, principally basketball and football.
Chapter 6
illustrates the use of statistical analysis to penetrate the business
of baseball, particularly its effects on competitive balance.
Chapter 7
assesses sabermetrics’ success, or lack thereof, in improving team performance.

Finally, it is useful to clarify some vocabulary before proceeding.
Sabermetrics
means the use of statistical methods to analyze player performance and game strategy.
Baseball analytics
also means the use of statistical methods to assess player performance and game strategy, but it further involves the use of statistical methods to evaluate team and league business decisions. The term
analytics
as applied to sports has also come to include the interpretation of digital video images, often with associated quantity metrics. We use
moneyball
(with the lowercase
m
) to mean the application of sabermetrics with the goal of identifying player skills and players that the market undervalues.

Michael Lewis’s 2003 bestselling book
Moneyball
has sold well over a million copies. The 2011 movie
Moneyball
has exceeded $120 million in box-office sales and was nominated for six Academy Awards, including best actor and best picture. It is safe to assume that the story that Michael Lewis fell in love with back in 2002 has been widely assimilated by people who care about baseball as well as by many who don’t. The book was a significant catalyst in spreading the sabermetric gospel in baseball front offices, as well as feeding the growing popularity of sports analytics over the Internet, in academia, and in fantasy sports leagues. In a sense, the book brought into the mainstream the incorporation of sabermetric practice within the baseball industry, much as Bill James had popularized new statistical ways of understanding the game and its players.

Yet, for all its storytelling virtues, the book, though containing an underlying truth, substantially misrepresents baseball reality, and the 2011 movie, as movies are wont to do, distorts reality still further. Thus, before we begin our discussion of the intellectual state of baseball analytics, its application in the industry, and its future prospects, it is important to clear away the popular debris that has been left behind by the two versions of
Moneyball
.

The film has the same basic storyline, stripped of its emotional embellishments and flourishes, as the million-copy-selling book. The Oakland A’s, a
small market team with a parsimonious owner, needed to find a way to remain competitive after the 2001 season. The team was going to lose three of its star players (Jason Giambi, Johnny Damon, and Jason Isringhausen) to free agency (and to the Yankees, Red Sox, and Cardinals, respectively), and the owner would not provide the cash to sign any worthy replacements. The A’s general manager (GM), Billy Beane, travels to Cleveland to discuss a trade for relief pitcher Ricardo Rincon and discovers that Cleveland GM Mark Shapiro is paying close attention to the opinions of a dorky-looking Yale grad on his staff (called Peter Brand on screen). After the meeting, Beane corners Brand in the parking lot and presses him to reveal how he approaches valuing baseball players. An enthralled Beane hires Brand and adopts a unique strategy to assemble a winning team based on Brand’s philosophy. (Brand’s character was based on the real-life Paul DePodesta, a tall, slender Harvard grad. Once he saw what the screenplay did with his character, DePodesta did not give permission to have his name used for the film.)

A central tenet of this unconventional philosophy is that teams pay too much attention to a hitter’s batting average (BA) and not enough attention to a player’s on-base percentage (OBP, roughly batting average plus walk rate and hit by pitch rate). The basic idea is that walks were dramatically undervalued; just like a hit, a walk puts a runner on base, avoids an out, and brings another batter up to the plate. Of course, many have also observed that players with a good eye at the plate help to run up the pitch count of the starting pitcher and accelerate getting into the opposing team’s bullpen.

In the movie, Peter Brand’s approach, in turn, is represented as being derived from that of Bill James. By focusing on OBP, the A’s could identify undervalued players and assemble a winning team on the cheap. (This is the idea of a market inefficiency. The actors in the market are making decisions based on incomplete or wrong information which means that some inputs—players in this case—are systematically paid more and others less than they are worth.)

The movie and the book both make the case that the A’s implemented this philosophy. Further, it is represented that the strategy worked and explains why the team won an American League record twenty straight games and the AL Western Division title in 2002.

One of the most dramatic events in the movie occurs when Billy Beane walks into the team clubhouse after a loss and sees the players dancing to music, led by outfielder Jeremy Giambi. (Beane acquired Giambi via trade before the 2000 season, but he is represented as having been acquired by the A’s during 2001–2002 offseason due to his high OBP and in order to help replace the lost OBP from his brother Jason.) The next morning Beane arrives at his office and, against the advice of his guru Peter Brand, in a fit makes two trades: sending Jeremy Giambi to Philadelphia for John Mabry and trading the A’s first baseman Carlos Peña to the Detroit Tigers for cash. Peña was reputed to be a leading candidate for Rookie of the Year honors, but A’s manager Art Howe was playing him instead of Scott Hatteberg, whom both Beane and Brand had designated as their first baseman. Jeremy Giambi, at the time of the trade, had an impressive OBP of .390, which only rose further, to .435, after he went to the Phillies. Meanwhile, John Mabry had a below-average OBP of .304 at the time of the trade and a subpar OBP of .322 with the A’s. (The average OBP in major league baseball hovers around .333, with small variance from year to year.) Mabry had only been weaker in earlier years, with an OBP below .300 in 1999, 2000, and 2001. Why would Beane make this trade?

After the trades, according to the screenplay, the mood in the clubhouse changes and the A’s suddenly become a winning team. The Giambi trade is shown as occurring on May 22, and the turnaround takes place after the team’s loss to the Orioles, 11 to 3, on May 23. The A’s record was 20 wins and 26 losses at this point. Following the loss on May 23, the A’s won five consecutive games and 24 of their next 29. By June 24, the A’s record was 44 and 31.

While it isn’t elaborated, the viewer is led to believe that the substitution of Hatteberg for Peña at first base was a key element to the team’s newfound success. One of the problems with this presentation is that Peña was not, in fact, traded until six weeks later, on July 5. Thus, the whole notion that the trades of Giambi and Peña were (a) based on sabermetric principles and (b) responsible for the team’s turnaround is belied by the facts.

Of lesser significance, a few other matters of artistic license in the film should be mentioned. First, the A’s frugal owner, Steve Schott, actually bought the team from the estate of Walter Haas in 1995 and had been enforcing a
tight budget from the beginning. The imperative to build a winning team on a shoestring budget did not begin in 2002. Second, no GM, and certainly not one on a tight budget, would fly across the country to discuss a potential trade for a relief pitcher, and when GMs discuss trades they don’t do so with eight other people in the room (possibly excepting a rare occasion at baseball’s winter meetings). The meeting was dreamed up as a way to introduce Paul DePodesta (aka Peter Brand) into the story. After the meeting Beane calls Brand and tells him that he’s been “bought” by the A’s. In fact, what happened is that in November 1998 (fully three years earlier) the A’s asked the Indians for permission to make DePodesta (Brand) their assistant general manager and the Indians agreed. Third, at the time of the apocryphal meeting, Mark Shapiro was not GM of the Indians, as shown in the movie.

Fourth, there are various sins of omission, but perhaps the most glaring is that Sandy Alderson is left out of the film altogether. Alderson served as GM of the Oakland A’s from 1983 through 1997. He hired Billy Beane and made him his assistant GM in 1993. More important, it was Alderson who introduced sabermetrics into the A’s organization in the mid-1980s (more on this below) and to Beane in the mid-1990s. Alderson was in the film’s early script, but his character was excised before the final version.

Fifth, there is, of course, the use of hyperbole throughout for cinematic effect.
¹
Listen to Paul DePodesta reflect on the film and describe the actual relationship between the stats analysts and the traditional scouts in the A’s clubhouse: “I think it’s overblown. . . . Surely there were spirited debates at different times internally, but they were always very respectful, with everybody largely on the same page. . . . Even some of the metrics we came up with were things that were born out of conversations we had with longtime scouts. So it really was an organization which I think was bound together much more than has been portrayed.”
²

Sixth, Billy Beane, in a heated exchange with Art Howe, asserts that “I don’t care about righty/lefty.”
³
Yet, if Beane were a true student of sabermetrics, he would know that there is ample evidence, going back at least to the statistical work of George Lindsey in the 1950s, that platooning can have a substantial effect on hitting success. (Lindsey, based on a statistical analysis of 400 games from the 1950s, found facing a pitcher of the opposite hand raised batting averages by a mean of 32 points.)
⁴

Seventh, Beane quips to the A’s players, “No more stealing.” This is a bastardization of the sabermetric wisdom. Although different analysts come to slightly different estimates, the basic sabermetric conclusion is that in an average game situation, a runner should only attempt to steal second base if he has at least a 65 percent chance of success.
⁵
Runners with a lower success percentage reduce the expected number of runs scored in an inning by attempting base thievery. This inference, of course, also varies by the game situation, the pitcher’s move to first, speed of delivery, and pitch velocity, the catcher’s release and arm strength, among other factors. Further, the sabermetric wisdom is itself limited by an inadequate consideration of how the threat of a steal affects the pitcher’s concentration, selection of pitches, and stamina. The notion that a saber-savvy GM would outlaw base stealing is, at best, misguided.

We discuss baserunning at greater length in subsequent chapters, but it is interesting here to note that the A’s seem to have shifted strategy on stolen bases in recent years. During the twenty-two-year period from 1990 through 2011, the average major league team attempted 147.5 stolen bases per season. During the early moneyball years, 2002–2007, the A’s attempted only 66 stolen bases on average; yet during the three-year period 2009–2011, the A’s were well above the long-term MLB average with 178 stolen base attempts per year. However, considering the five years since 2007, the A’s average attempts were just below the MLB average at 142 per year.
⁶
What these variations suggest is that stolen bases vary as much according to team personnel as to team strategy. One reason why the A’s have attempted more stolen bases since 2007 is that they have more effective base stealers on their roster: the team’s average stolen base success rate jumped from 70.3 percent during 2003–2007 to 76.9 percent during 2007–2011. (The major league average success rate since 1990 is 69.3 percent.)