The Sabermetric Revolution: Assessing the Growth of Analytics in Baseball (26 page)

6
. This average applies to the years 1985–2011.

7
. Even if one took out the contributions by Johnny Damon and Jason Giambi, the A’s OBP in 2001 and 2000 was .331 and .346, respectively. That is, without their contributions, the 2002 A’s still only exceeded the OBP in one of the two previous years. Nonetheless, the point is that Beane set out to at least replace the lost OBP of these two players and he failed to do so.

8
. Lewis,
Moneyball
, p. 248.

9
. For now, suffice it to note that WAR suffers from a variety of deficiencies. Among other issues, different authors and websites have different definitions of WAR. The one we employ here is from
Baseball-Reference.com
, and, hence, more precisely should be designated as either bWAR or rWAR. We use WAR on occasion because it is standard in the sabermetric vernacular and there are no comprehensive measures of player performance that are clearly superior.

10
. See
http://www.baseball-reference.com/teams/OAK/2002.shtml
.

11
. In Chapter 5, we discuss the barren state of error bounds for WAR. In the case of the 2002 A’s, every pitcher other than the five mentioned above had a WAR between –0.5 and 0.5, which even WAR proponents acknowledge is within the margin of error.

12
. Of these pitchers, Zito is perhaps the only question mark because he was not a particularly hard thrower and Lewis suggests that only a team eschewing the traditional measures of pitching talent would have drafted him. However, if Zito were a diamond in the rough that only the sabermetric eyes of Billy Beane could spot, then there would have been little reason to waste a first-round (ninth overall) pick on him in 1999. The A’s could have waited until the second, third, or a later round to draft Zito. The fact of the matter is that Zito was well-scouted and highly touted; see, for instance, Sheldon Hirsch and Alan Hirsch,
The Beauty of Short Hops
(Jefferson, N.C.: McFarland, 2011), p. 18. Indeed, in 1998 the Texas Rangers drafted Zito in the third round, though Zito didn’t sign and returned to finish his college career at USC, where he was a first-team All-America selection by
USA Today, Baseball Weekly, Collegiate Baseball
, and
Baseball America
. With a 12–3 record, a 3.28 ERA, and 154 strikeouts in 113⅔ innings, Zito was named Pac-10 Pitcher of the Year in 1999.

13
. Lewis,
Moneyball
, p. 231.

14
. Lewis appears to misrepresent the speed of another pitcher’s velocity. In discussing Jason Grimsley’s pitching prowess in 2002 Lewis writes (p. 260): “With a pitcher like Grimsley you always know what you’ll be getting: 96-mph heat.” Grimsley may have reached that velocity occasionally at the beginning of his career, but arm injuries and time conspired to make him a slower sinkerball pitcher by 2002. According to data on
Fangraphs.com
, for instance, his fastball in 2002 came in at 93 miles per hour.

15
. ERA+ is a pitcher’s earned run average, adjusted for ballpark factors, relative to the league average. WHIP stands for walks and hits per inning pitched. WHIP, while increasingly used, is deficient in that it does not adjust for fielding and factors of chance.

16
. In 2001, the A’s drafted Bobby Crosby, Jeremy Bonderman, Neal Cotts, Dan
Johnson, and Mike Wood, all of whom played at least three years in the big leagues. It is hard to think of this output from one draft as a disaster. According to our model detailed below, the A’s earned 3.2 WAR beyond what would have been expected in this draft. Unfortunately for the A’s, about two-thirds of that was accumulated by Bonderman while wearing a Tigers uniform.

17
. All of these ten at bats were with the A’s, and during September call-ups at that. It is not likely that any other team would have called him up even for such a short cup of coffee. Some in the baseball industry even suggested that Beane called him up to increase the number of major leaguers he drafted in 2002.

18
. Blanton has had a solid major league career. He was the 24th pick in the 2002 draft. The 25th pick that year was the Giants’ all-star pitcher Matt Cain. Teahen was not a bad 39th pick. He was used to facilitate a useful trade for the team and he played seven years in the majors. Five picks after him, however, at 44th, Cincinnati selected first baseman Joey Votto, who became the 2010 National League MVP.

19
. Lewis,
Moneyball
, p. 116.

20
. We chose 220 players so as to include Colamarino, the 218th pick in the 2002 draft. Both models were fit to a negative exponential distribution, where the response variable was WAR, and the explanatory variable was the overall draft pick slot. The only difference between the two models is whether minor league players were assigned a WAR of 0, or excluded from the data set. Our model is largely consistent with the findings reported in Rany Jazayerli, “Doctoring the Numbers: The Draft,”
Baseball Prospectus
, May 13, 2005.
http://www.baseballprospectus.com/article.php?articleid=4026
.

21
. To be more precise, the net WAR of 6.4 earned by the A’s in the 2002 draft ranks in the seventy-first percentile among the 464 team drafts between 1990 and 2005.

22
. In this case the A’s net WAR of 2.0 ranks in the 65th percentile.

23
. Lewis,
Moneyball
, pp. 103–106.

24
. Lewis,
Moneyball
, p. 20.

25
. Lewis,
Moneyball
, p. 109.

26
. Lewis,
Moneyball
, p. 109.

27
. Lewis,
Moneyball
, p. 17.

28
. Bonderman also started Game 4 of the 2006 ALCS against the A’s, leaving a tie game in the seventh inning. The Tigers scored three runs in the ninth to sweep the A’s out of the playoffs.

29
. Lewis,
Moneyball
, p. 112.

30
. Bryan Bullington (1) and Jeff Francis (9) were college players. Adam Loewen (4) was drafted out of junior college.

31
. Lewis,
Moneyball
, p. 112.

32
. Of the thirty players selected in the first round of the 2002 draft (one of whom did not sign that year), the fifteen high school players have accumulated nearly 63 percent of the cumulative major league service time and 76 percent of the WAR. Only one
of the five players who never played in the major leagues was a high schooler (Chris Gruler at 3), while two others were selected by the A’s (McCurdy and Ben Fritz) and a third was on their list (Bobby Brownlie). To be fair, Swisher and Blanton are probably the two most accomplished college players selected (Jeff Francis, Jeremy Guthrie, and Joe Saunders have also had decent careers), but the list of high school players is loaded with superstars (B. J. Upton, Fielder, Greinke, Matt Cain, Cole Hamels, Kazmir) and players who have had careers approaching those of Swisher and Blanton (Denard Span, Jeff Francoeur, Khalil Greene, James Loney).

33
. Schwarz,
The Numbers Game
, pp. 124–126. James believed, as did Chadwick and many others, that the commonly used “fielding percentage” failed to measure a fielder’s ability to get to a batted ball. The range factor metric was an (ultimately flawed) early effort to rectify this deficiency.

34
. Quoted in Schwarz,
The Numbers Game
, p. 34.

35
. Arguably, Rickey acted even before this. One source has Rickey hiring Travis Hoke to do a statistical analysis of the St. Louis Cardinals in the 1930s. See Dayn Perry, “Extra Innings: Can Stats and Scouts Get Along?” in Jonah Keri (ed.),
Baseball Between the Numbers
(New York: Basic Books, 2006).

36
. SLG is slugging percentage, where the numerator is the number of total bases on base hits divided by the number of at bats. That is, a single counts as one base, a double as two bases, a triple as three, and a home run as four.

37
. Lindsey’s baseball work was published in two articles in the journal
Operations Research
in 1959 and 1963.

38
. Frank Deford, “Baseball Is Played All Wrong,”
Sports Illustrated
, March 23, 1964, pp. 14–17.

39
. Schwarz,
The Numbers Game
, and interview, June 22, 2012. Schwarz speculated that Earl Weaver was also aware of Cook’s writings, both because of Deford’s piece and because there were several articles in the Baltimore newspapers on Cook’s work. Weaver, however, says that his approach was based entirely on his experience as a minor leaguer and learning from experimentation as a manager (interview, June 27, 2012). Lou Gorman also served as the assistant GM of the Orioles and the Mets, and was the farm director of the Kansas City Royals.

40
. Also part of the Orioles (and Reds) system in the 1960s and 1970s was head scout Jim McLaughlin. McLaughlin had a systematic and statistical method for scouting players that included computerizing player data, using professional psychological testing methods to assess the makeup of each prospect, systematizing the measurement of scout performance, and employing physical testing to quantify each player’s skills. See Kevin Kerrane,
Dollar Sign on the Muscle: The World of Baseball Scouting
(Lincoln, Neb.: Bison Books, 1999).

41
. According to an article in SABR’s
Baseball Research Journal
, “1977: When Earl Weaver Became Earl Weaver,” Fall 2011, by Bryan Soderholm-Difatte, Weaver did not put his new philosophy fully into play until 1977.

42
. Although Eric Walker has been out of baseball since the 1990s, he has continued to write about the game. Most notably, he maintains a website,
steroids-and-baseball.com
, in which he rants against the Mitchell Report on steroids prepared by the committee led by George Mitchell for the commissioner’s office and avers that there is no statistical evidence linking steroid use to increased power in the game. For a critical discussion of Walker’s statistical analysis, see A. Zimbalist, “Performance-Enhancing Drugs and Antidoping Policy in Major League Baseball,” in Zimbalist (ed.),
Circling the Bases: Essays on the Challenges and Prospects of the Sports Industry
(Philadelphia: Temple University Press, 2011).

43
. Lewis incorrectly states: “Palmer had written a book back in the 1960s revealing all this. The manuscript was still gathering dust on his desk when Bill James came along and created a market for it” (p. 80). Yet according to both Pete Palmer and John Thorn, there was no such book in the 1960s or the 1970s. There was just statistical work that Palmer had done. The book was written largely by John Thorn in the early 1980s.

44
. Schwarz,
The Numbers Game
, pp. 220–225.

45
. Interviews with Dan Duquette on June 11, 2012 and with Larry Lucchino on June 12, 2012.

46
. Schwarz,
The Numbers Game
, p. 227.

47
. Lewis also reminds the reader on p. 69: “The cause was the systematic search for new baseball knowledge.”

48
. Lewis,
Moneyball
, p. 14.

49
. Lewis,
Moneyball
, p. 193.

50
. The A’s front office had conducted “a systematic scientific investigation of their sport.” Lewis,
Moneyball
, p. xiv.

51
. Lewis,
Moneyball
, p. 201.

52
. Lewis’s confusions do not end here. For instance, he both extols Beane’s instilling a culture of patience at the plate throughout the A’s minor league system and yet writes that Beane believed that plate “discipline can’t be taught” (
Moneyball
, p. 148).

53
. Lewis,
Moneyball
, p. xiv.

54
. Lewis,
Moneyball
, p. 277.

55
. Lewis,
Moneyball
, p. 277. Also see,
Baseball-Reference.com
on J. P. Ricciardi and Keith Law’s commentary on Baseball Think Factory, September 15, 2011.

56
. Of course, team payrolls rose between these periods, from an average median payroll of $60.5 million during 2000–2001 to $78.6 million during 2002–2009. So the Blue Jays’ payroll rose slightly less than that of the average team, but Ricciardi’s reported pledge was to decrease payroll in absolute terms.

57
. Lewis,
Moneyball
, p. 122.

58
. As perceptions catch up to realities and some market inefficiencies disappear, it is always possible that new market inefficiencies will emerge and the perspicacious GM will be able to stay one step ahead of the field and exploit these. Indeed, as we shall see, Beane’s strategy does seem to have shifted over the years.

59
. On p. 205, Lewis, in reference to the negotiations for the 2002 collective bargaining agreement (CBA), cites Beane as believing that the only way the draft pick compensation for free agents would be changed is if the players were to allow either a salary cap or team revenue sharing. In fact, baseball’s modern revenue sharing system was introduced in the 1996 CBA, and by 2002, the amount of sharing from the rich to poor teams had grown to over $165 million. Beane must have known this, but Lewis should have known, too, given that he was writing a book about how to resolve MLB’s competitive balance problem. Lewis commits a similar gaffe on p. 3 when he states that the players “had been granted free agency by a court of law.” Free agency was granted by an internal baseball arbitrator, Peter Seitz, on December 21, 1975, in a sixty-one-page decision. The Seitz ruling was later upheld in a court of law.

60
. The idealized standard deviation is a hypothetical concept that would apply under the assumption that all teams had equal talent and, therefore, a 0.5 chance of winning each game. The ratio shows how far the actual deviates from the idealized standard deviation. It is a useful metric for cross league and longitudinal comparisons.