reviews extra Moneyball: Brad Pitt, the statistician and the movie

A major Hollywood release has an unlikely hero. In Moneyball, Brad Pitt, Hollywood heart-throb and twice voted sexiest man alive, plays not a glamorous spy, nor a hard-bitten detective but … a statistician. Ray Stefani and Jim Albert say the movie hits a home …

The release of the movie Moneyball has been an When history is rewritten for entertainment exciting event for statisticians. It is new for a there can be some casualties. Paul DePodesta, statistician to be the hero of a major Hollywood the person who actually convinced Beane that film, and even more encouraging that the stat- statistics could revolutionise the way istician is played by Brad Pitt. Perhaps a whole teams are constructed, chose not to be associ- new image of our profession is developing. What ated with the movie. Hence the movie created next? Shall we see Angelina Jolie and Anne the fictional statistical protagonist Peter Brand, Hathaway competing for roles as statistician played wonderfully by Jonah Hill. Statisticians super-heroines, saving the world through Bayes- are seekers after, and sticklers for, truth; but they ian analyses? We can but hope. must perhaps forgive such departures from it in Moneyball is based on the book of the same the interests of entertainment, or art – or of con- title by Michael Lewis and tells the true story of veying statistics through entertainment and art. Billy Beane, who as general manager of the fail- As Peter Brand explains in the movie, one ing Oakland Athletics baseball team transformed should value a player by how many wins he creates the team and its fortunes by selecting players for his team, and baseball games are won when using sports statistics (or “” as they teams score more runs than their opponents. The are known – SABR being the Society for American traditional average (hits per at bat) is at Baseball Research). best an imperfect measure of runs production. It is ironic that the basic storyline of the If one looks at team hitting data, say for the Moneyball book is about building a baseball 2008–2010 seasons, then the has team with impersonal statistics, whereas the a modest 0.74 correlation with runs scored. In the Moneyball movie is intensely personal from start movie, when Oakland loses the prestigious hitter to finish. Moneyball immerses the viewer in the Jeremy Giambi, Billy Beane says that Oakland can- world of baseball, and Brad Pitt brings the com- not replace Giambi but they can replace Giambi’s plex and compelling personality of Billy Beane on-base percentage (OBP). Indeed, the OBP (the correlation with runs scored. Bill James, pioneer to life. Indeed the movie is Brad Pitt. So do the number of times on base per plate appearance) of sabermetrics and the man who coined the word, statistics come through? is a superior offensive measure -- it has a 0.88 in one of his Baseball Abstracts explained that a

Baseball and cricket Baseball and cricket share much common his- on the fundamentals of for in cricket); being by a pitch (if the tory. Baseball is mentioned in Jane Austen’s those who know only cricket. ball impacts his body rather than the bat); Northanger Abbey, written in 1798. The first making a “sacrifice hit”, intended to help a international cricket match was played between Baseball for cricketers team-mate complete a run while sacrificing the USA and Canada in 1844 (Canada won by • A hit, also called a base hit, is credited to his own chance; or being awarded a first 23 runs). a batter when he safely reaches first base base due to or obstruction. Henry Chadwick (1824–1908), often called after hitting the ball. A hit involves reach- • A batter is on base when safely on any one the founder of baseball, was an English-born ing only first base; a run is a circuit of all of the four corners of the diamond-shaped American brought up on cricket. four bases on the diamond. infield. Not everyone in the world understands the • The plate is the base at which the batter • A batter’s on-base percentage is the number rules, or the statistics, of baseball. Cricket fans stands to receive the ball. of times he gets on base, expressed as a may be equally surprised that some of the arcana • A player’s plate appearance is the number percentage of his total number of plate of that game are not universally known. Fans of of times he comes to bat. appearances. each, however, share a common interest: the two • The number of times at bat is the same as • A player’s is calculated games are probably the most statistically analysed plate appearance – but with deductions as the total number of bases reached by the of all sports. In the interests of multiculturalism for: receiving four pitches that the umpire player divided by the number of times at we therefore provide the following information calls “balls” (the equivalent of “no balls” bat.

© 2011 The Royal Statistical Society december2011 185 batting measure should include the ability to get himself, which shaped his desire to work at a Whether the viewer likes sports or not, or sta- on base and the ability to advance runners to higher level in management. Yet Beane had such tistics or not, this movie will provide an intensely home. A simple good “sabermetric” measure of a intense dislike of failure that he couldn’t bring compelling story that hits a (or hits batting performance is the “on base percentage himself to attend a game; he spent many games for six, as would be said in cricket). For those plus slugging” or OPS; for team data, the OPS has in the exercise room of the team locker area. who want more, the Moneyball book provides a a 0.96 correlation with runs scored. In cricket, a Ironically, Oakland reached the 2002 playoffs fuller insight into the sabermetrics movement in similar debate rages as to batting average (runs with a major league record 20-game winning baseball. Chapter 6 describes in detail Beane’s per dismissal) being inferior to strike rate (runs streak – which was relying on the improbable dealings in what he calls the “science of winning per 100 balls faced). to prove the probable. The team did not win the an unfair game”. Beane, as general manager of the Oakland World Series, which discouraged Beane who felt Athletics, was faced with severe financial restric- that his quest had failed. The “failure” was in tions. He said the team must “adapt or die”, a re- his own mind only. The most telling lines in the Ray Stefani is Professor at California State University, markably liberal philosophy for a sport that does movie come from the owner of the Boston Red Long Beach, and a specialist in sports statistics. Jim Albert is Professor of Maths and Statistics at Bowling things “by the book”. He embraced Peter Brand’s Sox who tried unsuccessfully to hire Beane. He Green State University, Ohio (read Paul DePodesta’s) innovative approach and tells Beane that Oakland spent $250,000 per win Moneyball is on general release in the USA; it through the power of Beane’s personality insisted while the spent $1.4 million per opens in the UK on 25th November 2011. that the scouts and team on-field manager com- win: Beane’s method had worked. Boston (with Moneyball – The Art of Winning an Unfair Game, ply, with much resistance from the traditional enough money invested) went on to win a World by Michael Lewis is published by W.W. Norton (2003), baseball people. Beane had been a failed player Series – and they did it by playing moneyball. ISBN 0-393-05765-8.

Writing with Significance!

A joint writing competition hosted by the Young Statisticians Section of the Royal Statistical Society and Significance magazine Illustration: Tom Boulton

Calling all budding writers within the first 10 articles are not peer-reviewed (see www.sig- done, or it could explain the work of others. years of their statistical career! nificancemagazine.org/). Articles should Only submissions in English will be considered. be accessible to a wide and non-specialist audi- Manuscripts must be original and not under • Do you have an idea for an interesting ence and should be about an area or application consideration for publication elsewhere, though article which would be suitable for of statistics that is of broad relevance or has we welcome magazine articles based on work in Significance? an important and topical application, in a way theses or in papers that have been submitted • Have you always wanted to try your hand that lives up to both meanings of the tagline to or accepted by academic journals, provided at writing in a more journalistic style? “Statistics making sense”. Articles should be the two are sufficiently different. The closing Are you good at interpreting data? between 1800 and 3000 words long and can date for entry is 1st March 2012. All articles include tables, figures and photographs. Their will be assessed by a review committee made Then why not write an article and submit it style should be clear and easy to read – avoid up of representatives from both the Young to our competition? This is the first writing the formal layout of an academic report – and Statisticians Section and Significance. The win- competition hosted jointly between Sig- technical terms and mathematics should be ning article will be published in Significance. nificance and the Young Statisticians Section used sparingly if at all, and suitably explained. Runner-up articles will be published on the of the Royal Statistical Society. Significance is End references are optional, but should be Significance website, or in Significance at the published by the Royal Statistical Society and limited to three or four at most. editor’s discretion. the American Statistical Association, and is for Anyone is welcome to enter, regardless of Please email your submissions, in Word or as anyone interested in statistics and the analysis membership or affiliation. The only stipulation a PDF, to [email protected]. We and interpretation of data. It is primarily a is that you should be “young” (in career terms, hope you enjoy writing your piece. general interest magazine for statisticians, not necessarily in age) – that is, you must be a users of statistics and all those interested in student or within the first 10 years of your ca- Laura Gray, RSS Young Statisticians Section statistics. It is not a research journal, and reer. The article could be on work that you have Julian Champkin, Significance

186 december2011