Vol. 42, No. 2, March–April 2012, pp. 105–108 ISSN 0092-2102 (print) ISSN 1526-551X (online) http://dx.doi.org/10.1287/inte.1120.0633 © 2012 INFORMS Introduction to the Special Issue on Analytics in Sports, Part I: General Sports Applications Michael J. Fry Department of Operations, Business Analytics and Information Systems, University of Cincinnati, Cincinnati, Ohio 45221, [email protected] Jeffrey W. Ohlmann Department of Management Sciences, University of Iowa, Iowa City, Iowa 52242, [email protected] Sports is a rapidly growing application area for analytics. The use of analytics is pervasive in the professional sports community as evidenced by the increased role for those practicing sports analytics in front-office manage- ment and coaching. Part I of this Special Issue on Analytics in Sports is devoted to the application of a variety of methodologies to a broad range of sports. The papers analyze golf, hockey, baseball, motorcycle racing, and college football. The methodologies employed draw from many areas of analytics including optimization, probabilistic modeling, and choice models. Key words: sports; analytics. ports are big business. Most estimates place the example of applying analytics to baseball—Bill James Stotal market value of spectator sports in the United (James 2001), George Lindsey (Lindsey 1959, Lindsey States in the range of hundreds of billions of dollars. 1961, Lindsey 1963), and many others preceded the As one paper in this special issue notes, the mar- work described in Moneyball—this book was the cata- ket size of spectator sports in the United States has lyst for introducing the broader sports community to been estimated to be double that of the automotive the potential benefits of quantitative analysis. Money- industry and is easily one of the top-10 business mar- ball became a best seller; in 2011, a movie of the same kets globally. Given the size and impact of the sports name (starring Brad Pitt) achieved considerable box industry, it should not be surprising that this is a office success. fertile application area for operations research (OR) Recently, a new breed of management has become models. more prominent in sports organizations. There has Sports analytics has also been gaining widespread been an increase in the number of front office person- notoriety in the general populace. Michael Lewis’ nel with quantitative training and (or) the apprecia- entertaining story about the use of data analysis in tion for the power of analytics to help improve the baseball in Moneyball: The Art of Winning an Unfair performance of their teams both on the field of play Game (Lewis 2004) is arguably the most visible and from a business perspective. Paul DePodesta, one account of sports analytics. Many of the strategies of the central figures in Moneyball, earned an eco- documented in Moneyball, which were then employed nomics degree from Harvard; he has worked in the by the small-market Oakland Athletics team to help it front offices of several MLB teams and currently is the compete with teams with much larger payrolls, have vice president of player development and scouting for been adopted in some form by many other Major the New York Mets. Daryl Morey, general manager of League Baseball (MLB) teams and teams in other the Houston Rockets in the National Basketball Asso- sports. Although Moneyball is clearly not the earliest ciation (NBA), has degrees in computer science and 105 Fry and Ohlmann: Introduction to Special Issue on Analytics in Sports, Part I 106 Interfaces 42(2), pp. 105–108, © 2012 INFORMS statistics from Northwestern University and an MBA Although several dedicated sports analytics jour- from MIT; yet, he has little actual basketball play- nals have recently been established, Interfaces has tra- ing experience. Morey has been an active crusader ditionally been a leader in publishing papers on for sports analytics through his involvement with the this topic. This, and other findings, are highlighted annual MIT-Sloan Sports Analytics Conference, which in “Identifying the ‘Players’ in Sports Analytics brings academics and practitioners together to discuss Research” by B. Jay Coleman. This paper provides an the latest advances in sports analytics. Another pro- overview of the sports analytics research landscape: ponent of quantitative methods in the NBA is Mark which universities are contributing research on this Cuban, owner of the 2011 NBA Champion Dallas topic, where the papers are appearing, and which Mavericks. Cuban has worked directly with Wayne papers are most commonly cited. This paper is similar Winston, John and Esther Reese Professor of Decision in spirit to the Rothkopf rankings (Fricker 2011) pub- Sciences at Indiana University’s Kelley School of Busi- lished by Interfaces, which identifies the universities ness, to apply sports analytics concepts to on-court and organizations that are the leading contributors to and off-court decisions made by the Dallas Mavericks. OR practice. Coleman shows that sports analytics is As these and other examples show, the sports world is a fast-growing research area, outpacing other appli- teeming with opportunities for analytics to contribute cation areas in OR. The author finds that a relatively to sports management. small group of researchers works on sport analytics, This special issue of Interfaces seeks to add to and that many of these researchers are based outside the academic literature examining sports applications. the United States. Historically, Interfaces has been a leading outlet for Most team-based professional sports leagues assign sports analytics research, but this is the first spe- newly eligible players to teams through some form cial issue of Interfaces dedicated to this topic. Stick- of sports draft where each team selects players for ing with the overall spirit of the Interfaces journal, its team from the pool of all draft-eligible players. we present practice-oriented sports analytics papers. In MLB, this is a particularly difficult task because As evidenced by the overwhelming response to the 1,500 players are selected each year over 50 draft call for papers for this special issue (we received rounds. All these players must be evaluated by the over 40 submissions), many researchers in OR-related baseball teams, but player evaluations are inherently fields are working on sports-related applications. imprecise. Multiple scouts evaluating the same player often have varying opinions on the player’s pro- jected performance in the major leagues. In “A Major Papers in Part I of This Special Issue League Baseball Team Uses Operations Research to Because of the large number of high-quality submis- Improve Draft Preparation,” Streib et al. present an sions to this special issue, we publish it as a double application of OR to improve an MLB team’s process issue. The first issue focuses on general (nonschedul- of assessing baseball players. The authors develop an ing) sports applications; the second issue focuses optimization tool that helps a team come to a consen- solely on sports scheduling, which has emerged as sus on rankings for draft-eligible players. The algo- a highly successful application area for OR. As seen rithm is based on voter-preference research, including by the variety of sports examined in the first issue, the classic Kemeny-Young method; however, the size the application of analytics is not limited to any of the problem faced by MLB teams requires new particular sport. Three of the papers examine golf methods and solution approaches. The authors’ tool applications. Sports analytics applied to golf has expe- has been used by one MLB team for several years, and rienced increased interest recently because of the other MLB teams have expressed interest in develop- wealth of data made available through the PGA ing a similar tool. This paper presents an excellent TOUR’s ShotLink database (Deason 2006). The other example of using analytics to better utilize the invalu- papers in this special issue discuss hockey, baseball, able subject matter expertise of personnel (scouts in motorcycle racing, and college football. this case). Fry and Ohlmann: Introduction to Special Issue on Analytics in Sports, Part I Interfaces 42(2), pp. 105–108, © 2012 INFORMS 107 The goal in team sports is to create an optimal academic papers and contradicts the commonly held lineup of players at different positions to defeat the belief characterized by the oft-repeated quote from opposing team. However, this problem is difficult golfer Bobby Locke that golfers “drive for show, but because different player positions require different putt for dough” (Riach 2007). skills and a collection of the best individual players In “A Proposal for Redesign of the FedEx Cup does not necessarily produce the best lineup because Playoff Series on the PGA TOUR,” Hall and Potts of interaction effects. In “Quantifying the Contribu- propose a new design for this season-ending event. tion of NHL Player Types to Team Performance,” The FedEx Cup playoff series was established in 2007 Chan et al. examine the question of how to determine to bring the professional golf season to an exciting the optimal player lineups in ice hockey by using a end, similar to the Super Bowl in the National Foot- clustering technique to define distinct player types for ball League or the World Series in MLB. The FedEx offense, defense, and goalies. These clusters are used Cup system is a season-long points competition that in regression models to determine the relationship ends in a playoff series of four events where the top between team performance and the clustered player 125 point-scoring players are reduced each week until types. The authors show how the results from their models can be used to analyze trades in the National 30 players compete in a final 72-hole stroke-play tour- Hockey League, and they provide an Excel-based tool nament.