The September 2016 issue (volume 12, issue 3) of the Journal of Quantitative Analysis in Sports (JQAS) consists of three articles with applications to soccer, tennis, and tournament scheduling. The first two articles described here represent work presented at the 2015 MathSport International conference in Loughborough, UK.
“Searching for the GOAT of Tennis Win Prediction” by Stephanie Kovalchik is the Editor’s Choice article for this issue and available for free download for a year. The article describes 11 forecasting models used for predicting professional tennis matches and compares their predictive performance. The approaches considered—which were categorized as regression-based, point-based, and paired comparison models—developed models based on rolling 52 weeks of data and applied to forecast games immediately after the 52-week window. The comparison of these methods was applied to professional tennis matches played in 2014, with prediction accuracy achieving 70% by the best of the approaches examined.
“A Combined Approximation for the Traveling Tournament Problem and the Traveling Umpire Problem” by Marco Bender and Stephan Westphal considers the task of designing a double round-robin tournament schedule in which no two teams compete against each other in consecutive rounds, minimizing the total travel distance by the umpires who must referee at least one game at every team’s home venue. The problem addressed in this article essentially combines the traveling tournament problem and the traveling umpire problem into one overall optimization problem. The authors develop an approximation algorithm to address the problem and provide assurances of its near-optimality.
Finally, “Analysis of Substitution Times in Soccer” by Rajitha Silva and Tim Swartz study the problem of determining optimal substitution times in soccer games. The paper was inspired by a 2012 JQAS paper by Bret Myers that proposed a particular substitution time algorithm. The authors review the paper by Myers, provide a critique of the results, and offer a different perspective of substitution time guidelines based on an analysis through Bayesian logistic regression modeling. The manuscript is accompanied by a comment provided by Myers and a rejoinder by Silva and Swartz.
These articles are available to all members of the Section on Statistics in Sports and to everyone else on a subscription basis from the JQAS website. Prospective authors also can find the journal’s aims and scope, as well as manuscript submission instructions, on the website.
Leave a Reply