Lead by Tim Swartz, Professor, Department of Statistics and Actuarial Science, Simon Fraser University.
The Sports Analytics Group (SAG) at Simon Fraser University (SFU) is an informal unit of academics that carry out research in the nascent fled of sports analytics. This Collaborating Research Team in Sports Analytics will assist SAG to expand its research activities involving interuniversity collaborations and will provide students with experiences and opportunities to work in this growing space. The three general research themes are:
- Casual inference in sport
- Contextual analysis in sport
- Advanced physiological models in sport
Progress in these three themes will advance methodologies in statistics and will address important problems in sports analytics. The commonality between the three themes is that they all share computational and statistical modelling challenges resulting from large and complex datasets. Methodological contributions include the proposal and fitting of complex models, clustering spatial settings, the development of optimization algorithms and the development of prediction methods. On the sports side, all of the research themes address important problems in sports analytics which will benefit teams and athletes, affect the way that sport is played, and generate quantitative insights in sport.
Sports analytics has traditionally been a niche topic in the academic world. However, the sports analytics landscape has changed dramatically in the last decade. A reason for this is the growing attention provided to “big” sport where the financial aspects related to sport have reached unprecedented levels, and consequently, the need to win has never been greater. However, the major reason why sports analytics have taken off is the availability and complexity of sport related data.
In the past, sports analytics were at best constrained to box score data, and the scope of analyses that could be entertained was limited. Today, almost all of big sport have “event” data. Event data is detailed data that goes well beyond box score data and consist of a chronological record of well-defined events that occur during a match which are relevant to the match and are recorded with a time stamp. With event data, every time an event occurs during a match (e.g. goal, assist, tackle, etc.), characteristics of the event are recorded (e.g. location on the pitch, players involved, etc.). Even though event data are highly informative, the gold standard for sports data is “player tracking data”. Player tracking data consist of the (x, y) coordinates of the ball and all players recorded at regular and frequent time intervals (e.g. 10-25 Hertz). Player tracking data in sport are the catalysts for big data analyses where a single match in our soccer datasets comprise approximately one million rows. Tracking data is sometimes supplemented with physiological data which measures how an athlete’s body reacts to physical activity.
The contribution of sports analytics extends to various domains. On the sports science side, areas of interest include nutrition, sleep, injury prevention, training, physiology, exercise, biomechanics, health, medicine and biochemistry. On the economics/business side, researchers study topics such as variable price ticketing, fan retention, growth, efficiency of gambling markets, marketing and salary cap issues. On the competition side, problems include fairness of competition, scheduling, tactics, the identification of key performance indicators, drafting, rules, ranking, decision making, handicapping and player evaluation.
Tim Swartz, Professor, Department of Statistics and Actuarial Science, Simon Fraser University
Oliver Schulte, Professor, School of Computing Science, Simon Fraser University
Dave Clarke, Associate Professor, Department of Biomedical Physiology and Kinesiology, Simon Fraser University
Alex LeBlanc, Associate Professor, Department of Statistics, University of Manitoba
Tianyu Guan, Assistant Professor, Department of Mathematics and Statistics, Brock University