1.3.3: Sports

From simply keeping score to gathering data to compare players and teams, the use of data and statistical analysis in sports has a long history. The professional sport often most associated with the use of statistics is baseball. Statistics like batting average, number of home runs, number of strike outs, and number of runs batted in have been on the minds of baseball fans for over a century. Ask any baseball fan who their favorite player is and then ask them what makes them a great player. Chances are they will respond with statistics.

In the modern era the use of statistics and data science has produced the field of sports analytics. Sports analytics uses historical statistics to predict player and team performance in a way that can provide a competitive advantage to a team or individual. The results can be used to help players, coaches and staff make decisions both during and prior to sporting events.

Sports analytics was popularized following the release of the 2011 film Moneyball, in which a baseball team relies on the use of analytics to build a competitive team on a minimal budget. The essential idea behind this use of sports analytics is based on solving an optimization problem based on past performance data. Given the resources, that is, the allowable budget given to a team to pay player contracts, how should one find players to compose the team such that the winning potential of the team is the largest? The answer is not always obvious. For example, it may not be worth paying a large amount of money for a single high-impact player when the same money could be spent of several medium-impact players.

The use of statistics and data science goes well beyond the application of building a team. For example, in baseball, many analyses have considered the number of runs that are expected for a team under many different types of circumstances. This work traces its beginning back to 1963, where play-by-play data was used to compute an estimate of the average number of runs scored that could be expected in the remainder of the inning for each of the 24 possible states of an inning (Lindsey 1963). Later analyses became more complex and now provide insights into the general strategy of baseball. For example, it has been shown that a sacrifice bunt is rarely advantageous as most teams would be expected to score more runs with a runner on first and no outs than they would with a runner on second and one out (Tango et al.2007).

In American football one of the most prominent recent developments is an estimate of the probability that each team has of winning. This probability is updated after each play for the duration of the game. Many of these estimates are available during online streams. An early approach to this type of analysis used data based on the current score, the point spread, the number of timeouts remaining for each team, and statistics on the status of play to estimate the win probability (Lock and Nettleton 2014). More complicated methods have been developed that use sophisticated statistical modeling (Yurko et al. 2019; Carl and Baldwin 2022).

Of course, statistical methods and mathematical modeling are applied to every conceivable sport including basketball, soccer, ice hockey, rugby and even e-sports (Baumer et al. 2023). With the new prevalence of online wagering in many states, these analytical methods have increased importance not only for those making wagers but also more importantly,for those who set the odds for the games. As more data is available about each game and player, one can expect that the world of sports statistics and analytics will continue to develop and gain importance.