This is a review of the openWAR baseball analytics research conducted by Benjamin S. Baumer, Shane T. Jensen, and Gregory J. Matthews.
The purpose of sports is to win. Analysts use statistical methods to study key areas of interest. One challenge is quantifying the contributions of individual players to their team and the number of games the team wins. Major League Baseball uses wins above replacement, or WAR, in this determination.
Wins above replacement is comprehensive and easy to understand but has two problems, namely a lack of uncertainty estimation and a lack of reproducibility. In order to combat these issues openWAR is developed, which is based on a conservation of runs framework with uncertainty values estimated by resampling methods. Central to the model is the idea that the positive and negative consequences of all runs scored must be allocated across four aspects of baseball performance: batting, base running, fielding, and pitching.
In openWAR, the win above replacement is defined as the sum of all contributions from each of the four aspects, compared to a hypothetical replacement level player after controlling for factors like ballpark, handedness, and position. OpenWAR is not a statistic to be used for forecasting but instead is a retrospective measure of player performance. The credit or blame for hits on balls in play is shared between the pitcher and fielders. The extent to which each is responsible is determined by the location of the batted ball.
A conservation of runs framework is based on the idea that for every run value gained by the offense a corresponding run value is lost by the defense. Offensive run values are adjusted for several factors beyond the control of the players including the ballpark and handedness of the pitcher and batter. Base runners are only given credit for advancing beyond what would be expected given their starting position, number of outs, and the hitting event. Hitting performance is evaluated relative to the expected hitting performance based on all players at the same fielding position. The defensive run value must be divided between the pitcher and other fielders involved. The division is determined by how difficult the batted ball was to field. The entire value is assigned to the pitcher if the plate appearance does not result in a play such a strikeout or home run. Any value not assigned to the pitcher is divided among the fielders who were potentially responsible for the ball.
For each plate appearance, the run values for each base runner, fielder, hitter, and pitcher are calculated. The overall run value for a player is determined by adding those run values across all plate appearances involving that player as a hitter, pitcher, base runner, or fielder.
Team management is able to accurately evaluate players in regards to their contribution to the team when looking at offering contracts or making trades. Players in different positions can be ranked according to their openWAR score. This would facilitate comparison between high and low ranking players, allowing coaches to determine weaknesses within their players that can be improved upon in order to increase team effectiveness.
Instead of putting all the blame for a loss on the pitcher or all the credit for the win on the batter this statistic allows all players to take their part of the blame or the credit, holding them more accountable for their actions.
Analytics Used: WAR, openWAR, Conservation of Runs Framework, Resampling
Find out how Sports Analytics Expert Victor Holman can give your team the competitive advantage.
How mature is your team’s analytics program? Take the Sports Analytics Maturity Assessment.
Discover the Groundbreaking Sports Analytics Application and Framework coaches and sports analysts are talking about!
Learn all about sports analytics in Victor Holman’s Sports Analytics Blog.