This is a review of the player injury analytics research conducted by Hisham Talukder, Thomas Vincent, Geoff Foster, Camden Hu, Juan Huerta, Aparna Kumar, Mark Malazarte, Diego Saldana, and Shawn Simpson
Player injuries significantly affect the overall performance of a team and therefore are very concerning for team management and fans. A player injury model has been developed to help teams predict the likelihood that any given player will be injured during an upcoming game. This is done through a quantitative and systematic approach.
Data for the analysis was gathered using play-by-play game data, SportsVU data, player workload and measurements as well team schedules for two years. Testing the model using this data proved that it is able to predict the probability of a player being injured in the upcoming week.
The research demonstrated that the most relevant factors that increase the risk of injury were (in decreasing order): (1) the average speed at which a player ran during games; (2) the total number of games played; (3) the average distance covered by a player; (4) the average number of minutes played; and (5) the average number of field goals attempted. Total number of games played, average number of minutes played and average number of field goals attempted are all related to workload. Increased workload would naturally be associated with a greater risk of injury. The average speed a player ran and average distance covered deal with playing style which is also related to an increased risk of injury. However, the number of back-to-back games and the number of games played during a 14-day period did not significantly increase the risk of player injury. This contrary to the thought process of many, including NBA officials. Acting on this data could change the direction the NBA has been taking to schedule fewer back to back games in order to minimize injuries.
Combining these results with team schedules would allow team management to identify when would be the best time for a team to rest their star players and reduce the risk of long-term injuries. If the top 20% of high risk players were rested it could be possible to prevent 60% of all injuries.
Rather than resting players randomly during the year, it could be done strategically using this model. When the model indicates that a player has probability of 0.15 or higher of being injured that player should be rested for the next game. Resting for one game dramatically reduces the risk of injury. The team could take into account both the probability of injury and the importance of the upcoming game in making decisions regarding resting players. It is not always possible or feasible to rest a player for an entire game. An alternative approach would be to decrease the number of minutes a player plays over several games rather than resting them for an entire game.
In conclusion, reducing the risk of player injury would enhance team performance and fan enjoyment as well as minimize the financial cost associated with injured players missing games.
How mature is your team’s analytics program? Take the Sports Analytics Maturity Assessment.
Learn about the Groundbreaking Sports Analytics Model coaches and sports analysts are talking about!
Learn all about Sports Analytics here.