NBA ANALYTICS HELP EXPLAIN RACEHORSE INJURIES

A news release from the American Jockey Club (March 22, 2016) reported a significant drop in fatal racehorse injuries–defined as those occurring within a race or 72 hours afterwards–between 2014 and 2015.

Statistics from The Equine Injury Database revealed a 14% decrease in fatal injuries. Dr. Tim Parkin, the veterinarian and researcher who performed the data analysis, attributed the improvement to “wide-ranging safety initiatives embraced by tracks, horsemen, and regulators…”

The Jockey Club related the specific statistics, as follows:

“Across all surfaces, ages, and distances, the fatality rate dropped from 1.89 per 1,000 starts in 2014 to 1.62 per 1,000 starts in 2015. The overall fatality rate of 1.62 per 1,000 starts is the lowest since the Equine Injury Database started publishing annual statistics in 2009.

The fatality rates associated with each racing surface were as follows:

  • On turf surfaces, there were 1.22 fatalities per 1,000 starts in 2015, compared to 1.75 in 2014.
  • On dirt surfaces, there were 1.78 fatalities per 1,000 starts in 2015, compared to 2.02 in 2014.
  • On synthetic surfaces, there were 1.18 fatalities per 1,000 starts in 2015, compared to 1.20 in 2014.”

In early March 2016, researchers Hisham Talukder and Thomas Vincent presented a paper titled “Preventing In-Game Injuries for NBA Players” at the MIT Sloan Sports Analytics Conference that focused on injury prevention in the National Basketball Association. Some of its findings are intriguing in terms of reducing racehorse injuries. The co-authors found that resting the 20% of NBA players who are at most risk for injuries could prevent 60% of injuries. Who are the 20%?

A Wall Street Journal article summarized key findings:

“What surprised the authors most were the factors that were least important: the number of games in the last 14 days and the number of games on consecutive nights in that time….Better indicators were a player’s speed, usage, and overall season workload. That helps explain why the injury risk right now is someone who plays like a complete maniac: Oklahoma City guard Russell Westbrook.”

In 2015, Talukder and Vincent were able to compute the day-to-day probability that Chicago Bulls’ guard Derrick Rose would be injured.  With a predictive model that incorporated variables like distance run, player speed, points scored, and rebounds, the probability that Rose would be injured gradually escalated from low risk on February 7 to medium risk on February 20. Then, on February 21, 22, and 23, the risk profile elevated like a rocket…and indeed Rose was injured on February 23.

Such findings about human athletes should have relevance to training racehorses. For instance, the Jockey Club news release stated: “An analysis of 2015 race distance statistics shows that shorter races (less than 6 furlongs) were again associated with higher injury rates versus middle distance races (6 to 8 furlongs) and long races (more than 8 furlongs). This has been consistent each year over the seven-year span.”

A plausible explanation from the study of NBA players is that horses in shorter races most fit the description of running “like a complete maniac.” Perhaps injuries could be curbed if trainers gave sprinters more time between races and lighter workouts than horses that race at longer distances. At first glance, this seems counter-intuitive but makes sense in terms of “speed, usage, and overall workload.”

Empirical findings from the Jockey Club’s Equine Injury Database are invaluable in saving racehorses from fatalities. Analytic research from other sports suggest ideas for further experimentation. Trainers and veterinarians who understand the purpose and application of analytics can, working together, further reduce horse-racing fatalities.

Copyright © 2016 Horse Racing Business