Big Data May Not be Enough to Anticipate Injuries In NFL
Sat, April 10, 2021

Big Data May Not be Enough to Anticipate Injuries In NFL

Injuries are hard to predict during football games / Photo Credit: Jamie Lamor Thompson (via Shutterstock)


The NFL hopes that big data tools can minimize the number of concussions, ligament tears, and other injuries sustained during football games, wrote Nicole Wetsman of The Verge, an American technology news portal. Presently, the injury count per match is at a steady average of six or seven. Hence, league engineers are collaborating with Amazon Web Services (AWS) to integrate machine learning and AI tools to player data to find “in-game situations that lead to injury,” as stated by Andre Beaton of The Wall Street Journal, an English-language international daily newspaper. 

That means the NFL can identify and predict injury risk scenarios and formulate innovations that will make football games safer for their athletes without comprising the quality of play, said Jeff Crandall, chair of the NFL’s engineering committee. Amazon and the NFL may have adequate resources at their disposal, but injuries can be hard to predict in chaotic sports like football, as noted by Zachary Binney, an epidemiologist and consultant who has worked with Major League Baseball and college sports teams on injury prevention. 

The AWS partnership will attempt to bridge the gap with league-level data from the NFL’s “Next Gen” stats, which collects location data, acceleration, and speed for every player on the field hundreds of times a minute via the microchips embedded in their pads. It also includes video footage of games, information on playing surface, and environmental factors. However, it doesn’t collect information on how hard body parts hit the ground or other players, Binney said. But it can see “how and what speed a player ran a play, hanged direction, or made a tackle” with granular detail. 

The league-level data includes some measures of player activity, while individual teams can track more granular-level data like heart rate, fatigue, etc., which can contribute to injury risk for a player. However, most of the player-specific data stays at the team level to avoid relaying potentially useful information to their opponents. Health data of players will not be included in the injury prediction program, an NFL spokesperson told The Verge via email. This might affect the system’s predictive power. 

Still, the NFL’s efforts might present a “road map to other sports” if their injury prevention and prediction system prove to be a success.