How Olympic athletes use machine learning and data analysis to reach peak performance levels

• Read original article here

For the first time ever, Ireland will have a field hockey team participating in the Summer Olympics. To make sure its athletes are performing at the highest level, the national team is getting some help from an Ireland-based startup that develops biometric measurement technology to identify players at risk for injury.

Kitman Labs, based in Dublin with offices in Silicon Valley, is working with both the Irish national field hockey team and the South African rugby team as they compete in this month’s Summer Olympics.

The company’s “Athlete Optimization System” analyzes athlete data collected from multiple systems including wearable trackers that show workload information and other data related to sleep, hydration, diet, mood, stress, and perceived muscle soreness. Coaches can look at the analytics to help drive decision-making related to the amount of training an athlete should be doing.

Here’s an example of the analysis Kitman Labs produces:

Kitman Labs uses machine learning that helps coaches see individualized data for each athlete, based on variables like fitness levels and overall capabilities specific to that player. This takes it to another level beyond just tracking data, explained Stephen Smith, a former Irish rugby trainer who founded Kitman Labs in 2012.

“The key to truly actionable data is the individual nature of the analysis,” he told GeekWire. “Not all athletes are the same, so machine learning is a critical component of the system. The technology learns about each athlete and flags any meaningful, high-risk changes in the athlete.”

Kitman Labs works with teams ranging from the Seattle Sounders to the Miami Dolphins. But Smith noted that the unique nature of the Olympics — high intensity and high stress for individuals and a team over a short period of time — highlights the need for regular monitoring and analysis.

Furthermore, Smith cited data showing that 25 percent of injuries from the 2012 Summer Olympics were related to overuse.

“These are some of the most avoidable injuries,” he said. “Stress, fatigue, dehydration are all huge factors in overuse injuries and any changes in these outside of normal fluctuations for athletes — which is where Kitman can help — are critical to identify early.”

Smith added that the Olympics are also unique because of high stress created from long travel, lots of walking, late nights, an unfamiliar environment, and poor conditions at venues.

“This stress can hurt player performance at the Games,” he said. “Continuing to monitor player well-being and status even once the teams have arrived in Rio can help coaches keep their athletes on track and ready to go for their big moment.”

While it may seem like common sense for athletes and teams to take advantage of this type of technology and analytics, not all players and coaches agree that this innovation helps improve performance.

“I don’t think anybody has come up with technology that athletes particularly think will help,” Los Angeles Clippers owner and former Microsoft CEO Steve Ballmer said at our recent GeekWire Sports Tech Summit. “It’s not like there aren’t some uses of technology — don’t get me wrong. On the other hand, if you said, ‘what is it that the athletes believe in that they must do?’ You’re not going to find much.”

Smith said the “real purpose of the technology can feel mysterious” to teams and athletes who are just learning about data, technology, and sport science. But he expects that disconnect to disappear once players and coaches see the benefits.

“Our goal is to lengthen player careers and maintain player value and public interest,” he said. “The athletes on the teams we partner with see the benefits, and as more and more teams experience these benefits, the mystery of the technology and the use of the science will become norms. And they’ll likely at that point want more data about themselves, and more frequently.”