Jan. 12 was a good day for Philadelphia’s basketball fans. Their 76ers ended a five-game losing streak with a win over the Houston Rockets at the Wells Fargo Center.
The 76ers’ Jrue Holiday sank two free throws late in the game to hit the 30-point mark, leaving him just three points shy of matching his career high. Holiday also finished the game with nine assists, four steals and with a perfect percentage from the line (he also had six turnovers, but why needlessly tarnish the cup of victory?).
The sports media promptly trumpeted Holiday’s stats following the game, but there was a slew of other metrics generated by him and his teammates for the first time that night that were never reported.
A new use for tracking cameras and software
The week before, the team had installed a player-tracking system that was originally developed to track and analyze missile trajectories. This game marked their first use of the technology, and they became the 15th NBA team to deploy it.
Six small high-definition cameras, about half the size of a basketball, are now positioned in the rafters above the arena’s court. These cameras are the backbone of a system called SportVU, which tracks players’ every move using optical recognition and generates data for analysis in real time.
“We are able to capture data you couldn’t capture without this technology even if you wanted to,” SportVU’s Brian Kopp tells Txchnologist. “Twenty-five times every second our system is recording the time, the exact position of each player on an X/Y axis and the ball on an X/Y/Z axis. The software uses complex algorithms to turn that raw data into something that can be digested.”
Hardcore processing and analytics
The system uses color and lighting to identify objects and jersey numbers to recognize individual players — team lineups and player numbers are stored in a database that the system uses to distinguish athletes as they play.
Complex algorithms take the millions of raw data points collected to create player and team profiles. As the game progresses, the tracking software compiles and logs how far players run, their average speed and the areas of the court where they perform best in terms of rebounds, steels, shooting, assists and more. At the end of a game, a report then breaks down player and team performance.
Kopp says the system analyzes speed zones and factors in distances to every stat to reveal deep insights into performance.
Game reports generated by the technology don’t just look at the number of rebounds that a player had — the complex algorithms built into the system can assess rebound opportunities based on a player’s distance from the basket, how a player rebounds in traffic based on nearby teammates and opponents and even how quickly a player recovers and gets back after making a rebound.
Individual profiles can be analyzed to understand athletes’ conditioning and performance in a detailed way that was impossible five years ago.
Kopp says his company is looking to recreate the vocabulary that media analysts, coaches and fans use to understand basketball. He believes teams will soon generate more data, sports analysts will report in greater detail and coaches will field their teams differently. “This is Moneyball, only more so,” he says.
If using advanced sensors and statistical analysis to revamp sports sounds familiar, that’s because it is. SportVU is only one example of many products entering the player-tracking market across professional sports. Similar technology is being used to analyze athletes as they throw, kick, shoot and run in sports from rowing to football to running.
Now and in the future
The NBA’s Minnesota Timberwolves were early adopters of the SportVU system, deploying it at the start of the 2011-2012 season.
Matt Bollero works in operations for the Timberwolves and confirms they are using the system. “At this point, we use it primarily for our own players to analyze individual and team performance,” he says.
When probed for specifics about how it is used, however, he answered only vaguely. “Many of the areas the cameras can focus on do not show up in the box scores,” He says.
Bollero does say that the data generated by the system is proving to be a useful coaching tool, adding that he personally sees it doing much more in the future. “In five years, front offices and coaching staffs will be more aware of its usefulness and capabilities. It will likely be mainstream intel for teams,”
The way that this technology will influence players, teams, coaches, recruiters, broadcasters and consumers going forward is still not entirely known, but with half of all NBA teams now using the technology, it will undoubtedly shape the future of basketball analytics.
Top Image: A view from the SportVU tracking camera operating above a pro basketball game. Software has superimposed players’ numbers that it recognized from their jerseys and a red circle around the ball that it’s tracking. Courtesy SportVU.