Back in Fall 2006, I began my Ph.D. journey by looking at how complex network analysis could be applied to basketball. Drawing data from two women’s basketball games, I laboriously mapped every ball movement (passes and advancing by dribble) and related them to their outcome at the end of the possession. My hope was to find natural patterns that could eventually evolve into game planning and real-time halfcourt adjustments.
The work—which I iterated over the course of the semester, but then dropped to pursue other things—identified some issues at the time with data collection and analysis. Working from video tapes of each game was labor intensive, taking several times longer than the game clock to process the data: which player in which position on the court directed the ball in which way. If I had continued with this line of research, one of the major initiatives for a dissertation would have been finding other ways to collect my data automatically.
The options I explored at the time included video game data (i.e., getting EA to generate a ball-movement network from NFL Live game play), creating my own simulator in NetLogo (made some progress on this front), computer-aided video analysis (i.e., detecting player and ball position from the regular TV coverage), or turning the basketball into a sensor. As it turns out, that last idea has become a reality in the interim.
InfoMotion Sports Technologies has teamed with University of Michigan researchers to create a smart ball that assesses shooting skills and ball-handling mechanics. Their five-gram circuit board affixes to the inside of a basketball without altering its flight or bounce. Players run through dribbling or shooting drills while the board gathers data from its accelerometers and gyroscopes. The device wirelessly transmits the data to a computer, which produces a breakdown of each shot and analyzes your performance.
Source: “The Embeddable Coach” (PopSci, Jan. 21, 2011)
The focus for the tech is on improving mechanics. However, add similar sensors to a player’s uniform, and this system could conceivably get me the proximity detection I would need to know who is doing what with the ball at which area of the court.
The PopSci article claims the technology is being used in youth leagues and five top-20 basketball programs (presumably the University of Michigan basketball team is a guinea pig, too). Plans to spread into other sports—namely soccer, golf and baseball—make this an intriguing industry to watch.
One option I hadn’t considered in 2006 was the gesture-based technology, like the kind behind the Kinect game system. This season, a few NBA and NHL teams started using CrowdWave, a large-scale motion-tracking system that enables massive group games. Eight high-def cameras in the rafters feed computer software that translates waving into actions. It might be possible to train these HD cameras on the court and differentiate between dribbling, passing, shooting, and rebounding. This technology was the kind of thing envisioned by the CHI Student Design Competition in 2004, by the way, proving designers sometimes need to think about five years ahead of business before new innovations become attractive to monetize.
