Science of Moving Dots: Pattern Recognition and Machine Learning in Golf

The science of spatiotemporal pattern recognition aims to find patterns and to uncover meaning and insight from analyzing “moving dots.” An important tool in developing this science is the use of machine learning. By giving the machine specific examples of movements, we can teach the computer to find features that enable it to distinguish particular movements and to discover the relationships between these movements.

So what does this have to do with learning golf?

Most golf swing coaches will agree that logistics of swing analytics follow 3 basic flow of questions: First, what is the ball doing? Second, what did the club do to influence the ball? And third, what did golfer do to influence the club?

Launch Monitors such as Foresight and Trackman allowed us to better understand the first 2 questions in the recent years. It’s safe to assert that all instructions prior to launch monitors were hearsays, based on what coaches thought were correct and based on what good players thought they did.

With the advent of “big data” and the machine learning, we can now connect the dots and start to understand the question of what did golfer do to influence the club.

“Dots” represents our understanding of golf swing from the setup to the finish – the movement of the club head, the shaft, hands, arms, shoulders, hips, etc. We can now capture and track these dots every fraction of a second with incredibly fine details, and turn it into “moving dots” that machine can understand.

The impact is almost revolutionary in the way that people coach, play and watch golf. The new science allows coaches to provide fact-based diagnosis, prediction based on motion measurements and video analysis, and data-driven solutions to pin point relevant issues. From the player perspective, it can provide them with immediate feedback, better understanding of the shot outcome and the shot probability, thus leading to better shot selection. Spectators will also benefit by added enjoyment of more accurate game predictions.

Contact @leejhl for more information on how the science can help your game

- Lee J.H. Lee