Moneyball, Machine Learning, and Why I Seek the Holy Grail of Sports-Tech
I claim that anyone with a math background and a love of baseball has at one time given serious thought whether they could be “the one” to unlock the equations that make Moneyball a reality. I am no different. This is the story behind the inspiration of my mid-career vision quest in search of the holy grail of sports-tech.
June 2017. Frankfurt, Germany. A hot, sweaty, un-airconditioned ballroom with about a 1000 supercomputer guys. I was one of a handful of women (and an engineer, to boot!) in a sea of male PhDs. I felt invisible, alone, and my monumental case of jet lag was not doing me any favors!
The 3rd cup of bitter, strangely thick coffee was no help at all. I soon found myself drifting off and daydreaming of baseball. 2017 was the perfect year to be into baseball back home in Houston. (I never heard the infamous trash-can banging, but our math did uncover anomalies.) My lack of attention went beyond boredom and jet lag, though. I was mentally taking a break from an unaccepting and exclusionary environment to one that I knew would always welcome me with open arms: the world of being a sports fan! It should be no mystery why I was mulling over Jose Altuve’s 350 batting average instead of the latest product on the Intel roadmap.
Then, the next speaker changed my whole world!
Enter Dr. Eng Lim Goh, the new CTO for the High-Performance Compute (HPC) group of my organization. A speaker that forces audiences to perk up by making a lot of eye contact and asking for audience agreement, nodding, and hand-raising; his secret power that day was an insistence that we all ‘get’ the philosophy of AI. Imagine the eye rolls amongst the crowd that was anxious for deep mathematical concepts. But the philosophical approach got my attention. “AI/Deep Learning/Machine Learning is all about letting the data tell us the story and generating the right equations to solve problems. Forget the equation first method. Work the data.”
These concepts were not new to me, but with baseball on the brain, I started to wonder. Can we apply deep learning to sports data and make better sense of stats and trends in baseball? To better understand the game? In every sport? Oh my! It was not lost on me that EVERY other person around me wanted to use these mathematical tools to solve complex weather problems, cure cancer, make jet engines, and manage nuclear stockpiles. Me? I just wanted to see whether deep learning could help me enjoy and better understand baseball.
It was that day that I began my lonely quest to use math to crack the world of sports. And Oh! The stories I can tell from then on!
Before I do, I must thank Dr. Goh.
PS — Like any ‘proud parent’ here is a snapshot of our live dashboard that tells the story of the game in real-time with each pitch and each down! Isn’t she pretty? I will have a lot more to say about this in the future & how it is different than anything else in the industry! The way we do the predictions is new to the industry and something I’d love to share!
