Computer science major Tad Berkery’s hockey analytics research was featured in The Washington Post.
Suspecting that coaches and general managers woefully undervalue the significance of faceoff win and the skills they require, computer science major Tad Berkery and his peers in Hopkins Engineering’s sports analytics research group built an AI-driven model to analyze millions of hockey plays. Their conclusion: Each faceoff win is worth about .015 goals on average, which is enough to cost teams goals, wins, and championships. The world of professional hockey is taking note, and Berkery will present his findings to at least one NHL team this summer.
Berkery collaborated with undergraduates Chase Seibold, Justin Nam and Max Stevens on this project, under the mentorship of faculty members Anton Dahbura and Donniell Fishkind.
Read “This Johns Hopkins student believes AI can bolster hockey analytics.”