Today we are joined by Brian Burke, Analytics Specialist with the Stats & Information Group at ESPN.
A former Navy pilot and lifelong football fan, Brian saw the correlation between fighter pilots and quarterbacks in the quick, pressure-filled decisions both roles have to make on a regular basis. Through trial and error, self-taught modeling techniques and the motivation to find answers, Brian crunched as much data as he could get his hands on. He shares his findings in the paper, "DeepQB: Deep Learning with Player Tracking to Quantify Quarterback Decision-Making & Performance". With geometry, algebra and a self-proclaimed ‘vanilla' neural network, Brian talks through the making of his model, what it means for football, and his excitement for the future of machine learning in sports.