The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) DeepQB: Deep Learning to Quantify Quarterback Decision-Making with Brian Burke - TWIML Talk #297
Sep 5, 2019
Brian Burke, an Analytics Specialist at ESPN and former Navy pilot, connects the worlds of aviation and football through an analysis of quarterback decision-making. He discusses his innovative model, DeepQB, which leverages player tracking data to evaluate performance. Burke shares insights on the evolution of football analytics, the intricate challenges of quantifying decisions under pressure, and how machine learning can transform coaching strategies. His unique journey from jet pilot to sports analyst showcases the power of data-driven decision-making in football.
AI Snips
Chapters
Books
Transcript
Episode notes
From Jets to Football Stats
- Brian Burke transitioned from a Navy pilot to sports analytics due to boredom and a passion for football.
- He noticed a lack of sophisticated statistics in football media and began applying his military statistical training to the sport.
Next Gen Stats Data
- ESPN gained access to NFL player tracking data, including position, velocity, acceleration, and orientation data at 10 Hertz.
- This data, collected via chips on players' shoulder pads, presents a significant "big data" challenge and opportunity for analysis.
Rams' Tactical Advantage
- The Los Angeles Rams used tight formations and quick snaps, giving defenses less time to react.
- This tactical advantage, measurable through simple geometry, is an example of insights derived from player tracking data.
