Advanced Football Analytics

Trey Causey (@TreyCausey) is a data scientist, NFL statistical consultant and writer. He joins Dave on this episode to discuss what he's learned during his time working for an "unnamed" NFL team. He explains how the structure of a team's analytics department is crucial to their ability to promote change in an organization, and how the unique time pressures of professional sports make creating a successful long-standing analytics operation a tough challenge. Trey goes on to explain some of his research on k-means clustering and how it's a useful way to organize players into groups. Dave and Trey close out the episode with a discussion of the 4th down bot and the challenges Trey has faced in updating its win probability algorithms for end of game situations. Links from this episode www.treycausey.com www.thspread.us Trey's article on K-means clustering Trey's article on putting together an NFL analytics department Trey's post on fine-tuning win probability Trey's twitter Boris Chen's fantasy football k-me