For the 2016 UNC SPEIA Basketball Analytics Summit case competition sponsored by the Charlotte Hornets, Eli Shayer, Travis Chen, Nick Canova, and I created a model to determine how knowledge about offensive systems can provide insight into how NBA draft prospects might perform in the NBA.

Some key findings:

• The similarity between collegiate and NBA offensive systems has no effect on a player's performance.  That is, a statement like "<Player X> is a good fit for <Team Y> because the offensive system he played in college is similar to the one he would join if drafted by <Team Y>" is not quantitatively justified.
• The proportion of certain play types (e.g. isolation, transition, post up) that a collegiate offense runs can be good predictors of whether players that come from that offense will be stars or busts in the NBA.

Nick Canova wrote a series of blog posts about our work (Part 1, Part 2, & Part 3). I also went on The Sports Shop Radio on WRAL (Raleigh) to talk about our work.

Our submission won the case competition and has been presented at a meeting of the Stanford Sports Analytics Club, the 2016 INFORMS Annual Meeting in Nashville, TN, and in a session of GSBGEN 360 Sports Business Management at Stanford University.