Evan Green


Next Gen Stats xComp Thoughts: Evan Green ‘17 provides some thoughts on the newest NFL Next Gen Stats metrics, breaking down some key flaws and proposing future changes. 


NFL Timeout Tendencies: Evan Green ‘17 and Daniel Tokarz ’20 examine timeout strategy in the NFL.

NFL Timeout Analysis: Harrison Tracy ‘18 and Evan Green ‘17 examine timeout usage as a way to measure home field advantage in the NFL.

MIT Sloan Hackathon: Defining Player Vision: Evan Green ‘17 creates two new metrics, entropy and clustering, to measure NBA players passing ability using SportVU data. This is a summary of the presentation Evan presented during the conference as a finalist for the Hackathon. 

BY THE NUMBERS: Counting Yale’s Losses: Evan Green ‘17 breaks down shot charts from the Yale Men’s Basketball team’s graduating seniors and explores how team tendencies have changed this season.

BY THE NUMBERS: Home-field advantage in the Ivy League: In another collaboration with the Yale Daily News, Evan Green ‘17 investigated the size of and factors behind home field advantage in different Ivy League sports. Despite the differences in fan attendance, travel accommodations, and skill level of athletes, home field advantage was not significantly different in the Ivy League than that seen in professional sports leagues. 

Total Defensive Pressure (TDP): Analysis of Off Ball Defense in the NBA: In the inaugural NBA Hackathon, we took the opportunity to analyze SportsVU tracking data and create a new metric for off ball defense which had mostly been neglected by the current  research. The metric was based on the sum of squared distances from each offensive player to the defender closer to him. We found that closer off ball defense was related to more missed shots and fewer three points attempts. We also noted individual differences in effort based on differences in on ball and off ball defense.


The Success of Short Yardage Play Types on Fourth Down: The quarterback sneak is a play in football in which the quarterback pushes forward behind the offensive line immediately after the center hikes the ball. It is often employed in short yardage situations, especially when the defense has not had time to set their position. We sought to determine whether this play in under or overused by NFL coaches in such situations.

Player Clustering in the NBA: As a result of position, physical attributes, or basketball ability players in the NBA naturally gravitate toward certain roles within a team. We attempt to formalize this notion by clustering NBA players based on a combination of traditional and advanced statistics. These clusters can then be applied to examine how teams construct lineups.

NBA Daily Fantasy Lineup Generator: We sought to program a system to automatically created daily fantasy basketball lineups. We used penalized regression to predict players’ expected points and variances for a game based on their stats and opponents. We then used a probabilistic solution to the weighted knapsack problem to generate lineups.  Since we targeted “double up” games we created high expected points lineups with low variance.