This Week in Sports Analytics: February 8, 2019

1 minute read


This Week in Sports Analytics:’s weekly roundup is here.


Alex Caravan and Dan Aucoin walk through results of combining HitTrax (swing outcome) and BlastMotion (swing characteristics). They were able to find some very strong relationships between data gathered from both systems. Looking forward to part 2.

I enjoyed Austin Rochford’s code for generating park factors.


Justin Jacobs walks through using Kullback-Leibler Divergence to compare how shooting frequency distributions differ between players. As always he’s thorough on not just the underlying theory and math, but also the applications where KL-Divergence could be useful in the realm of player decision making.


SeanfromSeaback reported on one of the most exciting plays from Super Bowl LIII. The longest punt in Super Bowl history from former Oregon State Beaver (Go Beavs!) Johnny Hekker. His analysis from the NFL Data Bowl demonstrates how punts that are able to make a returner move more are more likely to end in a downed punt.


Jacob Beckett @jacobbeckett22 took a crack at modeling the odds for the CONCACAF Champions League at American Soccer Analysis. As a Sounders fan, I’m bummed we don’t get to participate but I will be rooting for the MLS clubs to be the first to win the league.


David Miller published a breakdown of GAMs called “Bayesian views of generalized additive modeling”. The paper is focused on the Bayesian interpretation of the mgcv R package implementation. Other implementations include BAMLSS, an example with brms, and a python implementation called pyGAM. GAMs can be a useful tool in baseball analysis as Jim Albert showed back in January 2018.