Some interesting results from a player projection model I've been tinkering with. Here's an example position. Measurable speed does matter!
Each Combine/Draft season I try to improve on my player projection models. A fun exercise is to see how accurate a model can get using measurables alone (apart from scout grades). I think I've settled on a really good model type--Bayesian regression trees.
@bburkeESPN Interesting. I’d think weight would correlate with bench reps, but weight/AV and bench/AV track differently
@bburkeESPN What is the Y axis? This is an intriguing analysis and very fun. It would be cool to evaluate the movement in context of the inherit trait. 40 yard dash for RBs within a certain height weight threshold. It may b helpful when considering the system the player will play in
@bburkeESPN Love this. Is there an explanation or a theory why there is that bulge/wave in the shuttle data?
@bburkeESPN any reason you did 3-year AV here instead of 4-year to represent the entire rookie contract?