Regularized Adjusted Plus-Minus (RAPM)
With RAPM becoming more common across the NBA, I wanted to bring this powerful tool to the college game.
RAPM calculations rely on knowing how performance changes when players are on and off the court. Unfortunately, anyone who has played around with NCAA data knows that this is not an easy process (large number of teams, inconsistent player names, missing substitutions, etc.). Nevertheless, I have done my best to convert raw NCAA play-by-play data into possession "stint" data (similar to this).
While RAPM is an objective calculation, there are endless enhancements that you can perform to improve performance or learn unique insight. I have tried to keep my RAPM calculations as vanilla as possible, however it is worth noting that I am (1) controlling for home court advantage and (2) filtering "garbage time" possessions. Over time I may explore other extensions (e.g. coach impact, team specific HCA, Bayesian prior adjustment, further fine-tuning regularization).
NCAA Basketball RAPM Reliability
Given the smaller samples sizes for college basketball (relative to the NBA), I was unsure how noisy the RAPM results would be. One quick sanity check is to look at year-to-year consistency of RAPM across seasons, which I have been pleasantly surprised with. For example, across each season of his college career, RAPM consistently had Draymond Green as a highly-rated player.
For anyone who is still skeptical, I have calculated "multi-year" RAPM results which utilize all available years of data (currently 9) to improve stabilization.
Download complete RAPM results (currently updated for 2010-2018 seasons):
NBA Draft Picks
Select your favorite draft class to see how RAPM viewed future NBA players (in their final college season):
Check out the top-rated players in O-RAPM, D-RAPM, and overall RAPM:
Select your favorite team1 to see if RAPM results align with your expectations:
The team views are currently limited to the top 50 teams (based on highest cumulative RAPM). Feel free to download the raw data to look closer at a team which wasn't included or reach out on Twitter if you think we should build an interactive view which allows you to pick your own custom filters. ↩