http://repository.stcloudstate.edu/cgi/viewcontent.cgi?article=1002&context=stat_etds
Abstract
This is a study of predicting success in the NBA based on college experience, and repredicting
after the rookie year of professional basketball. All first and second round picks
from the 1985 draft through the 2005 draft are included in overall analysis, with 841 players
having had at least one year of NBA experience and have played on a Division I NCAA team.
The primary endpoints used in analyses are Player Efficiency Rating (PER), win shares, and
win shares per 48 minutes. This paper will predict the success of picks using their draft pick,
college statistics (both qualitative and quantitative), and physical qualities (height and
weight). Also, rookie year statistics are used to update the analysis to determine if any
additional information is gained after one year of professional basketball. This study
concludes that a statistical analysis of college statistics predicted performance well using win
shares per 48 minutes, but did not improve predicted performance with PER and win shares
for first round draft picks. In addition, one of the predictive formulas was able to predict
performance of the top 100 NBA prospects with higher accuracy than the actual draft.