Prediction of the NBA and the NFL Draft Picks for 2001 and 2002

• Activity Number: 168
• Type: Topic Contributed
• Date/Time: Tuesday, August 13, 2002 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Sports*
• Author(s): Todd Bishop* and Byron Gajewski+
• Affiliation(s): St. Cloud State University and St. Cloud State University
• Keywords: Applied Statistics ; Classification

Abstract

Whether to apply for the NBA draft or continue one more year of college basketball eligibility is an important question. A move to stay in college is becoming less and less popular. I wish to further explore the decision of a player to continue their collegiate eligibility. I have collected multivariate data from all division-one players eligible for the 2001 NBA draft. This database includes potential draft predictors for all college seniors and any early entry candidates. Using logistic regression and principal components, I came up with ten different models, choosing three in the end. With these three models I performed cross-validation to calculate the probability a player is to be drafted (their "bling" number). These "bling" numbers will allow athletes a statistical evaluation of their draft likelihood.

I will show the results of this work on the 2001 NBA draft, comparing the accuracy of my models against the actual outcome. I case-study individual players, who might have improved their draft position by staying in school. Finally, I will look ahead to ranking current players based on my prediction of the 2002 draft using the 2001 model. We also predict the NFL.