Date

Spring 5-2-2016

Document Type

Honors Project

First Advisor

Davies, Stephen

Degree Name

Bachelor of Science

Major or Concentration

Computer Science

Department or Program

Computer Science

Abstract

College basketball is a highly popular sport and, in the wake of March Madness, one wonders what will happen to key players next year. The National Basketball Association (NBA) currently restricts players from entering the draft until they are 19. This leads some players to enter college simply as a practice and "waiting area" for the NBA. These players - often termed "one-and-dones" - stay for a year and then at their earliest chance, enter the draft. I wondered if staying in college longer allows NBA-bound players more practice and experience playing under pressure, or if players were better off by leaving college early to play while they were still young. Using R, I gathered individual player data from both college and the NBA. This required significant work in gathering, fusing, and cleansing electronic data from multiple sources into a usable form. I then investigated various accepted performance aggregation metrics, and settled on efficiency (EFF) which is a relatively simple measure that consolidates a player's yearly performance (including points, rebounds, assists, etc.) into a single number. Using machine learning techniques, I divided the players into "clusters" (small groups of statistically "similar" players) based on their freshman-year data and then examined each cluster individually. For each cluster, I analyzed whether there was a significant difference between the one-and-dones and the others. In this way I could examine the likely effect that additional college experience would have had on a player's NBA career. This analysis found little significance between the "one-and-dones" and the "more-and-dones", meaning perhaps a player's NBA performance is not hurt by coming out early.

Language

English

Share

COinS