Moneyball in the NBA (2008)

Undergraduate: Peter Wang

Faculty Advisor: Peter Mucha
Department: Economics

---

Moneyball is a book written by Michael Lewis detailing the Oakland A's statistical approach to running a Major League Baseball franchise. In using statistics such as stolen bases or walks, a $41 million dollar player salary organization (the Oakland A's) can compete with a $100 million dollar team (New York Yankees). APBRmetrics (Association of Professional Basketball Research) is Moneyball's professional basketball counterpart. Through the use of multiple regression techniques to calculate adjusted +/- values and observing underlying networks, this study aims to develop a player ranking system that can pinpoint undervalued players that are not traditionally (via subjective scouting techniques or box score statistics) considered productive.
Calculations were performed using MatLab on a series of play-by-play information provided by ESPN.com to mine data specifically showing changes in lineups and points scored in every game. A multiple regression analysis was performed on the entire collection of possessions grouped by 10-man matchups to show which players were the most productive on a per possession level with 0 being the "average" NBA player. This method accounted for the quality of the players played against as well as the quality of his teammates. The player ranking chart shows a list of the most productive players and the dendrogram shows the time each player spent playing on the same court as another player. Many players traditionally considered "good" topped the list.