Sign In

Communications of the ACM

ACM TechNews

New Algorithm Ranks Sports Teams Like Google's Pagerank

View as: Print Mobile App Share:
Football fan


Sandia National Laboratory researcher Ed Feng has developed PowerRank, an algorithm for ranking sports teams that is similar to how Google's PageRank ranks Web sites. PowerRank requires just two parameters--the game score and the home field advantage.

"There is no human bias, no memory of last season, and no style points," Feng says. The PowerRank algorithm is set up like a network, with the different teams acting as nodes and the games played acting as the links between nodes. When two teams play each other, a number is assigned to the link based on the score and location of the game. A larger margin of victory leads to a higher number. After all the games in the league have been played, each team receives a value. The teams are ranked by their values, much like how Google's PageRank algorithm ranks Web pages. Teams earn a higher rank by beating highly ranked teams.

Feng has ranked all 120 NCAA Bowl Championship Series football teams, all 32 NFL teams, and hopes to create a ranking list for international soccer before the World Cup in June 2010. PowerRank is not meant to be used to predict the outcome of a single game because there are so many arbitrary factors to take into account, Feng says. "The PowerRank is truly about ranking teams based on how they have performed in the past," he says.

View Full Article


 Abstracts Copyright © 2009 Information Inc., Bethesda, Maryland, USA


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account