Home Help
 Academic Programs
 Medical Center
 Events Calendar
 Prospective Students
 Faculty & Staff
 About Columbia
 A–Z Index
 E-mail & Computing

Columbia News
Search Columbia News
Advanced Search
News Home | New York Stories | The Record | Archives | Submit Story Ideas | About | RSS Feed

Jan Vecer Turns the Art of Predicting Sports Scores into a Science

By Colin Morris

Vecer's program depicts the 2003 U.S. Open Semi Final Match. Roddick and Nalbandian's signatures can be seen on the graph.

As every coach, player and fan knows, professional sports have become a battleground in which psychology, technology and science play an increasingly important role. The sports community has especially focused on athlete psychology, turning-point scenarios and adaptive strategy to gain an edge. Thanks to the ingenuity of Jan Vecer, assistant professor in the Department of Statistics, the sports world may soon have a tool that can measure all of this crucial data -- as well as statistically predict an outcome of a game with an accuracy never before achieved.

An avid tennis fan, Vecer developed the idea for a new program while watching the Wimbledon tournament. During the broadcast Vecer found himself at odds with the television commentators' analyses. The match points deemed pivotal by the sportscasters made little sense from Vecer's statistical perspective. To confirm his intuition, Vecer developed a statistical model capable of measuring the importance of individual points in predicting the outcome of a game, set and match. As the game progressed, the model could instantly recalculate the chances of an outcome with each rally.

Vecer took the idea to Tony Mauriello, associate director of Columbia Science and Technology Ventures (S&TV) -- the department responsible for identifying and patenting the many innovative creations of the Columbia community. After listening to Vecer's statistical description, Mauriello suggested that Vecer create a user-friendly version by which to convey this data to the public. After a few weeks spent on programming, Vecer returned with a functioning display of his invention.

After successfully testing the invention, the collaborators realized their program could be even more useful if it took into account the psychological dimensions of a match. Because the program measures the point statistics through the narrative progression of the game, a player's psychological approach could be quantified with a "signature." This signature, displayed in the form of a graph in Vecer's program, can reveal how a player planned and reacted to a particular situation -- information of intense interest to coaches, players and all sporting types.

Vecer used Andy Roddick's 2003 U.S. Open semifinal victory over David Nalbandian to illustrate how the program functions. He plugged the individual points into his program to generate a statistical portrait of the game. When Roddick faced a match point in the third set, Vecer's program gave Roddick only a 1:20 chance to win the whole match. Nalbandian, who had been dominating the first part of the contest, let up at moments that proved to be statistically significant. Roddick, meanwhile, converted the statistically important points in the remaining sets, which led to his victory in the match. All of this is depicted in the player's signature graph.

"The match itself was interesting in the way that Roddick had only 1:20 chance of winning the match -- quite a rare event, which makes this match so memorable," Vecer said.

Vecer and Mauriello are currently developing other ways to market the tennis match analysis tool as well as how to apply the technology to other sports.

The tennis program is one of many inventions Mauriello and his team has developed. S&TV works to identify and patent new inventions by the Columbia community. The group accepts proposals from students, faculty and staff. S&TV helps to steer clients through relations with industrial companies to license inventions and technology developed in Columbia laboratories. Through the Innovation Enhancement Fund, a pilot program, S&TV provides limited funding to faculty who have significant potential for advancing intellectual property with commercial value.

Published: Jan 30, 2004
Last modified: Apr 26, 2004

Tell your friend about this story