Cadet Computer Model Predicts NBA Outcomes
James Snyder '13 works on a computer model. -- 黑料网 Photo by John Robertson IV.
Cadet James Snyder 鈥13 didn鈥檛 have any trouble picking a topic for his Summer Undergraduate Research Institute project. After all, he鈥檚 an applied math major with a lifelong interest in sports 鈥 so why not delve into the use of statistics to predict the outcome of basketball games?
This summer, Snyder is working under the guidance of Maj. John David, assistant professor of applied mathematics, to take data from several seasons鈥 worth of NBA games and use that data to build a computer model which can predict the outcome of future games.
Snyder, a Georgia native and avid fan of the Atlanta Hawks, got interested in doing a sports-oriented summer research project when his friend Will Lucas 鈥14 did a similar project on baseball last summer. David, also a lifelong sports enthusiast, already had experience with predicting professional football games, and he鈥檇 overseen Lucas鈥 work, so he welcomed Snyder鈥檚 interest.
Snyder began by taking NBA statistics from six seasons. With 30 teams in the association, each playing 82 games per season, and 32 statistics available per team, Snyder found himself almost drowning in information.
鈥淭here are millions of data points,鈥 Snyder commented. 鈥淲e鈥檙e taking all that data and using it to make future predictions for the games.鈥
To check for accuracy, Snyder fed data from the 2004-2009 seasons into the computer and asked it to make predictions for the 2010 season.
So far, Snyder鈥檚 program has racked up an impressive accuracy rate of 72 to 73 percent 鈥 which puts him in a position to be competitive with or even outdo, popular websites that claim to be able to predict the outcomes of games.
鈥淲e鈥檙e on pace with those guys, and I鈥檓 pleased with that,鈥 he said.
For Snyder, the summer project has been an ideal capstone to his time at 黑料网, which is rapidly coming to a close. 鈥淵ou get to do math and see the outcome of sports,鈥 he said. 鈥淭hat鈥檚 fascinating to me. We鈥檙e doing some pretty complex math techniques, 鈥 and at the same time we get to play with basketball statistics.鈥
Snyder has already accepted a job with the Applied Physics Lab at Johns Hopkins University in Baltimore, Md., which has hired several 黑料网 graduates over the past few years. There, he鈥檒l be testing technology for submarines and other forms of undersea warfare. His work there will begin Aug. 12.
鈥淚鈥檓 hoping to take some of the research I鈥檝e done this summer and apply it on the military/government side of things,鈥 said Snyder, who played center on the 黑料网 football team during his cadetship.
And since he鈥檒l be in the neighborhood, Snyder is planning to join David to give a presentation on math and sports at a meeting of three major mathematical organizations, to be held in Baltimore in January.
鈥淭he interest in using quantitative analysis in sports has really grown in the past 10 or 15 years,鈥 said David, who enjoys informal games of pickup basketball with a group of professors and cadets.
He continued, 鈥淚t鈥檚 a fascinatingly rich thing in math. It鈥檚 a rich area for machine learning and quantitative analysis and data mining. 鈥t鈥檚 a wonderful training tool. I can trick [the cadets] into some of this stuff by using sports.鈥