For five years now, college football has crowned its champion through its much-debated Bowl Championship Series (BCS) system. The BCS uses opinion polls, strength-of-schedule calculations, seven statistical ranking systems and its own mathematical formulas to determine the best team in the land. (See “Choosing Huskers over Ducks,” this page.)
It’s more Nasdaq than Knute Rockne, but don’t think this is an IT dream come true for a CIO with a football jones. The BCS, administered by the National Football Foundation and the College Football Hall of Fame, has only a $200,000 budget. That’s popcorn money relative to the tens of millions handed out to the six athletic conferences and the four bowl games that make up the BCS. Add to that the nearly $1 billion the BSC gets from its 5-year TV contract with ABC and you’d think the BCS folks would hire a big IT vendor to calculate those critical rankings.
“Mine’s like 50 lines of code,” says Richard Billingsley, describing his BCS ranking application that, in fact, he devised 35 years ago when he was 16. “I keep the ratings on my desktop and just e-mail them to the BCS.” For fun, Billingsley, a stress-management counselor in Hugo, Okla., created all-time rankings for all college football teams, applying his formula to the 90,000 or so games played since Rutgers lined up against Princeton in November 1869. (His all-time top team is Notre Dame. Shake down the thunder.)
CIO talked to five of the seven owners of the computer ratings that go into the rankings (Jeff Sagarin and The New York Times did not respond). None is paid. Most have advanced degrees. Their programs range from Billingsley’s 50 lines of code to Kenneth Massey’s nearly 20,000 lines.
“No supercomputers needed,” says Massey, a PhD candidate in numerical analysis at Virginia Tech (his Hokies are a bowl contender this year). Massey, like the others, was motivated by love of the game and disgust over the old poll system, which relied solely on opinion.
Wes Colley, an astrophysics professor at the University of Virginia, created his formula “for fun” while he was at Princeton. Peter Wolfe, MD, is “an infectious disease guy” at UCLA. He uses a spreadsheet and a generic probability algorithm for his rankings. “I love the game, and the math’s interesting too,” he says. Jeff Anderson, a political scientist at the U.S. Air Force Academy (with his partner Chris Hester) says, “The key to developing an accurate computer ranking is applying a sound philosophy to it.”