Recruiting by computer

Here's an interesting read from Andy Staples at SI.com. Seems that three economists with nothing better to do designed a computer to predict where high school football stars would commit to playing college ball. Staples goes up against the computer for the last 17 uncommitted recruits for 2009.

The economists think a 58-68% average is good. But is it really that difficult at this point in the process to figure out where these guys are going to end up considering the overwhelming amount of information that Staples has access to, including one on one discussions with the recruits?