The ESPN/TopCoder Winning Formula Challenge asks for algorithms to predict NCAA football scores from historical data. I missed the pre-season myself: I tried to sign up with a very simple algorithm yesterday evening, intending to tune it up to something better in a couple of weeks, but I’d forgotten my TopCoder log-in credentials, and by the time I got my password reset the deadline had passed. (I’d describe my algorithm for you, but that’s against the rules. Hard-coding picks is also against the rules.)As you can see if you click on the image at left to blow it up, the challenge is to write an algorithm to predict the winners and scores of this year’s NCAA football games, based on historical data assembled by ESPN. First prize is $50,000. Think of it as “programming derby meets office football pool,” with a much bigger pot than any office I’ve ever worked in.Since many of TopCoder’s, umm, top coders hail from Eastern Europe, education has played a big role in the preparation for the challenge. Some examples: No, American Football isn’t the same as soccer. A fumble is important because the other team gets an unexpected chance to score and goes on offense on the next play. pass_att != pass_cmd+pass_int because a pass attempt can be completed, incomplete, or intercepted; further, a completed pass can sometimes be a touchdown pass. According to the contest rules, solutions can be in one of five programming languages: C++, C#.NET, VB.NET, Java, and Python. There are both CPU time and memory limits. Apparently the best algorithms submitted so far get about 75% of their win/loss predictions correct, which would be more than enough to make real money in a state such as Nevada that allows gambling on sports. Software Development