Risk: Medium

Results: 2005 ? 97th percentile in ESPN Tourney Challenge; 2006 ? 49th percentile; 2007 ? 90th percentile; 2008 ? 20th percentile; 2009 ? 38th percentile; 2010 ? 97th percentile; 2011 ? 43rd percentile. OVERALL AVERAGE ? 62.0 percentile.

Strategy: Advance top seeds three rounds, No. 2 seeds two rounds, and Nos. 3 and 4 seeds one round. For the remaining matchups, average out the coaching, team and conference PASE values for each squad. In toss-up games, always take the higher PASE. In upset games, only take the lower seed if their combined PASE is 0.500 better.

Outlook: This approach did very well in 2005, stumbled in 2006 as many of the traditional overperforming coaches stumbled, performed solidly in 2007, then cratered in 2008 and 2009 before rebounding in 2010. Last year yielded middling results, tied for fifth best among 12 models?shows you how weak the competition was.

This model assumes that past overachievers continue their overperforming ways. Call it the "Legends Bracket"...or "Your Father's Bracket." It might be a reasonable assumption to factor in historical coaching performance, but I?m starting to wonder whether a school?s or conference?s track record really comes into play when all the players keep changing. Of course, the Colonial League keeps spitting out Cinderellas like George Mason and VCU. So I?m going to keep spitting out this model.