As compelling as possession-based efficiency numbers are, I don't think they tell the whole story of tourney performance. My research has shown that a number of factors beyond pure game play have influence on whether a team overachieves or fails to live up to seed expectations in the tourney.
|Do Mike Krzyzewski and Nolan Smith have a Duke repeat in them? (Getty Images)|
How do you take into consideration the value of coaching, while respecting the importance of possession-based data? That's what this bracket model is all about.
Bracket strategy: This methodology combines Pythag values and coaching PASE to take into account what happens both on and off the court. We eliminate the 14 through 16 seeds, then take each of the top 13 seeds' Pythag values and add their coaches' PASE performance, adjusted to normalize the small differences in Pythag and wide variances in PASE. Then, we rank the teams and advance the squads with the highest Pythag/PASE values in every matchup.
Bracket results: How do you feel about a Duke repeat? Or Kansas and Bill Self's penchant for underachievement going down to Louisville and Rick Pitino's overachieving ways? Do you think Wisconsin can handle Pitt -- and that Billy Donovan can take his Gators to the Final Four? Those are the kind of results you get when you factor in the historical performance of coaches.
You can't argue with Duke as a potential champion, but things could've been much, much crazier. The second-ranked team in this scenario was San Diego State. That's because Steve Fisher has such a strong record of overachievement in the Dance (and, no, I didn't discount the Fab Five years). This could've been a Blue Devils-Aztecs final ... except for the fact that the teams were in the same region.
There are certainly things about this bracket that I like. It forces me away from the chalkiness of some of the other models ... and this tourney is not going to go by the book. On my own, I don't think I would've elevated Louisville over Kansas ... or Wisconsin over Pitt ... or Florida into the Final Four. But none of those possibilities is out of the question.
Some random thoughts and wild predictions about the Field of 68. Bracket overview >>
|More Bracket Science|
If you believe that coaching makes a difference -- and that we're in for a crazy tournament -- this could be the model for you. I believe both of these assertions ... but that doesn't mean I'll be penciling these picks into my personal bracket.
Here's the round-by-round breakdown:
• All of the one through five seeds advance. (I'm thinking not.)
• Missouri is the only 11 seed to beat a six seed.
• In the 7 vs. 10 matchups, the model likes Washington, Texas A&M, Penn State and Michigan State (call it the Izzo factor).
• In the 8 vs. 9 matchups, advance George Mason, Illinois, Michigan and Butler.
• This goes completely by the seed ... with all of the top four seeds advancing.
• Ohio State and Syracuse, Duke and San Diego State, Louisville (yes, Louisville) and Purdue and Wisconsin and Florida advance to the Elite Eight.
• Ohio State and Duke uphold the honor of top seeds, while Louisville bests Purdue and Florida gets by Wisconsin ... no doubt in a mudfight.
• Duke and Coach K prevail over Ohio State and Thad Matta.
• Pitino's Cardinals end the run of Donovan's Gators.
• Duke wins the championship ... as another last-second heave from midcourt, this one by Preston Knowles, rims out. No ... the model didn't predict that.