Tuesday, April 07, 2009

How to pick a successful NCAA tournament bracket

Pictured below is my NCAA "sheet of integrity, " on the NY Times website. I was playing for a Kindle. No econometric model (which I usually do), no long hours of analysis, no money spent, and not many games watched this year. You'll see I finished #258 out of how many tens of thousands (not sure how many played, but it is the #1 newspaper in the world and the game was featured prominently on the front page). I missed 15 games. The winner missed 7. This is better than I've ever done with my model.


I also entered a bracket in a Yahoo family group, lost to my mother-in-law (congrats, Jan).

I feel like Taleb saying "All econometrics is useless." My train of thought:

1. Last year was an anomaly with four #1 seeds. We rarely see three #1 seeds, so I picked 2.

2. Last year was an anomaly that there were 4 NBA-ready dominant teams. Pomeroy's pythag was something like 99% accurate last year, so everyone jumped on that bandwagon. That's how financial markets create bubbles and busts. That strategy picked Memphis this year.

3. There are rarely upsets of 1-4 seeds in the first 2 rounds.

4. I didn't watch much basketball. From games I watched, North Carolina was ridiculously good and Louisville only beat Kentucky by 3.
I only picked Pitt in the Final 4 because Seth Davis and Clark Kellogg seemed to think it was a no-brainer on selection Sunday. The last UConn team this talented lost to George Mason, so I didn't think they'd proceed to the F4.

And that's about it. The last 2 years have seen relatively few upsets and some boring games. Will that trend continue, or are we due for a crazy bracket next year? I don't know.

No comments: