Thanks for asking. I decided not to predict the bracket this season because the model I was using (not my own) may have done a good job at explaining the way the committee USED TO choose teams, but it did very poorly the previous two seasons. Until I can come up with something that works better, I'll focus on other things.
It actually does better than most early on (a month or more before selection sunday, when the projections from my site give me an advantage), but by the day of the announcement, human judgement seems to do better. In particular, the average of the "Bracket Matrix" does surprisingly well. I don't have to time to go through all 68 teams every day, but at some point, I may put something up.
2 comments:
This isn't where I wanted to put this comment, but since there is no other post to put it in, when will you start doing your bracket predictions?
Howard,
Thanks for asking. I decided not to predict the bracket this season because the model I was using (not my own) may have done a good job at explaining the way the committee USED TO choose teams, but it did very poorly the previous two seasons. Until I can come up with something that works better, I'll focus on other things.
It actually does better than most early on (a month or more before selection sunday, when the projections from my site give me an advantage), but by the day of the announcement, human judgement seems to do better. In particular, the average of the "Bracket Matrix" does surprisingly well. I don't have to time to go through all 68 teams every day, but at some point, I may put something up.
Thanks for visiting!
Post a Comment