In the spirit of The Matchup Zone's theme of office supply related backronyms, I've decided to christen my tournament prediction model the Silly Tournament Algorithm for Predicting the Likelihood of Entrance.
This year, its first in existence, STAPLE nailed 34 of the 36 at-large selections to the Tournament field. It predicted that Southern Methodist (79% chance of selection) and Minnesota (68%) would be in the field, while Dayton (45%) and NC State (3%) would be in the NIT.
In terms of seed line predictions, the biggest miss was Kentucky, which STAPLE had pegged as a 4-seed, but the Committee dropped down to the 8 line. The average error for seed line predictions was 0.74, with 58 of the 66 seed line predictions within 1 seed of the actual field.
For next season I'll look at the variables to see if there are any tweaks that could be made to make the system a little bit better. I'll also add a report that'll be updated daily with all of the information all in one place.
No comments:
Post a Comment