ESPN didn't have a ton of success predicting the field this year. Here's a summary of the analysis:
1. Accurately predicted 65 of the 68 teams (95.6%). Or, more appropriately, accurately predicted 34 of the 37 at-large teams (91.9%).
2. Missed the selections of Georgia, VCU and UAB by predicting Colorado, Virginia Tech and St. Mary's.
3. Of the 65 teams that were accurately picked, ESPN (1) predicted the correct seed 30 times, (2) was 1 seed off 25 times, (3) was 2 seeds off 7 times, (4) was 3 seeds off 2 times and (5) was 4 seeds off once.
4. Biggest seeding error prediction was Utah St. as an 8-seed (instead of a 12-seed). Next biggest misses were Michigan (predicted to be an 11-seed, but came in as an 8-seed) and Missouri (predicted to be an 8-seed, but came in as an 11-seed).
5. Nailed two matchups exactly right (each from the MWC): SDSU/Northern Colorado and BYU/Wofford.
6. Amazingly, for the seven Big 10 teams in the tournament, ESPN accurately predicted the seeds of two teams (OSU and Wisconsin), but predicted worse seeds for the five other teams. As mentioned, Michigan was predicted to be 3 seed-lines worse than they did and also Illinois and Penn St. were predicted to be 2 seed-lines worse than they did. Purdue and Michigan St. were 1 seed-line off. What conference is the Chairman of the selection committee from?
7. Unusually, of the top 4 seed-lines in each region (i.e. the top 16 teams), ESPN missed HALF of the seeds of those teams. Missed one 1-seed, two 2-seeds, three(!) 3-seeds and two 4-seeds.