Are we up to revision three already?
A week ago I gave you a simple simulation and tool for the NCAA tournament. Then I made it better and it did well out there.
Now, I want to make some corrections before we get to the Sweet Sixteen.
The first is that we are eliminating ESPN BPI from the model. It was the worst of the initial five models and did not add anything to justify its inclusion. This version is built using Sagarin, Ken Pomeroy, Sonny Moore, and my own DYI Ratings. The injury model has been updated with the latest news, but it still uses a Wins Produced model for each player and uses it for working out the impact of injuries and rotations (thanks Dave!).
The result of all that work is right here.
The sheet has six tabs:
- All the ratings for every Division 1 school
- The ratings for teams in the tournament
- The adjustable bracket
- The lookup table
To build your bracket you need to:
- Go to the bracket tab
- Select your model (cell E3)
- For every matchup select your winner (the cell that contains either Tm1 or Tm2- pick one)
- Once you're done, print your bracket.
- Profit!
Because I'm nice, the Tm fields will change colors when you pick team2, and the odds will change to red when you pick an upset.
As a bonus, here are the odds for every possible matchup:
| Updated Odds Table | |||||||||||||||||||
| Opponent | UVA | M St | I St | Uconn | Lville | Mich | UK | UT | UF | UCLA | Stan | UD | Az | Wi | SDSU | Ba | |||
| Team | Region | Seed | PM | 19.2 | 18.1 | 18.8 | 15.7 | 23.3 | 16.5 | 15.5 | 17.3 | 21.1 | 16.7 | 12.1 | 10.8 | 21.3 | 18.5 | 14.6 | 15.3 |
| Virginia | East | 1 | 19.2 | 53.9% | 51.4% | 62.3% | 35.6% | 59.6% | 63.0% | 56.8% | 43.2% | 58.9% | 73.8% | 77.5% | 42.5% | 52.5% | 66.0% | 63.7% | |
| Michigan State | East | 4 | 18.1 | 46.1% | 47.5% | 58.5% | 32.0% | 55.7% | 59.2% | 52.9% | 39.4% | 55.0% | 70.5% | 74.4% | 38.7% | 48.6% | 62.3% | 59.9% | |
| Iowa State | East | 3 | 18.8 | 48.6% | 52.5% | 61.0% | 34.3% | 58.2% | 61.6% | 55.4% | 41.8% | 57.5% | 72.6% | 76.4% | 41.1% | 51.1% | 64.7% | 62.3% | |
| Connecticut | East | 7 | 15.7 | 37.7% | 41.5% | 39.0% | 24.8% | 47.1% | 50.7% | 44.3% | 31.4% | 46.4% | 62.7% | 67.0% | 30.8% | 40.1% | 53.9% | 51.4% | |
| Louisville | Midwest | 4 | 23.3 | 64.4% | 68.0% | 65.7% | 75.2% | 72.9% | 75.8% | 70.5% | 57.8% | 72.3% | 84.3% | 86.9% | 57.1% | 66.7% | 78.3% | 76.4% | |
| Michigan | Midwest | 2 | 16.5 | 40.4% | 44.3% | 41.8% | 52.9% | 27.1% | 53.6% | 47.1% | 34.0% | 49.3% | 65.4% | 69.6% | 33.3% | 42.9% | 56.8% | 54.3% | |
| Kentucky | Midwest | 8 | 15.5 | 37.0% | 40.8% | 38.4% | 49.3% | 24.2% | 46.4% | 43.6% | 30.8% | 45.7% | 62.0% | 66.3% | 30.1% | 39.4% | 53.2% | 50.7% | |
| Tennessee | Midwest | 11 | 17.3 | 43.2% | 47.1% | 44.6% | 55.7% | 29.5% | 52.9% | 56.4% | 36.7% | 52.1% | 68.0% | 72.0% | 36.0% | 45.7% | 59.6% | 57.1% | |
| Florida | South | 1 | 21.1 | 56.8% | 60.6% | 58.2% | 68.6% | 42.2% | 66.0% | 69.2% | 63.3% | 65.4% | 79.0% | 82.2% | 49.3% | 59.2% | 72.0% | 69.9% | |
| UCLA | South | 4 | 16.7 | 41.1% | 45.0% | 42.5% | 53.6% | 27.7% | 50.7% | 54.3% | 47.9% | 34.6% | 66.0% | 70.2% | 34.0% | 43.6% | 57.5% | 55.0% | |
| Stanford | South | 10 | 12.1 | 26.2% | 29.5% | 27.4% | 37.3% | 15.7% | 34.6% | 38.0% | 32.0% | 21.0% | 34.0% | 54.6% | 20.4% | 28.3% | 41.1% | 38.7% | |
| Dayton | South | 11 | 10.8 | 22.5% | 25.6% | 23.6% | 33.0% | 13.1% | 30.4% | 33.7% | 28.0% | 17.8% | 29.8% | 45.4% | 17.3% | 24.5% | 36.7% | 34.3% | |
| Arizona | West | 1 | 21.3 | 57.5% | 61.3% | 58.9% | 69.2% | 42.9% | 66.7% | 69.9% | 64.0% | 50.7% | 66.0% | 79.6% | 82.7% | 59.9% | 72.6% | 70.5% | |
| Wisconsin | West | 2 | 18.5 | 47.5% | 51.4% | 48.9% | 59.9% | 33.3% | 57.1% | 60.6% | 54.3% | 40.8% | 56.4% | 71.7% | 75.5% | 40.1% | 63.7% | 61.3% | |
| San Diego State | West | 4 | 14.6 | 34.0% | 37.7% | 35.3% | 46.1% | 21.7% | 43.2% | 46.8% | 40.4% | 28.0% | 42.5% | 58.9% | 63.3% | 27.4% | 36.3% | 47.5% | |
| Baylor | West | 6 | 15.3 | 36.3% | 40.1% | 37.7% | 48.6% | 23.6% | 45.7% | 49.3% | 42.9% | 30.1% | 45.0% | 61.3% | 65.7% | 29.5% | 38.7% | 52.5% | |
Note: odds were updated to reflect the observed standard deviation in the first two rounds.
Enjoy!