So the whole goal of this project was to do one thing: chase perfection. The brackets have been filled out. Now all there is left to do is wait.
I have decided to run the four models that I went over during the conference tourneys. The composite model that I referenced ended up being too similar to the other models (since it is a composite) and I didn't find it to be that particularly interesting. So four models enter the perfection gauntlet and it is very likely that no models will survive tomorrow.
But to see where our models might fall, I have provided each model's first eight games and their winner and discuss where each model might fall.
| Game | EDM | SPM | MASPM | EDLM |
|---|---|---|---|---|
| 7 Louisville vs 10 Minnesota | 7 Minnesota | 7 Minnesota | 7 Minnesota | 7 Minnesota |
| 3 LSU vs 14 Yale | 3 LSU | 3 LSU | 3 LSU | 3 LSU |
| 5 Auburn vs 12 New Mexico State | 5 Auburn | 5 Auburn | 5 Auburn | 5 Auburn |
| 4 Florida State vs 13 Vermont | 4 Florida State | 4 Florida State | 4 Florida State | 4 Florida State |
| 2 Michigan State vs 15 Bradley | 2 Michigan State | 2 Michigan State | 2 Michigan State | 2 Michigan State |
| 6 Maryland vs 11 Belmont | 11 Belmont | 6 Maryland | 6 Maryland | 11 Belmont |
| 4 Kansas vs 13 Northeastern | 4 Kansas | 4 Kansas | 4 Kansas | 4 Kansas |
| 5 Marquette vs 12 Murray State | 12 Murray State | 5 Marquette | 5 Marquette | 5 Marquette |
The most worrisome game is the first game of the day. All four models pick Louisville and if Minnesota wins, well, that was disappointing. LSU should win over Yale, but Yale is a pretty strong 14 seed. Then some scary 12 and 13 seeds in New Mexico State and Vermont that could axe all my brackets at once. All models match in their predictions in the first five games.
So let's assume that we achieve perfection through the first five. That's not out of the question. What about game six? Well, we are forced to lose two models since there is a difference of opinion here. Some models pick Belmont while others pick Maryland. If we somehow make it to eight games, there is another difference in opinion between the EDM model and the rest of the models. If the EDM is right to this point, it will be alone.
There's a decent chance that if the set of models get their first five games correct, there will be a perfect bracket 8 games through the tournament. Which leaves 55 games to go. But that Minnesota game scares me. And the New Mexico State one. Those ooze upset potential. But we'll see. Knowing the tournament, something random will happen like Michigan State losing to Bradley. Which keeps things fun.