https://fantasy.espn.com/tournament-challenge-bracket/2021/en/entry?entryID=48742608
So at the time of writing my last post, the round of 32 had ended and the tournament was heading into the sweet sixteen. At that point, my bracket was in the 99.4 percentile, and unfortunately, that isn’t the case anymore. My model definitely had some misses, and my human errors also cost me, but in some cases they actually helped. So lets pick things up at the sweet sixteen for each region and see where I went wrong.
westregionalthird %>% select(Team, .pred_class, Opponent, .pred_L) %>% kable() %>% kable_styling(bootstrap_options = c("striped", "condensed"))| Team | .pred_class | Opponent | .pred_L |
|---|---|---|---|
| Gonzaga | W | Creighton | 0.3309778 |
| Iowa | W | Southern California | 0.4358532 |
I was half right on the sweet sixteen in the west, Gonzaga moved on (pretty easily according to the model as well) but Iowa was already eliminated thanks to Oregon, who ended up losing to USC who then got stomped by Gonzaga on their way to the championship game.
eastregionalthird %>% select(Team, .pred_class, Opponent, .pred_L) %>% kable() %>% kable_styling(bootstrap_options = c("striped", "condensed"))| Team | .pred_class | Opponent | .pred_L |
|---|---|---|---|
| Brigham Young | L | Alabama | 0.5938540 |
| Florida State | L | Michigan | 0.6405587 |
So remember, I made a mistake here that had BYU moving on when Alabama should have in their place. Ultimately UCLA went on a run that made that not really matter, but still worth noting. Michigan beat Florida state but got knocked off in a close game against UCLA.
southregionalthird %>% select(Team, .pred_class, Opponent, .pred_L) %>% kable() %>% kable_styling(bootstrap_options = c("striped", "condensed"))| Team | .pred_class | Opponent | .pred_L |
|---|---|---|---|
| Colgate | W | Ohio State | 0.3635714 |
| Baylor | W | Villanova | 0.4004643 |
The South proved to be a pretty bad region in my bracket, thanks to Ohio State losing to Oral Roberts and my model loving Colgate. I got one half of the sweet sixteen right though, and ultimately called Baylor winning the region.
midwestregionalthird %>% select(Team, .pred_class, Opponent, .pred_L) %>% kable() %>% kable_styling(bootstrap_options = c("striped", "condensed"))| Team | .pred_class | Opponent | .pred_L |
|---|---|---|---|
| San Diego State | L | Houston | 0.5743786 |
| Loyola (IL) | W | Tennessee | 0.4593992 |
The midwest region proved to be the worst in my bracket, due to the mistake I made with Loyola Chicago moving on. San Diego had no business being at this point in the bracket, and I made the mistake of having them move on instead of Houston. Although, if I had done this, my model would have Houston beating Baylor in the final four, and ultimately winning the tournament, which would have been a miss that hurt a lot, so I guess that was lucky.
finalfourresults %>% select(Team, .pred_class, Opponent, .pred_L) %>% kable() %>% kable_styling(bootstrap_options = c("striped", "condensed"))| Team | .pred_class | Opponent | .pred_L |
|---|---|---|---|
| Baylor | W | Loyola (IL) | 0.3792516 |
| Gonzaga | W | Michigan | 0.4289167 |
Not bad, got the championship right, and with decently confident margins too.
champs %>% select(Team, .pred_class, Opponent, .pred_L) %>% kable() %>% kable_styling(bootstrap_options = c("striped", "condensed"))| Team | .pred_class | Opponent | .pred_L |
|---|---|---|---|
| Gonzaga | W | Baylor | 0.3790897 |
Well, Gonzaga was supposed to win this game handily, and I really thought they would. Gonzaga was cold at the start, and Baylor just came out hot and didn’t give Gonzaga a chance to come back. Still a very solid bracket, I finished tied for 5th in the class, and in the top 10.7% brackets in the nation, which is pretty good for not having the right team winning the tournament. I used the model to help determine some wins in another bracket I did, but ultimately the bracket purely based on the model beat my personal bracket and my entire family. Going forward, I might have to make some minor changes to attempt to perfect it, and I will definitely use it in future bracket pools and challenges. Thanks for giving this a read, more content to come soon!
-B