Welcome to Badger Bracketology 2.0! We have an updated methodology this year that relies on Markov chains, logistic regression, and simulation to forecast the four team college football playoff. Read more about it here. We have two versions of the model that weigh the games slightly differently, and we will therefore have two flavors of forecasts this year. Both forecasts are very similar: one uses (truncated) point differentials (mLRMC) and the other uses ln(point differentials) (ln(mLRMC)) since a log-scale accounts for large score differentials without giving credit for running up the score without having to truncate the scores.
Both models like Clemson, LSU, Ohio State, and Stanford. The odds drop off quickly after that. However, last year Ohio State was on our radar even though they were not very high in the rankings. Surprises may lie ahead. We are sad that Wisconsin isn’t on our radar this year, but we are happy that several of our B1G friends (frenemies??) are in the Top 10. We are not at all sad that Minnesota isn’t forecasted to make the playoff.
We are surprised that Clemson is the favorite right now, but it looks like the combination of quality wins and an “easy” path to the playoff (sorry, ACC!) is a dangerous combination.
A note about the forecasting: our method uses the network of game matchups. With more data, we will have a more connected network, which make better for better predictions. Each week will be more exciting.
Thanks for following! See you in the coming weeks.
———- mLRMC Method: Top Teams to make the 4-team playoff ————
2 Louisiana State
3 Ohio State
7 Notre Dame
8 Michigan State
11 Texas Christian
13 Oklahoma State
17 Florida State
———- ln(mLRMC) Method: Top Teams to make the 4-team playoff ————
3 Ohio State
4 Louisiana State
6 Michigan State
9 Texas Christian
11 Notre Dame
12 Oklahoma State
16 Florida State