College football bowl season has rapidly changed in recent years with the rise of transfers and opt-outs. That could become even more exaggerated with the expansion of the playoff to 12 teams next year. For now, it means projecting these games is a mess.
My college football model has projected scores and point totals for every bowl game, including the two College Football Playoff semifinals. The model takes in past performance, returning talent and recruiting rankings. We then compare the model’s projections to the BetMGM lines for extra context. However, there must be a major disclaimer with the numbers during bowl season.
I did some of my best guesswork to project who will play or not and the value of each player. Basically, the process is that I have projected values for each position, and I adjust based on who isn’t going to play. On top of that, there are games where I have to guess whether they play or not. We still haven’t heard from Marvin Harrison Jr., for example, on whether or not he will play in the Cotton Bowl, so for now, he is at 50 percent value. Then you have games with major coaching staff changes, so you have no idea how much (or how differently) they’re prepping.
Essentially, I have to regress everyone closer together while increasing variance.
All that’s to say bet with these numbers at your own peril. These are probably more for entertainment purposes than normal.
Full bowl game model projections
XMOV is the model’s projected margin of victory for the game with the BetMGM spread as of publish time listed to the right of it. XTOTAL is the projected combined point total with the BetMGM total (over/under) listed as of publish time to the right of that number.
(Photo of Luther Burden: Wesley Hitt / Getty Images)
2023-12-13 18:37:04
#College #football #bowl #projections #Predicted #scores #bowl #game
:quality(85)/cloudfront-us-east-1.images.arcpublishing.com/infobae/5VWBFN6IKJA5HJQNNRV2C6J7NE.jpg?fit=300%2C300&ssl=1)
