College Football Prediction Models
We finally made it! We are back with another season of college football and I’m so appreciative you’re choosing to make reading my work and including TSI into your weekly college football betting regimen.
First, a refresher on what TSI is: TSI stands for T Shoe Index, a nod to my (admittedly uncreative) nickname. It is my ratings and projections system that I’ve created to assist with my sports betting ventures across multiple sports. It is an opponent-adjusted, tempo-adjusted, team strength metric that aims to answer the question, “How many points would Team A be expected to score and allow vs. Team B”? Seems simple enough, but I think sometimes in sports betting and the analytics space, people try to get overly complicated and create super advanced metrics, that are very cool but perhaps not very predictive when you’re trying to beat a spread or an Over/Under.
This will be my weekly Thursday piece throughout the season where I will demonstrate where TSI and other top predictive models (SP+ and FPI) align the most – which could be a bet-on signal – or where they disagree the most – which could be a sign to stay away. Obviously, with just a handful of games this week, the volume will be a little lower, but moving forward I’d anticipate 3-5 games per week in this article.
Note: SP+ Rankings are the work of CFB analyst Bill Connelly (ESPN) and FPI is the Football Power Index from ESPN Analytics.
Model Consensus
Kansas State -3
In probably the biggest game of the weekend, Kansas State and Iowa State will do battle in a Dublin, Ireland edition of “Farmageddon”. I wrote this game up in my Monday article as a best bet because TSI projects Kansas State -7, but adding more gasoline to the flame is the fact the both SP+ and FPI predict Kansas State -4, which is still comfortably on the TSI side of the -3 line that is widely available.
Stanford +2.5
This line has moved a ton over the summer, where Stanford was once a 2.5-point favorite both domestically and offshore, they are now a 2.5-point underdog to the Rainbow Warriors in a huge steam move; however, TSI projects Stanford -4, while SP+ and FPI project Stanford -4 and -8, respectively. This will be a very good early season test of market knowledge vs. the nerds, but this much alignment on a huge discrepancy tells me I need to get down some Stanford money.
Model Dissension
With just a few games this week, there is actually no real model dissension. All three models are pretty aligned on Kansas State, Sam Houston State and Stanford.
Again, this will be a weekly feature so starting in week 1, there will definitely be some noteworthy discrepancies between the models. Remember, not all “sharps” or models are always on the same side of a game!
