ESPN’s Football Power Index (FPI) doesn’t buy Colorado in 2023.

The Buffs are favored by the predictive model in just one game this upcoming season — a Sept. 16 showdown with Colorado State. The model gives CU less than a 10% chance of winning four games on the schedule — TCU, Oregon, USC, and Utah — and less than a 30% chance of winning four others.

Here’s the schedule breakdown from FPI:

  • Sept. 2 at TCU — 8.3% chance of winning
  • Sept. 9 vs. Nebraska — 28.5%
  • Sept. 16 vs. Colorado State — 59.1%
  • Sept. 23 at Oregon — 5.4%
  • Sept. 30 vs. USC — 6.0%
  • Oct. 7 at Arizona State — 27.9%
  • Oct. 13 vs. Stanford — 42.3%
  • Oct. 28 at UCLA — 13.1%
  • Nov. 4 vs. Oregon State — 16.5%
  • Nov. 11 vs. Arizona — 28.2%
  • Nov. 17 at Washington State — 25.4%
  • Nov. 25 at Utah — 5.4%

It’s worth pointing out that predictive models like FPI are built on quantitative metrics that will seriously undervalue this specific Colorado team. Returning production, recent history, and recent recruiting success all play large roles in building out the preseason rankings for things like FPI and SP+.

Colorado has none of that. First-year head coach Deion Sanders dramatically overhauled the roster this offseason, replacing a huge percentage of the team that went 1-11 last fall with the largest crop of transfer players in the FBS this season.

How does the experiment work? That’s anyone’s guess.

CU averaged just 15.4 points a game last season, the fifth-worst mark in the FBS. That mark figures to improve under Sanders.

But Colorado also gave up 44.5 points a game to rank dead last in the country. How much ground gets made up there? That’ll determine how many games CU can, judging by FPI’s predictions, “steal” in 2023.