There is a lot of things to consider when building a rankings system for college football. When I first had the idea for this project a few years ago, I wanted to build a computer rankings that did two things:
- Eliminate the preseason rankings entirely and only count on the field results
- Remove the concept of "poll inertia", where a team can only drop or gain so many spots week to week
In my opinion, when you are building a rankings list each week of the season should count the same as the last. This doesn't necessarily give you the information about what team is the hottest currently, but it gives you a good overview of the season. I always thought it was weird how the AP Poll would sort of arbitrarily decide to drop teams x spots week to week. Or how they would overweigh preseason rankings, so an 0-2 Notre Dame was still ranked after the start of this season. Even though I have them ranked #11 now, they actually had to earn that spot rather than relying on the eye test or recruiting rankings. On September 14th after they had just lost to Texas A&M, the Index had them ranked 72nd. As the season went on they proceeded to continue winning while the two losses to Miami and A&M began to age better.
So I set out to build my rankings system that would count each week evenly (except for conference championships/playoff wins) and only count the on the field metrics. The main metrics I look at in no particular order are:
- W/L
- Strength of schedule
- Yards per game
- Yards per game allowed
- Points per game
- Points per game allowed
- Number of wins and losses
- A few more various stats
I have weights that count some of these stats more than others. Each team is compared against the best in the country and points are awarded based on that. So the team with the most YPG will get the full amount of points, then every other team gets some fraction of those based on their YPG. All of these points are added up and normalized so that the team with the most is at 1000. As of today that is the Indiana Hoosiers who are very closely followed by OSU. Here are the rankings for 11/18/25:
| Rank | Difference | Team | Win | Loss | Score |
|---|---|---|---|---|---|
| 1 | - | Indiana | 11 | 0 | 1000 |
| 2 | - | Ohio State | 10 | 0 | 991 |
| 3 | ↑ 1 | Texas Tech | 10 | 1 | 937 |
| 4 | ↓ 1 | Texas A&M | 10 | 0 | 926 |
| 5 | - | Georgia | 9 | 1 | 925 |
| 6 | ↑ 2 | BYU | 9 | 1 | 906 |
| 7 | ↑ 2 | Oregon | 9 | 1 | 892 |
| 8 | ↓ 1 | Ole Miss | 10 | 1 | 889 |
| 9 | ↑ 2 | USC | 8 | 2 | 881 |
| 10 | - | Utah | 8 | 2 | 878 |
| 11 | ↑ 1 | Notre Dame | 8 | 2 | 876 |
| 12 | ↓ 6 | Alabama | 8 | 2 | 872 |
| 13 | - | Oklahoma | 8 | 2 | 867 |
| 14 | ↑ 2 | Miami | 8 | 2 | 848 |
| 15 | ↓ 1 | Michigan | 8 | 2 | 847 |
| 16 | ↑ 2 | Virginia | 9 | 2 | 830 |
| 17 | ↓ 3 | Georgia Tech | 9 | 1 | 820 |
| 17 | ↑ 4 | North Texas | 9 | 1 | 820 |
| 19 | ↓ 2 | Vanderbilt | 8 | 2 | 818 |
| 20 | ↑ 5 | James Madison | 9 | 1 | 808 |
| 20 | ↑ 2 | Illinois | 7 | 3 | 808 |
| 22 | - | Tulane | 8 | 1 | 801 |
| 23 | ↑ 11 | Arizona | 7 | 3 | 789 |
| 24 | ↓ 4 | Texas | 7 | 3 | 786 |
| 25 | ↑ 7 | Missouri | 7 | 3 | 782 |
This week is a bit controversial, as the rankings intentionally do not count head to head losses when putting a team in its spot. This can lead to seemingly strange behavior, like Alabama being ranked above OU who just beat them. Naysayers will call me an Alabama homer for this (and they would be right), but there is a thought process. In my rankings the winning team is already getting a boost (via W/L and assumed good stats for the win) and the losing team is getting punished at the same time (for the opposite reasons). It feels weird to then also tack on an arbitrary rule that forces one team to be higher than the other. But in a game like the Alabama OU one, Alabama wouldn't be punished as hard as they would for the FSU one. This is because OU has a solid record and SOS, and Alabama statistically didn't play all that bad (except for turnovers).
I prefer to have the rankings extend over the entire season, which means that head to head isn't valued directly as much as other stats. But I understand how people will have a problem with this at first glance, and don't mind the criticism.
If you have hot takes about the rankings please send them to me and I'll give you a guest blog spot.