I prefer a system similar to the old BCS

I believe the best way to rank college football teams is to take the geometric mean of three complementary ranking systems. For simplicity, I use:

((AP + Coaches)/2 × Colley × Sagarin)^(1/3)

Each component captures a different, legitimate way of evaluating teams:

• Colley: Wins and losses against the schedule actually played — what you've earned.
• Sagarin: Performance-based evaluation of team strength — how you'd be expected to fare today.
• Human polls: Subjective judgment and heuristics, incorporating context such as injuries and other factors that mathematical systems may miss — how knowledgeable observers assess you.

These are three substantially different approaches to ranking teams, each with its own strengths and weaknesses.

The geometric mean is particularly appropriate because it rewards excellence in any one category without allowing a poor ranking in another to dominate. In a field of 130-plus teams, the difference between #1 and #5 is far more consequential than the difference between #51 and #55.

As of this past weekend (November 1), my top 15 are:

1. Miami 1.587
2. Georgia 3.302
3. Notre Dame 3.803
4. Utah 4.718
5. UCLA 5.066
6. Indiana 6.240
7. Texas 6.804
8. Alabama 7.958
9. Kentucky 8.143
10. LSU 8.772
11. Ohio State 9.628
12. Ole Miss 10.483
13. BYU 10.801
14. Wake Forest 12.515
15. Texas Tech 14.753

For Sagarin, I use his primary overall rankings, not the separate Predictor or Elo-Chess variants. For teams outside the AP or Coaches top 25, I extend the rankings using the others receiving votes totals.