Polls are subjective. There’s no wrong way to do it, unless you consciously downgrade a specific conference and refuse to justify it. My goal was to create a poll that excised what I hate about the existing poll system and emphasized what I feel is valuable. I tried to eliminate the hype and the hypothetical and to focus on what tangibly occurred on the field this season. The methodology and results are after the jump.

Preseason polls have undue influence through inertia, so I began after a reasonable sample size in week four. Not all won-loss records are created equal, so rather than slotting the 4-0 team automatically ahead of the 3-1 team. I qualified wins and losses. I disagree with FCS patsies on principle, so they weren’t counted, unless negatively.  Essentially, my system rewards teams that performed well in a difficult environment, disregarding my subjective assessment as much as possible.

Teams were graded on a point system. One point for a win. Teams received an additional point for an emphatic win (roughly 21 pts or more) or a win over a big opponent (roughly top 25). Teams also received a half-point for a road win. UCLA beat a tough opponent (Texas), emphatically (34-12), on the road. Losses were -1. A “bad loss,” either to a low-ranked opponent or a blow-out, lost an additional point. So, they earned 3.5 pts for that win. But, they also lost two pts for being blown out at home by Stanford.  Ties were broken based on subjective strength of schedule.

The “Duffy” Top 25

1. Alabama (9.5)
2. Ohio St. (8)
3. LSU (7.5)
4. Florida(7.5)
5. Utah (7.5)
6. Stanford (7)
7. Boise St. (6.5)
8. USC (6)
9. Oklahoma (5.5)
10. Michigan (5)*
11. Oklahoma St. (5)
12. Arizona (4.5)
13. Auburn (4.5)
14. Oregon (4.5)
15. TCU (4.5)
16. Nevada (4.5)
17. Nebraska (4.5)
18. North Carolina St. (4)
19. Michigan St. (4)
20. Northwestern (4)
21. Army (4)
22. Wisconsin (3.5)
23. Kentucky (3.5)**
24. Kansas St. (3)
25. Florida St. (3)

Five Out: Texas A&M (3), Toledo (2.5), Arkansas (2.5), Texas (2.5), Texas Tech (2.5)

*I docked Michigan a half-point for the “win” vs. UMass.
** forgot Kentucky when I transferred teams to different document.

Some of the disparities should work out in conference play.  The system is rewarding teams based on conference schedule.  I would probably argue that Arkansas is among the top 20 teams, but they also haven’t proven it.  Against FBS opponents, they beat Louisiana-Monroe (+2), had a close win at Georgia (+1.5) and lost to Alabama (-1).  Florida and LSU played FBS opposition and, though subjective interpretation might suggest differently, beat them emphatically.

I also realize the math is ham-handed.  Happy to accept suggestions from the more mathematically inclined.