It's obviously not perfect lol, but I think the results are somewhat reasonable, with a clear few outliers(appalachin state is a joke, but the way they've played "games" the model, they have their blowout loss to Clemson, but otherwise have dominated every other game inflating their off/def even with their 122 ranked sos adjusted.
Iowa has the #13 SOS in my model since their OW record is 30-14(.682) and their OOW is 197-126(.61) and that gives them a SOS of .6676 since it's weighted 80-20. They most likely are going to be undefeated going into the B1G championship game(Ohio State will blow them out, MSU will be close, and Michigan who can technically win it would also be a close game, or Michigan could blow them out possibly)
Iowa's offense is adjusted to be 19th and their adjusted defense is 4th. and being undefeated which my model being 20% win% 40% offense and 40% defense they do extremely well since they are top 20 Offense and top 5 defense with an undefeated record.
Personally I do think Memphis is a top 10 team, and they have upcoming games to prove it against Houston and Temple(especially if they knock off ND), in the model they benefit from having the #1 offense(Baylor SOS is worse than Memphis so while statistically they are better their SOS multiple lowers them behind memphis) their defense is clearly weak coming in at #83, but they are undefeated.
Maybe I should make it 10% win% 45 offense 45 defense, but I want to see how this works for at least a week or two.
2012 Ohio State had #21 Offense and #32 defense. which for the 12th spot is perfectly fine. Remember they only beat a 3-9 Cal team by 7, MSU by 1, Indiana by 3, Purdue in OT by 7, Wisconsin by 7 in OT, and Michigan by 5. They really only blew out 3 teams in Miami of Ohio, Nebraska "can't play big games", and world beater Illinois. You might be remembering them too strong since they finished undefeated and #3. Personally they would be underdogs against each of the fina top 13 teams.
2013 Baylor has #1 Offense and #28 defense, remember they were #6 in the polls before the Fiesta Bowl loss. They blew out #10 Oklahoma and beat a ranked Texas by 20 as well. Hitting 60 points in 6 games(2 G5 and 3 P5 and 1 FCS) with 4 being 70 or better. I do not think they were #2 I think MSU was easily #2 and would've had a similar shot that Auburn had against FSU.
2014 FSU had #17 Offense and #33 Defense similar to #12 2012 Ohio State.
Here is current top AP top 10 and my AP top 10:
1 Ohio State ----- Clemson
2 Baylor ----- Iowa
3 Clemson ----- LSU
4 LSU ----- Ohio St
5 TCU ----- Memphis
6 Michigan State ----- Baylor
7 Alabama ----- Oklahoma
8 Stanford ----- Alabama
9 Notre Dame ----- Michigan
10 Iowa ----- Notre Dame
2014
1. Ohio State ----- Ohio State
2. Oregon -----Oregon
3. TCU ----- Alabama
4. Alabama ----- Michigan St
5. Michigan St ----- TCU
6. Florida St ----- Georgia
7. Baylor ----- Mississippi
8. Georgia Tech ----- Boise St
9. Georgia ----- Marshall
10. UCLA ----- Georgia Tech
2013
1. Florida St ----- Florida St
2. Auburn ----- Baylor
3. Michigan St ----- Alabama
4. South Carolina ----- Michigan St
5. Missouri ----- Missouri
6. Oklahoma ----- South Carolina
7. Alabama ----- Auburn
8. Clemson ----- Clemson
9. Oregon ----- Louisville
10. UCF ----- Stanford
2012
1. Alabama ----- Alabama
2. Oregon ----- Texas A&M
3. Ohio State ----- Oregon
4. Notre Dame ----- Florida
5. Georgia ----- Notre Dame
6. Texas A&M ----- Georgia
7. Stanford ----- South Carolina
8. South Carolina ----- Florida St
9. Florida ----- LSU
10. Florida St ----- Stanford
Now it's generally the same idea,the top team has been correct each season( was going to do pre-bowl and after bowl), but one of the components I couldn't get pre-bowl data for, so I would've wanted to see the Pre-Bowl and Final Rankings consideration.
For a starting point I think its a solid poll, I agree with the comments on how some teams are over/under ranked, but remember it's a computer poll and a few opinions were by watching not statistically.
I do agree that things like turnovers need to be accounted for, I plan on adding turnover margin and probably something like Opponent short field drives to measure field position. I think I'd like to include 3rd down conversion rate and Yards per play into my offensive and defensive equations and weight rushing offense 55-45 over passing offense and same with defense.
I'll probably try to have that done after this weeks games and see how that looks.