Removing the bias: Anonymous College Football Rankings

[BobJoeJim]

Here's the game. Below I have listed ten college football teams, with their schedules and results. The catch is, I have removed all team names. Opponents are listed in terms of win-loss record and their opponents' winning percentage (or their division in the case of 1AA teams).

Place these ten teams in order from best to worst. I will tally the results of all ballots on Thursday, prior to the BC/VT game. I think it will be more fun if you don't actively try to figure out who is who until after you vote, though some teams may be obvious. Yes, these are the current top ten teams in the BCS presented in random order.

TEAM A (6-1)
@ (6-1, .523) - L, 21-13
@ (4-3, .500) - W, 31-14
(4-4, .578) - W, 48-7
(4-4, .501) - W, 38-13
(3-5, .464) - W, 62-24
@ (2-6, .567) - W, 55-14
@ (0-6, .545) - W, 48-23

TEAM B (7-0)
@ (4-3, .597) - W, 30-24
@ (4-4, .583) - W, 19-14
(4-4, .492) - W, 52-7
(3-5, .531) - W, 58-10
(3-5, .478) - W, 45-13
(0-7, .657) - W, 55-3
(2-5, 1AA) - W, 62-0

TEAM C (7-0)
(4-3, .597) - W, 44-32
(4-4, .583) - W, 33-14
@ (3-4, .710) - W, 41-3
(3-5, .466) - W, 45-3
(2-5, .746) - W, 44-20
@ (2-5, .597) - W, 23-20
(2-5, .463) - W, 34-14

TEAM D (7-1)
(6-1, .493) - W, 41-31
vs. (6-2, .464) - 28-21
(5-3, .378) - W, 51-13
@ (4-3, .436) - W, 62-21
@ (4-4, .583) - L, 27-24
(1-6, .404) - W, 79-10
@ (1-7, .532) - W, 17-7
(0-7, .631) - W, 54-3

TEAM E (6-1)
@ (6-2, .495) - W, 39-7
(5-2, .433) - L, 31-24
(5-2, .412) - W, 52-21
(4-3, .415) - W, 48-27
@ (3-4, .710) - W, 55-31
@ (2-5, .746) - W, 55-34
(2-5, .597) - W, 53-7

TEAM F (7-0)
(5-2, .571) - W, 38-28
@ (5-3, .555) - W, 24-10
(4-3, .536) - W, 55-24
(3-5, .491) - W, 37-17
(2-5, .583) - W, 37-17
@ (1-7, .746) - W, 27-14
(6-1, 1AA) - W, 24-14

TEAM G (8-0)
@ (6-2, .407) - W, 23-7
(5-3, .486) - W, 24-17
(5-3, .448) - W, 58-7
(3-4, .584) - W, 20-2
(3-5, .494) - W, 48-3
@ (2-5, .746) - W, 33-14
@ (1-7, .577) - W, 30-7
(5-3, 1AA)- W, 38-6

TEAM H (7-1)
(6-1, .445) - W, 48-7
@ (6-2, .532) - L, 43-37
(6-2, .514) - W, 28-16
(5-2, .593) - W, 28-24
(5-3, .619) - W, 30-24
(4-4, .501) - W, 45-0
(3-5, .507) - W, 44-0
@ (2-5, .459) - W, 34-9

TEAM I (6-1)
@ (7-1, .594) - L, 48-7
@ (5-2, .443) - W, 41-23
(4-4, .578) - W, 17-7
(3-5, .399) - W, 28-7
(2-5, .654) - W, 17-10
@ (1-6, .722) - W, 43-14
(4-3, 1AA) - W, 44-3

TEAM J (6-1)
(6-1, .444) - W, 21-13
@ (5-2, .505) - L, 30-27
@ (5-3, .619) - W, 26-23
@ (4-3, .477) - W, 35-23
(4-3, .470) - W, 64-12
(2-5, .654) - W, 37-10
(5-2, 1AA) - W, 28-13

[MyTurn2Raise]

[img]/images/graemlins/frown.gif[/img] ... I wish I didn't know the scores of games played this year like the back of my hand

good exercise though

[Semtex]

Here are the strengths of schedule, which is the first step. So many numbers to deal with. The first column is that teams win loss record. The second is the combined opponents win loss record.

1. J (6-1) (.620)
2. H (7-1) (.607)
3. E (6-1) (.540)
4. F (7-0) (.500)
4. I (6-1) (.500)
6. G (8-0) (.484)
7. D (7-1) (.450)
8. A (6-1) (.442)
9. C (7-0) (.392)
10. B (7-0) (.377)

EDITED to include win loss records for the teams

[Semtex]

OK now to clear up all the information, included information about the teams loss. We should be good to go.

1. J (6-1) (.620) (@ 5-2, 30-27)
2. H (7-1) (.607) (@ 6-2, 43-37)
3. E (6-1) (.540) (5-2, 31-24)
4. F (7-0) (.500)
4. I (6-1) (.500) (@ 7-1, 48-7)
6. G (8-0) (.484)
7. D (7-1) (.450) (@ 4-4, 27-24)
8. A (6-1) (.442) (@ 6-1, 21-13)
9. C (7-0) (.392)
10. B (7-0) (.377)

given all this, my list goes like this:

1. F
2. G
3. J
4. H
5. E
6. C
7. B
8. I
9. A
10. D

F and G played decently hard schedules, and managed to go unbeaten. J H and E played the toughest schedules out of the entire group, and their one loss games were close against winning teams. While team I played a much harder schedule than the remaining two unbeatens, C and B, I punish them for having the only blowout loss out of all the 1 loss teams. Finally are D and A who played tougher schedules than C and B but lost once. I punish D for losing to a worse team than A, even though D played a slightly tougher schedule.

[PokerFink]

And then we can add in margin of victory, which is the second important statistic in determining team strenght.

And then, voila! We have a computer ranking, which are >>>>>>> the human polls.

[Semtex]

[ QUOTE ]
And then we can add in margin of victory, which is the second important statistic in determining team strenght.

And then, voila! We have a computer ranking

[/ QUOTE ]
this isn't saying much. computers are just doing what we do: objectively assimilating, analyzing and quantifying the relevant data. they assist us in making the decisions we want to because there is a ton of data that we can't go through, but would organize a certain way if we had the time to do it by hand. it seems like what you are saying is given this data everyone would end up with the same ranking, which is not true, because everyone would have their own rules for organizing it.

[Semtex]

[ QUOTE ]
Strengths of schedule. 1st column is the team's win-loss, 2nd is combined opponent's win-loss, 3rd is details of the loss:

1. J (6-1) (.620) (@ 5-2, 30-27)
2. H (7-1) (.607) (@ 6-2, 43-37)
3. E (6-1) (.540) (5-2, 31-24)
4. F (7-0) (.500)
4. I (6-1) (.500) (@ 7-1, 48-7)
6. G (8-0) (.484)
7. D (7-1) (.450) (@ 4-4, 27-24)
8. A (6-1) (.442) (@ 6-1, 21-13)
9. C (7-0) (.392)
10. B (7-0) (.377)

Given all this, my list goes like this:

1. F
2. G
3. J
4. H
5. E
6. C
7. B
8. I
9. A
10. D

F and G played decently hard schedules, and managed to go unbeaten. J H and E played the toughest schedules out of the entire group, and their one loss games were close against winning teams. While team I played a much harder schedule than the remaining two unbeatens, C and B, I punish them for having the only blowout loss out of all the 1 loss teams. Finally are D and A who played tougher schedules than C and B but lost once. I punish D for losing to a worse team than A, even though D played a slightly tougher schedule.

[/ QUOTE ]

This list was:
1. BC
2. tOSU
3. USF
4. LSU
5. ORE
6. ASU
7. KU
8. VT
9. WVU
10. OU

I don't know, looks interesting. If this system and my interpretation is valid it would indicate USF is severely underrated and Oklahoma is overrated. Oklahoma makes sense they got the worst computer ranking of anyone in the top 10.

Here are the BCS computer rankings:

1. BC
2. LSU
3. tOSU
4. ASU
5. USF
6. UV (huh?)
7. KU
8. VT
9. ORE
T10. WVU
T10. UF

I'm guessing Oregon gets no love from the computers because they haven't reached the really hard part of their schedule yet. The analysis I did using the Sagarin Predictor showed that Oregon's total schedule was significantly tougher than LSU's. However for the most part in the Sagarin ratings the SEC teams are ranked higher while in the Predictor the Pac 10 teams are, and I used the Predictor. If the different systems used for the BCS resemble the Predictor more than the normal Sagarin ratings Oregon should overtake LSU in the computers if they win out. The question of whether they will in the polls is another matter. Also one cannot factor in the SEC championship game, but the Predictor showed Oregon's schedule to be significatly tougher, enough to where one game, even against a number 1 wouldn't be enough. I did average opponents rank though and I don't know how valid that is. LSU's non conference opponents were all in the low 100s and even 200s (minus VT, though the Predictor ranks nowhere near the top 10). This killed their strength of schedule using that method.

[BobJoeJim]

Problem is that the BCS computers don't use margin of victory, and the predictor does, so the predictor is not actually a good tool to use when analyzing how tough a team's schedule will be in the "minds" of the BCS computers.