SRS: With Win Over Florida, LSU Grabs Top Spot; Auburn No. 2

So after reading all of the comments and thinking through the list, I looked at the formula a little closer, and found a tragic, egregious, major error.  I found it because I really looked at the two teams that made the least sense to me (to others, I'm not sure):  Hawaii and Temple.  Both teams had something in common:  They had both beaten Army and received a ton of points for it.  But why?  Why was Army rated so highly?  

I looked into it and found out that losses were (for the conference rankings) weren't being applied equally.  Fixing this error changed quite a bit.  Unfortunately, there are probably some other issues that you all are going to be upset with, but that's part of the fun.  I've included in my rankings this week what they were ranked last week, and what they should have been ranked last week, as well as what their rating is, and how many points they earned for their victory (or how many they points they lost if they did not win).

Another new part of the column is what should I look into changing for next season (or even next week if it truly makes the ranking system seemingly better).  Right now, I think the biggest problem with the system is that when measuring the difference between conferences, it is not accurate enough.  The best conference in the country is exactly 1 point better than the second, and exactly 16 points better than the worst.  

What if the best conference has all of the best teams in the nation?  They shouldn't be rated as exactly 1 point better than the second conference, and they'd certainly be better than just 16 points above the worst conference.  I'm working on creating a rule to reflect this, once I get it in, and looking right, I'll let you know what happened--but with school, my job, and my business, this may take a while (I'm already late for this week's rankings!)

So how are the rankings created? 

First, the easy one: The whole point of me starting a ranking system 15 years ago was to remove any and all bias (that’s what they all say). The only teams I might have any sort of bias towards are the small Boise States of the world (no, I don’t have them No. 1) or my favorite team, Fresno State (not ranked). 

I’ve tried to come up with the rationale that every sports announcer I’ve ever heard brings up: Conference rankings (check), Road vs. Away (check), and who did you play (double-check).

As I said before, I’ve been doing this privately for over 15 years, and every year there have been slight changes to it. Usually by the end of the season, I see more improvements I can make, and I do that for the following season. By no stretch is it perfect, but I think it’s getting there.

So what are the specifics? It starts by generating a ranking for each conference. This is done by generating a win-loss record for each conference against the other conferences. For the purpose of this system, I consider each division a conference (for example, the SEC West is a conference—so are the independents).

This is modified by determining who those wins are against. For example, the Big Ten loses some of its power by scheduling lots of MAC teams. I take the total and divide it by how many teams are in the conference to normalize the value, and then take a ranking of that value.

As of Oct. 12, 2010, the conference rankings are as follows:

17. MAC East (LW: 16, LW with Bug Fix: 17)

16. C-USA East (LW: 15, 15)

15. Sun Belt (LW: 17, 16)

14. MAC West (LW: 13, 14)

13. Big East (LW: 14, 13)

12. ACC Atlantic (LW: 12, 12)

11. WAC (LW: 11, 11)

10. C-USA West (LW: 10, 10)

9.  MWC (LW: 6, 9)

8.  ACC Coastal (LW: 8, 7)

7. Independents (LW: 7, 12)

6. SEC East (LW: 9, 6)

5. Big Ten (LW: 5, 5)

4. Pac-10 (LW: 4, 4)

3. Big 12 North (LW: 3, 3)

2. Big 12 South (LW: 2, 2)

1. SEC West (LW: 1, 1)

After determining the conference rankings, I generate values that a team would receive for playing another team. Values are generated for a Home Win, a Home Loss, a Road Win, and a Road Loss.

Each of the values is derived from the Home Win, which is formed by taking the team's place in the conference (first, second, third, etc.), inverting it so more points are awarded for beating the team, and multiplying it by the inversed conference ranking. This is then normalized again for the number of teams in the conference.

For example, if the first place team in the first place conference is Alabama, and Arkansas beats Alabama, they will earn 102 points. This is because Alabama was in first place (inversed = 7 multiplied by inversed conference ranking = 17, 7 * 17 = 102). 

Losses are similar, though there is no inverting. Thus an Arkansas loss will only cause them to lose two points (Place in Conference = 1, Conference Rank = 1, 1 * 1 = 1). I added an additional modifier of two if you lost at home, so 1 * 2 = 2. A Road Win is the same as a home win, except the value is multiplied by two, and a Road Loss takes the Home Loss and divides it by two. 

Did you follow all of that? If not, don’t worry about it; just know that it is under constant improvement, and again, this where I think they will be if they continue to do what they did last year. I’ll go in reverse order, like a Top 10 list:

25. Texas (3-2), 105: (LW: --, --), BYE; Best Win = @Texas Texas Tech; Worst Loss = UCLA

24. Arizona (4-1), 109: (LW: 11, 10), Lost vs. Oregon St. (-24); BW = Cal; WL = Oregon St.

23. Michigan (5-1), 112: (LW: 13, 15), Lost vs. Michigan St. (-16); BW = @Indiana; WL = Michigan St.

22. Temple (3-2), 114: (LW: 10, 24), Lost AT N. Illinois (-14); BW = @Army; WL = @Penn State

21. Florida (4-2), 119: (LW: 16, 8), Lost vs. LSU (-2); BW = @Tennessee; WL = @Alabama

20. Missouri (5-0), 122: (LW: --, --), Won vs. Colorado (45); BW = Colorado; WL = N/A

19. Oklahoma St. (5-0), 126: (LW: 24, 17), Won @ Louisiana-Lafayette (28); BW = Texas A&M; WL = N/A

18. Stanford (5-1), 148: (LW: 9, 11), Won vs. USC (34); BW = @UCLA; WL = @Oregon

17. North Carolina St. (5-1), 150: (LW: 20, 14), Won vs. BC (18); BW = @Georgia Tech; WL = Virginia Tech

16. Air Force (5-1), 150: (LW: 8, 18), Won vs. Colorado St. (18); BW = @Wyoming; WL = @Oklahoma

15. Florida St. (5-1), 154: (LW: 21, 22), Won AT Miami (FL) (80); BW = @Miami (FL); WL = @Oklahoma

14. Arkansas (4-1), 154: (LW: --, --), Won AT Texas A&M (96); BW = @Texas A&M; WL = Alabama

13. Nebraska (5-0), 156: (LW: --, --), Won AT Kansas St (120); BW = @Kansas St; WL = N/A

12. Hawaii (4-2), 161: (LW: 18, --), Won AT Fresno St (56); BW = @Army; WL = USC

11. Ohio St. (6-0), 162: (LW: 17, 12), Won vs. Indiana (35); BW = @Illinois; WL = N/A

10. Nevada (5-0), 174: (LW: 4, 9), Won vs. San Jose St (9); BW = @BYU; Worst Loss = N/A

9.  Oregon St. (3-2), 177: (LW: --, --), Won AT Arizona (134); BW = @Arizona; Worst Loss = @Boise St.

8.  Oklahoma (5-0), 196: (LW: 7, 6), BYE; BW = Texas A&M; Worst Loss = N/A

7.  Alabama (5-1), 200: (LW: 2, 2), Lost AT South Carolina (-6); BW = @Arkansas; Worst Loss = @South Carolina

6.  Oregon (5-0), 205: (LW: 1, 1), Won AT Washington St (17); BW = @Arizona State; WL = N/A

5.  Michigan St (6-0), 237: (LW: --, 25), Won AT Michigan (128); BW = @Michigan; WL = N/A

4. TCU (6-0), 240: (LW: 5, 5), Won vs. Wyoming (30); BW = Baylor; Worst Loss = N/A 

3. Boise State (6-0), 269: (LW: 3, 4), Won vs. Toledo (20); BW = @Virginia Tech; Worst Loss = N/A

2. Auburn (5-0), 282: (LW: 12, 7), Won AT Kentucky (96); BW = @Kentucky; WL = N/A

1. LSU (6-0), 360: (LW: 6, 3), Won AT Florida; BW = @Florida; Worst Loss = N/A

My college football playoff preference:

The detractors of a playoff indicate that the reasons they don’t like a playoff is that there would be too many games included in the season, and it would be hard to transport as many people as would be needed to each and every neutral site game.  So to solve this, the playoff needs to let a minimum number of teams in as possible, and only have a few of the games be a neutral site game. My thoughts are: Make it like the NIT tournament, each of the higher seeded teams get home games until the final four.  Also, in order to give everyone a chance, have only conference champions invited.  That gives us 11 teams, and in order to make it an even number, I’m okay with one “wild card”.  So given that, let’s take the Current Standings and choose our 12 schools.

ACC—Florida St

Big XII—Oklahoma

Big East—Syracuse

Big Ten—Michigan St

C-USA—SMU

 MAC—Northern Illinois

MWC—Air Force

Pac-10—Oregon

SEC—LSU

Sun Belt—Troy

WAC—Hawaii

Wild Card—Boise St (BCS Ranked No. 1, but Hawaii currently ahead of them in the conference standings)

Taking these 12 schools, I’d then seed them by conference strength.

1. LSU
2. Oklahoma
3. Oregon
4. Michigan St
5. Florida St
6. Air Force
7. SMU
8. Hawaii
9. Syracuse
10. Northern Illinois
11. Troy
12. Boise St

In this scenario, the first round of games would be Boise St @ Florida St, Troy @ Air Force, Northern Illinois @ SMU, and Syracuse @ Hawaii.  I think the only road team that wins in those 4 would be Boise St.  Similar to the NFL, I’d then make the lowest seeded team that survives play the highest seeded team on a bye.  That would result in:  Boise St @ LSU, Hawaii @ Oklahoma, SMU @ Oregon, and Air Force @ Michigan St. 

That would leave us with a Final Four Saturday-Sunday Contests in a Neutral Site to be:  Michigan St-LSU, and Oregon-Oklahoma, with a Final of LSU-Oklahoma.

Playoff Dreams are fun, too bad you have to wake up.