Back to threaded view
Full conversation · 12 posts

My top 12 ranking. What is yours? Post and I'll tabulate.

MarkHarman 8162 posts Original UHND Member
Posted
  1. Georgia
  2. Utah
  3. Texas
  4. Notre Dame
  5. Miami
  6. Tennessee
  7. UCLA
  8. Alabama
  9. Indiana
  10. Wake Forest
  11. BYU
  12. Texas Tech
  • Level 1

    Re: My top 12 ranking. What is yours? Post and I'll tabulate.

    UGA

    UTAH

    ND

    UCLA

    MIA

    IU

    TN

    TT

    BYU

    TEX

    ALA

    ORE

  • Level 1

    Here is Truth

    GA

    ND

    MIAMI

    UCLA

    UTAH

    IU

    TX

    BYU

    WAKE

    TENN

    TEX TECH

    BOISE

    ALABAMA

  • Level 1

    Re: My top 12 ranking. What is yours? Post and I'll tabulate.

    I've revised my methodology a bit and now bring you ~BCS-Lite 2.0:

    Score = [(AP × ½ + Coaches × ½) × (Colley × ½ + SOR × ½) × (Sagarin Predictor × ½ + FPI × ½)]^(1/3)

    Rank Team AP CP COL SOR SAG FPI Score
    1 Georgia 1 1 4 4 1 1 1.587
    2 Utah 2 2 1 1 14 12 2.962
    3 Notre Dame 3 3 7 7 2 2 3.476
    4 Miami 5 6 5 3 7 3 4.791
    5 Texas 8 8 3 5 3 6 5.241
    6 UCLA 4 4 2 2 16 25 5.474
    7 Alabama 7 7 6 6 5 4 5.739
    8 Indiana 6 5 8 8 9 8 7.205
    9 Tennessee 9 9 12 10 10 11 10.130
    10 Wake Forest 10 10 11 9 11 10 10.164
    11 Ohio State 16 17 15 14 4 5 10.249
    12 Oregon 14 14 14 15 6 9 11.504
    13 Texas Tech 11 11 10 11 17 14 12
    14 BYU 12 12 13 12 21 15 13.925
    15 Iowa 13 13 17 17 13 17 14.911
    16 LSU 27 23 24 23 8 7 16.394
    17 Ole Miss 25 16 21 19 12 13 17.241
    18 USC 18 19 18 18 18 19 18.332
    19 Boise State 15 15 9 16 46 31 19.327
    20 Kentucky 17 18 16 13 30 34 20.100
    21 Louisville 21 22 25 27 20 16 21.589
    22 Florida 23 25 20 20 19 26 22.104
    23 Oklahoma 26 38 22 22 15 20 23.096
    24 Arizona 19 20 28 25 29 22 23.620
    25 Houston 20 21 23 26 37 30 25.624

    This considers three fundamentally different approaches to evaluating college football teams, each represented by two ranking systems:

    • Human judgment (AP and Coaches): Subjective assessments incorporating the collective judgment of people who follow college football, with all the benefits and biases that entails.

    • Accomplishment (Colley Matrix and ESPN Strength of Record): What has a team actually accomplished given the opponents it has faced?

    • Predictive strength (Sagarin Predictor and ESPN FPI): How good is a team, irrespective of its actual win-loss record, and how would we expect it to perform going forward?

    I use the arithmetic mean within each category because the two systems are ostensibly measuring the same thing, but the geometric mean across the three categories because they are measuring fundamentally different things. The latter gives greater credit for excellence in any one category without disproportionately penalizing a team for relative weakness in another.

    The objective isn't to eliminate subjectivity from the rankings, but rather to balance subjective judgment with two distinct forms of objective analysis through a transparent, consistent and reproducible methodology.

    • Level 2

      Along those lines, I'm fine with the committee picking five at-large teams

      Then let computer select teams 11 and 12. Had they done this last year, ND would have been #11 and Miami #12.

      • Level 3

        Re: Along those lines, I'm fine with the committee picking five at-large teams

        Even better: Let the committee rank the conference champions (and champs only) to determine which 5 get automatic bids; let a BCS-Lite style poll rank everyone and the top teams get auto-bids until 10 total auto bids are granted, then let the committee fill the remaining slots with an explicit preference (but not a mandate) for conference championship game losers, excluded conference champions, and exclusion of additional teams from any conference with 4 or more automatic qualifiers.
        THEN the committe ranks those 12 and those 12 only to seed them,

        Last season this would have given:
        Champs: JMU, Georgia, WMU, Duke, Indiana, TexTech, BoiseSt., Tulane, KennSt.
        Presumably Georgia, Indiana, TexTech, Tulane, JMU

        BCSLite (estimate): Indiana, OhSt., Georgia, Oregon, TT, ND, TAM, Bama, OleMiss, Miami, OK, Texas
        Presumably: OhSt., Oregon, ND, TAM, Bama.

        Then committee picks from: Duke, WMU, BoiseSt., KennSt., BYU
        Presumably: BYU, Duke.

        Who would have been left out? OK, Ole Miss and Miami in favor of ND, Duke, and BYU. BUT, they could have used their discretion to exclude Duke and include Miami instead, and likewise BYU for Ole Miss.

  • Level 1

    I'd say:

    1. Georgia
    2. Notre Dame
    3. Miami
    4. Utah
    5. UCLA
    6. Indiana
    7. Tennessee
    8. Texas
    9. BYU
    10. Alabama
    11. Wake Forest
    12. Oregon
    13. Texas Tech
    14. Iowa
    15. Kentucky
    16. Boise St.
    17. USC
    18. Ohio St.
    19. Arizona
    20. Louisville
    21. Houston
    22. SMU
    23. Duke
    24. Liberty

  • Level 1

    Yours is fine.

    (no message)
  • Level 1

    Computer composite of Sagarin, ESPN, Massey, Colley Matrix and Congrove says...

    1. Georgia
    2. Notre Dame
    3. Miami
    4. Alabama
    5. Texas
    6. Utah
    7. Indiana
    8. Wake Forest
    9. Ohio State
    10. Tennessee
    11. Oregon
    12. UCLA
    13. Texas Tech
    14. BYU
    15. Iowa

    This might help explain the committee's love for Alabama besides human bias (hey, as their lawyer I don't have much to work with). UCLA is the paper bruin we thought they were. What is 100% obvious is this committee is completely ignoring computer metrics and data when it comes to Notre Dame.

  • Level 1

    I agree with your selections, Mark. Well done!

    (no message)
  • Level 1

    Once Tennessee get into Alabama pod we ll be behind both

    (no message)
End of conversation