Power Rankings?

Viewing 4 reply threads

- 09/17/2017 at 3:26 PM #41524
  5
  
  shakeitsugaree
  Ute Fan
  
  @shakeitsugaree
  
  Stupid, or stupid awesome?
  
  What’s a good win? What’s a bad loss? Can we quantify these things to give us a methodology to rankings, so that I don’t want to puke every time USC gets ranked #1 just because of their pretty five star recruits?
  
  Here’s a thing:
  
  I ranked Pac12 teams by record; all 3-0 teams are at the top, while teams with a conference win are ahead of those that have not yet played a conference opponent. The 2-1 teams are next, followed by the 1-2 teams and bringing up the rear is the Beaver punching bag at 1-3. I then resolved the ties using this formula to evaluate this week’s game:
  
  – W=5, L=0
  
  – road game=2, home=0
  
  – opponent’s 2016 record (this is included because there haven’t been enough games played to evaluate teams on
  
  their 2017 record, and will be dropped either this coming week or the next): 13-9 wins=3, 8 & 7 wins=2, 6
  
  wins=1, <6 wins = 0
  
  – OW% 2017 (this will replace 2016 record, but is still considered this week): 100=3, 67=2, 50=1, <50=0
  
  – Conference: P5=2, G5=1, FCS=0
  
  – Margin of victory (this scale is also applied in reverse for losses): >21=2, 21-14=1, 13-7=0.5, 6-4=0.25, <4=0
  
  – 2017 wins: # of wins on the season
  
  – Bonus: rivalry=1, shutout=1
  
  This gives us these rankings:
  
  1) WSU (3-0, 1-0, 11 pts in my ranking system)
  
  2) USC (3-0, 1-0, 10 pts)
  
  3) Oregon (3-0, 0-0, 15 pts)
  
  4) Cal (3-0, 0-0, 12.5 pts)
  
  5) Colorado (3-0, 0-0, 12 pts)
  
  6) UW (3-0, 0-0, 11 pts)
  
  7) Utah (3-0, 0-0, 10 pts)
  
  8) Zona (2-1, 0-0, 11 pts)
  
  9) UCLA (2-1, 0-0, 10 pts)
  
  10) ASU (1-2, 0-0, 7.5 pts)
  
  11) Furd (1-2, 0-1, 10 pts)
  
  12) Beavs (1-3, 0-1, 8 pts)
  
  What say you?
- 09/17/2017 at 3:30 PM #41526
  2
  
  Utah
  Ute Fan
  
  @utah
  
  Honestly, your rankings really seem to really pass the eye test. I like them. Keep doing this all season.
- 09/17/2017 at 3:33 PM #41527
  1
  
  ladyinred
  Ute Fan
  
  @ladyinred
  
  I like your system, and honestly I seems like a fair assessment for Utah given the body of work so far. I don’t think cal and wazzu will fare as well in PAC 12 play, but have to admit their wins so far are good, especially cal.
  
  I am also highly unimpressed by UW’s ooc schedule, so dropping them that low seems correct to me.
- 09/17/2017 at 4:14 PM #41533
  
  Utahute72
  Ute Fan
  
  @utahute72
  
  I’m not sure how you are applying those numbers. Are the points multiplicative or additive? If they are multiplied to get a result I would probably use fractions to downgrade the results rather than zeros. I like that winning is critical, and any system where the key is winning is positive (not really a fan of quality losses), but I do think teams need to be given credit for winning against good teams.
  - 09/17/2017 at 7:14 PM #41553
    
    shakeitsugaree
    Ute Fan
    
    @shakeitsugaree
    
    Points are additive, and the weighting system is purposeful – meaning, I tried to distribute the points based on what I consider to be most valuable (the ‘win’ is worth the most, against a good team is next, conference affiliation is worth less than the first two factors, etc). Here’s a breakdown of how I calculated the points:
    
    Oregon = 5(win) + 2(road) + 2(Wyoming 8-6 in 2016) + 1(Wyoming 1-1 going into the game) + 1(Wyoming G5) + 2(+36 margin of victory) + 2(Oregon 2-0) = 15
    
    Oregon State = 0(loss) + 2(road) + 2(WSU 8-5 in 2016) + 3(WSU 2-0 going into the game) + 2(WSU P5) – 2(-29 loss margin) + 1(OSU 1-2) = 8
    - 09/18/2017 at 10:52 AM #41601
      
      Utahute72
      Ute Fan
      
      @utahute72
      
      I’m still not sure how you get to the numbers you have. On the surface, if Cal has three wins the lowest rating they could have over three games is 15. Unless you are doing an average or a by weekly matching.
      - 09/18/2017 at 1:40 PM #41628
        
        shakeitsugaree
        Ute Fan
        
        @shakeitsugaree
        
        Maybe I have been unclear about the process. The points are only evaluating the most recent game, and are used to evaluate the ‘strength’ of a win or ‘severity’ of a loss. So:
        
        1) rank teams by record
        2) resolve ties by ‘strength’ of win this week
        
        Cal is tied at second because they have three wins, no losses and no conference wins or losses:
        
        3-0, 1-0: WSU & USC
        3-0, 0-0: Oregon, Cal, UW, Colorado, Utah
        2-1, 0-0: Zona & UCLA
        1-2, 0-0: ASU
        1-2, 0-1: Furd
        
        To evaluate which team from the ‘second place’ group should be first, we look at the points system. Oregon earned the most points, then Cal, Colorado and finally Utah. Here’s a breakdown of Cal’s points this week:
        
        Win 5 pts + At home 0 pts + Ole Miss 5-7 in 2016 0 pts + Ole Miss 2-0 going into the game 3 pts + Ole Miss is in the SEC, P5 conference 2 pts + Margin of victory 11 pts 0.5 pt + Cal 2-0 goint into the game 2 pts
        
        5+0+0+3+2+0.5+2=12.5
        
        Or, maybe I’m not understanding your confusion?
      - 09/18/2017 at 1:48 PM #41630
        
        shakeitsugaree
        Ute Fan
        
        @shakeitsugaree
        
        OK, I think I understand where you are going. You are adding points across weeks, which I wasn’t doing. However, I think it’s a good idea – I will do that next week.
- 09/17/2017 at 5:37 PM #41539
  
  EagleMountainUte
  Ute Fan
  
  @battlegroundute
  
  Oregon is number one then?? I am confused on the point system yet you ranked them differently.
  - 09/17/2017 at 7:19 PM #41556
    
    shakeitsugaree
    Ute Fan
    
    @shakeitsugaree
    
    So, I ranked teams first on their overall win-loss records, then on their conference win-loss records, which gave me a bunch of ties:
    
    3-0, 1-0: WSU & USC
    3-0, 0-0: Oregon, Cal, UW, Colorado, Utah
    2-1, 0-0: Zona & UCLA
    1-2, 0-0: ASU
    1-2, 0-1: Furd
    1-3, 0-1: Beavs
    
    Then, I used the points system to evaluate each team’s performance in this week’s game. I used the resultant scores to resolve the ties.
    - 09/17/2017 at 9:02 PM #41569
      
      EagleMountainUte
      Ute Fan
      
      @battlegroundute
      
      I see that now I didn’t read so good on the first go.