Wook
11-09-2008, 05:28 AM
Been mulling this over. Even if a playoff system were enacted there would have to be some sort of selection comittee to identify the 8, 16, or 32 best teams. So how to do that has been floating around in the back of my head. I hate the BCS methodology as it helps to encourage blowout style poor sportsmanship and scheduling raggmuffin tune up games. I figure a proper rankings system needs to account for the following:
1. Winning - The teams that win the most, in the 2 losses or less category, should score out towards the top of the rankings consistently.
2. Strength of schedule - Not power division vs. victim division, that's another factor that's separate, but simply how good your opponents were in absolute terms.
4. Power of schedule - This is the recognition that a MAC team isn't a Big 10, SEC, or Pac 10 team.
From this you should get a pretty clear picture which teams are the best in the country IMHO. Here is a first pass at the numbers:
Winning:
10 pts per win
+5 pts per road win
-10 pts per loss
-5 pts per home loss
Power:
+1 per week that an opponent receives vote in the AP or Coaches poll.
+2 per week that an opponet is ranked in the Top 25.
+3 per week than an opponent is ranked in the top 10.
Caveat one: You can not receive this bonus more times for any given team than games your team played during the season.
Caveat two: This is a compound bonus. So a top 25 team is worth 3 pts/wk and a top 10 team is worth 6 pts/wk.
One time bonus: +5 per win against a ranked opponent
One time penalty: -5 per loss to an unranked opponent
Strength:
Final value at end of season is adjusted by opponents total combined winning percentage as a multiplier.
Note that there is nothing about margins of victory or holding down the score in there. How big you win does not impact you standing. Playing a bunch of cremepuffs hurts you in the long run. In some ways more than losing does. Playing solid opponents boosts your standing.
All this is intended to do is give a mathematical model to compare relative merit between teams. After this you'd want a selection committee to be making informed choices based on the particulars of the games and teams themselves since it's inevitable that there will be to close to call measures by this mechanic.
1. Winning - The teams that win the most, in the 2 losses or less category, should score out towards the top of the rankings consistently.
2. Strength of schedule - Not power division vs. victim division, that's another factor that's separate, but simply how good your opponents were in absolute terms.
4. Power of schedule - This is the recognition that a MAC team isn't a Big 10, SEC, or Pac 10 team.
From this you should get a pretty clear picture which teams are the best in the country IMHO. Here is a first pass at the numbers:
Winning:
10 pts per win
+5 pts per road win
-10 pts per loss
-5 pts per home loss
Power:
+1 per week that an opponent receives vote in the AP or Coaches poll.
+2 per week that an opponet is ranked in the Top 25.
+3 per week than an opponent is ranked in the top 10.
Caveat one: You can not receive this bonus more times for any given team than games your team played during the season.
Caveat two: This is a compound bonus. So a top 25 team is worth 3 pts/wk and a top 10 team is worth 6 pts/wk.
One time bonus: +5 per win against a ranked opponent
One time penalty: -5 per loss to an unranked opponent
Strength:
Final value at end of season is adjusted by opponents total combined winning percentage as a multiplier.
Note that there is nothing about margins of victory or holding down the score in there. How big you win does not impact you standing. Playing a bunch of cremepuffs hurts you in the long run. In some ways more than losing does. Playing solid opponents boosts your standing.
All this is intended to do is give a mathematical model to compare relative merit between teams. After this you'd want a selection committee to be making informed choices based on the particulars of the games and teams themselves since it's inevitable that there will be to close to call measures by this mechanic.