Premise: As we know the polls are largely based on a merit system and pecking order determined by a preseason poll. This is akin to a ladder system where you win and move up, you lose you drop down a random position. These polls remain largely static and while the end result usually results in the correct order, they are by and large flawed.
Hypothesis: Polls should be FLUID INSTEAD OF STATIC. Meaning they should be BASED ON RESULTS UP TO THAT POINT in the season rather than future projections or biased preseason rankings. All wins and losses are not created equal and need to be accounted for.
Solution: Given the current BCS parameters, a mathematical formula can be determined to resolve this issue. Criteria should be sorted on a weighted basis with priority for more substantial factors over less substantial factors. Historical data can be used as evidence for this criteria.
Thus, we derive the equations for proper poll weighting:
TW= Total wins- C wins
Zero L> 1 L..., where L stands for loss.
2 L = Null set.
C= 0, where C stands for non BCS conference teams (specifically C-USA, MAC, MWC, Sun Belt, WAC, Non-ND independents)*
RW>HW>RL>HL, this is the essence of the formula where road wins are greater than home wins which are greater than road losses which are greater than home losses. In using this, we can establish a proper hierarchy or mathematically accurate poll.
*C losses negate 2TW thus having a double loss effect on the data.
Once again, please note this only measures the PERFORMANCE TO DATE of the teams. We are not in the business of forecasting the future (then again the polls should not be as well) and feel the data will sort itself out over time.
In the next post I will post the mathematical poll along with the appropriate TW-RW-HW-RL-HL for each team so as to provide some reasoning.
Comments and questions are welcome but please keep it on topic and with some reasoning.
Hypothesis: Polls should be FLUID INSTEAD OF STATIC. Meaning they should be BASED ON RESULTS UP TO THAT POINT in the season rather than future projections or biased preseason rankings. All wins and losses are not created equal and need to be accounted for.
Solution: Given the current BCS parameters, a mathematical formula can be determined to resolve this issue. Criteria should be sorted on a weighted basis with priority for more substantial factors over less substantial factors. Historical data can be used as evidence for this criteria.
Thus, we derive the equations for proper poll weighting:
TW= Total wins- C wins
Zero L> 1 L..., where L stands for loss.
2 L = Null set.
C= 0, where C stands for non BCS conference teams (specifically C-USA, MAC, MWC, Sun Belt, WAC, Non-ND independents)*
RW>HW>RL>HL, this is the essence of the formula where road wins are greater than home wins which are greater than road losses which are greater than home losses. In using this, we can establish a proper hierarchy or mathematically accurate poll.
*C losses negate 2TW thus having a double loss effect on the data.
Once again, please note this only measures the PERFORMANCE TO DATE of the teams. We are not in the business of forecasting the future (then again the polls should not be as well) and feel the data will sort itself out over time.
In the next post I will post the mathematical poll along with the appropriate TW-RW-HW-RL-HL for each team so as to provide some reasoning.
Comments and questions are welcome but please keep it on topic and with some reasoning.
