Originally Posted by hawkoooo
I think what you're saying is that in the first few years they could be off using their standard formula IF the committee picks teams they wouldn't have based on the switch from RPI to NET (that happens regardless because they go against historical trends and Dance Card gets some wrong, but it could happen MORE).
Sure, that could happen, but over time that will correct itself as the formula takes in more data where the committee leans on the NET. They really don't have to change the formula or the use of the RPI, as they can still measure and do the formula using historical trends against the RPI.
I'm still probably not explaining this right, but someone smarter than me might.
|
You are right but typically you don’t switch cold turkey unless you have some specific metrics you can validate against (i.e more P5 at large bids)
What you more typically want to do is evaluate how the changes will affect the entire population before cutting over to it - thus testing against a smaller challenger sample set - say 10% of the entire universe.
Since you can’t do that here, what makes sense to do is pick the at large teams with the new and old methods and them look only at teams that either are in one but. It the other or teams that had significant seeding changes. In an over simplified way, what you want to ensure is that teams affected by the switch are because of flaws in RPI not in NET. Without knowing the affects, it is much harder to measure that.
That will allow better analysis of the real impact NET has on what is trying to be done with other switch. I guarantee if the new method had significantly less P5 teams in the final field this year they wouldn’t just go with it and let it correct itself over time.
Posted via Mobile Device