Re: Regional Rankings - Part III
IMO, Champs, SOS and win % are easily (and roughly equally) the two most salient factors in this statistical comparison. RNK has a cliff, which you never want in a statistical model, and COP is already accounted-for via win%/SOS. L25 is just silly. I can see using H2H when two teams are in a virtual tie, but H2H data is generally too sparse to be given much weight in the context of an entire season.
Somewhat agree to an extent, however Win % isn't as pure as you want to believe. I invite you to go over and read some of the banter that Webb and others were discussing in regards to the SOS and the advantage playing a 3-game series conference has with regards to elevated SOS. With that being said, you make "somewhat" valid points when dealing with the other criteriea, the issue is that no ONE criteria works alone. When you look at all 5 criteria together it makes much more sense. Team A has a high winning %. Okay who did they play? Well they have a decent SOS. Okay but once again who did they play? Well they played good teams, but they also played some really good (ranked) teams and did this. Okay, that's great and all, but Team B didn't play those teams, but they did play these teams (COP). How did they do against common opponents? IMO unless there is substantial data (more than "x" amount of games played), the criteria shouldn't be looked at. For example: Endicott 1-0-0 RNK....is that better then a 6-1-2?? Criteria says yes. Common opponents: is 2-0-0 better than 5-1-1? Criteria says yes. RNK : is 2-0-1 (but 2 of the 3 games vs #10 RNK and a tie vs #9, better then 5-2-0 but with wins over #1, #3 x2, #5x2, and a loss to #2 and #4? Once again we come back to SOS: is a 17-4-2 NEHC school who's played 8-9 games vs ranked teams better then a 19-2-2 MASCAC school who hasn't played one? Yes you go to OOWP but even the MASCAC schools play ECACNE teams who can have high winning%.
There is just to many questions to strictly go by SOS and Win%, and while I agree somewhat that the other criteria either run into each other or can make little to no sense, when you look at everything as a whole, it makes a little sense.
Honestly I wish we didn't have 60+ teams and we were not limited to a 25 game schedule so more teams couple play one another. Then a KRACH or RPI ranking would be much easier. Too many teams play too many different styles of schedules, both conference and non-conference.
Said it a million times, but the AQ's are the biggest wrench in these gears. Not throwing you a bone here, but if Platty doesn't get a berth this season (for instance) it'll be a crime against reason.
Again, somewhat agree, but only because unlike DI bouncyball 11 spots with 7 auto bids is just too small to get a lot of well deserving teams. Having said that, while it would be nice to see more Pool C bids awarded, only a couple times over the past 2 decades has a team receiving a Pool C bid won the title. Even more rare is 2 Pool C teams making it to the Title Game. And again, while having an 11 team tournament with no autobids would be cute and all, there are just too many teams who can maneuver their schedule to beat the math system with out going into really big detailed fair math solutions. (How everyone is up in arms about Top FBS schools playing cupcakes and padding their win totals)
