Re: 2016 Pairwise Calculator
Is that in your projections or with actual results to date?
It's actual results to date.
Worth noting, by the way --
Minnesota's RPI before removing bad wins is .6364
Wisconsin's RPI before removing bad wins is .6281.
If you throw out all those games, then the only thing that prevents Wisconsin and Minnesota from being identical is that a) they played each other; b) we are including two extra wins for UM over Yale, and one more win over St. Cloud State.
The Yale games are having a minimal effect on Minnesota's RPI. Minnesota's RPI before QWB is .6493. The wins against Yale are only counting for .6496, virtually identical and a couple extraneous results from being dropped from the RPI entirely.
The extra win against SCSU is counting for .6624. Helping a bit, yes, but it's only one game.
What's really driving this is the UM/UW head to head.
UM's "Game RPIs" for the games against Wisco are:
2 W vs. UW = .7500 -- (Wisco's W% = .9531,
OW% = .4810)
2 L vs. UW = .4500
UW's "Game RPIs" for the games against Wisco are:
2 W vs. UM = .7668 (Minny's W% = .9516,
OW% = .5184)
2 L vs. UM = .4668
So -- and if you weren't annoyed already, you will be now -- the reason that UW is ranked higher than UM right now is actually
because Minnesota has played tougher competition. Because so many games are thrown out, Minnesota's not getting the benefit of its own higher OW%, but Wisconsin
is getting credit for Minnesota's high OW% in the H2H games, bolded above. So paradoxically, Wisconsin gets credit for Minnesota's tougher schedule and Minnesota doesn't.
Amazing.
To prove our theory is correct, if we take out the 4 H2H games in there, Wisconsin should no longer benefit from Minnesota's tougher OW% and should fall behind Minnesota even with removing bad wins -- and they do:
UM's RPI after deleting the 4 games vs. UW: .6716
UW's RPI after deleting the 4 games vs. UM: .6707