NCAA 2016: Sweet 16 projections

Using the Pythagorean Win % statistic from Ken Pomeroy's site, I've estimated the probability of victory for each sweet 16 team advancing further into the tournament:

Team	Pyth	Elite 8	Final 4	Finals	Champs
Kansas 1	0.9503	73.6%	45.2%	30.0%	18.2%
Maryland 5	0.8725	26.4%	9.6%	4.0%	1.5%
Miami FL 3	0.8971	37.7%	14.2%	6.8%	2.9%
Villanova 2	0.9352	62.3%	30.9%	18.5%	10.0%
Oregon 1	0.9123	59.7%	30.4%	12.9%	6.0%
Duke 4	0.8752	40.3%	16.6%	5.6%	2.1%
Texas A&M 3	0.8923	41.9%	20.2%	7.5%	3.1%
Oklahoma 2	0.9199	58.1%	32.7%	14.7%	7.1%
North Carolina 1	0.9407	64.8%	49.7%	28.8%	16.0%
Indiana 5	0.8959	35.2%	22.6%	9.9%	4.0%
Notre Dame 6	0.8131	45.5%	11.7%	3.4%	0.9%
Wisconsin 7	0.839	54.5%	16.0%	5.2%	1.6%
Virginia 1	0.9482	68.9%	53.3%	34.0%	20.1%
Iowa St. 4	0.8919	31.1%	18.9%	8.6%	3.4%
Gonzaga 11	0.8609	59.0%	18.1%	7.1%	2.4%
Syracuse 10	0.8111	41.0%	9.7%	3.0%	0.8%

However, as George Box observed, "all models are wrong, some are useful",  so maybe it's more meaningful to look at how these probabilities rank:

Team/Seed	Conference	Pyth	Champs
Virginia 1	ACC	0.9482	20.1%
Kansas 1	B12	0.9503	18.2%
North Carolina 1	ACC	0.9407	16.0%
Villanova 2	BE	0.9352	10.0%
Oklahoma 2	B12	0.9199	7.1%
Oregon 1	P12	0.9123	6.0%
Indiana 5	B10	0.8959	4.0%
Iowa St. 4	B12	0.8919	3.4%
Texas A&M 3	SEC	0.8923	3.1%
Miami FL 3	ACC	0.8971	2.9%
Gonzaga 11	WCC	0.8609	2.4%
Duke 4	ACC	0.8752	2.1%
Wisconsin 7	B10	0.839	1.6%
Maryland 5	B10	0.8725	1.5%
Notre Dame 6	ACC	0.8131	0.9%
Syracuse 10	ACC	0.8111	0.8%

With the record number of ACC teams, it might be interesting to compute the probability of an ACC champ.  I've been using a lot of complement logic lately, so may as well stick with the trend:

P(no ACC win) = (1-.201)*(1-.160)*(1-.029)*(1-.021)*(1-.009)*(1-.008) = 0.628

Therefore,  P(ACC win) = 1 - P(no ACC win) = 1 - 0.628 = 0.372.

So Ken Pomeroy's model of win % estimates the probability of an ACC champ at 37.2%, where most of that probability is wrapped up in either Virginia or North Carolina (32.8% to be exact).