And to illustrate how different the RPI formula is, here are the current RPIs for all B1G teams (with difference from kenpom):
10 Minnesota (+16)
24 Michigan State (+23)
31 Purdue (-17)
32 Maryland (+23)
41 Illinois (+21)
47 Northwestern (-7)
48 Wisconsin (-37)
52 Nebraska (+35)
70 Penn State (+14)
85 OSU (-32)
88 Michigan (-41)
97 Indiana (-68)
124 Rutgers (+10)
Indiana really jumps out at you--#3 in the league per kenpom; #13 in the league per RPI! Northwestern are Rutgers are the only teams with reasonably comparable ratings in both indexes.
"Race to the Middle" was a great way to describe it. Outside of Rutgers, that's exactly what we've got here.
It may be worth noting that the Sagarin numbers are actually quite similar to the Ken Pom ratings (Cats at 42 there). Which I suppose shouldn't be shocking. I think Sagarin is better for CFB and Ken Pom better at handling the relative wealth of data in CBB, but regardless they both have similar results.
http://sagarin.com/sports/cbsend.htm
And no need to rehash the RPI debate again, but I think most of us would agree that it's at least somewhat arbitrary and Ken Pom is much more accurate. I think the Committee has largely come around to that view too... while I believe they still have RPI and top 50 / 100 wins on their sheets, I suspect they are starting to look more at other metrics. Also it's worth noting that RPI will start converge a bit with the others over the course of the conference season as the level of competition (and WL % of opponents) comes up or down, relatively speaking, for those who were flattered by beating bad opponents that had decent records, or hurt by playing good opponents with bad records (it's more complicated than that, but that's a simplification of it).