Here’s the updated brackets heading in to Friday’s games:
Last Four In: Providence, Kansas State, Rhode Island, Illinois State
First Four Out: USC, Illinois, Syracuse, Iowa
Notes: There’s a lot of movement in the bottom 1/3 of the at-large pool, mostly because the lines between these teams are razor-thin. Iowa and Illinois both had chances to move up (or in) to the field, but bad losses in the Big Ten Tournament hurt both of their causes.
Will the committee reward large win totals with a lack of impressive top-end victories, or will they reward big wins and bad metrics? That’s the big question right now. I’m wavering back and forth between those two realities, but right now I’m guessing they’ll reward the larger win totals. Bad as the RPI is, it’s what the committee uses, and the RPI is favoring a team like Illinois State over a team like USC.
There were some questions asked on Twitter:
Interesting question. From a competition standpoint, the answer is a no-brainer YES. But even if you went straight s-curve, there are other balancing issues you have to take into account, like keeping conference members apart from each other in the bracket. If you were to also do away with those restrictions, then I’d say go ahead and switch to straight s-curving. The regional orientation is a vestige of an era when plane travel wasn’t as easy as it is today, and while the current system does benefit some higher seeded teams by not sending them across the country, it disadvantages teams that are maybe one or two arbitrarily decided places below them.
SMU, St. Mary’s, Wichita State…lots of teams fit into this category of having a small number of top 50 wins but a gaudy overall record. SMU is in better shape than the other two teams I mentioned, with eight top-100 wins and 9 road/neutral wins. The committee has rewarded teams for regular season conference titles before, and I think that will happen here as well. Right now I have them in the 6/7 seed range, but wouldn’t be shocked if they ended up on the 5 line. I can’t see them rising any higher than that, though.