Florida vs. Ohio St.
This is a very interesting matchup, if only because Florida beat them so badly last year in the title game. UCLA is a better team now than they were last year, though I'm not sure that Florida is really much better (that's not a knock on Florida, they were just playing GREAT basketball last year at the end of the Tourney).
The big advantage for Florida would seem to be down low, but I don't think it's as bad for UCLA as it seems. I think Mata/Aboya will be able to defend well against Noah 1-on-1, and they will bring a double on Horford with the other post defender. That's risky because Horford is a really good passer, so it's vitally important for the Bruins guards to stay on the outside shooters, especially Humphrey.
That's why, as I wrote a few days ago, I think it's very important that UCLA gets a lot of full-court pressure on the guards and forces Donovan to maybe take Humphrey out of the game and bring in another ballhandler. This would make the kickouts from Horford less lethal and slow down the Florida offense.
So who do I pick to win? Well, if you're a regular reader you know my pick... UCLA.
Ohio St. vs. Georgetown
I think this game will boil down to whether or not Greg Oden can stay out of foul trouble. If he can, the Buckeyes can go to a zone with him as the anchor and have a shot at slowing down the Hoyas. But if Oden does get into foul trouble, I really have a hard time seeing how they'll stop Green, Hibbert, and Ewing. With the way the last few games have gone, I'm a bit skeptical that he can stay in the game.
For the Buckeyes, they'll be in the game because of Conley. He controls the pace of the game and gets everyone else shots. He can get open shots for guys like Ron Lewis, Jamar Butler, and Daequan Cook, and they've got to hit them.
But in the end, I like GEORGETOWN here, I really think that unless Oden plays extensive minutes, the Hoyas are too good offensively. Everyone can score, everyone can handle the ball, and everyone can pass well. That makes them really tough to handle.