I pick based on “feel”, usually… it tends to be a bit hit and miss, because my inherent bias gets involved too often. I’ve done money pools since… 1999, I think? So this will be my eighth year. Of the past seven, I’ve finished in the money (top 3) three times, but the other four times I’ve done pretty abysmally (bottom 25%). That’s a bit less impressive because two of the three times I’ve done well, UConn won it all as a 1 seed where I had seen almost all of their games - so I was able to make a homer pick that I had a really good feel for. I tend to think that (at least in pools with no upset bonuses), how many you get in the first round doesn’t really matter as long as your deeper picks are still in it. Getting three of the final four, and both of the final two, is worth a lot more than picking a couple extra upsets correctly in the first round. That said, the very first year I won my money pool, the margin of victory was my pick of some 10 seed beating Stanford and going all the way to the regional final, that no one else picked. Oh yeah, Gonzaga 
For the opposite example, my amazing bracket last year where I finished with like 12% in Yahoo-scoring. I did pretty much standard for my first-round picks (75% correct or so)… but I only got like 7 of the Sweet 16, and one of the final four. “Feel” didn’t help nearly so much, especially when I was too busy with life to stay up on some of the better teams, never mind the underdogs.
The other thing is I like to take a champion pick outside the “mainstream”. I have UConn this year in the SDMB pool, but in my money pool I have 'Nova. There’s nothing worse than running a pretty good bracket, sitting just outside the money going into the final four… and having no chance of moving up because the entire top ten has Duke and UConn in the final in some permutation.
Also, this year, I had the time to follow a LOT more non-Big-East stuff than usual… so theoretically I have a better “feel” for it. We’ll see if that actually helps or not.
As for the Big East going 0-3 yesterday, Marquette is the only one of those losses that really surprises me. Everyone knew Seton Hall was perfectly capable of laying down a bad one against a decent team, and most of us though Cinci would have been a better pick in that spot anyways… and though I certainly WANTED to believe in Syracuse, there were plenty of visions of McNamara against Vermont last year in the back of my mind (after a similar, though less miraculous, Big East run). Marquette, though, was the type of team that could have caused some trouble, and Alabama didn’t look to be the kind of team that could keep up with them. To 'bama’s credit, though, they played one of their absolute best games of the year, and kept Marquette’s “miracle” opportunities down to just that one shot (when early tourney losses are so often the result of letting the trailing team have multiple shots at that last three-pointer to tie or win).
In the Friday afternoon games, the Big Ten is having some first-half issues: Arizona is absolutely running it up on Wisconsin, Ohio State is losing to Davidson at halftime, and NW St is keeping it close to Iowa. I imagine all of those except for the Badgers will work themselves out, but who knows.
My average bracket is going to have something like 12 correct today… give or take, since none of my pools are exactly the same.