We do it every year, and it fails every year but we do a statistical analysis called a profile analysis where we create a winner and loser profile for each round based on each teams stats from all the teams that have been in the tournament over the past ten years and use different decision rules in case both teams are rated as winners or losers.
In the first one, the one that is traditionally more conservative, it was conservative in the first round only a couple of upsets. However, the rest of the way it went kinda crazy. Here is the final four:
Villanova Gonzaga, Oklahoma State (!), North Carolina, Final: Villanova over Oklahoma State (!)

The only explanation we can see is that OK St scores a lot of points, has a good 3 point % and good FT %.

The second one is more liberal and shows another problem with our 'system':
Final four: Villanova, Gonzaga, Oregon, North Carolina; Final: Villanova over Oregon

The fact that when doing this we cannot take injury into account allows Oregon to get this far.

In other words, do not pick Villanova because we never win.

_________________

Gonzaga ain't getting past the Cats!

_________________
O judgment! Thou art fled to brutish beasts,
And men have lost their reason.

One of the trends was that the stats didn't like the Pac 10, Other than Oregon's run, every other pac 12 team was losing when they were the underdog. The Big 12 underdogs were the ones to pay attention too.

_________________