One of the keys to nailing your brackets is to correctly pick the ever-popular 12/5 upset picks. The math says that the five seed has won 65% of the time historically, which is less than both the four and six seeds. There's always at least one per year it seems, and last year, the 12 seeds won three out of their four first round matchups. Not too shabby.

Now that we know the bracket, which of this year's 12/5 games have the best chance at an upset? I will tackle each one, going from the least likely upset to the most likely.

For the following tables, I will use the following statistics to analyze the games (rank is national rank):

ORtg: points scored per 100 possessions
DRtg: points allowed per 100 possessions
Pace: average number of possessions per game
nERD: our own stat, see detailed explanation here

Note: The ORtg and DRtg ranks are not adjusted.

#5 Virginia Commonwealth versus #12 Stephen F. Austin

Without factoring specific game situations and historical comparables like we do, you can get a general idea of how well a team will compete versus another by using our nERD metric. For example, VCU has a nERD of 13.43 and SFU has a nERD of 6.68, so in theory, we would expect VCU to win by about 6.75 points in this contest. And of the four 12/5 matchups, that is the biggest spread.

But keep in mind that there are other factors that go into predicting the outcome of games, and you can find our specific predictions by looking at our game simulators, here.

As you can see by their ORtg and DRtg above, this game will feature a good offense (SFU) against a solid defense (VCU). It could get interesting if SFU can manage to slow it down and limit the possessions - on average, VCU gets five more possessions per game than SFU. If you do the math with their ORtg, that means the Rams get an extra 5.26 points (though less efficiently) than the Lumberjacks, just due to the increased pace.

If SFU can slow the pace down and play efficiently, they could definitely pull the upset. But given nERD, VCU will be favorites.

#5 Oklahoma versus #12 North Dakota State

North Dakota State is similar to Stephen F. Austin in that they have a prolific offense - ranked third nationally in efficiency - but play at a slow pace. By the math, the difference in pace is even more drastic than the VCU/SFU game, and would account for 8.83 more Oklahoma points due solely to pace. That's a huge margin, and if Oklahoma can use their fast pace (ranked 14th nationally), NDSU could be in for a long afternoon.

Where NDSU has a chance is their own offense - they have a better shot than SFU has of slowing the game down and playing their style of ball. It doesn't hurt either that Oklahoma's defense is pretty terrible and ranks 195th nationally in efficiency. OU is really banking on an up-and-down affair where their defense doesn't have to match up against Bison's offense in the half-court.

Which ever team can exert their specific style more will likely take this affair.

#5 Cincinnati versus #12 Harvard

Now this should be quite an interesting game. The numbers say that we have two top-15 defenses in the nation, and that both play at a slow pace. However, on offense, Harvard is a lot more efficient, scoring 4.2 more points per 100 possessions than the Bearcats. Their pace is almost identical, so there should be no clash of styles in this game - instead, it will be a matter of who is just better on Thursday.

This could be a close game, and the reason for this is the numbers on Harvard. Although they're a 12 seed, their numbers are that of a nine, whereas Cincinnati has the numbers of a seven seed. When you look at it that way, it's less about who is seeded where, and you can look at the fact that these two teams are pretty even.

I expect a tightly contested game - one that could come down to one of the legendary buzzer-beaters that we've come to love about March Madness. If you're like me and will have six different TVs going on Thursday, make sure that Harvard and Cincy is on one of them.

#5 Saint Louis versus #12 Xavier/North Carolina State

You get an extra team special here since we won't know whether Xavier or NC State will be taking on Saint Louis until late Tuesday night. Instead of guessing, I gave you the numbers on both teams, and you can predict for yourself who you think will win the Musketeers and Wolfpack matchup.

By the numbers, Billikens fans definitely want NC State to win on Tuesday night, as Xavier looks to be the much tougher matchup. The one area where Saint Louis would have an edge on either team is defense. In that regard, the Billikens are as good as it gets - they allow only 90.3 points per 100 possessions, which is good for fifth in the entire nation.

Offensively, none of the three teams are particularly great, with Xavier being a little bit better than NC State, but still not very impressive. Meanwhile, while the Billikens defense is top notch, their offense is, well, pretty offensive. They rank a putrid 226th in the nation in ORtg and have struggled very often on that side of the court this year. Pace wise, the Billikens are a bit faster than either team, but their terrible offense probably makes that a wash.

From a nERD perspective, we would expect a Saint Louis and Xavier matchup to be pretty much dead even. An NC State matchup, on the other hand, would be a different story, as Saint Louis would figure to be a favorite. Billiken fans will be quite invested at 9:10 PM tomorrow night watching to see who their opponent will be.

Our brackets give you a 71% higher chance of winning your pool. Want to learn more? Click here and get started for FREE.