Now that the bracket has been revealed for the 2013 NCAA Tournament, everyone wants to know: where will the upsets come from this year?  March Madness isn’t mad without top teams going home early.

I’ve studied data from the NCAA tournaments of the last 10 years to find trends that might help identify the best candidates for 1st-round upsets this year.  It turns out that key stats to look at when examining a match-up are; how frequently a team forces turnovers, how well they rebound on offense, and how much the underdog relies on 3-pointers.  However, the most telling stat of all is Efficiency Margin.

I’ve already looked at the correlation between Efficiency Margin and NCAA tournament success for teams that excel at this metric, but beyond overall tourney success, EM can also be used to predict 1st round match-up winners.  When the higher-ranked team also had the higher EM, they won 82% of the time.  But when the higher-ranked team had the lower EM, they only won 59% of the time.

Re-running this analysis to exclude 1 vs. 16 and 8 vs. 9 games dulls the significance of EM a bit, but the result is similar: high-seeded teams win 80% of the time when they bring the higher efficiency margin and only 61% of the time they don’t.

So to identify potential first round upsets this season, let’s start with the games in which the higher-seeded team has a lower EM than their underdog opponent (EM in parenthesis):

• #6 Memphis (15.2) vs. #11 Middle Tenn. St. (16.0) or St. Mary’s (16.0)
• #3 Michigan St. (13.9) vs. #14 Valparaiso (14.4)
• #6 Butler (8.4) vs. #11 Bucknell (15.6)
• #3 Marquette (9.6) vs. #14 Davidson (11.5)
• #7 Illinois (5.8) vs. #10 Colorado (6.6)
• #6 UCLA (8.2) vs. #11 Minnesota (10.7)
• #6 Arizona (14.2) vs. #11 Belmont (17.5)
• #7 Notre Dame (11.3) vs. #10 Iowa St. (11.8)

Interestingly, every single 6-seed makes this list, along with a pair of 7-seeds and 3-seeds.

After identifying these match-ups as possible upsets, I looked deeper to rank them in order of likelihood to occur (see the raw data for all 1st-round match ups):

1. Arizona vs. Belmont: Belmont fits the mold of classic upsetter.  They force turnovers at a high rate (11th in the nation), rely heavily on 3-pt FGs, and shoot well from the free throw line, all indicators of upset potential.  On the other side, Arizona is an average ball-handling team, turning the ball over on almost 20% of their possessions, and allows opponents to make 36% of their 3-pointers, good for 277th nationally.  This game has the ingredients for an upset (and bonus Sweet 16 alert – Belmont very could well shoot their way past New Mexico in the second round).

2. Marquette vs. Davidson: Davidson protects the ball well and relies heavily on threes, making them dangerous should the shots start falling.  Rebounding will be the key in this matchup, as both teams figure to go for Marquette’s misses (MU is 25th in offensive rebounding, Davidson is 37th defensively) while seemingly ignoring Davidson’s (265th offensively to Marquette’s 243rd defensively).  Also of note – Davidson is the nation’s top free throw shooting team.

3. UCLA vs. Minnesota: Minnesota is the nation’s top offensive rebounding team while UCLA ranks in the 200′s in rebounding on both ends of the court.  Neither team is great at forcing turnovers, placing additional importance on pulling in missed shots, giving Minnesota the advantage in this one.

4. Notre Dame vs. Iowa St.: This game is basically a toss-up.  Statistically, these teams are similar with one exception – Iowa St. shoots a lot more threes (9th nationally in 3-pt reliance).  If the threes are falling, Iowa St. will come out on top.  If not, Notre Dame will advance to take on (probably) Ohio State.

5. Butler vs. Bucknell: Bucknell protects the ball tremendously well, placing 2nd nationally in not turning the ball over.  This will be important if they are to pull off the upset, because they don’t figure to get many second chances (Butler is 13th in defensive rebounds, while Bucknell is 249th in offensive boards).  Butler is just an average free throw shooting team, which could doom them late, but overall, it may be tough for Bucknell to send the Bulldogs packing.

6. Illinois vs. Colorado: These teams are just about even in many respects, but Illinois’ reliance on the three may make them susceptible to an upset if the shots aren’t falling.  Colorado meanwhile, isn’t particularly adept at protecting the ball or forcing turnovers, so they aren’t a huge threat to do significant tournament damage.

7. Michigan St. vs Valparaiso: Despite the high EM, Valparaiso doesn’t register high on the upset meter.  They don’t force many turnovers (207th nationally), though they do rely heavily on 3-pointers.  If their shooters get hot at the right time, Michigan St. might be in trouble, as they’d be prevented from asserting their dominance on the glass (22nd overall).

8. Memphis vs. MTSU/St. Mary’s: This one is last because we don’t yet know who Memphis will face.  The Tigers are susceptible to the upset bug based on their inability to take care of the ball (226th in Offensive Turnover Rate) and their poor free throw shooting (258th in the nation).  A late lead may not be safe as turnovers and missed free throws are a sure way to let your opponent back in the game.  As for their potential opponent, St. Mary’s presents the most dangerous foe.  They rely heavily on 3-pt FGs and rebound well (5th in the country in Rebound %), both indicative of upset potential.  Middle Tennessee, on the other hand, forces more turnovers than either Memphis or St. Mary’s.

So there you have it.  Starting with efficiency margin, we can use statistics to identify potential 1st-round upsets.  Certainly, there is upset potential among match-ups not mentioned in this article, but if you need to identify the single most likely upset (such as in a survival-style tournament pool), I would choose from among those listed above.

Which 1st-round match-up do you think is most likely to produce an upset?