March Madness: Ranking the Most Likely 12/5 Upsets
The 12-5 matchups in the NCAA Tournament are always the public’s favorite to pick upsets. It makes sense -– statistically, over the last 30 years or so, the numbers are very favorable for the numbers 1 through 4 seeds, going from 100% win percentage for 1 seeds to about 80% for 4 seeds in the first round. However, the 5-12 matchup is when the statistics dip significantly -– the 5 seeds have won about 65% of the time, which is about the same percentage as 6 seeds and only slightly higher than 7 seeds.
As you can see, picking the correct 12-5 upsets are key to having a winning bracket. Thankfully, our high-powered computers here at numberFire have already done the hard work, and we can look at good ol' math to gain an edge in picking which 12 seeds can pull the upset on Thursday and Friday.
A quick statistical key and we'll start, beginning with the least likely upset:
ORtg: Points scored per 100 possessions (using adjusted)
DRtg: Points allowed per 100 possessions (using adjusted)
Tempo: Possessions per 40 minutes
nERD: Our own metric, measuring the number of points a team would be expected to win by against an average opponent on a neutral court.
(5) Arkansas versus (12) Wofford
It's pretty fascinating how the brackets shaped up this year regarding the 12-5 matches. In terms of nERD, both Arkansas and Northern Iowa were much worse than the other two 5 seeds (in comparison to Utah, they were much lower, but we'll get to that in a bit), but they also got easily the two worst 12 seeds in the tournament.
Wofford is outside the top 75 in both offensive and defensive efficiency, and they also play at a very slow pace. Meanwhile, Arkansas is one of the fastest teams in the country, led by sophomore stud Bobby Portis, who, despite all of the 35 future NBA players that play for Kentucky, took home SEC Player of the Year honors this season.
When simulating this game, our algorithms didn't see many ways Wofford could keep up with Arkansas. We simulate a win percentage for the Razorbacks of 87.69%, easily the highest of all the 12-5 games. Wofford will have to slow the game down and try to frustrate Arkansas by limiting possessions, but the chances of that succeeding are quite slim.
(5) Northern Iowa versus (12) Wyoming
Northern Iowa is also lower in our power rankings than Utah and West Virginia, but they also get an easier matchup in Wyoming, who stole a bid by beating San Diego State in the Mountain West Tournament to get an automatic selection.
Part of March Madness is watching as many games at the same time with multiple TVs set up. Sometimes you have to make sacrifices though and put a game on the smaller TV -- this is that game. Just take a look at the tempo of these two teams. Both rank at the bottom of the nation in terms of pace, and will likely both try use that to their advantage.
For that reason, I wouldn't be surprised if Wyoming hung in the game. In general, the more possessions you have in a game, the more likely it is for the better team to come out on top -- they have more opportunities to take advantage of their superiority, in a sense. So while Northern Iowa is certainly the better team, you never know what will happen when it comes to low-scoring affairs.
(5) Utah versus (12) Stephen F. Austin
Both Utah and Stephen F. Austin boast good offenses, ranking in the top 20 in the nation. However, the difference between the two teams is on the defensive side of the ball. Utah is even better on that end -- eighth in the nation in defensive efficiency -- while Stephen F. Austin is outside of the top 100.
Our numbers are actually really high on Utah this year. We have them as our eighth-best team overall with a 16.75 nERD, which would put them right on par with the 2 seeds of the tournament. They were incredibly underrated this season.
In fact, I found that over 85% of teams to make the Final Four have finished the season with a top-10 Pythagorean Rating (Pythagorean Rating courtesy of kenpom.com). Utah, despite being a 5 seed, is one of those teams. They do get one of the tougher 12 seeds in Stephen F. Austin -- their nERD of 9.03 is way higher than Wofford or Wyoming's -- but Utah is in another class this year. If they can get past first-round jitters and avoid losing to the Lumberjacks, they could make a long run in the tournament.
(5) West Virginia versus (12) Buffalo
And finally we get to our most likely 12-5 upset this year. Both of these teams like to play fast, ranking in the top 30 nationally in terms of possessions per 40 minutes. And they both have top 50 offenses, with a slight edge going to West Virginia.
Our algorithms still have West Virginia winning about 75% of the time, but that's much lower than the other 5 seeds. Buffalo is coming into the tournament after winning their conference tournament over Central Michigan, while West Virginia lost in the quarterfinals of the Big 12 conference tournament to now 3-seeded Baylor.
Despite this game being the most likely upset of the 12-5 games, it's still not likely. Really, the selection committee did a good job in this regard -- last year the Saint Louis and Cincinnati games were much closer in terms of nERD difference between the 5 and 12 seeds. There could still be upsets for sure, but if there was a year to lean more on the 5 seeds and go contrarian, this is definitely the one to do it.