All scores →
Loading…
Loading…
Loading…

The 12-5 upset: history, odds and the pattern

By 4Basketball Staff · · 11 min read

Why 12 seeds beat 5 seeds about a third of the time, how the matchup became the most famous first-round upset, which 12 seeds are worth picking, and how the odds work out across four games.

If you ask a group of basketball fans which first-round upset they would circle on a bracket, most will say the same thing before you finish the question: a 12 over a 5. It has become the signature early upset of March, the one that shows up in newspaper columns, television segments and office-pool advice, the one that people pick even when they pick nothing else. The 5 seed is usually a proud team from a major conference. The 12 seed is usually a conference champion nobody outside its region has watched. Somewhere around the second half of the first Thursday, the two meet, and about a third of the time the 12 wins.

The popularity of the pick can make it seem like folklore, something fans repeat because it sounds wise. But the 12-5 upset is one of the best-supported patterns in the tournament, and it has a clear explanation. It is not luck, and it is not a trick of the numbers. It comes from how the committee builds the bracket, from the type of team that ends up on each of these seed lines, and from the arithmetic of a single game. All of that can be laid out in plain terms.

This article gives the historical record, explains why the matchup is so rich in upsets, tells the stories of some of the best 12 seeds, and then turns to the practical question: how many of these upsets should you pick, and which ones? It builds on the general explanation in why upsets happen, and it is a companion to the broader treatment of bracket-picking strategy.

The record

Since the tournament settled on a 64-team bracket in 1985, the 12 seeds have beaten the 5 seeds in a little more than a third of their meetings. That is the figure that has drawn all the attention. It is far higher than the rate for the matchups on either side of it. To put the 12-5 game in context, here are the approximate win rates of the higher seed in each first-round pairing across the 64-team era.

Matchup Higher seed wins (approximately)
1 vs 16 99 percent
2 vs 15 93 to 94 percent
3 vs 14 85 percent
4 vs 13 79 percent
5 vs 12 64 to 65 percent
6 vs 11 62 percent
7 vs 10 60 percent
8 vs 9 about 50 percent

These are rounded, and they move by a point or two as each new tournament is added. The pattern is the thing to notice. From the 1 line down to the 4 line, the favorite's advantage falls steadily, and then it steps down sharply at the 5 line. The 5-12 game is the first matchup in which the underdog wins roughly one time in three. After that, the 6-11 and 7-10 games continue at similar levels, and the 8-9 game is a toss-up.

There is another pattern worth noting. In many years at least one 12 seed beats a 5 seed, and in some years several do. In 2014, three 12 seeds won: Harvard beat Cincinnati, North Dakota State beat Oklahoma in overtime, and Stephen F. Austin beat VCU in overtime. In 2019 three won again: Oregon beat Wisconsin, Liberty beat Mississippi State, and Murray State, led by the future NBA star Ja Morant, beat Marquette. Other years have produced one or two. A year with none is the exception.

Why this matchup produces upsets

Several forces combine at this seed line, and each one is simple.

The gap is smaller than the numbers suggest

A seed number is an ordinal label. It tells you which of the sixteen groups a team is in, not how far apart the groups are. The distance between the 5 line and the 12 line, measured in efficiency ratings or the point spread, is much smaller than the distance between the 1 line and the 8 line. The 5 seed is a good team, but it is not a great one, and the 12 seed is a good mid-major team, not a poor one. In terms of point spread, a typical 5-12 game is a contest in which the 5 seed is favored by a handful of points, not a dozen. Using the rule of thumb from the earlier article, that corresponds to a win probability for the favorite in the mid-60s to low 70s, which matches the record. The upset rate is what you would expect from the size of the gap.

Who the 12 seeds are

The composition of the 12 line is the most useful piece of context. Because of how the field is chosen, the teams on the 12 line come from two sources. Some are conference champions from leagues that receive an automatic bid but are not among the best conferences, and who earned a seed in the middle of the field because their records and ratings were strong. Others are at-large teams from the major conferences that were among the last to be admitted. Those last teams often sit on the 11 and 12 lines, and some of them play in the First Four, which we describe in the First Four: why there are play-in games.

The first kind of 12 seed is the interesting one. A team that won 28 games in a mid-major league and is seeded 12th may be better than its seed suggests, because the committee has less evidence about teams from smaller leagues. Such a team may have played a limited schedule against strong opponents, so its rating has a wide margin of error. That uncertainty cuts both ways, and the best of those teams turn out to be quite good.

Who the 5 seeds are

The 5 seed, by contrast, is typically a team from a major conference with a record that is good but not spectacular. It has played a tough schedule and has taken some losses, and its rating is more reliable because it is based on many games against strong opponents. A team in this position is solid. It is not a team that dominates opponents, which is why the 12 seed has a fighting chance.

Style and experience

Many 12 seeds are older than their opponents. Mid-major programs often rely on juniors and seniors who have played together for years, while power-conference teams may feature freshmen, transfers and one-and-done players. Experienced teams tend to turn the ball over less and to execute in the half court. Many of the 12 seeds that win also have an unusual style, such as a slow tempo, a pressing defense or a heavy reliance on three-point shooting. As we explain in the article on upsets, any of those features can add variance to a game, and variance helps the underdog.

Stories of the 12 seeds that won

The 12 seeds that beat 5 seeds have produced some of the best stories of the tournament. A few are worth recalling because they show the different forms an upset can take.

  • Cornell, 2010. The Ivy League champions beat Temple, a 5 seed, and then Wisconsin, a 4 seed, to reach the Sweet 16. They showed that an Ivy League team could win more than once in March. They lost to Kentucky in the regional semifinal.
  • Oregon State, 2021. The Beavers won the Pac-12 tournament to get an automatic bid, entered the tournament as a 12 seed, beat Tennessee, Oklahoma State and Loyola Chicago, and reached the Elite Eight before losing to Houston. They are among the highest-reaching 12 seeds in the history of the event.
  • Murray State, 2019. The Racers were built around Ja Morant, who put up a triple-double in the first-round win over Marquette. The performance made a national audience aware of a player who would soon become a top draft pick. It is a reminder that a single star can swing a game against a deeper opponent.
  • Oregon, 2019. The Ducks were a 12 seed that beat Wisconsin and then UC Irvine to reach the Sweet 16. They were an example of a team from a major conference that had been seeded low, then won a conference tournament.
  • New Mexico State, 2022. The Aggies, champions of the WAC, beat UConn, a program with a famous name and a big following, and sent the Huskies home in the first round.
  • Richmond, 2022. The Spiders beat Iowa in the same round, and their win gave that year's bracket several surprising first-round results.

In each case the story is not that a bad team got lucky. It is that a good team, seeded lower than it deserved, got an opportunity against a team that was a little less good than its seed implied.

The odds across four games

The tournament has four 5-12 games every year, one in each region. That makes it natural to ask how many 12 seeds should win. If we treat each game as having a 36 percent chance of an upset, and assume the games are independent, we can calculate the distribution.

  • The chance that no 12 seed wins is about 17 percent.
  • The chance that exactly one wins is about 38 percent.
  • The chance that two or more win is about 45 percent.

These are estimates built on an assumption, and real years vary around them. But they explain why the pattern feels so reliable. In about four years out of five, at least one 12 seed should win, and in nearly half of all years two or more should. A bracket that picks no 12 seeds is almost certainly wrong somewhere in this line, and a bracket that picks all four is almost certainly wrong too.

This calculation also exposes a common misunderstanding. The fact that a 12 seed will win somewhere does not tell you which one. Each individual 12 seed is still a roughly two-in-three underdog, and it is easy to pick the wrong team. The pattern is a statement about the group, not about any one game. That distinction is the core of using this knowledge well.

Which 12 seed to pick

If you are going to pick one or two upsets on this line, you want the ones with the best chance. There is no formula, but there are some features that add up.

  1. A strong rating for the seed. Check the efficiency ratings, which we describe in college efficiency metrics. A 12 seed with a rating that would fit comfortably on the 9 or 10 line is a better candidate than one whose rating matches the seed.
  2. A conference champion rather than a last-in at-large team. A team that won its league tournament on a hot streak is usually a better bet than one that sneaked into the field.
  3. A guard who can create a shot. Teams with good guards can keep a game close even when they are outplayed in the paint.
  4. Experience. Look for juniors and seniors in the rotation and a team that has played together for several seasons.
  5. A vulnerable 5 seed. Check the other side. A 5 seed with a weak defense, a short bench, a recent injury or a poor finish to the season is a better target than one that has been playing well. Check the latest injury reports before you finalize.
  6. A style that changes the game. A team that plays slowly, presses, or shoots a lot of threes adds variance, which is what an underdog needs.
  7. A geographic or travel edge. A 12 seed that plays close to home may have more fans in the building than the 5 seed.

No team will have all of these, but a team with four or five of them is a serious candidate.

How the 12-5 pick fits into a pool

The 12-5 upset is the most popular early upset pick, which has an important consequence for pool strategy. Because so many entrants pick one, the leverage you get from picking a 12 seed is smaller than the leverage you would get from a less obvious upset. If you pick the same 12 as everyone else and it wins, you gain little relative to the crowd. If you pick a 12 that few others chose, and it wins, you gain more.

The value of the pick also depends on what happens next. A 12 that wins faces either a 4 or a 13, and a 12 that wins twice reaches the Sweet 16. The deeper the run you pick, the more points are at stake, but the lower the probability. In most pools, one first-round upset in this slot, backed by a reason, is enough. We explain the reasoning behind that in bracket-picking strategy, including the idea of placing your upsets away from the favorites you are counting on deep in the bracket.

Notice also how the 12-5 game fits in the structure of the region. The winner of 5-12 meets the winner of 4-13, so a 5 seed that wins gets a game against a team that is likely to be a 4 seed. A 12 seed that wins usually has to beat that 4 seed next, which is a much taller order than the first game. We describe that structure in how the bracket is built.

Is the pattern changing?

Fans often wonder whether the 12-5 upset rate has changed over time. The evidence is that it has been fairly stable at around a third across the decades, with ups and downs from year to year. The changes in the sport, from the three-point era to the transfer portal, have not erased the pattern. But the sample is small: four games a year, and a few dozen years. A run of a few years with fewer upsets or more tells you very little.

The best way to treat the pattern is as a base rate, a starting point for estimating the chance of an upset in each year's games. You then adjust that base rate for the specific teams. A 12 seed with a strong rating and a 5 seed with a weak profile should be seen as a much better chance than the average. A 12 with a weak rating against a 5 with a strong profile may be less likely than the average.

What to take away

The 12-5 upset is real, and it has a clear explanation. The gap between a 5 seed and a 12 seed is small compared with the gaps above it, the 12 seeds are often better than their seeds suggest, the 5 seeds are good but not great, and a single game carries a lot of randomness. The result is that the underdog wins about one time in three, which means that across four games a year, at least one 12 seed wins in most years.

For your bracket, that means one or two 12 seeds is the sensible number, chosen on ratings, style and the weaknesses of the 5 seed rather than on hope. You can follow the whole first round on the March Madness page and see how each year's pairings are laid out in the men's bracket. And when the first 12 seed of the year wins, you will know that it is not a miracle. It is the tournament doing what the numbers say it should.

Search more on this

Keep reading

The First Four: why there are play-in games

Why the NCAA tournament starts with four play-in games in Dayton, which teams play in them, how the format came to be, and what the teams that survived the extra game have done since.

4Basketball Staff · Aug 1, 2026 · 13 min read

Tournament expansion: from 8 teams to 68

The NCAA tournament began with eight teams in 1939 and now invites 68. This is the story of every major expansion, why each one happened, what it changed in the bracket, and how the women's tournament grew alongside it.

4Basketball Staff · Aug 19, 2026 · 11 min read