Which programs are best and worst in March Madness?

Well, it’s here. We’re finally at the one truly magical time in college basketball – March Madness. The eyes of the country fall upon college basketball – and their brackets – as every game becomes win or go home. Heroes and underdog stories will rise, just as favorites and scapegoats will fall.

As teams go through the conference tournaments – one last chance to jockey for bids and seeding in The Dance – I decided to take a look at the teams who have performed best – and worst – in the harsh spotlight of March Madness. It felt like something worth quickly going through, as we all aim to calibrate expectations after Selection Sunday. Looking for a program, or coach, to semi-rationally trust or doubt as you fill out your bracket? You’re in the right place.

The research took a tack similar to some previous historical studies I’ve done on NCAA tournament teams (see here for extreme teams, on offense or defense, and here for the effect of short benches). The data stretches back to 2002, as that’s the extent of KenPom’s adjusted efficiencies record. Expected wins were calculated in two ways. First, simply by seed – how many wins a typical 1-, 2-, 3- seed (etc.) have won in tournaments since 2002. And then by simulations that use their adjusted efficiencies – which was also used to classify whether a team was a favorite or underdog in any given game.

Without further preamble, here’s a look at the programs – and coaches – that have best outperformed expectations in the NCAA tournament since 2002.

Built for March

There were a couple different means used to tease out the best March squads. I looked at most victories as the underdog, best total differential (expected wins vs. actual wins), and best average differential (or, most extra victories per tournament appearance). The list we’re left with satisfies the following: the team has appeared in at least three NCAA tournaments since 2002, has performed well recently, and stands out in one or more of the above measures. The first: a team that’s tasted a lot of unexpected success in March, but needs to prove it still can under different management.

Butler

It’d be tough to build this list without mentioning Butler. In consecutive seasons (2009-10 and 2010-11) under current Celtics head coach Brad Stevens, the Bulldogs made it to the national title game. Though they lost both games, it still stands as one of the best runs in modern tournament history. Both teams were talented – both Gordon Hayward and Shelvin Mack currently play for the Utah Jazz, and Hayward in particular has become something very close to a superstar. But they certainly weren’t favorites. Butler made it to the finals as a 9 seed in 2009-10, and returned as a 5 seed the next year.

Those two runs are the basis of their extremely strong numbers in this analysis. In two consecutive years where the Bulldogs were barely expected to win one game (either by seed or efficiency margin), they won five instead.

Butler’s margins of 10.16 (per log5 expectations) and 10.7 (by seed) wins above expected in the NCAA tournament is good for 2nd best in the entire country since the 2001-02 season. They’re the only team in the top 10, by either measure, that has played in less than 10 NCAA tournaments in that time (9). On average, they win more than 1 extra game each tournament than a typical team would be expected to.

But it was almost entirely built by coaches and players who are no longer there. Though Butler has made the NCAA tournament in both 2015 and 2016 under Chris Holtmann, they won a single game each year before bowing out. Though it was technically more than they were expected to – 6 and 9 seeds, on average, have won less than 1 game in the Dance since 2001-02 – it’s marginal at that point.

However, Butler will have its best chance in a while to add a slew of tournament wins this year. Barring a first round Big East exit, Butler will almost certainly land a top-4 seed. While this regime shouldn’t have quite earned your trust yet, they’ll be an interesting team to follow as the tournament progresses. There may yet be some magic left.

North Carolina

The optics here might come as something of a surprise – North Carolina is one of college basketball’s blue bloods. Typically highly seeded and blessed with strong squads, postseason success is more expectation than surprise in Chapel Hill. So how does North Carolina rank as one of the strongest postseason teams, relative to expectation, in the country? Let’s take a deeper look.

North Carolina has compiled an incredible record as the favorite in the NCAA tournament. In their twelve NCAA tournament appearances in the last 15 years, they’ve gone 35-4 as the favorite, good for a winning percentage of 89.74 percent. That number is the third highest for any NCAA tournament team with more than ten victories as the favorite, with just Xavier and Oregon doing better. Perhaps even more incredible, North Carolina was the stronger team – based strictly on pre-tournament adjusted efficiencies – in 39 of its 46 tournament games over that span.

Quite simply, North Carolina dominates when it’s expected to. Of those twelve appearances, nine have come with a seed of four or higher. In all but one of those appearances (four seed or higher), North Carolina has exceeded or marginally underperformed its expected wins when measured by seed or adjusted efficiency. The one outlier was a round of 32 crashout for the 2006 team that pulled a 3-seed. Three runs to the title game with two championships certainly helps North Carolina in these measures, but they’re remarkably consistent in the Tournament when they’re good. And, as they’re slotted for at worst a 2-seed this year, history pegs this program as one of the more reliable bets to make the Elite Eight in your bracket.

Be careful when they’re matched up against a squad that brought a better efficiency margin than them to the tournament, though. North Carolina is just 1-6 as the underdog in the NCAA tournament since 2001-02, and they lost their first game to a stronger squad in each of their five tournaments preceding last year.

Michigan State

Tom Izzo is the other stereotypical pick for expected success in March. I’m sorry to disappoint readers here looking for unforeseen insight – it’s true. Michigan State has been one of the top 7 teams in the country over the past 15 tournaments in extracting wins where we wouldn’t expect them. If you’re curious about some lesser known teams, see the quick hitters at the bottom of this section.

What’s interesting about Izzo’s postseason success, though, is that it typically only comes when you’re not expecting it. The Spartans are rarely seeded too highly – since 2001-02, they’ve only been higher than a 4-seed five times. But those high seeds have not turned out well – twice, Michigan State failed to ‘justify’ its seed, and just barely made it in two other instances. Thanks especially to last season’s upset loss to Middle Tennessee State in the first round, Izzo’s highly seeded teams have actually lost about a game and a half more than they’d be expected to according to efficiency margin.

Where Izzo and Michigan State have worked their magic is as an underdog. If we just counted Izzo’s teams that have received a 5-seed or lower, they’d rank as the 5th best program in both overall differential of expected wins to actual wins and added wins per tournament.

While Michigan State’s winning percentage of 78.26 as the favorite ranks somewhere in the middle of the pack, they’ve won half of their 20 games as the underdog in the tournament. This is tied for the most wins as the underdog, and in the top ten for winning percentage among all teams with at least 3 victories as the underdog.

Michigan State may be somewhat close to the bubble at this point, but history tells us to fear Izzo most when his team is counted out. Though I’m sure this is hardly news to most of you, keep Michigan State in mind when looking for first weekend upsets.

Quick Hitters

UConn’s just on another planet in most of these measures – they win nearly 70 percent of their games as the underdog in the tournament… Syracuse rates out well, but it’s almost entirely due to last year’s run, the Melo championship, and 2014 – you’re either hitting big or missing entirely… UCLA lands in the top 10, but mostly thanks to teams from 2006 and 2007… We miss you, George Mason… Kentucky, Florida, Oregon, and Wichita State round out the best among high profile tournament teams this year.

Underperformers

Looking for probable tournament teams that you may want to think twice before trusting? We’ve got you covered there, too. These are the teams that have historically done poorly in March, ranked and judged much as the best teams were.

Virginia

Tony Bennett has been building something in Charlottesville. After a non-descript first four seasons, in which Virginia made just a single NCAA tournament appearance (a first-round exit to Florida), the Cavaliers have secured a top 2-seed in each of the past three tournaments. They’ve put together another strong season by adjusted efficiency this year – KenPom currently has them at 5th, based on adjusted points scored and allowed per possession. But a 21-9 record is likely to doom them to something between a 4 and 6 seed.

It may be for the best – they certainly haven’t worn the mantle of favorite all too well. While Virginia finally managed to break through to the Elite 8 last year, they collapsed in the 2nd half against a Syracuse press once there and blew a double digit lead. And getting to the Elite 8, for a 1 or 2 seed, is simply justifying your seed; Virginia hadn’t managed that in either of the previous two years.

It all adds up to some bad juju for Virginia’s numbers in this analysis, Even though they’ve made just five appearances in the tournament in the last fifteen years, their negative differential (-4.13) in actual wins vs. expected wins (by log5 simulation) is the 7th worst among any team that made the tournament during that time. They haven’t won a game as the underdog (0-2), and are a pedestrian 7-3 as the favorite.

While it’s certainly possible Virginia is one hot streak away from breaking through in a real way – just ask Villanova fans – the last few tournament years have not been kind to Tony Bennett and his squad. Virginia’s efficiency numbers will undoubtedly look attractive as a 4- or 5-seed, but consider carefully before relying on them to go too far.

Duke

We go back to North Carolina for a blue blood program – but this time, it’s one that hasn’t consistently lived up to expectations in the Big Dance. While Duke has multiple national titles under its belt during this stretch – and another just out of reach in this sample – they rate as one of the most disappointing programs relative to postseason expectations in the last 15 years. Let’s dive into why.

One of the biggest is, quite simply, the ‘burden’ placed on Duke at the beginning of each postseason – they should win a lot of games. The Blue Devils have made every single NCAA tournament in the sample, and their average seed during this time is 2.07. 1-seeds and 2-seeds are expected to win 3.30 games and 2.35 games over the time span of the sample, respectively. This about matches up with what’s seen under Duke’s ‘expected win’ column in the log5 samples.

But while Kansas – one of the other teams that have been in each tournament, and has a high average seed – still manages to outperform its expectations, Duke falls well short. So there’s something else going on.

Duke’s first round exits are perhaps more celebrated than their championships, simply because a lot of people love to hate Duke. The Blue Devils have lost in the first round twice as a 2-seed, and once as a 3-seed. These losses absolutely kill Duke’s margins. Other culprits include Sweet 16 losses in 2005 and 2006, when Duke held a 1-seed.

Excluding the years in which they won the national championship, Duke has won an average of just 1.7 games per year, while playing with an average seed of 2.23. Their record as a favorite drops to 22-10, and they have no wins as an underdog. That’s bad.

While I’m sure Duke fans will live with the ups and downs – mainly because the ups are championships – you need to be as dispassionate as a bracketeer. Duke’s history is an all-or-nothing type of pick. If you don’t have them going to the championship, think hard before putting an Elite 8 appearance down in pen.

Quick Hitter

The Old Big East dominates both ends of the spectrum, as Pitt is by far the worst in actual vs. expected differential… Maybe hold off on Cincinnati as your dark horse – they’ve underperformed their expected wins by log5 in all but three of their ten tournament appearances, and three of their last four… More than a few bubble-ish teams land here, including Clemson, Oklahoma State, Creighton, Wake Forest, Iowa, and Providence.