NCAA Tournament upset picks: Can Grand Canyon and Yale continue their Cinderella run into the Sweet 16?

Our model had a terrible, horrible, no good, very bad day for most of Friday’s action. Its most likely first-round upset, New Mexico over Clemson, turned into a blowout win for the Tigers. It was not ideal, but the whole market was on the Lobos: They entered the game favored. We can live being wrong when it is aligned with conventional wisdom.

But then came the true stunner: Yale over Auburn. Slingshot gave Yale just a 5 percent chance of winning. It saw Auburn as severely underseeded and viewed Yale as a good team that doesn’t have any statistical resemblance to typical giant killers. None of the 10 most similar games in our model’s database ended in upsets. But you know what happened. If Slingshot had hands, it would have buried its head in them and wept quietly.

If that wasn’t enough, James Madison wiped the floor with Wisconsin. It’s not that Slingshot didn’t like the Dukes — it gave them about a 25 percent chance of winning. But the market was even more bullish on their chances, so we suggested that picking Wisconsin would be a nice chance for leverage in a bracket pool. By the time that game ended, any remaining cockiness from our Oakland call had evaporated, and we were looking into moving to a cabin in the wilderness, from which we’d never return.

But then came the mighty Grand Canyon Lopes to save the day. We’ve been writing about the Lopes. Slingshot loved them, and it loved the matchup with St. Mary’s. And Tyon Grant-Foster, Ray Harrison and the rest of the mighty Canyon Crew swarmed the Gaels on defense, attacked the rim and flew to a 9-point win.

So with the dust having settled on Round 1, here are some quick Slingshot stats:

• Five of its 10 most likely upsets occurred, including three of the top five.
• It gave Oakland a 19 percent chance of beating Kentucky, compared to roughly 12 percent implied odds on the money line. That’s a big win from just outside our top 10 upsets.
• The only other upset that occurred was the aforementioned Yale/Auburn game. That’s a big loss from well outside our top 10 upsets.

But enough about the past. It’s time to look at Sunday’s slate, where a whopping seven of eight games have Bracket Breaker implications (seed differential of at least five). On to the numbers!

Odds are from BetMGM. For more Underdogs, listen to Peter and Jordan’s podcast. For all our March Madness coverage, check out our content hub. 

No. 2 Marquette Golden Eagles vs. No. 10 Colorado Buffaloes

Upset Chance: 31.8 percent

Spread: Marquette favored by 3.5 points

At halftime of their first-round game against Western Kentucky, it looked like the Golden Eagles might not even make it this far. Marquette got it together in the second half, but that scare came against a team that didn’t even possess strong giant-killing credentials. Things are about to get much tougher.

Five of the eight games most similar to this matchup ended up in an upset, according to our model’s database. Marquette profiles as a “Generic Giant” (the most vulnerable of our model’s four families of overdogs), while Colorado is part of a Killer cluster that has success on the offensive boards: These teams typically cause problems for Generic Giants. The two teams are also separated by only nine spots in our model’s basic power ratings, yet somehow are eight seeds apart. That’s weird.

But if you want to figure out who will win, focus on two areas: rebounds and turnovers. If this is a battle on the boards, things will go well for the Buffaloes. They grab 32 percent of their own misses and only allow a 25 percent rate for opponents. Marquette has opposite splits: 26 percent on offense and 31 percent for opponents. So Shaka Smart’s crew will have to look elsewhere to gain a possession edge, and that comes from turnovers. They force miscues on 21 percent of opponents’ possessions (20th in the country) and turn it over only 14 percent of the time themselves (26th). Colorado, on the other hand, ranks 253rd in holding onto the ball and 300th in forcing turnovers (15 percent).

Rarely do you see such a clear stylistic clash, so this game should be a lot of fun. One other area to watch: Colorado doesn’t shoot a lot of threes (only 30 percent of shot attempts). But the Buffaloes rank sixth in the nation by making 39.4 percent of them. Marquette cuts off the paint on defense and forces teams to bomb away from deep, allowing 43% of opponents’ shots to come from downtown. So, in a funny way, Marquette may increase Colorado’s ability to pull off an upset by encouraging them to take a higher-variance shot that they happen to make at an excellent rate.

Put it all together, and Marquette fans should be very, very nervous.

No. 4 Alabama Crimson Tide vs. No. 12 Grand Canyon Lopes

Upset Chance: 22.3 percent

Spread: Alabama favored by 5.5 points

Can the Lopes do it again? Our model thinks they have a solid chance. The task will be tough but not complicated: Slow down Alabama’s high-powered offense (third in adjusted efficiency, ninth in effective field goal percentage, seventh in adjusted tempo) enough to expose its leaky defense (117th in adjusted efficiency). If you watched Grand Canyon’s win over St. Mary’s, you saw that they have the athletes to do it. And the Lopes excel in two key areas in a game where extra possessions will mean a lot: forcing turnovers (20 percent) and offensive rebounding (34 percent).

Our similar-games model gives some reason for pause: The eight most similar matchups went to the favorite, and most were blowouts. (Games nine and 10 were both upsets, though!) Still, the Lopes have a shot to keep things rolling into the Sweet 16.

No. 1 Houston Cougars vs. No. 9 Texas A&M Aggies

Upset Chance: 20.1 percent

Spread: Houston favored by 10.5 points

We’ve been looking forward to this one for a long time. Well, not necessarily this specific matchup, but more the chance for Texas A&M to challenge any top seed. The Aggies are the classic Wounded Assassin from a power conference and have the most interesting (and enjoyable) statistical profile in the country. They cannot shoot: The Aggies rank 339th in effective field goal percentage and make fewer than 30 percent of their three-pointers. But they are the country’s best offensive-rebounding team (42 percent). Announcers often talk about a team’s best offense being a missed shot. It’s trite and typically misdirected. But in this case, it’s accurate.

And now along comes Houston, sporting the country’s best defense. They clamp down on teams, harass opposing guards, choke off the lane and generally make everyone miserable. But the Cougars have one weakness: They are a below-average defensive rebounding team, allowing opponents a 30 percent offensive rebound rate. So imagine what Texas A&M will do with missed shots.

Houston’s mission is clear, then. The Cougars need to use their extreme ball pressure, which forces turnovers on a quarter of opponents’ possessions (third in the country), to keep Texas A&M from getting a first shot attempt, let alone a second. Do that, and they should have enough success against the Aggies’ 63rd-ranked defense to cruise to a comfortable win.

But if the Aggies can hold onto the ball and get shots up, things could get dicey for Houston. And there’s one more factor to consider: The Aggies are shooting the ball better of late. Here’s their three-point shooting in their past six games:

It’s still a small sample size, and shooting percentages are volatile by nature. But Texas A&M may have turned a corner, and even going from “totally awful” to “decent” makes a huge difference. Wade Taylor and Tyrece Radford have shown that they can cook. Now let’s see if they have the recipe for success against Houston’s loaded backcourt.

No. 4 Duke Blue Devils vs. No. 12 James Madison Dukes

Upset Chance: 21.1 percent

Spread: Duke favored by 7.5 points

James Madison entered its game against Wisconsin riding the nation’s longest winning streak. But questions remained, given the weakness of their schedule. As we wrote in our first-round preview, outside of their season-opening win at Michigan State, the Dukes played just two games against teams ranked in the top 100 at KenPom.com. Both were losses to Appalachian State.

But that’s less of a concern after watching James Madison control the game against the Badgers. Now, we can focus even more intently on their outstanding statistical profile. The Dukes own edges over their opponents in every key giant-killing category: offensive rebounds, turnovers and three-pointers.

That last category will be particularly important in this matchup. James Madison holds teams to 28.8 percent shooting from beyond the arc. But the Blue Devils knock down 37.6 percent of their 3s. Will James Madison double Kyle Filipowski in the post like Vermont did and try to close out on Duke’s shooters (and hope they miss)? Will they try to choke off the perimeter and live with Filipowski trying to finish inside? It will be an interesting strategic decision.

Losing Caleb Foster has hurt Duke more than many experts have acknowledged — they now have hardly any perimeter depth and barely any scoring off the bench. But Duke’s defense looked connected against Vermont, and if the Blue Devils can cut off James Madison’s transition offense and force them to score in the half court, they can author a repeat performance.

One other factor to watch: James Madison’s perimeter pressure. Duke is typically a low-turnover team. But so was Wisconsin, and the Badgers turned the ball over 19 times in their first-round loss. James Madison gets after it on defense and will challenge Duke’s ball-handlers, especially without Foster.

No. 5 San Diego State Aztecs vs. No. 13 Yale Bulldogs

Upset Chance: 19.8 percent

Spread: San Diego State favored by 5.5

If Yale can beat Auburn, they can beat San Diego State, which sits 19 spots lower in Slingshot’s power ratings. Yale flashed some atypical traits on Friday. The Bulldogs are typically a low-volume three-point shooting team, but took 20 threes against Auburn and made nine. They also prefer to play a packed-in-defense rather than pressure the perimeter, but they forced the Tigers into 14 turnovers. So were those functions of a strategic shift that Yale can make against more talented opponents, or just Auburn playing terribly? That remains to be seen.

However, our model likes this matchup much more for Yale. Three of our database’s 10 most similar games were upsets, including the most similar one: Oregon State’s 14-point walloping of Tennessee in 2021. The Aztecs play slowly, which should allow Yale to hang around. And the Aztecs allow teams to shoot lots of threes (40.4 percent of shots). Now, those are usually tough threes: They hold teams to 30.6 percent shooting. However, simply allowing more attempts is dangerous, as it increases the variance for an underdog. Let’s see if Yale takes advantage of those opportunities.

No. 1 UConn Huskies vs. No. 9 Northwestern Wildcats

Upset Chance: 11.7 percent

Spread: UConn favored by 14.5 points

Hey, here’s some breaking news from our model: UConn is good. The Huskies have no obvious weakness(our model would prefer if they played faster and shot fewer threes, but our model likes to nitpick; it also frequently comments when its children forget to put their napkins on their laps). To win, Northwestern will have to embrace some serious Chameleon energy. The Wildcats make 39.3 percent of their threes (seventh in the country) but only take 36.7 percent of their shots from deep (198th). Of course, UConn’s defense at the arc is great, limiting both attempts (33.2 percent) and accuracy (31.4 percent). The Huskies also own a huge edge on the glass. So it’s threes or bust for Northwestern even to have a chance at an upset.

No. 1 Purdue Boilermakers vs. No. 8 Utah State Aggies

Upset Chance: 7.3 percent

Spread: Purdue favored by 11.5 points

We’re honestly shocked that Utah State’s chances are this low. The Aggies are a good team, ranking 37th in Slingshot’s basic power ratings. And Purdue has a bit of a “history” when it comes to upsets.

But this year’s Boilermakers are well-rounded and tough. They dominate the boards at both ends, don’t rely on threes, but make 41 percent of the ones they do take (second in the country). To beat the Boilermakers, teams have to make life difficult on their guards and force turnovers (Purdue gives it up on 16.5 percent of possessions). But that’s not Utah State’s game: The Aggies rank 209th in the nation in forcing turnovers.

Only one of our database’s 10 most similar games ended in an upset, and Slingshot doesn’t expect this one to follow suit. As good as the Aggies looked against TCU, beating Zach Edey will be a much, uh, taller task.

(Photo of Gabe McGlothan: Steph Chambers / Getty Images)