Predicting The Number of Upsets in Round 1 of March Madness Based on Kalshi Markets

How many upsets will shake up the NCAA Division I Men's Basketball Tournament Round of 64? A data-driven look at Kalshi markets, historical trends, and 2026 bracket volatility.

March madness graphics at center court before a practice session ahead of the first round of the men's 2026 NCAA Tournament at Xfinity Mobile Arena. / Bill Streicher-Imagn Images

The NCAA Division I Men's Basketball Tournament tips off in just about an hour, making this the ideal window to evaluate real-time pricing across prediction markets like Kalshi and Polymarket, where activity has surged as the bracket locks in. Every year, the tournament delivers chaos, and naturally, attention has shifted to one of the most popular markets on the board: “Number of Upsets in the Round of 64.”

In this context, an “upset” is defined strictly by seed-line matchup—so if a No. 9 seed defeats a No. 8 seed, it is considered an upset for this market, regardless of which team is favored by sportsbooks.

Kalshi

Leading up to Selection Sunday, expectations hovered as high as nine upsets in the opening round. That number has since come down, reflecting a growing sentiment that this year’s field is more top-heavy. Last year offered a reminder of how chalky things can get, with limited early-round surprises and all four No. 1 seeds advancing to the Final Four. That context appears to be influencing current expectations, with the market now clustering around seven.

Number of Upsets in the Round of 64 Historically

Let’s take a look at how many upsets there were in the first round in each of the previous 10 tournaments. Looking back, the volatility becomes clear:

• 2025: 7
• 2024: 11
• 2023: 7
• 2022: 10
• 2021: 10
• 2019: 12
• 2018: 9
• 2017: 6
• 2016: 13
• 2015: 5

So despite there not being a ton of upsets last year, the average over the last 10 tourneys is a solid 9.0. Yet the market this year says that there is only a 19% implied probability that there will be 9+ upsets in this year’s first round. So what gives?

How Many Upsets Will Occur in Round 1 of This Year’s NCAA Tournament?

A closer look at this year’s bracket suggests there seems to be a few mispriced games on Kalshi. Not all of the mispriced favorites are going to lose, but the market is a bit too high on a few teams such as Nebraska and Kansas. 

Kansas Jayhawks forward Flory Bidunga (40) reacts to game play on the bench during the second half against the Houston Cougars at T-Mobile Center. / William Purnell-Imagn Images

There are also several tightly lined matchups that could swing either direction. Games featuring narrow spreads and minimal separation between seeds often drive the upset count, and this year is no exception.

Some of the more competitive first-round matchups include:

11. South Florida vs. 6. Louisville (-4.5)

11. VCU vs. 6. North Carolina (-2.5)

11. Texas vs. 6. BYU (-2.5)

10. Texas A&M vs. 7. Saint Mary’s (-2.5)

10. Santa Clara vs. 7. Kentucky (-3.5)

10. Missouri vs. 7. Miami (-1.5)

That’s a substantial group of games where outcomes are far from certain. And that doesn’t even account for the usual volatility in 8–9 matchups, where the gap between teams is often negligible. Upsets in that range tend to add up quickly. Some of the nine seeds are actually favored on both sportsbooks and on Kalshi. Just look at Utah State-Villanova or Iowa-Clemson.

Not to mention, we typically see at least one 5 seed fall in the first round. While most of the No. 5 seeds seem better positioned this year than in year’s past, don’t be too surprised if High Point (+10.5) shocks the world and defeats the Wisconsin Badgers. Akron (+7.5) also has a shot against Texas Tech. 

So how many upsets will actually occur in Round 1? 

Kalshi

The market currently centers around seven upsets, but history suggests a wider range of plausible outcomes. The nature of March Madness is unpredictability, and even in years perceived as “top-heavy,” variance tends to show up quickly once games tip.

From a trading perspective, it’s less about certainty and more about understanding that distribution of outcomes. While seven may be the consensus midpoint, the broader range—particularly between seven and ten—remains very much in play given historical trends and matchup dynamics. We've witnessed 10+ upsets far more than the 9% implied probability listed on Kalshi (40% over the last 10 tournaments). That’s a significant gap. That's where the value lies.

In a tournament defined by chaos, even the most data-driven expectations can shift quickly once the ball is tipped.

The Verdict

Safe Pick: 7+ (50%)

My Pick: 8+ (31%)

Longshot: 10+ (9%)

Trading is risky, always trade responsibly. If your activity is becoming a problem, support is available by calling 1-800-522-4700.

More Prediction Markets On SI News:

feed