Will a 15-seed or 16-seed win a Round 1 game? Prediction market analysis
Published:
It’s finally time for the 2026 NCAA Tournament.
As part of the lead-in to Thursday, I’ll be surveying some of the most interesting NCAA Tournament props on Kalshi.
Today’s question: Will any 15- or 16-seed win a Round of 64 game?
There are 2 angles we can use to solve this. The latter is better than the former, but we’ll use both.
First, the quicker data point: History.
Over the last 25 NCAA Tournaments (since 2000 but excluding 2020), 15- and 16-seeds are a combined 10-190. That’s good for a winning percentage of just 5% over the last quarter century.Â
It’s perhaps worth noting that most of those wins are clustered between 2012 and 2023. All but 1 to be exact, with the exception being a lone win for those seeds way back in 2001. If you only consider games since 2012, the winning percentage spikes to 8.7% (9-95).Â
So according to recent history, we can say the expected winning percentage for these seeds is somewhere between 5% and 8.7%.
Moving on. The next way we can approximate the right pick for this Kalshi prop is through implied probabilities with money lines. NCAA Tournament betting markets are efficient, so using the money lines to determine win likelihood will get us to a reasonable place.
Here are the (de-vigged) implied probabilities derived from DraftKings money lines:
| Favorite | Win % | Underdog | Win % |
|---|---|---|---|
| Duke | 97.28% | Siena | 2.72% |
| UConn | 95.78% | Furman | 4.22% |
| Houston | 96.85% | Idaho | 3.15% |
| Iowa State | 95.78% | Tennessee St. | 4.22% |
| Arizona | 97.28% | LIU | 2.72% |
| Purdue | 96.85% | Queens | 3.15% |
| Michigan | 97.28% | UMBC/Howard | 2.72% |
| Florida | 97.28% | Lehigh/PVAM | 2.72% |
Note: Betting odds are not yet available for Michigan and Florida’s first-round games, so I used the same figure (2.72%) that Duke and Arizona’s opponents received as 16-seeds.
Given those figures, the math suggests a 22.9% chance of at least one of those underdogs winning outright. We arrive at that number by multiplying the individual win probabilities for all 8 favorites together and subtracting that number from 1.
Now, this outcome is pretty unusual. Kalshi and the betting markets are typically not this far apart on props like this — at least not for very long. Experts may have differing theories, but I’d guess the gap between Kalshi and the betting apps is that underdogs are typically overpriced at the NCAA Tournament due to the event’s reputation. Bettors want to back dogs with big payouts and prices are set accordingly. This is more speculation than anything on my part, but that’s my best guess at what’s going on here.
If you’d rather rely on a projection system than Vegas, the picture does change a bit. Here are win probabilities for all of these games, per EvanMiya:
| Favorite | Win % | Underdog | Win % |
|---|---|---|---|
| Duke | 99.80% | Siena | 0.20% |
| UConn | 96.60% | Furman | 3.40% |
| Houston | 99.50% | Idaho | 0.50% |
| Iowa State | 98.40% | Tennessee St. | 1.60% |
| Arizona | 99.60% | LIU | 0.40% |
| Purdue | 99.80% | Queens | 0.20% |
| Michigan | 99.75% | UMBC/Howard | 0.25% |
| Florida | 99.80% | Lehigh/PVAM | 0.20% |
Note: Estimated win probabilities were used for 1-seeds who do not yet have an opponent finalized.
The odds of a 15- or 16-seed pulling off at least 1 upset are significantly lower by EvanMiya’s estimation. It’s down to just 6.6%, which is in line with historical outcomes mentioned above but still far away from Kalshi’s market (12%) and even farther away from the sportsbooks’ implied probability (22.9%).Â
In general, I’m a believer in top-down analysis. That means my initial instinct is to side with the sportsbooks’ prices, which imply a massive edge taking the “Yes” side of this prop. But the reality is that these money lines are too extreme to trust at the same level we would if we were looking at an NFL spread market. These are not hyper-efficient lines that have gone through the price discovery process because there’s not a significant appetite for sharps to bet Duke -100000 against Siena or even Arizona -20000 against LIU.Â
Therefore, I think history and EvanMiya’s projections are closer to what we can truly expect from the 15- and 16-seeds this year. And when you compare those expectations to the Kalshi market, there appears to be value on the “No” side of this prop.Â
Pick: No 15 or 16 seed to win in the Round of 64 (89 cents per contract)
Spenser is a news editor for Saturday Down South and covers college football across all Saturday Football brands.
