The 2026 NCAA tournament odds
And why you should ignore them and pick Purdue to win it all
Hey, we have odds. But first, a few random thoughts.
The committee’s terrible SMU rationale
Pretty good job by the committee, who went by WAB for inclusion with the exception of swapping in SMU for Auburn. Look, you shouldn’t feel sorry for Auburn, but if you are going to use a system you have to use it. The reason given for SMU’s inclusion was the terrible, awful, injury excuse. Presumably, BJ Edwards will be back for the tournament after missing the last six games for the Ponies, a stretch in which they went 1-5. I mean, all the big college hoops media guys are saying he’s definitely back. So he’s definitely back, right?
Right?
OK, cool!
Well, wait a second. “Expected to return” isn’t giving me a warm fuzzy feeling he’ll be back. I mean, SMU has every incentive to say this. And it’s not like there’s any penalty for being wrong. But hey, I’m sure they’re being honest!
Um, what? Like what are the chances he even plays?
Evaluating roster strength or the effect of absences should not be in the committee’s domain. For one thing, many basketball people suck at this task, and the people on the committee are not even basketball people. They don’t understand things like 3-point luck, which has killed SMU recently and has little to do with Edwards’ absence.
And furthermore, three of those losses while he was gone were by 8, 13, and 20 points. How many points do you think BJ Edwards is worth? No doubt SMU has played worse without him, but you just can’t ignore the last six games if you want to get into the roster evaluation business (which to be clear, you shouldn’t). BJ Edwards is not Michael Jordan. If you believe betting markets, Edwards is worth about 2 points to SMU. Hardly enough to explain all or even most of SMU’s issues down the stretch.
But aside from all that, why are we doing this to determine who qualifies for the postseason? No other serious sports league does this. Neither should the NCAA.
Purdue is an interesting case
The top three teams own 48% (OK, 58%, roughly in line with last season.) of championship odds. That is high, but it’s not like it’s that much different from the past. In 2019, the top three had about 50%! Though 2022 and 2023 were around 40%. So I think I’m going to zig while the pundits zag and talk about how dominant the one-seeds are. No doubt the top teams are increasing their distance from the conference champs of one-bid leagues. But their dominance over teams in the 5-15 range is overstated.
Further cluttering matters is that it’s rare for a team to win the title while being better on defense than offense. And yet, the top six teams are worse on offense than defense. (I mean, it’s close for some of them and surely within the margin of error, but let me build a narrative here.)
Given that fact, Purdue is a pretty intriguing contrarian pick for the title. (Unfortunately, it’s less contrarian now that they won the Big Ten tournament.) The Boilermakers have the highest-rated offense in the country, but the opinion of them among the hot-take industry has been heavily clouded by utterly terrible defensive 3-point luck in their home losses to high-profile opponents. Mackey Mystique is dead, people said.
Yeah, it’s dead when the opponent makes half of their 3’s. And that’s what happened in home losses to Illinois, Iowa State, Michigan, and Wisconsin, who combined to make 60 of 118 3’s (50.8%). But let’s consider an alternate universe where Wisconsin goes 4-for-25 from 3, or Michigan goes 7-for-24 from 3. In fact, this isn’t an alternate universe. It actually happened and Purdue won those games kind of easily.
And now the chart…
Note these odds do not consider roster issues. If the Ngongba Show has truly been cancelled, then Duke’s actual chances are a bit lower, for instance. I’m far less concerned about LJ Cason’s absence for Michigan, though.
The ordering of the best probabilities to win a title roughly follows the ordering of the underlying ratings used. Which is more or less proof that seeding, while a nice status symbol of season-long accomplishments, doesn’t really affect a team’s chance of winning a title much.
Finally, the way to evaluate a tournament forecast from a system is not to plug the highest-probability team into the each round and see how the favorites do. Most systems will have very similar brackets using that method. The best way is to create thousands of brackets using a Monte Carlo method and take the average points scored using some scoring system. (I kind of hate the traditional point systems which are heavily backloaded, but whatever.) This is the right way to do it as it rewards forecasts that get the probabilities right.
It looks like the top 3 teams have a 58%, not 48% chance of winning the title. But I'm not here to nitpick. In terms of "narrative," the big story is that the committee is getting better and better at its job the more it makes itself aware of the work of you and your epigones. Hats off!
The math still isn't mathing. The probabilities of the semifinals between the West/Midwest and the East/South should each total out at 100%. The numbers are noticeably off. The West/Midwest is sitting at 103.89%
Ken, Ive seen these numbers edited like 3 times over already since you first posted this about 9 hours ago. Com'n, let's get it right.