How stats and history would pick the 2025 men's NCAA Tournament
Live free, pick Houston
PAST EDITIONS: 2021, 2022, 2023, 2024.
If you’d like to know more about how this field shapes up historically, read this.
On with the show.
After a tournament I thought would be pretty calm last year and generally was, we still ended up with a 4 and 11 seed in the Final Four. (It was indeed historically calm, at least in one aspect.) Somehow, we might be tracking for something even less exciting, at least if you’re really into good teams losing early.
This year has 4.28 projected Round of 64 upsets, the second-fewest in my 24-year database behind 2017 (4.22). That’s despite Michigan, Memphis, and BYU all being handed brutal Round of 64 games as 5/6 seeds, especially Memphis, who is an underdog to their 12-seed opponent. 2012 (which had 6.71! projected upsets and had 7) or 2006 (6.51, had 6), this is not. If it heeds to expectation and we do only get four, it would be tied with 2015, 2017, and 2023 (strangely) for the fewest since the all-time low of two in 2007.
Now, one note on that: we do have a total of 1.25 expected wins from the 13-16 seed lines. The problem is that this is the fewest ever measured, at least in my personal database. The only years within half a win of it are last year (one 13-16 seed win), 2019 (also just the one), and 2017 (none). Along with the “God, this is dominant” feeling, the 1 seeds have their highest collective expected representation in the Final Four (1.87 teams) in modern history, and only 2019 carries lower odds of a 6+ seed getting to April. Look, it’s not my job to tell you how to do a bracket; I just share the info.
For whenever this gets picked up by people who don’t normally read this website, many of these picks will be wrong. Even the very best brackets miss on an average of 13-15 picks out of 63 total a year. If I missed on 15 total picks, I would be beyond thrilled. I missed 21 last year; maybe that can get below 20 this year. Who knows. I look forward to seeing how it all unfolds.
As a reminder, here’s how all of this works: What this is is simply a game-by-game projection of the field of 68 based on a document I’ve put together since 2018. I’ve accumulated KenPom odds, based on pre-tournament data, from 2002 to present.
Lastly, this post is paywalled. I’ll unlock it after the games begin, but I’m a little past the point of this either getting aggregated or people begging for me to make bracket picks for them. Guys, when I do a bracket I spend about 10 minutes on it and fart it out into the family group. I like watching the games. (Stephen A. Smith voice) BUT! If you find these helpful, I really think you should sign up. Our picks finished in the 98th-percentile in 2021, 95th-percentile in 2023, and 80th-percentile in 2024. Pretty good when exactly zero of those years, I had the right champion. Please do not ask about 2022.