A historical analysis of the 2025 NCAA Tournament field

Hope you hate upsets

Well, they just released the bracket. I am a little confused by what they did with it, but we’ll touch on that at the very end. Facts only to start.

Before I forget, this is the last non-paywalled post of the entire NCAA Tournament. Sign up for $20/year.

Everything noted below is over the last 23 NCAA Tournaments, from 2002 to present.

1 seeds

• At an average AdjEM of +36.12 this is the strongest batch of 1 seeds in history, demolishing the previous record of +32.7 in 2014-15.

• Out of 92 1 seeds from 2002 to now, Duke is tied for the 5th-highest Final Four odds of any 1 seed in history.

• This is the first year since 2019, and just the second in history, that all 1 seeds have a 35% or higher chance to make the Final Four.

• Three of the four 1 seeds have an 80% or higher shot to make the Sweet Sixteen, a thing only exceeded by the 2019 1 seeds. (All four 1 seeds made the Sweet Sixteen that year.)

• Don’t expect much in the way of Round of 64 or 32 action. The +42.86 AdjEM gap from the average 1 to the average 16 is the highest ever seen, and the +15.25 AdjEM gap from the average 1 to the average 8/9 seed is only exceeded by 2019 (no upsets) and 2015 (one).

2 seeds

• Well, you guessed it: at +28.82 AdjEM, this is the second-strongest batch of 2 seeds ever, only behind 2015.

• At 26.3%, no 2 seed has better Final Four odds than Tennessee, though they’re just tied for 20th-best among the 92 2 seeds in my 23-year database.

• If you don’t like that, perhaps avoid St. John’s, who sits 11th-lowest in history at 11.4% to make the Final Four. No one below 14% has ever made it (0-for-15).

• One minor note: 2 seeds at 50% or better to make the Elite Eight are, well, 6-for-12 in getting there. Hey, beats the average 42% rate for 2 seeds. (Tennessee’s at 51%.)

• Of the seven 2 seeds to eat it in the Round of 64 since 2002, all were 94.4% or lower to win their R64 game, and five of the seven had a Shot Volume Index of +2 or lower. No one qualifies this year, FWIW.

• This is the second-highest gap (+26.84 on average) from 2 seeds to 15s in history, barely behind 2015’s.

• But! The gap between 2s and 7s/10s is fairly normal. Largest since 2021, but only 7th-biggest in modern history.

3 seeds

• For the first time since 2022 - yes, seriously - the 3 seeds are better than the 4 seeds on average. They are the strongest 3 seeds ever at +26.44 AdjEM, surpassing the previous record set in 2019.

• For the first time ever, three of the four 3 seeds are at 55% or better to make the second weekend. The only non-qualifier, surprisingly (to some), is Kentucky.

• The gap between 3s and 14s, at +20.96, is the highest in my database.

• BUT: the gap between 3s and 6/11s, at +6.13, is squarely at the historical average and is actually smaller than last year. Perhaps more havoc will arrive for this group in the Round of 32.

4 seeds

• A tremendous group this year with an average of +24.98, the highest ever. EIGHT seed lines set records this year of the good variety. (We’ll get to one putrid one.)

• For the first time since 2011, every 4 seed is inside the KenPom top 20. That year did see a 4 seed make the Final Four…but it also saw one lose in the Round of 64.

• There are two gigantic positive outliers this year: Arizona and Maryland, who are at 61.2% and 65.8% to make the Sweet Sixteen, respectively. They are the fourth and fifth 4 seeds in history to have >60% odds for the Sweet Sixteen, and Maryland has the second-highest odds in history behind 2013-14 Louisville.

5 seeds

• Here’s your area of opportunity. At +20.26 AdjEM on average, this is the 13th-strongest batch of 5 seeds out of the last 23 Tournaments. That is important because…

• At a gap of just +4.43 AdjEM on average, this is the third-smallest 5-12 gap I have on record. The only smaller ones were 2007 (…oops. No 12 over 5s!) and 2011 (one 12 over 5, but that 12 made the Sweet Sixteen).

• This group actually rates out pretty similarly, against the rest of the field, to 2013 (three 12-over-5s) and 2014 (also three). Ya never know.

• At 44.6% to win their first game, Memphis has the fifth-worst odds of any 5 to beat a 12 in the last 24 years of March Madness. The benefit for them: two of those four (2010 Texas A&M, 2011 Kansas State) did beat their first-round opposition.

• For the first time since 2012 and just the sixth time ever, two 5 seeds (Michigan and Memphis) have a <60% chance to win their first game.

6 seeds

• Back to dominance: +22.43 AdjEM, strongest ever.

• At +5.62 AdjEM, this is the largest average gap between 6s and 11s since 2019 and the sixth-largest since 2002.

• All 6 seeds are inside the KenPom top 25 for the first time since 2021 (two upsets!) and the second since 2009 (one upset!).

• BYU’s odds to beat VCU, at just 54.1%, are the lowest by far of any of the 6 seeds…but they’re just the 28th-lowest on record of the 92 6 seeds. Nothing that nuts.

• Missouri, at 72.7%, has the 8th-highest odds ever to beat an 11 seed opponent in Drake. The issue: the seven teams ahead of them went 3-4.

7 seeds

• Well, this is getting boring. +22.34 AdjEM, the strongest ever by over two full points. The only year that comes close is 2021, which did see a 7 seed make the Sweet Sixteen.

• The gap from 7s to 10s is +5.83 AdjEM, the largest we’ve ever seen. The previous largest was +4.00 in 2019, when 7s…uh, went 1-3.

8 seeds

• At a +22.34 AdjEM, you guessed it: strongest ever. When everyone is strong, does that mean anything? TBD.

9 seeds

• +19.39 AdjEM, the strongest ever. It’s getting boring, almost. Also, at a +2.95 AdjEM gap, the 8-9 gap is the second-highest in history behind 2013, when 8 seeds went 2-2. Alas.

10 seeds

• Now here’s something: a good, but not elite group. A +16.6 AdjEM is not only below that of the 11s and near-equal to the 12s, it’s just the 7th-highest we’ve seen from this seed line. The group in 2023 was rated higher, and when everyone is doing really well efficiency-wise this year, this could be a group to fade.

11 seeds

• …oh boy. With all First Four groups included, this group averages a +17.5 AdjEM, the strongest ever. Because of this, the +8.94 gap from 3 seeds to 11s is only the seventh-highest ever, but it’s very similar to last year (when 11 seeds went 3-1).

12 seeds

• Onions! A +15.82 AdjEM on average, the strongest ever. Just a +4.43 average gap from the 5s to 12s, the third-smallest ever. And, to top it off, a gap between 4s and 12s that’s three points shorter than last season. Huzzah!

13 seeds

• …well, sorry. It had to end eventually. This batch of 13 seeds rates out at an average of +8.52 AdjEM, the 16th-highest on record and the worst group since 2019. That group did get in one upset, but none beyond the Round of 64. This is the largest gap between 4s and 13s ever.

14 seeds

• At a +5.47 AdjEM, this is only the 18th-strongest batch of 14s on record out of 23 total. Again: huge gap between them and the 3s.

15 seeds

• The good news: this group is better than last year’s utter slop. The bad news: at +1.98 AdjEM, it’s just the 14th-best in history and is the second-worst of the last seven Tournaments. No tournament with a +2 AdjEM average or lower has seen a 15 beat a 2.

16 seeds

• Brace yourselves. This is the worst batch of 16 seeds since 2004. When combined with the greatest 1 seeds we’ve seen in at least a decade if not further, well, my hopes aren’t high for upsets. (Average scoring margin in 1-16 games that year: +25 per game in favor of the 1.)

What does it all mean?

As I’ve said here for two months, this is a huge bummer if you’re into upsets whatsoever. And honestly, it’s even more stark than I figured it would be.

• The gap between 1 and 16 seeds is +42.86 AdjEM on average, the highest ever.

• The gap between 2 and 15 seeds is +26.84 AdjEM on average, the second-highest ever.

• The gap between 3 and 14 seeds is +20.96 AdjEM on average, the highest ever.

• Lastly, the gap between 4 and 13 seeds is +16.48 AdjEM on average, which is not the highest ever. Just kidding: it’s the highest ever.

This means that, barring a serious reversal of the numbers, we are in line for one of the driest, least upset-filled Round of 64s we’ve ever seen. The 5-12 line looks to promise some chaos, at least, and I won’t rule out the chance of a true shocker somewhere, but 2023 or 2021 or 2018 this is not. This is probably going to be one of the chalkier NCAA Tournaments of our collective lifetimes.

This is not really the fault of the 13-16 seeds; it’s the fault, if anything, of a superpowered upper echelon of college basketball that created an amazing regular season but will likely end up giving us few March Moments we crave. The average AdjEM of the 13-16 seeds this year is +2.31, which is actually a little better than last year…but that’s barely a win at all, because last year’s 13-16 seeds were the worst in Tournament history by their average.

Now, perhaps interesting here is that the closest Round of 64 by correlation score is 2011, which actually does line up pretty well: no 14-16 seed upsets, one 13-over 4, one 12-over-5, and three 11-over-6s. Beyond that is 2019, which went similarly: no 14-16 seed upsets, one 13-over 4, three 12-over-5s, and an 11-over-6. What I’m saying is this: keep your expectations low for moonshots.

For the Round of 32, the strongest correlations were 2015, 2002, and 2004 in that order. These are intriguing for a number of reasons, but mostly that collectively, eight of a possible 12 1 seeds made the Sweet Sixteen and there were an average of 3.7 upsets in the Round of 32. (We’re talking true upsets in the sense of a 5+ seed gap.) The closest comparisons all seem to have pretty similar paths: a boring Round of 64 followed by a banger Round of 32. We shall see.

Overall, the top comparisons across history, by our correlation score, are 2019, 2024, and 2007. These have some commonalities. All three had a 1 seed win it all, the pre-tournament #1 team at KenPom. (That’s Duke.) All three saw a lower-than-average number of upsets. All three saw at least three of the Final Four teams as 3 seeds or better.

Anyway, here’s my takeaways: this Tournament, on face value, looks exactly like it has looked for two months. Probably not that many upsets, probably a lot of really good second weekend games, and probably a few moments where you’re like “man, I wish that team would’ve won.” So: a pretty good Tournament. Let’s basketball.