NCAA Weekly, Vol. 9.1: Yes, the 2022-23 college basketball field is unusually weak
but mostly at some particular upset-friendly spots
There’s been a lot made so far this college hoops season about how there’s no such thing as a Best Team; it’s more a collection of some pretty good ones masquerading as great sides. Look basically anywhere and you’ll find it. Ask five media members (preferably not AP Poll voters) who their number one team is this year, and you may well get five different answers. While I’d like to figure out the answer to the question of “why is it like this” before the end of the season, I think there’s another question worth discussing: is this the weakest college basketball field in recent history?
This has begun to pop up more and more as of recent, as people have taken note of an unusual amount of losses by ranked teams and a general lack of obvious title favorites. Per KenPom, there’s a breakaway pack (of sorts) of seven teams; following them, there’s a five-point efficiency gap from 8th to 26th. It feels like One Of Those Years, if you remember 2011 or 2018 or even 2006.
The good news for research purposes is that KenPom dot com has daily ratings dating back to 2011-12, with ratings for every season starting on February 1. We’ll check back in on this again at that time, but I’d like to offer three graphs to you. I’ve taken the ratings of each top 50 team over the last 11 seasons (2019-20 excluded for obvious reasons), which is just a rough way of showing the quality curve of sorts of the eventual Tournament field.
The below is a chaotic showing of where the 2022-23 group ranks in relation to every group to come before it on previous January 15s:
And then there is this, which shows you where yesterday’s ratings are in comparison to the historical average on this day:
And then, finally, there is this:
Let’s break down what these graphics mean, in order of 1-to-50.
The top 38 teams are all below the historical average at this time of year. I mean. That kinda sells it better than anything else, doesn’t it?
However, the top 5 teams are not the weakest ever. They are merely the third-weakest in this 11-year data set at an average AdjEM of +27.75, which is ahead of 2016 (+26.93) and 2014 (+26.54). Possibly worth noting that those two years featured Final Four teams with seeds of 7, 8, and 10.
The rest of the top 10, however, is the weakest to date. The average AdjEM of this year’s top 10 is +25.36, which ranks below 2014’s +25.67 and sets a new low for the data set. If you just want teams 6-10, it’s the worst average AdjEM by over a point at +22.97. The next closest is 2015 at +24.19, which…had a 7 seed in its Final Four.
The top 6, while weak, ranks out relatively okay. All of the 2022-23 top six avoid being rated out as the worst ever at their respective ranks, though sixth-place UConn came awfully close.
The problem: everyone ranked 7th through 21st offers the worst efficiency margin in at least 12 years. Everyone from 7th-place Kansas to 21st-place Arkansas is the worst team we’ve seen at that ranking in this data set. 8th-place Saint Mary’s, for example, has the worst efficiency margin by an 8th-place team by a full two points.
Among the 2022-23 field, an astounding 19 teams rank as the worst ever at their given ranking. That includes 7th through 21st as well as 24th through 27th.
However: the field gets stronger the deeper we go. This is where it gets weird. As mentioned above, teams 1-38 are weaker than the historical average. However, teams 39-50 are all stronger and edge closer to the historical max the lower we go.
The most similar year, using a basic similarity score, for each grouping:
Overall: 2018
Top 5: 2013
6-10: 2018
11-25: 2018
26-50: 2013
Because we still need more data - and because we have a lot of time left in the season for this to change - I don’t know that I’d be ready to take this too seriously just yet. However, I think we’re headed towards a uniquely strange and potentially pretty dangerous (for 3 seeds and below) tournament, based on what we’ve seen so far. Really, if you look at those similarity scores and how well they’ve honed in on 2013 and 2018, it gets interesting very quickly:
2013: 7 first-round upsets (seed difference of 5 or greater), 3 Round of 32 upsets, Final Four of 1, 4, 4, and 9 seeds, Elite Eight with 3 of potential 8 1 & 2 seeds, 15 seed beat 2 seed in first round
2018: 5 first-round upsets, 5 Round of 32 upsets, Final Four of 1, 1, 3, and 11 seeds, Elite Eight with 3 of potential 8 1 & 2 seeds, 16 seed beat 1 seed in first round
Both of those tournaments ended with a 1 seed winning it all, as is standard, but they also featured some pretty wild paths to get there. To go with this, 2022-23 is one of just four seasons in the 11-season data set with one or fewer +30 AdjEM teams. The others: 2015-16, 2013-14, and 2011-12. I don’t think we’re heading for a 2013-14 UConn event by any means, but the concept of this being a very weird season heading towards a very strange NCAA Tournament feels right. As usual, it all starts at the top, where the presumption that this is an unusually weak year for favorites and big boys is more or less an accurate take.
EDIT, January 17: Based on a request from Bart Torvik, I ran the numbers on his site as well. A similar story abounds - generally well underwater what we’ve come to expect - but with the note that #1 Houston and #2 Tennessee are actually better than the standard at their positions. We’ll see how long that lasts.
