The best of kenpom twitter in 2023-24
There is no place like twitter for measured discussion of college hoops analytics
We bookmark a lot of tweets during the season. Most of them have “kenpom” in them. It’s a good way to create a reference for review at the end of season. A lot of people are going to focus on the end-of-season ratings or the pre-tournament ratings, but every day of the ratings is important here. And so when someone pipes up and refers to the ratings being off in some way, we bookmark their tweet to see how things turned out once the season ends. What follows is the very worst of those tweets.
For example, we start with this fairly innocuous offering from November 29:
Comment: The top five offenses on November 29 (with a fair amount of preseason rating baked in) are currently 3rd, 6th, 7th, 2nd, and 14th (with no preseason influence). Not bad for data three weeks into the season. The ratings said those offenses would be awesome going forward and they ended up being really good.
And yeah, Tennessee ended up being pretty good as well, finishing 5th after an Elite Eight loss while Carolina finished 9th after a Sweet 16 loss. The ratings are predictive, kids, and there’s nothing stopping a 4-3 team from being a top ten team especially when the computer thought that team had top-ten talent to start the season. I mean, come on, the data was pretty great.
As we got deep into the season, the most popular kenpom slander was regarding Alabama. For example:
Comment: My man Justin has a very high bar for “terrible”. There were dozens of tweets like this over the course of the season and they’re fun to look back on.
Comment: Turns out Alabama was top 8 in its own league, actually finishing a game out of first. They were a decent team. Maybe even better than that.
Comment: Hey, a constructive observation on twitter. I could get used to this. Also, the AP poll can suck it.
Comment: There was nothing confusing about Alabama. How many years do we have to go through this? Adjusted scoring margin (plus a good and properly-applied preseason prior) provides a hell of an estimate of team strength. And also, we’ve long established that basketball isn’t played on spreadsheets. Thanks for reminding us.
Comment: Justin was still at it on January 20.
Comment: Oooh, me, me! I can explain! For one thing, Quad 1 wins are not a very useful predictor of the future and thus are not a factor in the ratings. Have I mentioned we are trying to make predictions here? For another, adjusted scoring margin is very useful. Alabama was very good at the latter and no so hot at the former. And look how it all turned out. Surely, MrRatable and Guy who bets on college sports will be more thoughtful in their comments next season.
Comment: At least it’s creative
.Comment: It was a hell of year for my ratings if that was the case.
Comment: In the early days of my site, Kansas fans were the largest group of subscribers. Things have changed a lot since then. Maybe it’s due to Jesse Newell moving on to covering the Chiefs. Maybe it’s due to twitter gradually sucking in general. But I was saddened by the number of bad Kansas takes.
Comment: I will never stop fighting for the fact that these ratings mean something 10 games into the season. Especially with all the preseason ratings baked in. Take a seat, Terry.
Comment: You sir, owe Kelvin Sampson an apology.
Comment: Well, Kansas didn’t end up being being way better than any of those teams. They might have been worse than all of them, though I’d probably concede BYU.
Comment: Oh hey, more Bama talk.
Comment: Would you settle for tied for 5th in the Big 12 with a kp ranking of 28th, 6th best in the league?
Comment: Finally, a quality twitter interaction. At about this time, we got some attention for writing about Memphis’s potential downfall despite almost universal belief that they were one of the best teams in the country. That turned out well. But buried in that piece were some words on Kansas, with more or less the same warning.
Comment: OK, some people started to figure it out. Speaking of a team that won the popularity contest this season…
Comment: I’m straying from my intended mission of focusing on kenpom-related tweets, but the Kentucky stuff is too good to pass up. I don’t want to be too harsh, but Kentucky has to be the biggest whiff of the eye test in quite some time. When the Wildcats’ offense was playing as the best in the nation they could hang with anyone, barely. When it wasn’t, they would struggle with bad teams and could lose to anyone. The defensive numbers were consistently mediocre all season. And yet…
Comment: I mean, those bigs weren’t a bunch of all-SEC guys. (And the SEC puts 20 guys on its all-conference teams.) What happened here? Is access to John Calipari so difficult that the national media needed to hype this team up all season? Is this where the eye test gets us in a year when Calipari coached the worst defense since his early UMass days? I need answers.
Comment: These are the wrong answers.
Alright, back to the kenpom tweets. There was like a week where Indiana fans were convinced they had a great team despite their poor rating on my site.
Comment: Apparently, Auburn was a six-point favorite over Indiana on December 9. The Tigers won by a score of 104-76. That’s 100k down the drain.
Comment: A random tweet representative of the furor over Memphis. Again, the AP poll can suck it.
Let’s close with the most tortured fan base in America this season.
Comment: The last time South Carolina was involved in a tournament run was in 2017, which may have predated the existence of modern analytics. Or maybe more likely, Cocksby90 and gamecock_trey were too young to remember much about that season. Alas, there was a contingent of Gamecock fans who couldn’t just enjoy their team’s surprising season and safe tournament bid. Unfortunately, the Gamecocks didn’t make anyone pay come tourney time.
This was doubly sad because I spent a few weeks during the offseason updating my preseason ratings method for our new transfer-abundant world. And one of the biggest outliers relative to others was South Carolina. After finishing #221 and having a player get drafted, my computer had them at #66 to start the season, higher than any known rating system. We were looking forward to making new friends. Maybe next time.
Comment: Indeed.