Revisiting preseason expectations
With the benefit of hindsight, how do some of my predictions look?
When I first tell people that I do a lot of data-driven basketball analysis, the first thing most people think of is that I must do a lot of sports betting or win a lot of bracket contests. Sadly neither of those things are true, because I don’t really think of the analysis I do as really trying to be especially predictive. College basketball is filled with small sample sizes, huge variability in opponent strength, and the unpredictability of 18-22 year olds. What I aim to do is use the data on what happened to tell a story of why it happened and what’s likely to continue, beyond the basic analysis you can read in any game recap. However, before every season I just can’t help myself in trying to predict what might happen based on limited data, some hunches, and anything useful I’m hearing. Sometimes it works out well and I look smart but, as you’re about to see, when there are a lot of unusual factors it’s all a lot of speculation.
Let’s revisit some of my thoughts over the offseason and see what I was right about, what I was wrong about, and what might still work out.
The importance of playmaking (and a lack of it)
In May, I wrote that UofL needed mroe playmaking:
even after their initial haul in the transfer portal. I noted that having 3 or more starters at an assist rate of 10% or more seemed to be a crucial benchmark for offensive success:
Mack teams had 3 or more starters with an assist rate of 10% or better in 10 of his 12 years coaching. The two exceptions were:
2021, when only Carlik Jones and David Johnson surpassed that mark
2013 at Xavier, when Semaj Christon and Dee Davis were the only Musketeers above that mark
Curiously enough, those are also the only two years in coaching where a Chris Mack team was not selected for the NCAA Tournament. Those two years were also the two lowest ranked offenses per KenPom, with the 2021 Cardinals ranking 92nd in offensive efficiency and the 2013 Musketeers at 116th. Lastly, those were also the two years when Mack’s teams ranked lowest in overall assist rate. The 2021 edition was 313th nationally, and the 2013 one placed 291st, per KenPom.
I noted that only Jarrod West and El Ellis seemed likely to be able to make plays for others, as Noah Locke had very low assist rates at Florida (<5%) and none of UofL’s returnees had any history of creating baskets for others.
Fast forward to today, where UofL currently has 3 starters above that 10% assist rate metric..but just barely. Assist rate reflects the percentage of teammate baskets you assist on, and West is at 22%, Dre Davis is at 11%, and Malik Williams is at 10.2%. As I expected, Locke is just under 7%. El Ellis and Mason Faulkner are at 20% and 32.5% off the bench, but most of the time Louisville is at a deficit of playmaking on the court. I ended that article saying that if Louisville doesn’t find one more playmaker “it won’t be pretty”, and so far it hasn’t been.
The McMains effect…uhh….
At the end of May I wrote about UofL’s new assists, Ross McMains:
I looked through data on how previous teams had played under him (New Zealand national team and Melbourne United), and made a few predictions for UofL:
Rank in the top 20 D-1 teams in tempo
Reduce their volume of midrange shots (2 pointers outside the paint) to less than 20% of total shots
Rank 250th or worse in % of made fields goals that are assisted
Rank outside the top 100 in offensive turnover rate
While UofL is just 119th in tempo per KenPom, they rank 33rd in possession length on offense after being 246th in that metric a year ago. The Cards are taking 23% of their shots from midrange, down from 35% a year ago but not quite to the level I projected. Louisville is just 168th in the rate of baskets that are assisted and 205th in turnover rate, although the offense hasn’t created as many driving opportunities as I thought.
I was bullish on Williamson and Withers in this offense because of their previous flashes of scoring in transition, at the rim, and in some jump shot situations, but that has just completely vanished this season. Overall, I was pretty far off in projecting how individual players would adapt. Maybe we’ll see some improvement as they get more comfortable, but I’ll take the L here.
Over/under predictions…boy, was I wrong
I also wrote a preseason column projecting some stat totals and predicting whether UofL would go over or under those:
Avert your eyes if you don’t want to think less of me. I projected:
Top 20 in KenPom tempo (not gonna happen)
Just oustide of top 30 in effective FG% (currently 221th, gulp)
Over 37% team 3pt FG% (currently 31%, 259th)
Under 40% of shots from three (currently 43%)
Jae’lyn Withers over 15 ppg (please don’t look it up)
El Ellis over 30mpg (he’s at 16 mpg)
JJ Traynor over 15mpg (I want to bury my head)
So, I went a solid 0 for 7, or maybe 1 for 7 if you’re EXTREMELY charitable and give me credit for saying over on 30th in eFG%. I expected improved shot distribution to mesh with capable shotmaking to really improve Louisville’s offense, but the shotmaking has been lacking. I thought UofL could become very good at scoring at the rim due to improved spacing, but they’re just average in that area. I also expected the returnees for UofL to lead the way due to familiarity, but I picked the wrong ones; Malik Williams and Dre Davis have been the only returnees to have a consistent role.
So, to recap, I was right about:
Potential offensive issues doe to lack of playmaking
Skepticism about Noah Locke
Optimism about Jarrod West
General offensive style (faster paced, less midrange)
I was very wrong about:
Offensive improvement
Pretty much every other player’s role
Literally everything I wrote about Withers and Williamson
I was way too optimistic that Louisville would score at a decent rate by improving their shot distribution, although I think everyone around the program (fans included) are probably pretty shocked that the Cards just can’t put the ball in the basket. The struggles in scoring at the rim and from three are pretty shocking, at the team at individual levels. I probably didn’t properly weigh my concerns about the lack of playmaking against the new offensive system that, ya know, requires playmaking. I wouldn’t be shocked to see some offensive improvement as the rotation tightens up, but 1 or 2 players from the Ellis/Faulkner/Williamson/Withers group needs to step up and show some ability to create off the dribble and get to the rim while not detracting from the excellent defense the Cards have built. If you want to know whether that will happen, ask someone who got some preseason predictions right!