I don’t think it’s a coincidence that you can’t spell “sticky” without “ick”. It’s a horrible feeling to have sticky hands, a body sticky with sweat, or even to eat something sticky when you’re not expecting that. But here’s the thing: both geckos and data scientists agree that “stickiness” can actually be a strength.
“Sticky stats” are statistics that tend to sustain their level over a period of time, and for that reason they are often fairly reflective of a player’s talent in that skill.
For instance, in baseball, batting average on balls in play (better known as BABIP) tends to sustain fairly well year-to-year for an individual and shows some combined effect of their hitting and baserunning ability. In football, the opposite is true of an individual touchdown rate (touchdowns per play), as they can fluctuate wildly year-over-year due to the quality of the defense they’re playing, playcalling, etc.
Many folks will say about an NBA prospect, “When his shot improves…” or “if he can just improve his rebounding”, but some of these qualities may not be that flexible. By identifying which stats are likeliest to sustain from college ball, we can get a benchmark for how players progress from the amateur game into the pros – figuring out which skills are static and which have a better chance to improve or evolve.
Which are basketball’s sticky stats, and what can they teach us for prospect valuation?
Stuck on You
First of all, let’s quickly define the process we used to analyze these statistics. That all starts with a little something called an R correlation.
If you’re not familiar with R correlations or the neat little number it spits out (the coefficient), the R correlation is simply a mathematical way of depicting the strength of a relationship between two sets of variables. While I’ll caution you to never confuse correlation for causation, this test can show us a benchmark of the strength of a statistical relationship. The closer the value of the R coefficient is to 1 or -1, the stronger the relationship between variables; the closer the value is to 0, the more chaotic or random the relationship.
So, I compared college basketball production rates per minute to the same player’s production in those same categories once they reached the NBA (e.g. college Total Rebounds per minute vs. NBA Total Rebounds per minute). After that, I assessed the average change in these statistical categories from college to the NBA so that we could get a sense of how much those skills developed (if they did) and whether they improved due to skill growth or declined thanks to tougher competition.
With that explanation of R correlation out of the way, let’s get to the data. The table below shows the correlation coefficients for a given statistic, comparing between college production and the NBA. Which per-minute rates tend to carry over from amateur ball to the pros?
There are a few clear relationships that do appear to cross into the “strong” category (assists per-minute and blocks per-minute), and a few moderately significant ones as well (total rebounds per-minute and 3-pointers per-minute, both made and attempted).
Let’s start from the strongest and work our way down: assists (passing the ball to teammates in position to score) and blocks (literally blocking an opponent’s shot) are two very integral skills to particular positions – guard and big, respectively.
Being able to distribute the ball tends to be a large reason why a guard is selected and earns playing time, as they tend to run the offense and handle the ball on most plays. The assist helps to capture a snapshot of their passing value and floor vision, which is hard to teach – or at least so critical for a player to already have when they enter the league that players with that skill already are prioritized.
The same goes for big men – forwards and centers, typically – and blocks. One of the primary responsibilities of a big is to defend, and elite shot blockers have physical traits like size and jumping ability as well as anticipation and body control to not foul the shooter. Those physical tools certainly don’t go away, and it may be that the instinct to react to another’s shot and knock it away is just very hard to develop.
Though it’s obvious, there are guards who are above-average blockers too (James Harden, Jrue Holiday, Matisse Thybulle) and bigs who are good at generating assists (LeBron James, Nikola Jokic, Giannis Antetokounmpo). The same skills do apply, and it’s even likelier that developing the skill later at their positions is a more difficult task.
Total rebounds seems to be a similar case to blocks (size, jumping, anticipation and reaction), and three-point shooting is something we sort of covered in the previous “sticky stats” article; if you are good at shooting threes in college, you likely retain that role into the pros (and if you shoot a lot of threes, you likely continue to believe in that shot).
Harder, Better, Faster, Stronger
The correlation isn’t the be-all, end-all of these skills, however; there is certainly the potential for a player’s skills and stats to change and grow and develop once they enter the NBA. The table below shows the average change in a skill’s production from college to the pros, listed in the form of a percent increase or decrease.
Surprising statistically – but perhaps not so much once we reason it out – is the drastic increase in both 3-pointers made and attempted per minute from college to the NBA. We saw a strong correlation between the college and pro rates in this category, so it seems like the increase or decrease should be minimal at best, but when we put into context how three-happy the NBA is, it should be no shock that on average trey attempts more than double from college. Important caveat, though: 3-point shooting percentage drops by about a sixth, which supports our theory on the increase being due more to volume than skill development.
Free throw numbers also drop by about a third on average, which can be somewhat linked to the proclivity to shoot instead of drive. Without contact on a shot, there’s no chance to draw a foul. More outside shots mean fewer shooting fouls, therefore, which means fewer free throw attempts – ergo, fewer free throws.
I also want to note that in every shooting category outside of 3-pointers, the makes drop by more than the attempts; the shots are still taken, but better defense in the pros means a slightly worse look on those shots. Points per minute played also likely ties into this in the inverse: the scoring dominators from college gain better talent around themselves in the NBA, which means there are more mouths to feed on offense. This helps a player’s assists go up and turnovers go down (due to fewer botched plays), but it’s harder to hog the rock in the pros.
There are certainly interesting things to learn from a basketball player’s progression out of their pre-NBA career to once they make the show, but some skills will tend to stick. Knowing which those are will help you set expectations for a prospect’s career trajectory and ensure that you’re building your dynasty lineups with the right information.