Solid Ground: Why Collin Sexton is Almost As Reliable As Kevin Durant

A statistical analysis of player consistency, from a certain point of view

Hopefully, I successfully roped you in with that headline, which, although slightly exaggerated, has statistical validity. Like Obi-Wan Kenobi, let’s take a look at things from a certain point of view.

What if I told you that RJ Barrett, who some seem to have forgotten about on the Raptors, is more reliable than Donovan Mitchell? What if, perhaps, Tidjane Salaun, the Hornets’ 6th overall selection from 2024, had a streakier season than Ben Simmons did? Or, to round out this set of three paradigms, what if Collin Sexton is almost as reliable as Kevin Durant, as I mentioned in the headline?

All three of these are true from a certain point of view, and that certain point of view is a fun little metric I’m calling SWSI:

S - Stability

W - Weighted

S - Scoring

I - Index

While it sounds fancy, it’s actually somewhat simple. One of the ways to measure consistency and reliability (from a scoring perspective) is to see how much a player’s points fluctuate from game to game. For example, a player who scores 10, 20, and 30 is likely less “reliable” than a player who scores 20, 20, and 20, because the latter is perfectly consistent and the former deviates from their average significantly from game to game.

However, there’s also the issue of giving everyone a fair playing ground to work with. Since some players get fewer minutes on some days and more on others, we need to be able to make sure that everyone’s opportunities are (statistically) the same. Hence, we’ll be using points per 36 minutes for this SWSI iteration, and we’ll throw in one last caveat: scoring volume matters.

Take, for example, two players. One scores 20, 20, and 20, while the other scores 30, 31, and 29. If we were strictly going off of standard deviation, the former would be more consistent than the latter, and there’s nothing inherently wrong with that. However — and this is where my opinion comes in — I’d posit that the latter player should be more deemed more reliable from a valuation standpoint when their scoring average is higher and their standard deviation is still extremely small.

Therefore, the final formula (for you formula junkies out there) looks like this:

Per36_SWSI = Avg_PTS36 / (PTS36_StdDev + 1)

But you don’t have to worry about that. The one-liner is that SWSI calculates a player’s reliability, taking into account their scoring volume. A higher SWSI equals a more consistent and reliable player, while the contrary is also true.

With the (likely too lengthy) explanation out of the way, let’s take a look at what the data says about some of the league’s best (and possibly worst) players…

P.S. All players here played at least 40 games in the NBA season, so as to further sift through outliers.