A refrain I’ve increasingly heard from casuals, like the talking heads on the NBA’s biggest broadcast partners, is that the league is too homogenous these days, that every team just goes out there and chucks up threes. Every game is the same.
Leaving aside the shortsightedness of famous, influential media analysts dumping on their own product, it’s a common enough complaint that I felt an urge to investigate. My eyeballs tell me the league’s never been more diverse and fun, thanks largely to the absurd wealth of stars today with skill sets as bespoke as their draft night suits. Jokic’s Nuggets look nothing like Tatum’s Celtics look nothing like Brunson’s Knicks, and to suggest otherwise is willful ignorance.
As I wrote about for Basketball Poetry yesterday, the current top five offenses are Boston, Cleveland, New York, the Lakers, and Dallas. Those five teams rank first, 14th, 17th, 24th, and 21st, respectively, in share of shots that are triples. You certainly don’t need to launch a million threes relative to your peers to score the basketball successfully.
Of course, that’s a rather superficial analysis. Regardless of rank, what if the absolute numbers are all getting closer together? That would represent a growing sameness.
I looked at shot data from this season, 10 years ago, and 20 years ago to see how much the share of three-pointers had changed. I then looked at the fifth, 15th, and 25th-ranked team’s percentage of shots that were threes (this is to eliminate any outliers on either end). While the absolute levels shift up significantly (teams are obviously shooting way more threes overall), we see only a slight convergence from 2004-05 to 2014-15 and virtually none at all from 2014-15 to this point of 2024-25.
The difference between the fifth and 25th-most prolific three-point teams in each season:
2004-05: 11.7%
2014-15: 9.2% (slightly down from 2004-05, suggesting a slight convergence)
2024-25: 9.1%
(We’re only 15 games into the NBA season, so there’s always risk in the data changing as the season marches on, but three-point attempt rate is one of the stats that stabilizes the quickest).
To get more League at Large articles each day, plus team-specific analysis, fantasy basketball tips, and more, click here to sign up for your favorite team’s Locked On Newsletter!
Given that the rise of the three-pointer is exclusively at the expense of the midrange, we’re simply trading one jumper for another (give or take the occasional post-up).
But there’s more we can look at! Synergy has tracked playtype data for two decades. While their data isn’t perfect*, it’s still a pretty good way to see what sort of actions teams use to generate shots.
How much variety is there between different playtypes?
As an example, if a playtype is called P&R Handler, that means the play ended with an action from the ballhandler in the pick-and-roll, like a shot or turnover. It’s not comprehensive — especially today, many of these actions are combined — but it’s still a reasonable directional view of how the league handles offense.
Let’s first establish the lay of the land by looking at what kinds of actions teams are using:
(Note that these won’t add up to 100% in a year because I excluded offensive rebounds and the catch-all “miscellaneous” bucket, the latter for hopefully obvious reasons and the former because I’m more concerned with how a team generated its first shot.)
While it’s interesting to see how teams are going away from isolations and post-ups in favor of transition and spot-up opportunities, this was mostly just for background. What I really wanted was to look at the standard deviations within each playtype. A higher standard deviation would indicate a wider range of usage of that playtype within the league, hinting at more diverse offenses. Let’s take a peek:
Immediately, you notice there’s a huge range in how often teams finish possessions with their pick-and-roll ballhandlers this year, but there isn’t as broad a spectrum within post-ups compared to previous decades, likely because so few teams feature post-ups with any regularity.
2024-25 has the largest standard deviation within five of the nine categories, while 2004-05 leads in the other four. Peculiarly, 2014-15 leads in nothing.
When it’s taken all together, we can put the broom away, because there are no sweeping declarations to be made. This isn’t quite rigorous or deep enough to say with certainty which era had more or less heterogeneity, but it does make me feel comfortable that, at the least, these eras had a similarly colorful viewing experience.
Further, these playtypes, while useful, don’t distinguish between the differences within a playtype; people have filled books with different ways just to run a pick-and-roll. Offenses have become more creative and complex than in the past, building upon what came before. That’s the natural evolution of things.
So yes, whether you prefer your jumpers taken from 14 feet or 24 feet is largely a matter of taste. But to say that the team-by-team viewing experience today has less diversity than it did in the past simply because there are more threes taken, that every team plays the same way, is clearly untrue. The data here suggests that although plenty of things have changed, there isn’t a noticeable stylistic convergence from the past ten or twenty years, and that’s before we even account for the explosion of variations within each grouping.
So stop whining, Shaq.
To get more League at Large articles each day, plus team-specific analysis, fantasy basketball tips, and more, click here to sign up for your favorite team’s Locked On Newsletter!
*The data from Synergy could be subject to slight methodological changes over the years, for example. This data also only measures what caused a play to end; if a pick-and-roll led to a spot-up shot, it would count only as the latter, not the former. So it’s not perfect, but it’s still fairly reliable overall.
