Statistics are one of the most divisive aspects of sports fandom (and elsewhere, but this is an NBA blog, so I’m shutting up and dribbling for now). It seems that the main issue with NBA statistics is that we as fans fall in love with particular ones. They become our cozy blanket for easy access in understanding; moreover, they’re great for simple conversations with coworkers, friends, and Twitter-folk. Regrettably, we forget that they’re nothing more than a man-made metaphor for understanding a complex game. Some become obsolete, and when that happens, we forget to replace them. Luckily, those replacements are floating around in the vast recesses of the internet. It’s time for NBA fans to uncover and use improved statistics.
Points, rebounds, and assists make up the trifecta of NBA statistics. They count towards triple doubles, impact fantasy basketball, and are easy to cite. Even Bill Simmons made up the 42-Club based on these three stats. However, taken out of context, they’re pretty dumb.
Toss Out Points Per Game
A seemingly simple fix for measuring scoring is to use points per 36 minutes or points per 100 possessions. While this makes sense theoretically, both still fall short of normalizing statistical comparisons across teams. The former still doesn’t adjust for pace, and the latter isn’t a realistic amount of individual possessions. Because of this, there has been movement towards using per 75 possessions for evaluating points (advocated by the likes of Ben Taylor and Andrew D. Bailey). Most starters will play about 75 possessions per game, so comparing scoring outputs will be more accurate. For example, the Hawks averaged 103.9 possessions per game vs. the Cavaliers 96.6. A player on the Hawks had many more opportunities to collect traditional counting stats.
Here are the top 10 scorers during the 2018-19 season by points/75 possessions vs. the typical points per game.
2018-19 | Points/75 | 2018-19 | Points/Game | |
James Harden | 36.2 | James Harden | 36.1 | |
Giannis Antetounmpo | 29.5 | Paul George | 28 | |
Joel Embiid | 29 | Giannis Antetounmpo | 27.7 | |
Stephen Curry | 28.8 | Joel Embiid | 27.5 | |
Kawhi Leonard | 28.1 | Stephen Curry | 27.3 | |
Devin Booker | 27.2 | Devin Booker | 26.6 | |
Kevin Durant | 26.8 | Kawhi Leonard | 26.6 | |
Kemba Walker | 26.8 | Kevin Durant | 26 | |
Paul George | 26.6 | Damian Lillard | 25.8 | |
Karl-Anthony Towns |
26.6 | Bradley Beal | 25.6 |
If you’re grumbling about this not causing any dramatic shifts, note that this is much more useful when comparing eras. Decades like the 60s were much faster than the early 2000s causing some incredible points/rebounds/assists totals. Considering the context of the statistics is definitely an improved way to analyze the game.
Get Rid of Rebounds Per Game

Because of the difference in pace among teams, rebounds per game also produce simplistic numbers. Instead, we should shift towards rebounding (both offensive and defensive) percentage. Simply put, this calculates the percentage of potential rebounds available a player grabs while on the floor. For example, a player with a offensive rebounding percentage of 10% means that he grabbed 1 out of every 10 possible offensive rebounds in his playing time. Here are the top 10 rebounders from 2018-19 based on rebounds per game and rebound percentage.
2018-19 | Rebounds/Game | 2018-19 | Rebound Percentage | |
Andre Drummond | 15.6 | Hassan Whiteside | 25.9 | |
Joel Embiid | 13.6 | Andre Drummond | 25.4 | |
DeAndre Jordan | 13.1 | Ed Davis | 25.2 | |
Rudy Gobert | 12.9 | DeAndre Jordan | 23.5 | |
Clint Capela | 12.7 | Rudy Gobert | 21.9 | |
Giannis Antetokounmpo | 12.5 | Joel Embiid | 21.4 | |
Karl-Anthony Towns | 12.4 | Enes Kanter | 21.1 | |
Nikola Vučević | 12 | Clint Capela | 20.8 | |
Hassan Whiteside | 11.3 | Domantas Sabonis | 20.8 | |
Russell Westbrook | 11.1 | Willy Hernangómez | 20.7 |
Viewing rebounding through this metric reveals that Whiteside grabs a higher percentage of rebounds than Drummond despite collecting four less rebounds a game.
Eliminate Assists Per Game

Unlike scoring and rebounding, using statistics to measure passing ability is a lot more difficult. Assists are super nebulous in that they don’t measure how effectively the pass helped the shooter. For instance, compare the value of these two passes from Gobert.
The first pass is a beautiful look-away pass for a Crowder layup while the second is a simple hand-off to Korver who dribbles three times, absorbs contact from Adams, and finishes. Both are considered assists even though the first is clearly a more valuable pass.
Ben Taylor’s box creation and passer rating stats are the two best statistics that I’ve found for replacing assists, but, unfortunately, they are only available to his Patreon subscribers. If people want a free replacement, stats.nba has a variety of other passing stats. My favorite of these is secondary assists which rewards the player who passes to the player who makes the assist. Stephen Curry is often near the top of the list for this stat.
There’s More Work to be Done
Breaking up our long-term relationship with the holy trinity of points, rebounds, and assists is a great step, but it’s not the end of our work. The NBA is filled with subtle nuances that haven’t been captured by numbers, but many have been, and thinkers like Daryl Morey exploit them to his advantage. If we can consciously stop using outdated and misleading statistical evidence in our conversations, the discourse surrounding the NBA will improve dramatically. We just need to embrace improved statistics. Or you can keep telling me that Kobe is the GOAT because he made a lot of tough shots. Whatever.
Follow Cody Houdek on Twitter for more NBA content and conversations. Check out his previous deep-dive on improved statistics.