What does data tell us about ageing in the modern NBA?

As many top-level sports grew and matured from their early days of bus rides and humble paychecks into big, multi-billion dollar business over the last few decades, so has sports science. Nowhere has this been more evident than in the NBA, a league where each team's success depends on a few star players more than any other team sports around the world.

To keep his body in peak physical condition, LeBron James is said to employ a team that would be the envy of royalty. His team is said to include a biomechanist, recovery coach, masseuses, and personal chefs, while each of his houses include a hyperbaric chamber to aid in recovery. With James still performing at MVP-calibre levels at age 35, it's hard to argue with the results while legends like Larry Bird retired at the same age.

But, is this true across the board? Let's take a look at the data to see what we can learn about aging in the modern NBA, from its best players, the make-up of the league in general and players' longevity and performance.

How old are the league's best players?

This chart shows statistics collected from 5-year periods (e.g. 1986-1990, or 2016-2020), plotting each player's BPM (box-plus-minus) statistic against their age. No one statistic is perfect to describe a player's performance, but BPM is as good a measure as any in capturing how a person has performed in relation to the rest of the league.

To concentrate on players who have had positive impacts, the chart only shows players with BPM stats of over 2, which translates to approximately 50 players per season.

More and more top NBA players are playing better, for longer.

Two trends are clear. The earlier years' scatter plots are far more "pointy" at the top, players with 4+ BPMs (all-star calibre players) being concentrated in ages 23-28. More recently (to the right of the page), age is less of a determining factor in individual performance levels and the data is more spread out.

In the years from 2016 to 2020, a number of players (Pau/Duncan/Ginobili/Dirk) have recorded 2+ BPM seasons at ages of above 35 (see bottom right corner of each graph). A 2+ BPM season is designed to be one for a "good starter", and these seasons being had for players aged 35+ was more or less unheard of in 1986-1990, except for Kareem's one last great season in 1986.

Is the modern NBA a young man's game?

Before we look at anything else, let's take a look at the age profile of players in the last (2019-2020) NBA season. The chart below shows the distribution of the players' ages in the NBA.

The majority of NBA players in 2019-2020 are aged between 22 and 26

Although the chart stretches all the way to 43 (congratulations to Vince Carter on an amazing career), the most common age for an NBA player is actually 22. In fact, more than half (55%) of the league's players are aged 25 or under. It's a young man's game indeed.

Now, let's look at whether that's always been true by taking a journey through snapshots in time, starting in 1990. Each subplot as we move down in this chart is a decade further in time than the previous subplot.

NBA players have been getting younger on average over time

As we move forward in time (and downwards in the chart), the columns showing percentage players at each age essentially shift leftward. This tells us that the proportion of young players breaking through has been getting bigger and bigger. The next chart simplifies the data into a histogram.

There are so many more players aged 24 and under in the NBA than ever before.

Each group shows percentages of players by age, and each colour bar shows each season that the data comes from.

The clear takeaway here is that the if you are a modern NBA player at 30 years or older, your job is increasingly under threat from under-24 year olds.  

But of course, this has a lot to do with the changes in rules relating to recruitment and who is eligible. It doesn't necessarily mean that young players coming through have gotten better. So next up, let's take a look at more nuanced stats rather than just statistics about who's in the league.

Youth vs Experience - Who's on the court?

The analysis above simply looked at who is on the roster. But it's a very different thing sitting at the bench all game to get maybe 5 mintutes of playing time a game, and getting 30-35 minutes a game as a starter. So in this section, we take into account how much time each players spends on the court.

Plotting players' playing time and grouping them by age, the below chart emerges.  

Older players typically get more playing time than younger players

Each column is made up of each players' playing time as a proportion of every minute played in the league. So a player who gets more minutes is marked in a more blue / darker colour, as well as taking up more height in each column. On the other hand, players who barely play might not contribute to a column, and are marked in a light colur.

When compared against the very first chart that we looked at (shown again below for reference), we can see that older players generally get more playing time per player than younger players.

NBA players by age (2019-2020 season)

At this point, let's expand our analysis of playing times to plot the data over time.

NBA players' have gotten younger, but also each player plays fewer minutes.

This chart is probably best described in two words - "load management". A trend that started in the late 90s and early 2000s with the San Antonio Spurs, and one adopted widely across the league now to minimise the wear and tear on each team's key players.

More specifically, the playing time chart for 1990 is covered with dark black bars, indicating high concentrations of playing times for individual players. Over time, these bars reduce in frequency to a point where hardly any players in 2020 are shouldering burdens that they used to in the 90s.

As a result the playing times look far more equitable across age bands, as well as between each player.

Still, this is a lot of data to digest and look at. Let's take a look at the data at a more granular level.

Ages of Stars vs Role Players (and bench-warmers)

It is undeniable that there are different groups of players in an NBA team from its stars to its bench players.  Accordingly, let's separate the data into groups for this analysis.

Here, I broke down the data to three "tiers" of players based on their minutes played - the rough theory being that simply better players get more minutes. The top tier represents on average top 2 players per team, second tier players 3 through to 7, and the remaining players being tier 3.

This data is plotted here for the 2019-2020 season.

Older players are over-represented in those playing heavy minutes

Even though the make-up of the NBA is dominated by players aged 26 and under, the majority of those players playing significant minutes are actually aged 27 and over.

We'll come back to why in a minute. But before we do that let's take a look at the historical data to see if that's always been the case, by comparing the same data from 2020 to those from 1990 and 2005.

The "tier 1" players are spread out far flatter across age groups in 2020 than in 1990.

This chart really tells a story that in 2020, each team's tier 1 (best) players are just as likely to be at any age between 21 and 30, whereas in 1990 they were most likely to be between 23 and 28, and really between 26 and 28.

And the tier 3 player group in 2020 has a longer tail of players aged 28 and over, whereas in 1990 they were few and far between, especially once they hit the magical 30 mark.

So, this chart shows that although players are on average younger, more and more players who are talented enough and good enough are able to extend their careers.

We marvel at Vince, Dirk and Kobe who have had 20+ year careers. I wonder if in 20 years, or even 10 years' time, how rare we would consider those career lengths to be.


But just before you go – if you liked this, say hi/follow on Twitter!

JP Hwang

JP Hwang

Data, numbers and analytics geek. Moreyball enthusiast.