Reference Series: Age

Back in January, I did a piece with Don Yates about the one-and-done rule. Buried in that piece, I wrote a section on age and development. I have referenced it so often since then that I am separating it out into its own piece here and expanding on it.

NBA draft prospects are typically discussed and separated by their college year. This is, quite frankly, a terrible system. Players in a given college year can be born more than a full year apart. This season, Markelle Fultz was born in May 1998 and Josh Jackson was born February 1997. Both are freshman. John Collins was born in September 1997 and Luke Kennard was born June 1996. Both are sophomores. Yet John Collins, the sophomore, is younger than Josh Jackson, the freshman. So why even bother talking about their college year?

I wanted to see if maybe there was some correlation between college year and future performance. I looked at it in two ways. First, I looked at the age and year top players (I used BPM and WS/48 to make this list, you may use a different method, but I think this is representative) entered the league. This was my “star search” – were superstars more predictable by age of first NBA season or draft year. Here were my results:

LeBron HSS 19
Durant Fr 19
Giannis Intl 19
Davis Fr 19
Westbrook So 20
Harden So 20
Paul So 20
Jokic Intl 20
Kawhi So 20
Lowry So 20
Cousins Fr 20
Conley Fr 20
Towns Fr 20
Hayward So 20
Jordan Fr 20
Wall Fr 20
Curry Jr 21
Gobert Intl 21
Griffin So 21
Butler Jr 22
Thomas Jr 22
Green Sr 22
Lillard Sr 22
M. Gasol Intl 24


4 19 year olds, 12 20 year olds, 3 21 year olds, 4 22 year olds, 1 24 year old. 1 high school senior, 7 freshmen, 7 sophomore, 3 junior, 2 senior, 4 international. Both year and age indicate that top players enter the league sooner rather than later, which is to be expected as most top players are identified early and come out early. There are many more players who play their first season at age 20 than age 19, so that could explain the difference, but looking at the four players on this list who entered at age 19, there is a common thread that points to something very interesting.

What do LeBron, Durant, Giannis, and Davis have in common? They are physical freaks of nature, guys who would have been top prospects by dint of physical attributes alone. Three of them were also considered generational prospects. Now, there’s too many variables to try to isolate any given variable, but I have often argued that prospects who enter the league young are judged before they have fully developed and are not given the opportunity they should be given. These four guys were going to be given every opportunity to succeed, and succeed they have.

A few years ago, a study was released regarding how hockey players born earlier in the year were given preferential treatment in youth hockey and in the draft even though there was no evidence that players born earlier in the year actually went on to be better players in the NHL. You can read about it here, here, and here (and elsewhere). Basically, youth hockey has a Dec. 31/Jan. 1 cutoff and when you’re 9, 10, 11 years old, the difference between being born in January and being born late in the year is massive. Players born earlier in the year are better simply by virtue of being older, and they are given more chances, better instruction, play against better competition, etc., while younger players are forgotten or left behind.

Looking at the above table, I wonder how many players drafted very young never developed into full-blown superstars because they were simply too young and were not given full opportunities to develop or shine. Maybe they weren’t ready emotionally or mentally. Maybe it’s just random noise, given that the NBA was dominated by guys like Kobe and Garnett for awhile, but at the same time, it was understood that straight from High School guys would take longer to develop than college players. It’s at least something to keep in mind when evaluating both draft prospects and young NBA players.

The second test I ran was looking at a wide range of players to see if there was any kind of pattern in terms of NBA growth. This included looking at stars, scrubs, players who entered young, players who entered old. I looked at about 50 players to try to get a good cross-section. Sometime during the upcoming season, I’m going to try to do a more detailed analysis, but it’s going to take time I don’t have right now to aggregate data, develop formulas, implement formulas, etc. So eying the data will have to suffice for now. Here is what I found:

  1. The age in which a player entered the league had no effect on growth curve. I can’t even confirm there was a higher rate of injury. It really looks like, as far as overall growth is concerned, it doesn’t matter what age you play your first NBA game.
  2. Pretty much without exception, players were bad before age 20 (there were a few age 19 backup quality seasons), took a leap at age 20 or 21, took another leap at age 22 or 23, and then basically all players showed either consistent growth through the late 20s or took another leap some time between 24 and 28.
  3. Age growth curves held steady among stars, midlevel guys, low level guys, HS guys, one-and-dones, multiple college years, and foreign guys.
  4. Age growth curves held steady across positions.
  5. Players occasionally had an “adjustment year” (year below what would be expected based on their age and the rest of their career) their first year, but it was not common and did not appear to follow any pattern.
  6. If you like using BPM, players with a BPM of -2.0 or better as a 19 year old typically go on to be stars, players between -2.1 and -4.0 have mixed outcomes ranging from 4th or 5th starter to washout (and there does not appear to be a pattern in that range – a -2.5 and a -3.5 appear to have the same probability of outcomes), and players below -4.0 are basically backups at best for their career (in part possibly because guys who play that poorly often stop receiving opportunities).

Now, I’m not going to say this is a perfect fit for every player, but as a prognosticator, I’m looking for high probability patterns, not 100% perfect fits, because they don’t exist. There will always be exceptions. But I’m comfortable using this as a projection fit for now.

In general, teams and fans are too quick to judge a player’s future value based on age 19-22 seasons without proper context and that do not represent a player’s peak talents. There is a bad tendency to look at what a guy does early in his career and use that as a baseline of his true talent level, when true talent level really is not established until age 23 or even later. So you have guys who end up with bad reputations (D’Angelo Russell and Tyus Jones come to mind as recent examples) despite doing absolutely nothing wrong and end up regarded far less than they should be because they just entered the league too early, and with guys who are not drafted particularly highly, often their careers end up on the wrong path for no reason other than being too young when they entered the league.

And it hurts everybody. Older (21-23 year old) players coming out of college are often not evaluated properly and are downgraded just because they weren’t highly touted and had a spikier growth curve. Younger (18-20 year old) players coming out of college with a lot of hype but before they’re really ready get tossed to the side before they’re fully formed and do not end up developing the skills they should have been, leading to a higher fail rate. The only players teams are really identifying well right now are true can’t miss guys.

Teams are discarding young guys for not being ready at an age nobody’s ready at, and teams are significantly underdrafting, underplaying, and giving far too few looks at older guys who had big talent spikes a little bit later than guys who had flatter growth. Teams are missing talent on both ends of the age curve due to misguided ideas (or, just as likely, a lack of thought altogether) about what players can and should look like and what improvement curves look like.

When you’re looking at prospects in this draft and around the league, put them in context. It can only help.

Leave a Reply