The Talent Distribution Question or "We need to talk about Jerry and Oscar."So the horse is out of the barn for this Top 100 but perhaps this may serve as food for thought for the next Top 100. I like era-relativity. In fact, I'm about to vote for George Mikan, which makes everything I'm about to say a bit odd I guess. But as we near possibly voting for Jerry West 14th and possibly Oscar or Mikan top 16 or 18, I can't help but wonder if we're really overrepresenting the guys from the First 25. Players whose careers were complete by the time we reached the NBA 25 year anniversary (ok, really 27th anniversary going to 1974, but functional primes done by the 25th anniversary). If all 3 make it consecutively, we will have 5 "First 25" players in the top 16. That's a lot. But of course, how a lot is a lot? After all, as pointed out, these are small sample sizes and prone to lumpiness.
So first, we have to acknowledge that the NBA talent pool has grown. We can roughly just look at the league having about 10 teams for the First 25, 20 teams for the Next 25, and 30 teams for the Last 25. Just taking team-seasons as a rough estimate, we would get that the First 25 represents 1/6 of league historical talent. I say lower, and we'll look into that, but we'll go with this for now. Earlier in the thread,
eminence took a crack at a distribution and came up with:
eminence wrote:Riffing on an f4p post from last thread - how would folks estimate Star+ talent distribution across the decades?
Lumping the incomplete decades in with the others, when guys entered the league, and rounding to the nearest 4% for now, my own estimations:
'47-'59: 4%
'60-'69: 8%
'70-'79: 8%
'80-'89: 16% (feels like a very talented decade that was a a leap from those prior)
'90-'99: 16%
'00-'09: 20%
'10-'23: 28% (benefitting from the extra years a bit, if I'd stopped in '19 I might have split off 4% somewhere else)
*Noting that the '10-'23 period will be less prevalent on a top 100 careers list like we're doing due to lack of longevity vs lacking talent.
Splitting the 70's in half and going to 1974 when Oscar and West retired, we get 16%, or about 1/6.
eminence was doing talent distribution but I think we can look at it as a good proxy of "For players whose career ended in year X, how much of league talent did their era represent?" Reasonable minds will disagree, but the first guess out of the box agreed with the back of the envelope. Granted, careers couldn't really "end" by 1947 so we're probably overrepresenting the First 25 again, but this is all approximation so let's go with it.
So assuming that was the distribution, how likely is it that we would see 5 out of the top 16 from the First 25? Well, doing a calculation where I just have excel generate 16 random numbers between 0 and 1, and then assign them to an era based on the cumulative historical talent to that point (i.e. 0.358 would line up with the cumulative total of 36% up to 1989 and be assigned to a career ending in 1989). So doing this 1000 times, we get an average of about 2.5 top 16 players from the First 25. We get only a 79 out of a 1000, or 7.9%, chance of seeing 5 top 16 players by random chance. Unlikely, but not necessarily statistically significant.
But what if you are like me and think we are probably at best talking about 10% of historical talent? Well, let's first take a stop by the 12.5% guess along the way (1/8 of historical talent by 1974):
That's a breakdown of:
'47-'59: 2.5%
'60-'69: 5.5%
'70-'79: 9.5% (4.5% by 1974)
'80-'89: 13%
'90-'99: 17.5%
'00-'09: 21%
'10-'23: 31% (trying to boost for international expansion)
Now we're already down to 1.9 Top 16 from the First 25 and only a 3.1% chance of 5 stars by random chance.
But what if we say all these guys aren't getting in at the Top 16. Maybe it's Top 20.
That's 2.35 Top 20 from the First 25 and 8.5% chance. so at least we're back to "not statistically significant" if they're only Top 20 and not Top 16.
But what about 10%?
that's a breakdown of:
'47-'59: 2.5%
'60-'69: 4.5%
'70-'79: 7% (3% by 1974, could probably ding the 50's harder and boost the 70's a little, but the 10% cumulative is what matters)
'80-'89: 14.5%
'90-'99: 18.5%
'00-'09: 21.5%
'10-'23: 31.5% (trying to boost for international expansion)
Now we're down to 1.5 Top 16 from the First 25 and only a 1.1% chance of 5 stars by random chance.
But what if we say all these guys aren't getting in at the Top 16. Maybe it's Top 20.
That's 1.8 Top 20 from the First 25 and 2.6% chance. so back to "statistically significant" even if they're only Top 20 and not Top 16.
I think there's a very strong chance we're overrepresenting the First 25.
But what if we're not even scratching the surface of how different the talent pool was in the past?
What to do with those plumbers or "How much can we talk about race?"So why do people make the plumber joke about the early NBA? Well, frankly, it's because they see a lot of white people. All white people in fact for the very early NBA. Heavily white well into the 60's. And well, they don't see that today. Certainly not American whites. The NBA was essentially 100% American back in the day (I'm sure there were exceptions, don't @ me) and started off all white. Based on census data for 1960 (we'll start in Jerry and Oscar's time), the population was ~180 million. Now it's 330 million. So multiply by 1.83, account for the 25% of the league that is international, throw in some increase due to money, and maybe we get a 3x-5x increase in the talent pool?
But based on an article from 2016, there were only 42 white Americans in the league in 2016. Take 450 players, lop off 120 give or take for international players, and that's 42 out of 330, or about 1 out of 8 for American players. A 100% white league is now 12.5% white from the American part of the talent pool. The census says there were ~160 million white Americans in 1960. There are 47 million black Americans today. So a population less than 1/3 of the effective talent pool population from 1960 is supplying something like 7 out of every 8 players today, in a league with almost 4 times the roster spots. The white population of America has only grown since 1960, up to over 200 million, and yet it is effectively cut off from the NBA by 47 million people. The previous leagues weren't just drawing from a smaller talent pool, they appear to have been drawing from the wrong talent pool. My 10% estimate from before is probably not even close to as severe as we should be.
So where does this leave us or "Should we talk about Babe Ruth?"The short answer is "I don't know". I probably believe the "severe" case is more true than not, but is that how we want to do an all-time ranking? Babe Ruth put up enormous numbers playing in a segregated league. Of course, black players never came to dominate MLB. In fact, they seem to often lament the lack of black talent in the majors. And the influx of Latin American talent seems to be less about integration and more akin to the NBA's international expansion. But setting aside the fact the situations aren't necessarily analogous, what if they were? Could we tell the history of baseball without Babe Ruth? I don't believe so.
Would an all-time list be fun if we just wiped away a huge chunk of league history? Probably not. But I think it's worth keeping in mind and possibly making some updates. And I think, more likely than not, that 5 Top 20 players from such a radically different league with a radically different talent pool is probably exceedingly unlikely. I have a soft spot for Mikan as the NBA's Babe Ruth to some degree and so I suppose I will vote for him in this thread. But two guys who did not dominate to even remotely the same degree, from only about a decade later, guys who aren't necessarily nearly as integral to telling the entire story of the NBA, the 3rd and 4th best players from their era? Now that I've thought about it more, I'm not sure I can really see them as all being worthy of a Top 15-20 slot. A reexamination may be in order.
