The conclusions of the Science study are a textbook example of collider bias: by studying only elite samples (where selection depends on both early and adult performance), the authors induced a spurious negative correlation that does not exist in the general population. Within the elite sample: if an athlete had low early performance but still made it into the elite sample, they must have had high adult performance (otherwise they wouldn't be in the sample). Conversely, athletes with high early performance could be in the sample even with moderate adult performance. This selection mechanism creates an artificial negative correlation. See here for more: https://zenodo.org/records/18002186
I was surprised by this result, but your explanation makes sense of it. This seems like it was really avoidable if they'd taken a sample of elite adult performers within a certain age bracket, then measured earlier performance. You could still find gradations within the adult elites.
This sounds like regression toward the mean as explained by Thomas Kahneman in Thinking, Fast and Slow. The people who peaked had nowhere to go but down. The people that didn’t were able to improve
Very nice observations on Mozart, especially interesting when comparing him to another prodigy, Mendelssohn, who indeed did not necessarily live up to the hype he garnered in his youth. There’s a lesson to be learned there.
That said, I’m not sure that becoming proficient at one thing is a deterrent to learning new things. Personally I found having learned to learn once properly as being a catalyst to delving into learning again. And in fact it made me learn new things faster and more efficiently, as I had a grasp of the system behind learning.
I was wondering about Mendelssohn as well, but from a slightly different direction: because, unlike Mozart, Mendelssohn in his mid-teens was composing works that still retain a solid place in the repertory (the Octet, the Midsummer Night's Dream music).
Mozart is surely the greater composer overall (nothing controversial about that), but one might reasonably argue that Mendelssohn was a high performer - even if not at the Mozart level - who was not a late bloomer.
Exactly. I do believe Mendelssohn did indeed peak much earlier, which hints at him being more of a classical prodigy.
Schubert, like Mozart, also doesn't fit the bill of the prodigy (as laid out here), as there is a clear ascension throughout his life until he peaked in his late works.
Yes, that's right, though one should be clear that "peak" here doesn't mean a falling off subsequent to it. Mendelssohn continued producing music of that quality throughout his working life - the Violin Concerto and his last string quartet come late in his (short) career.
Sure, don't get me wrong, I'm not saying he fell off a cliff afterwards. But he didn't reach significantly higher heights after his teenage years, which can't be said for Mozart. I think we're on the same page here (and this isn't meant as a sort of "bashing" of sorts, more like an observation).
Was there ever a time ageism was not a factor in how people are generally assessed? In my own experience and reading, I suspect early specialization is a major factor, as described above. What else can you bring to your work if your experience resides primarily in your work? There's also the system of rewards, which you allude to briefly, which I would be more interested to read about. If a "prodigy" receives universal acclaim, what do they have to prove to themselves or anyone else? There's also the erosion of the ability to self-critique---if you're told early on that you're a genius, it becomes easy to believe that everything you do is a work of genius.
This isn’t remotely true in chess. I don’t understand why this paper is getting such traction — the best players in the world were already extraordinarily good at very young ages.
The example the paper gives of a chess ‘late bloomer’ is Vishy Anand, who was playing on the Indian (senior) national team at 15, was Indian national (senior) champion at 16, world junior champion at 17, and a grandmaster at 18. That is not late blooming by conventional standards.
Brooks describes the life arc of certain people he calls Late Bloomers. You can read about his examples at the link. There’s an important difference in the life trajectory between Brooks’ Late Bloomers and Gladwell’s Outliers. The latter begin early in life, stay focused and practice their brains out. Late Bloomers, by contrast, lack focus, they try and fail (often repeatedly) at many disparate pursuits, are not infrequently a**holes and f*ckups when young, and are acknowledged only later in life as exceptional at something (an age well beyond when most Outliers are sitting back counting their plaudits).
How would you go about arguing that this isn't just selection? That is, you frame it as "high achievers who bloom late tend to be more talented." But couldn't another interpretation be "in order to become a high achiever late, you have to be more talented than those who started earlier." A more pessimistic conclusion.
Similarly, young prodigies are an incredibly tiny fraction of the overall population, so even if most absolute top achievers weren't prodigies, that still leaves room for potential enormous benefit to being a prodigy. This naturally leads to an alternative interpretation of your explore/exploit point. If different niches have different productivity but may be hard to find, then the vastly more numerous not-prodigies might by chance have their luckiest members fall into more productive niches than prodigies.
The conclusions of the Science study are a textbook example of collider bias: by studying only elite samples (where selection depends on both early and adult performance), the authors induced a spurious negative correlation that does not exist in the general population. Within the elite sample: if an athlete had low early performance but still made it into the elite sample, they must have had high adult performance (otherwise they wouldn't be in the sample). Conversely, athletes with high early performance could be in the sample even with moderate adult performance. This selection mechanism creates an artificial negative correlation. See here for more: https://zenodo.org/records/18002186
I was surprised by this result, but your explanation makes sense of it. This seems like it was really avoidable if they'd taken a sample of elite adult performers within a certain age bracket, then measured earlier performance. You could still find gradations within the adult elites.
Ouch
So you're telling me there's hope?
This sounds like regression toward the mean as explained by Thomas Kahneman in Thinking, Fast and Slow. The people who peaked had nowhere to go but down. The people that didn’t were able to improve
Very nice observations on Mozart, especially interesting when comparing him to another prodigy, Mendelssohn, who indeed did not necessarily live up to the hype he garnered in his youth. There’s a lesson to be learned there.
That said, I’m not sure that becoming proficient at one thing is a deterrent to learning new things. Personally I found having learned to learn once properly as being a catalyst to delving into learning again. And in fact it made me learn new things faster and more efficiently, as I had a grasp of the system behind learning.
I was wondering about Mendelssohn as well, but from a slightly different direction: because, unlike Mozart, Mendelssohn in his mid-teens was composing works that still retain a solid place in the repertory (the Octet, the Midsummer Night's Dream music).
Mozart is surely the greater composer overall (nothing controversial about that), but one might reasonably argue that Mendelssohn was a high performer - even if not at the Mozart level - who was not a late bloomer.
Exactly. I do believe Mendelssohn did indeed peak much earlier, which hints at him being more of a classical prodigy.
Schubert, like Mozart, also doesn't fit the bill of the prodigy (as laid out here), as there is a clear ascension throughout his life until he peaked in his late works.
Yes, that's right, though one should be clear that "peak" here doesn't mean a falling off subsequent to it. Mendelssohn continued producing music of that quality throughout his working life - the Violin Concerto and his last string quartet come late in his (short) career.
Sure, don't get me wrong, I'm not saying he fell off a cliff afterwards. But he didn't reach significantly higher heights after his teenage years, which can't be said for Mozart. I think we're on the same page here (and this isn't meant as a sort of "bashing" of sorts, more like an observation).
Was there ever a time ageism was not a factor in how people are generally assessed? In my own experience and reading, I suspect early specialization is a major factor, as described above. What else can you bring to your work if your experience resides primarily in your work? There's also the system of rewards, which you allude to briefly, which I would be more interested to read about. If a "prodigy" receives universal acclaim, what do they have to prove to themselves or anyone else? There's also the erosion of the ability to self-critique---if you're told early on that you're a genius, it becomes easy to believe that everything you do is a work of genius.
This isn’t remotely true in chess. I don’t understand why this paper is getting such traction — the best players in the world were already extraordinarily good at very young ages.
The example the paper gives of a chess ‘late bloomer’ is Vishy Anand, who was playing on the Indian (senior) national team at 15, was Indian national (senior) champion at 16, world junior champion at 17, and a grandmaster at 18. That is not late blooming by conventional standards.
As a category I think Late Bloomers exist, but I prefer the distinction of Late Bloomers from traditional exceptional achievers (e.g. Malcom Gladwell's "Outliers") as drawn by David Brooks in his essay last year: https://www.theatlantic.com/ideas/archive/2024/06/successs-late-bloomers-motivation/678798/ (may be paywalled)
Brooks describes the life arc of certain people he calls Late Bloomers. You can read about his examples at the link. There’s an important difference in the life trajectory between Brooks’ Late Bloomers and Gladwell’s Outliers. The latter begin early in life, stay focused and practice their brains out. Late Bloomers, by contrast, lack focus, they try and fail (often repeatedly) at many disparate pursuits, are not infrequently a**holes and f*ckups when young, and are acknowledged only later in life as exceptional at something (an age well beyond when most Outliers are sitting back counting their plaudits).
How would you go about arguing that this isn't just selection? That is, you frame it as "high achievers who bloom late tend to be more talented." But couldn't another interpretation be "in order to become a high achiever late, you have to be more talented than those who started earlier." A more pessimistic conclusion.
Similarly, young prodigies are an incredibly tiny fraction of the overall population, so even if most absolute top achievers weren't prodigies, that still leaves room for potential enormous benefit to being a prodigy. This naturally leads to an alternative interpretation of your explore/exploit point. If different niches have different productivity but may be hard to find, then the vastly more numerous not-prodigies might by chance have their luckiest members fall into more productive niches than prodigies.