The 10,000-Hour Rule for expertise popularized by Malcolm Gladwell is a funny case of citation telephone. Basically, the rule claims that anyone can become an expert if they spend ten thousand hours practicing. In other words, ten thousand hours of practice is necessary and sufficient to become an expert. It’s attributed to research by Anders Ericsson, the psychologist who pioneered the study of deliberate practice.

An interpretation which Ericsson published a three-page paper refuting. (You can read the whole thing here.)

The best violinists averaged ten thousand hours of practice, compared to their less accomplished counterparts who “only” averaged five thousand. (Keep in mind that Ericsson studies expert performance – he wants to know what makes people excel, so he’s not studying average populations.)

According to Ericsson, Gladwell claimed that 10,000 hours was the “magic number for true expertise” based on Ericsson’s work. This misrepresentation won popular appeal and spread. In internet mythos, the average is now misrepresented as a threshold after which people magically become experts. (I pity the poor stats professors weeping into their pillow at people abusing averages.)

This is another example of why you should take pop sci stuff with a grain of salt. The 10,000-Hour Rule was coined and popularized based on a game of citation telephone tracing back to a misunderstanding of a statistic. The guy it’s attributed to denies that he ever said or intended to convey that message. That should be enough to warrant a dollop of skepticism for the next bright claim to come along. 

Still, if the 10,000-Hour Rule is an urban myth, is there still anything to learn here? I think so.

1.

Becoming world class takes a heck of a lot of time. Ten thousand hours is twenty hours of practice a week for ten years. That’s how much the best violinists averaged by the time they were twenty, and it’s not like they stopped there. They probably continued that rate of practice during their actual career. In another example, winners of international piano competitions had around twenty-five thousand hours of practice under their belts.

Even assuming significant heterogeneity, it will probably take a lot of time to compete at the highest levels in a standardized field such as sports or music. If you can find niches where the competition doesn’t practice as much, you might be able to stand out with much less effort. For example, Ericsson found that students reached world class levels for memorizing long strings of numbers after only five hundred to a thousand hours of practice.

Either way, you should probably expect to measure your success across years. I’ve only spent about four hundred hours blogging, so I shouldn’t be surprised that I have oodles of room for improvement left. So don’t be disheartened if you only see tiny motes of progress on the scale of weeks.

2.

Raw time also isn’t sufficient. You also need to use that time effectively, which Ericsson claims is best done via deliberate practice. As someone who half-heartedly practiced piano for two years to appease my mother, I can attest that just putting in hours does not promise results. Deliberate practice tries to maximize learning per effort invested.

Importantly, deliberate practice commonly involves learning new and better methods, rather than repeating what you already know how to do. For example, one study subject started out able to remember a string of seven numbers. By learning better memorization techniques he got that number up to 18, more than doubling his original length. But then he struggled to continue learning. He had to learn even more memorization methods before he could break that plateau – and remember up to 79 numbers!

Deliberate practice is a slippery term, but tends to be characterized by:

• Focused practice on a specific subskill within the larger skill you want to improve. E.g., Ben Franklin improved his writing by transforming essays into poems to expand his vocabulary.

• Feeling pushed near the edge of your capabilities. You usually need to think hard about how to improve, learn challenging new material, or carefully practice something outside your comfort zone. E.g., practicing a tricky piano fingering until you can do it fluidly.

• Clear, rapid feedback on how well you did and what you need to improve. This can be self-directed, but ideally comes from a 1:1 instructor. E.g., an experienced cook telling a beginner cook what spices their dish is missing.

I suspect that both the expert instruction and intense focus benefit deliberate practice by helping you see ways to improve that would have slipped by unnoticed otherwise. Note, deliberate practice isn’t just about finding new methods – you also need to deliberately practice those methods until they become routine and your whole process works better.

Similarly, by focusing on one small skill at a time, you can give it the attention needed to find and practice ways to improve that piece of the overall skill. For example, I’m experimenting with Obsidian right now, because, for some people, a great note taking system acts as external working memory and brings new insights to their writing. If it works out, I might see a sudden jump in writing quality. Whereas I expect that just writing more will have a gradually diminishing learning curve. (Unless, of course, I learn new methods for writing.)

3.

In his paper, Ericsson addresses an interesting debate. Basically, some critics accused him of denying that genetics had any effect on success. Ericsson responds that he never argued genetics play no role, just that you can improve a lot with practice, particularly if you begin as a child.

What’s surprising here is how strong a claim Ericsson is still making. He believes you can change virtually every aspect of the human body and nervous system (except height and body size) with sustained intense training.

I actually had the chance to talk with Ericsson and some other positive psychology researchers about this topic a few years ago. The message I walked away with was that no one there denied that a ceiling probably exists on how good a particular person can become at something. They all accepted this was true.

It just didn’t matter.

They didn’t care about ceiling effects because it was like someone complaining that they could never beat Michal Phelps, so they never learned to swim. Sure, there is a ceiling there that they probably can’t beat. But they are so far from that ceiling that it just seems silly to care. If instead they focused on what they could do, they could probably become a pretty good swimmer.

Ericsson also writes about multiple domains where there seemed to exist a learning plateau, and then the person overcame it by learning new methods. E.g. people learning Morse code seemed to hit a limit, but better teaching methods were able to help them overcome it.

From that perspective, even the experience of feeling like you’ve hit a ceiling is suspect. A better teacher or new method could unblock even an already skilled person from improving.

Within effective altruism specifically, the ceiling people seem to worry the most about is intelligence.

I speculate that this is a backlash against IQ deniers. The internet spits out articles about how IQ is a myth and anyone can succeed if they just try. This causes people like Scott Alexander to write a bunch of articles defending IQ. Effective altruists see all this evidence -- and care about believing true things. (It probably doesn’t hurt that our monkey brains also notice these beliefs signal ingroup solidarity…)

This has the unfortunate side effect of pushing some effective altruists toward worrying that they aren’t smart enough. Then they obsess over a fixed trait and scramble for nootropics. (The resulting depression and anxiety also aren’t helpful.)

Which is rather strange. Scott has the excellent analogy that we’d look at someone funny if they said height didn’t matter one bit for becoming a professional basketball player or if they said that practice didn’t matter one bit for the tall person becoming a professional basketball player. Similar, intelligence has a fixed genetic component and your experience of intelligence is malleable in response to effort.

I’m particularly interested in the experience of plateaus that we attribute to limited intelligence. For example, when I asked Scott Alexander about his writing process a while back, he said that he composes the post in his head until writing is “a little harder than just transcribing the way I imagine it, but not *much* harder” (at least for short posts – like the length of this post). 

I can’t do that. I frequently find that I can’t hold more than a few paragraphs worth of an essay in mind at once. After that, they slip away and I lose the chain of thought.

Should I toss up my hands and despair of ever being a good writer? Well, if blogging is only worthwhile if my blog will be described as a “national treasure,” then it’s worth thinking about what this means for how much effort and time it would take to get to that level.

If, however, I just care about being able to built out useful thoughts? Then this probably doesn’t I say much, because the amount I can hold in mind at one time is a constraint on working memory. Fortunately, writing things down is itself a way around this problem. I can write down multiple threads of thought and address each one in turn. Writing is effectively a method for expanding working memory.

I expect many other cases that feel like limits in intelligence may actually be plateaus in working memory or processing speed or learning techniques. What do we actually experience that hints at lower intelligence? Slower learning? Less creative problem solving? Worse grades? Less work output?

These experiences are probably related to intelligence, but you will see an improvement in your learning speed and grades if you improve your note taking or learn better study techniques. That points to deliberate practice playing a role even in areas that we might traditionally attribute to intelligence.

And yes, there are some instances where it probably does matter if you’re in the top .01% of the population. If you need to succeed at that level for it to be worthwhile, it’s worth testing if you are likely to meet that threshold.

But in general, if you want to improve and are willing to invest effort finding the best methods to do so, you probably can.  

Special thanks to Nora Ammann for her feedback.