August 25, 2013 10:51

Gladwell on chess and the ‘10,000 hours rule’

Gladwell on chess and the ’10,000 hours rule’

Malcolm Gladwell, the author of the best-selling book Outliers, has written a piece on the website of The New Yorker dealing with some of the criticism he’s received over his popularization of the so-called ’10,000 hours’ rule that he describes in above-mentioned book.

This rule, which postulates that “the key to success in any field is, to a large extent, a matter of practicing a specific task for a total of around 10,000 hours” (Wikipedia), has been criticized for good and wrong reasons. (Gladwell himself notes in the first paragraph of his article that the number of 10,000 hours was actually suggested 40 years longer ago, by Herbert Simon and William Chase.)

As he notes in his New Yorker piece, many people have, incorrectly, interpreted the rule to mean that “with enough practice, anyone could achieve a level of proficiency that would rival that of a professional.” This is indeed not at all what Gladwell argued in his book – “achievement is talent plus preparation” he literally wrote in Outliers, in which chess plays a prominent role. Gladwell:

“In cognitively demanding fields, there are no naturals. Nobody walks into an operating room, straight out of a surgical rotation, and does world-class neurosurgery. And second—and more crucially for the theme of Outliers—the amount of practice necessary for exceptional performance is so extensive that people who end up on top need help. They invariably have access to lucky breaks or privileges or conditions that make all those years of practice possible.”

Gladwell is correct that his position is often wrongly interpreted, even in high-brow publications. A recent, oft-quoted article also seems to imply that the 10,000 hours of practice only, would be sufficient to reach a certain expert level.

Gladwell’s article can be read online here and is well worth reading in full. Unfortunately, when it comes to chess (as it inevitably does in discussions such as these!), Gladwell himself could be a lot clearer. One of the chess studies he mentions in the article is, as he says, about “chess masters” – a rather vague term in itself, as Gladwell notes as well. But then he goes on to say:

“Epstein [author of The Sports Gene – ed.] is talking about chess masters—the lowest of the four categories of recognized chess experts. (It’s Division II chess.) Grandmasters—the highest level—are a different story.”

This just doesn’t make sense. How are ‘chess masters’ in the lowest of four categories of recognized chess experts? Four? Which four? If we ignore the possibility that Gladwell and others refer to local categories, such as the American ‘master’ (a USCF rating of 2200+) or the Russian ‘Candidate Master’ (a bit more elastic), I can at best think of three ‘officially recognized’ ones: Fide Master (FM), International Master (IM), and International Grandmaster (GM).

Of these, FM is definitely not a fully ‘recognized’ title – you can buy it if at one point in your chess career you achieved a FIDE rating of 2300. Which is great, no doubt - but anyone flashing his FM title suggesting he’s some kind of real ‘chess master’ is in the real chess world gently dismissed: “see you when you have three norms and 2400, my friend!” This leaves only two ‘real’ recognized categories: IM and GM.

But there are more problems with what Gladwell writes. Whilst grandmasters are indeed officially the ‘highest level’, there are, of course, huge differences among grandmasters. This fact is ignored in Gladwell’s article (and also, it should be noted, in many other articles). In fact, these differences are, at least rating-wise, often bigger than the difference between IM and GM in the first place.

This is why it’s misleading to just state that grandmasters are ‘a different story’. An average GM is probably more similar to an average IM than he is to the likes of Vladimir Kramnik or Magnus Carlsen (whose picture accompanies the New Yorker article) – who are nevertheless also grandmasters. You just can’t lump the two together because they fit into the same official title category – especially not when you are criticizing someone else who’s doing the same. It’s simply a whole different ball game, and you need to make that distinction absolutely clear.

Similarly, he notes that “[t]he famous Polgár sisters (two of whom reached grandmaster status) put in somewhere north of fifty thousand hours of practice to reach the top.” This may be so, but there’s still a huge difference in level between Judit and Susan (about 200 points, which is again more that the difference between an average IM and an average GM). This important nuance is left out in Gladwell’s argument (even though it actually supports his broader theory that there’s more to chess expertise than just the number of hours put into it!)

Gladwell goes on to talk about running – a subject he claims to know a lot about as he’s done it himself. This might well be true, and what he writes certainly makes sense from an intuitive point of view, but it’s hard not to be skeptical when what he writes about chess is incomplete at best.

Arne Moll's picture
Author: Arne Moll


saturnz's picture

I'm just being devil's advocate, but maybe such detail was not required or would not have been appreciated by the target audience, nevertheless I do agree with the criticism.

Para's picture

Who cares.10,000 is just a guide to mastery, not a guarantee. I wouldnt be surprised if it takes 100,000 hours to reach the top.

Martin Matthiesen's picture

Maybe that's because you didn't calculate how many years that would take.

Fishy's picture

I played and studied chess for 10000 hours and am still not a GM.

Stephen's picture

Following the criticism maybe the title should be prefixed with "out and".

Soviet School's picture

I think the 10000 hours to mastery 'rule' is a little confusing in chess due to master titles. Personally I think will all the books, computers and competitive motivation chess players even just good club level are very skilled at chess compared to the level of skill that constitutes excellence in other fields.

For example maybe say ten thousand hours of study on the violin would produce a level of skill that was obvious to everyone while the competitive nature of chess means you have to be better than your peers to be seen as reaching mastery.

Anonymous's picture

You've identified the whole problem right here which is a semantic issue. There are rarely technical definitions of mastery, and when there are they differ from field to field and each person's interpretation is different. It is pointless to argue the 10,000 hour rule because it can never presented as an absolute so long as it describes something that is impossible to create a generalized explicit definition for.

Jon Doe's picture


That's a very good,succint examination of the generalisations about talent and practice that some 'pop' writers are currently getting away with in their books.Which on closer examination aren't quite as 'sound' as they are advertised to infer.

Good points you made.That was a refreshing read.Good article.

MDPhd's picture

I have an appreciation for Malcolm Gladwell raising provocative questions. His conclusions are acceptable (with modifications).However they suffer from the same fundamental problem that most all researchers in social science and a significant part of medical science suffer from. These are simply observations and all the conclusions are correlations. There is no fundamental principle elucidated. It is not like newtons 2nd law (f=ma) from which a reliable prediction can be made.

Therefore it is not surprising that we will find many exceptions to this 10,000 hour observation and it should be considered ,at best, a thought provoking concept.

me's picture

The whole story makes no sense in the world of chess. Rating does not indicate a lot about someone's strenght. It only tells you the chess-person got that many points, but you do not know if he/she got a large amount due to luck (blunders from opponent, just walking into a preperation, playing someone who desperately needs to win, or even someone who needs half a point for tournament-victory/or master-norm or whatever) or perhaps someone is not interested at all in ratingpoints and simply plays "art"-chess like Ivanchuk or Bronstein so he sometimes takes too much risks because he just likes it. Rating is a ridiculous concept. If you think about it, why is it continuously changing? Because they can never predict the relative strenght. By changing constantly you already proove it is worthless.

Anonymous's picture

"Rating does not indicate a lot about someone's strenght. It only tells you the chess-person got that many points, but you do not know if he/she got a large amount due to luck"

Agreed, Carlsen's rating doesn't indicate much about him being a strong chess player, but that he has had a large amount of luck.

RG13's picture

You miss the point, Carlsen's rating decrease does not mean that he is weaker. We know his relative strength from his RANKING.

Anonymous's picture

His ranking is based on his rating being higher than the rating of other players, Aronian's ranking is based on his having the second highest rating etc.

RG13's picture

You are conflating the method of determining ranking with the ranking itself. There is no standardized test to determine a player's strength; we determine a players strength by noting his performance relative to his peers. That performance could be measured using other methodologies than the Elo formula. RANKING however is completely objective. Calsen is the # 1 player by far. We don't think that he is 'somewhat weaker' because his rating has dipped from it's previous high. Otherwise we would have to assume that Carlsen at his peak rating is almost 100 points stronger than Fischer at his. This obviously is not the case. Fischer was the number one ranked player relative to those he competed against. When Lasker determined that Capablanca was the #1 player even before they played their match, how did he know it? It wasn't by Elo. Calculating Elo is just one methodology of determining ranking - not the only or the best way.

RG13's picture

@me, your point is well taken. Perhaps that is why tennis (which uses the Elo formula to calculate rankings) doesn't publish 'ratings' only the rankings.

Anonymous's picture

"Rating is a ridiculous concept. If you think about it, why is it continuously changing? Because they can never predict the relative strenght. By changing constantly you already proove it is worthless"

So ratings are worthless and ridiculous and can never say anything about relative strength because ratings aren't always static? Well, well

Jon Doe's picture

"Therefore it is not surprising that we will find many exceptions" Exactly,good point.

Too many exceptions.

The following link is a decent article too,but I post it largely (with respect and thanks to the original blogger) so that people can read K.Anders Ericsson's own views on developments,in his own words.

Considering the fact that a lot of these 'pop' books keep citing his work as the source material for their 'conclusions' it's an illuminating read:

K.Anders Ericsson in his own words is in the following-

2012 Ericssons reply to APS Observer article Oct 28 on web.doc

silvakov's picture

Well, when you don't know much about an specific subject, you generally go for the official bodies regulating such subjects, and indeed FIDE has 4 recognized masters categories: GM, IM, FM and CM. It's a different story that the chess world don't give any recognition for this last title for historical and practical reasons (since a CM isn't a "titled" player in many instances). On the other hand, we could easily support that a 2200 player is indeed a master: many national federations recognize 2200 as national master; initially, FIDE only considered 2200 performances, what we could argue is based on the fact that 2200 players already perform in a "decent" level; Jeff Sonas recent calculations indicate that only games between 2 2200+ players can have their outcome predicted with certain accuracy following the current ELO bases. So, from a generic point of view, you could argue that 10000 hours of "correct" work on chess will make you a master.

RG13's picture

"This just doesn’t make sense. How are ‘chess masters’ in the lowest of four categories of recognized chess experts? Four? Which four?"

These four:

1. 2200 USCF (equivalent to 2100 FIDE) is considered a 'National Master'

2. 2400 USCF (equivalent to a 2300 FIDE Master) is considered a 'Senior Master'

3. 2400 FIDE (equivalent to 2500 USCF) +
required norms is awarded the title of 'International Master'

4. 2500 FIDE (equivalent to 2600 USCF) + required norms is awarded the title of
'International Grandmaster'

Of course there are more levels going up but FIDE hasn't yet decided to award the 'Super Grandmaster' title even though everyone in the chess world knows who they are anyway.

Amos's picture

Why all this guessing what Gladwell or Epstein mean by "master"? Just google the study in question (Guillermo Campitelli and Fernand Gobet, 2007) -

"Three grandmasters (mean age = 31 years, standard deviation (±) 3.5), 10 international masters (29.1 ± 10.7), 13 FIDE masters (27.1 ± 8.9), 39 untitled players with international rating (30.2 ± 13.9), and 39 players without international rating (33.2 ± 17.8) filled in the questionnaire. The mean age of the sample was 30.8 ± 14.6 (range: from 10 to 78 years, median = 28 years). Since not all players had international rating, we used theTalent and Practice 15 national rating in order to measure chess skill. Note that the two ratings were closely related: for the 65 players having both international and national rating, the correlation between the two scales was .89.1

The range of the sample was 983 points (from 1490 to 2473), with a mean of 1990.8 and a standard deviation of 221.5. Since the Elo rating has a normal distribution with a theoretical standard deviation of 200, our sample had a range of nearly 5 standard deviations."

Amos's picture

And here is what they mean by "master":
"We tested this
hypothesis by calculating the cumulative number of hours spent in group and
individual practice until players reached 2200 Elo points (i.e., master level)."

2200 Elo in this case is Argentinian National rating.

Arne M's picture

Thanks for the information Amos. If the '10,000 hours' rule indeed aplies to 'national masters'I guess this opens a whole range of new questions but that's probably too much for the current scope. In any case, it's confusing to speak of 'masters' without specifying what is meant by it, as Gladwell (repeatedly) does.

Flaneur's picture

I tend to a bigger Chessvibes fan than a Gladwell fan, but your criticisms here are truly nitpicking at its most picayune. Gladwell's points about chess are -- for their intended general audience -- unusually accurate, in fact. An ordinary US master is in American parlance three levels below GM (National Master, FM, IM, GM), and as to differences among GMs or Polgar sisters, for a general audience, frankly, who cares? Anyway, keep up the (normally) good work!

Arne M's picture

@silvakov, Flaneur. I would normally agree with you but the point is that if we rate CMs or NMs to be 'masters' in the sense Gladwell and others assume it to be, then the 10,000 rule to achieve this doesn't make any sense anymore. This can easily be established by counting all the youngsters who effortlessly reach this level without putting half the amount of hours into it. My conclusion is they really did mean IMs when they were talking about the rule, not national or candidate masters - titles which are, frankly, meaningless.

Flaneur's picture

Arne, read Gladwell again: he agrees with you that a "mere" master title doesn't signify much. I really don"t know what your beef with him is, aside from some trainspotter-ish nitpicking.

(Although, as a longtime USCF 2100 who strives for 2200, I'd like to think the national master title is not totally "meaningless.")

Arne M's picture

I know he agrees with that, Flaneur. That's why I don't understand how he can be so confused about proper terminology. Anyway, I don't have a big beef with him (I usually like his books and articles) but I thought this couldn't be left unmentioned.

RG13's picture

2200 USCF is not totally "meaningless" it MEANS (according to Bill Goichberg) that you have achieved the equivalent of 2100 FIDE.

Sometimes candidate masters minimize their achievements because they are comparing themselves to extremely strong players instead of realizing that their percentile level is already very exclusive.

Arne M's picture

Yes RG13 but surely Gladwell didn't have 2100 FIDE in mind when he spoke about 'masters' - that would be ridiculous. Anyway, 2200 USCF certainly doesn't mean anything outside of the US and I'm pretty sure Gladwell intended his books and articles for a wider audience than just America. Hence my criticism,

RG13's picture

Well it means something to Kasparov since they are banned from his simuls! ;-)

Anonymous's picture

I love it when patzers speak and write about chess.

Isaac Hunt's picture

Sometime ago I estimated that it took me between 6 to 7000 hours (comprising study, training and playing) to reach 2200. I still don't have the slightest idea about what it should take to reach FM or IM, if possible at all.

me's picture

If all your opponents make the same hours, do you think you all can become 2200??? :-) In some ways doing your homework is some kind of cheating.

Brian Karen's picture

"Of these, FM is definitely not a fully ‘recognized’ title – you can buy it if at one point in your chess career you achieved a FIDE rating of 2300. Which is great, no doubt - but anyone flashing his FM title suggesting he’s some kind of real ‘chess master’ is in the real chess world gently dismissed: “see you when you have three norms and 2400, my friend!”"

Funny but insulting. The vast majority of chess players are not good enough to make FM. Many FMs have worked hard for the title. It is not a title that you can simply purchase.

But it is unnecessary to argue - they are by definition Masters.

Sune Berg Hansen (GM)'s picture

I kinda agree with Gladwell, but more with the talentcode and talent is overrated. Of course people are different and a certain IQ in some areas is nescessary to become a GM - still it is hard Work (the 10.000 hours) that will determine how far you will go. I like the term 'deliberate practice'a lot - it kinda explains why a lot of players stay at the same level forever - even though they do Work on chess. And of course Kramnik, Anand and Carlsen are fantastic players (and people) - but they did also Work very very hard (at some point). Which is why the term potential is much better than talent

Anonymous's picture

Note first: Not a chess player, but this article caught my eye. This is just my general analysis of the 10000 hour rule.

Another note: This doesn't directly address the article above.

I've always had a slight problem with the 10000 hour rule. While it's clear that in order to get good at anything, you need to practise for many hours, the rule seems to imply that every field would need that many hours. In fact, I daresay the 10000 hour rule is more important in say, something like chess (in which arguably far more than 10000 is needed) than various other sports simply due to the fact that chess is arguably the most competitive sport (I use this word loosely.) around. In no other sport do you see serious discussion about people needing to start before an age of around twelve just to hit an elite level (2700+), with a career of around twenty-five years (in which they can improve). It seems asinine to me to argue that every field has a competitive level high enough such that you would need 10000 hours to reach such an elite level. It just seems like a ridiculous claim.

Slightly off topic:

Note that being "good" is all relative. Staying in the realm of sports here, what if one were to design a sport where, besides the standard attrition from people not being too competitive/having poor genes/poor environment to train in, constant hazing occurs. We would then have a field of competitors who aren't necessarily as good as the same sport without hazing. Clearly then, it is easy to simplify things to the point where only a few hours are needed, given enough factors (say, your local rec league where everyone thinks they are good!).

While the 10000 hour rule seems like a useful thing to apply to life whenever you try to get good at something, as an actual statement it seems to be unfalsifiable. The rule just seems silly once you realize it can't even be a threshold number. As a life philosophy it seems useful, but as an actual rule it just fails. I am not surprised that such problems like the paragraphs above has caused such confusion in the above article and comments.

Anonymous's picture

Also, I'd like to add that while it seems obvious to me that a sport where you only need to put in ten hours in or so in order to be good is not "competitive", it is not clear to me at what number of hours are needed in order for a sport to be considered competitive, as there are far too many factors to consider.

Latest articles