Not That We’re Counting, But …

Our latest “Freakonomics” column in the New York Times Magazine, which is about how people get good at whatever they’re good at, is as of this moment No. 5 on the list of most e-mailed articles in the Times. Here’s our webpage with further information about the psychology professor Anders Ericsson and other researchers in the Expert Performance Movement. [P.S.: as of Monday morning, the column had hit No. 1 on the Times list. Three cheers for Anders Ericsson.]


It brings to mind the old joke about how many psychologists it takes to change a light bulb.

The research with sequences of digits also brings to mind the work of Simon and Newell on cognitive psychology of chess.

It was pretty easy to anticipate that clustering of early year birth dates among elite soccer players resulted form the the Jan 1 age cutoff affects in youth soccer. Hockey USA changed the age cuttoff from (I think) end of June to end of December several years ago. My observation was that It changed the relative ranking of the players. I don't have any statistics handy, but it there is a substantial data trove available should someone wish to mine it.

It's not just attention from the coaching staff as seemingly implied in the article (a factor I hadn't thought about). Players who excel are more likely to enjoy the game and thus more likely to invest time in practice.


Ericsson says in his article: "...individual differences in ultimate performance can largely be accounted for by differential amounts of past and current levels of practice."

Eleven men have won the last 16 Boston Marathons. Of these 11 men, 9 were from Kenya. And in this competition, which attracts runners from all over the world, that is because of deliberate practice?

I am not an economist or statistician, but to say something is a necessary condition does not make it sufficient. Additionally, the argument that talent is fixed is unproven.


when it comes to choosing a life path, you should do what you love — because if you don't love it, you are unlikely to work hard enough to get very good. Most people naturally don't like to do things they aren't "good" at. So they often give up, telling themselves they simply don't possess the talent for math or skiing or the violin. But what they really lack is the desire to be good and to undertake the deliberate practice that would make them better.

I found that all during high school I was told to chase after my dreams and do what I love because I would enjoy it and benefit from that decision for the rest of my life. This was great advice. However, I always found it a bit morbidly funny that I was told this while being made to take classes that I have yet to find useful in my life (even remotely). Moreover, this advice became a mantra in the educational facilities I came up in, even while they crushed and belittled the dreams of some others.

Take students who have a fascination with cars and mechanical devices, for instance. While mechanics can actually make a decent living and an even better one with starting a good business of their own, this is never and would never be promoted in our society's education. It's a blue collar job, or so many believe, and so we wouldn't dare want to promote it to what we believe should be a society filled with white collar workers and CEOs.

Boys (or girls) with this example fascination are instead told "that's a nice thought, but...", and then society proceeds to put them through four years of English, trigonometry, Spanish, etc. By the time they're 18, and decide whether or not they're going to try to attend college, they've already had four years of learning that will likely never come in handy in terms of the things they love. They never got the option to be apprenticed in things they actually enjoyed. They could have been preparing, at least some, for the things they loved. And it helps to have those with experience train you, but good luck getting that if your fascination isn't among the norm of society.

Is it not obvious that our mindset holds back millions of people? No child left behind, right? I think it's more "no child should deviate the norm."

It's hard to follow what you love when few people are in support of it. You'll find that many (though, granted, not all) of the people who "make it" to where they wanted to go usually had a good support system somewhere along the lines.



ok, that's now only something for soccer-fans. but:
maybe the german soccer youth players are mostly born at the beginning of the year. but the biggest stars in german soccer-history were nearly all born at the end of the year:

fritz walter: oct. 31
frank beckenbauer: sept.21
gerd müller: nov. 3
uwe seeler, nov. 5
günther netzer: sept. 14
paul breitner, nov. 5
karlheinz rummenigge sept. 25

and international there are also some "late-bloomers". especially between oct. 23 and nov. 5th you have stars like:

pele: oct. 25
garrincha: oct 28
maradonna: oct. 30
marco van basten: oct. 31
fritz walter: oct. 31
gerd müller: nov. 3
uwe seeler: nov. 5
paul breitner: nov. 5

do you get any better soccer stars in an 11 day period? i doubt it!
conclusion: maybe talent is still more important than an age-bracket...


Hasn't this "practice makes perfect" thing been around for a long time? The proofs are all around, the best of which I think are the results of standardized tests such as ToEFL, SAT, GRE etc... It you care to take enough time to prepare for them, you'll get good results.

But isn't this a contradiction of some theories in freakonomics? In the good parenting section, for example, the two Steves state that becoming good parents depends on who you are and not what you do. Doesn't this statement infer that there is some pre-determined talent base that each person is born with?


I believe Winter birth is also associated with increased risk of Schizophrenia. Is there a possible age grouping issue here as well? Schizophrenia usually has quite a late onset.

Ken D.

It is not news that in many sports, top level players tend to have specialized at an early age. The statistics on the prevelance of early-year soccer players in some countries tells me that those countries have unwisely crimped their pool of potential top performers by an early and rigid year-class system. The connection with practice-makes-perfect seems weak.
Also, following up on lesliew's point on the dominance of Kenyan and East African distance runners, it is also true that groups of West African origin, notably African Americans, have long domintated sprints well out of proportion to their numbers, but have produced very few top distance runners. Since both sprinting and distance running are among the most basic and universal of athletic events, this is relevant to nature vs. nurture discussion.


Try this on for size. Maybe the "early is better" theory works in the northern hemisphere and "later is better" in the southern hemisphere.


Maybe it has to do with the fact that kids who are born between January and March in the northern hemisphere usually start school (first grade) later (chronologically) than kids born in September - December. Since they are older (sometimes 8-10 months older than their classmates) and undoubtedly bigger (since growth rates in early years can often be significant), doesn't that give them an advantage from the beginning?

If true, and their early "success" in sports is encouraged and their ongoing development as athletes is put on a faster track than their peers, don't we wind up with top athletes?

Bill L. Lloyd

Freakonomics Guys:

I'm a little skeptical about your thesis here. The prevalence of early-year talent in youth leagues seems easily explained by the fact that teenagers' bodies are growing larger and more powerful by the month, so if you have, say, a team of soccer players who are 15 years and 10 months old each, they will hold a noticeable advantage over a team of players who are 15 years and 2 months, all other things being equal.

Sure, the 15 and 2 month kids might win a few games, but *on average*, the 15 and 10 months kids will be better.

For example, I used to work at a summer camp, where cabins were separated by age. If Cabin 17 played Cabin 16, say, in soccer (or anything else), Cabin 17 would usually win, because their kids would be, on average, 4-6 months older than Cabin 16 kids.

So I think this simple explanation, which you all don't mention, accounts for the early-year youth being so well-represented on all-star teams. If a coach can choose, say, the top 25 players from a pool 1,000 15-year-olds, you're naturally going to see more of them be 15 and 10 months than 15 and 2 months. Not saying you won't see any 15 and 2 months players, but they will likely be fewer in number.

Now, pros over, say 21 or so are a different matter. By that time, the body has stopped its adolescent growth spurt, so the month thing won't matter.

If Levitt and Dubner can really show that a high % of pro athletes have early-year birthdays, then I'll be impressed and think their practice-feedback loop has some validity.

Unfortunately for them, I'm got a free afternoon now so I'm going to go check 50 random NBA, NFL, MLB, and World Cup players to see if there's any correlation. If there isn't, I think we can safely flush this week's "Freakonomics" column down the toilet. If there is, I'll assume Levitt and Dubner are gods.

Back in a couple hours with the results...



I just did a quick analysis using data on Major League baseball players, a sport in which skill arguably is relatively more important than raw athleticism (at least cardio-vascular measures) as compared to other sports.

The publicly avaialable database has stats on players born from 1820 - 1986.

Certain birth months are not over-represented to the degree in this Freakonomics article, though the jump from July to August is curious.

bMonth Freq. Percent Cum.

1 367 8.56 8.56
2 315 7.35 15.92
3 347 8.10 24.01
4 333 7.77 31.79
5 311 7.26 39.04
6 296 6.91 45.95
7 300 7.00 52.95
8 470 10.97 63.92
9 403 9.40 73.33
10 412 9.61 82.94
11 381 8.89 91.83
12 350 8.17 100.00

I get this patern when I restrict the data to more recent years, or to players born in the US.

This may reflect the overall population birth trend however - I've read that August is the most frequent birth month - can't find the citation now though.


Bill L. Lloyd

No pattern in the NFL. I counted the birth months of the 50 first-listed wide receivers on this alphabetical list:

...and there is no correlation between birth month and being an NFL receiver. 21 were born in the first half of the year, 29 in the second half.

J 4
F 3
M 6
A 4
M 2
J 2
J 4
A 7
S 3
O 4
N 4
D 7

Off to count MLB, NBA, and World Cup. This isn't looking good for the authors of this piece.


I posted my thoughts on the article here under the authors' previous post on the article. As noted by many of you, the birth months do not prove the case the authors seem to be arguing.

To expand upon a couple of the items some of you have mentioned:
If Levitt and Dubner can really show that a high % of pro athletes have early-year birthdays...
This is what they're saying as I understand it. Amongst World Cup soccer players (pro level), the early birth months dominate. This makes sense, though. The younger kids quit pursuing soccer at an earlier age due to the slightly older always being chosen for the elite teams and getting the best training all the way up. By the time everything's equal, the better training and constant attention has caused the slightly older youths to actually be better at the skills that can be learned and taught.

biggest stars were nearly all born at the end of the year
This would make sense too. The absolutely most talented children born in the late months would have been good enough to make up for the size difference they may have been facing. Thus, they still made the team and got the same positive feedback and training the slightly older kids did. Plus, they always had to play against slightly bigger and older kids and so potentially had an even more rigorous training than even those kids did and ended up better. When a youth is elite, what do they usually do as a part of their training? Almost always, they're moved up an age-group, thus playing against kids who are older. This helps them compete against people similarly talented, but also helps them be better in the end.


Bill L. Lloyd

Unless my figures are wrong, I think the New York Times needs to run a correction on this column.

Levitt and Dubner's first line is:

"If you were to examine the birth certificates of every soccer player in next month's World Cup tournament, you would most likely find a noteworthy quirk: elite soccer players are more likely to have been born in the earlier months of the year than in the later months."

Well, I just looked up the birthdates of players on three 2006 World Cup teams (picked at random -- England, Holland, and Argentina), and Levitt and Dubner are wrong. 30 of the players on these three teams were born in the first half of the year, 32 were born in the second half.

There is no correlation. 12 of these 62 players were born in the first three months of the year.

Mr. Levitt and Mr. Dubner -- where did you get your information that the World Cup has more players born in early-year months than later-year months? This took me less than an hour of Googling to disprove.


Bill L. Lloyd

I just added Spain's team to my tally, and now it's 47-42 in favor of players born in the last 6 months of the year.

This is really amazing. It looks like Levitt and Dubner invented their thesis out of whole cloth. Jayson Blair, call your office...

Bill L. Lloyd

OK, I shouldn't have said the Jayson Blair thing. That was mean.

But anyway fellas, it would be nice to hear where you got it from that a noteworthily high % of pro (non youth league) soccer stars are born in earlier months of the year, since it doesn't appear to be true, and that fact is the keystone of your whole piece.

Bill L. Lloyd

More evidence that there is no correlation between what month you were born and whether you become a professional soccer player: I just tallied the birth months of a section of players from this Wikipedia list of famous soccer players:

As a random sample I chose players whose last names began with E, K, or R. There were 77 players total. 37 of them were born in the first six months of the year, and 40 were born in the last six months.

Again, I don't know where the authors got it that "elite soccer players are more likely to have been born in the earlier months of the year than in the later months," but it doesn't appear to be true, and their whole piece hangs on it being true.


Hey guys. I really liked this article, specially the correlation between the month you were born and whether you become a pro soccer player. really interesting, i have never thought about that. congratulations, great article :)


I liked the article, and may even agree with some of the findings. But in the US, you have your dates wrong.

Most soccer clubs in this country follow the US Youth Soccer Assoc (USYSA) guideleines, which mandfate July 31st as the age group cutoff. This way most children are able to play with thier school classmates, as this date is closer to most school district's age cutoff.

So August, September and October birthdays should be more represented in the elite of US born and raised soccer professionals, if you argument holds true.


The effecs in youth sports appear to be well documented.

The citations in the latter list indicate previous measured effects among professionals. I havn't checked the references.

I would discount MLB and NBA as they use a very different development model than youth hockey and soccer. They primarily use the highschool/college system in which selectively holding children back, often in the early years is the norm.

Analysis of World Cup teams is complicated by uncertainty of the "selection date". As with American hockey, these dates have changed over time and may not be consistent between countries. Indeed, the authors failure to find a bias in the U21 team may result from a cut date switch in that group.

Nonetheless, if a careful analysis of World Cup players fails to show the effect among players from countries with rigid age cut dates, then a hidden market efficiency is correcting the bias in the out years.



I thought the observation about medical diagnosis was the most interesting thing about the piece because it makes the most sense.

I work in IT consultancy, a field where seemingly endless learning is an absolute necessity to career survival. So learning is an obsession for many of us. I've thought long and hard about how to do it better, and come to some interesting conclusions.

I find that I have to do it myself. I have to write the program or at least make the things work, have to design my own curriculum, and have to teach myself. I might be able to learn in an intense lab environment but not that much from lectures. I get ideas and guidance from conventions and such but don't learn the actual skills that way.

Thinking about the mammogram problem the idea of practice sessions with instant feedback make a lot of sense. The problem is often one of slack. Is there time built into your schedule to actually improve or are you writing code/ making diagnoses 10 hours a day?

Some time ago I read a book titled Slack by Tom DeMarco, an internationally known management consultant, which advanced the theory that to some extent efficiency is the enemy of adaptability. DeMarco advanced the theory that the most efficient organizations tend to be the ones which die the quickest when the world changes in a significant way. I think I agree.

So how do we build incentives for organizations to be more adaptable when the immediate incentives drive them mostly toward being efficient?

Adaptability is surely a social good in the kind of world we live in; if US auto companies were better at adapting surely they would be a lot farther from going bankrupt than they are.