Rabu, 23 November 2011

Sports Science 2011: Talent vs training and Oscar P

Sports science in the media in 2011: Training, talent, doping and Oscar Pistorius

So yesterday was Day 1 of the fantastic UKSEM conference in London.  I gave a presentation on Sports Science in 2011, and that presentation is embedded in the post below.  I am a terrible judge of my own presentations, so I'll just say that mine went OK and hope that it did.  I always know instantly all the things I haven't explained clearly, when I was clumsy, when I repeated myself and when the point I was trying to make didn't quite come off!  But hopefully you can read quietly what I spoke about and it is better than the "live performance"!

I covered some of the more topical stories of the year, but given that I only had 30 minutes, I had to pick three, and they were:
  1. The Kenyan dominance of the marathon, which provided a nice lead in to the training vs talent debate
  2. Doping in cycling, in the context of how doping control changes doping behaviour
  3. Oscar Pistorius, and the scientific cover-up and hatchet job he and his band of "scientists" got away with

The presentation again lacks my voice-over - I may at some stage do a "voice-over" when I have more time, but for now, it should suffice as a read through.  Below, I elaborate on part of the talk (the talent vs training part.  I may, in the future, do the same for the Pistorius section).

Email subscribers click here to be taken to site to view presentation. All others, click on the grey button, wait for loading, then hover over "More" and click "Fullscreen"





The 10,000 hour concept

The biggest talking point, at least in the discussion I had with delegates afterwards, was the Training vs Talent debate (the first part of the talk).  Here, the only reason I included this was because I saw that Matthew Syed who wrote the book "Bounce" was on the programme after me, and his talk was called "The Science of Success".  So I decided that it would be good to have a little bit of science on the topic, because he doesn't provide it in support of his "training-sufficiency" position.

Effectively, Syed's thesis is this:  Genes and talent are over-rated, and great performers, whether they are sportsmen, doctors, musicians or businessmen, achieve expert performance not because of genetic factors or "talent", but because they accumulate enormous volumes of deliberate practice.  He has a few examples of this, and makes a compelling case, at least on the surface.

But when you really interrogate what he is saying, then you realise that the reality is that he is saying that in order to succeed at something at the highest level, to become an expert performer, you need to practice.  OK then... nobody should be surprised at this, and nor would they be.  The problem is that his (and Gladwell's) position seems to exist outside of a world where genetic factors also have an influence, and it's this exclusivity in his thinking that forces a closer look.

Unnecessarily polarizing the complexity of performance by ignoring genes and talent

So the issue is not that they advocate hard work and a lot of training, it is that they downplay the importance of talent or innate ability.  I emphasized this in my own talk, but it bears repeating - if Syed is correct, and the secret to success is training and accumulating many years and hours of practice, then Talent ID is a waste of time and money.  We should rather spend that money on getting 100 more children to train, because they should all (or most) become champions, provided they get through the required hours.

Note that this also completely overlooks the fact that children tend to do what they are good at, and that simply running a child through a "10,000 hour factory" is an imagined concept only.  I guess the real question is why are some children good at something almost within the first moments that they start it, thereby encouraging them to do it more?  It seems to me that this could be an innate difference too...

In a competitive sport, training is obviously a crucial determinant of success

But the theory that practice is important is so obvious it doesn't need emphasis.  As soon as you have competition, then within a narrow range of individuals (the top 10 tennis players, or the Olympic finalists, for example), training will become a crucial determinant of who wins and loses.

In "a small pond", where there is no competition, it's possible to succeed with talent alone.  Just think back to school level athletics, when there's no competition, a young athlete can show up on the day and dominate to win.  But the higher the level, the better the competition, the more important training becomes.  And those individuals who get attempt to by on talent alone are washed away in this more competitive landscape.  Syed made this point, and of course he's correct.  But the key is that the athlete who succeeds all the way to the Olympic podium is the one who dominated without training (that is, he's talented or genetically gifted), but then also trained incredibly hard to stay a champion as the competition intensified.  In otherwords, he has BOTH talent and training. 

In fact, I challenged him on this after his talk, and basically made the point that if he had walked into that venue today, with 200 people in the audience, and asked them to please raise their hands if they thought that sporting success was ENTIRELY genetic, he would have been the only person with his hand in the air.  He may have been laughed out the room had he tried to propose that the current belief is that success is all genetic.  Everyone knows that it is not.

Yet he seems to have arrived at this belief that someone out there believes that expert performance is achieved solely on the basis of genes and natural talent. Now, maybe I missed this in my studies, but I have not once heard this theory.  The established theory in sports science is that many, many years of training are required to hone and refine skills and physiology in order to become a world or Olympic champion.  The reality is that sports science does NOT believe that it's ALL in the genes, and nor do they believe that it's all about training.  So the first problem with the 10,000 hour concept is that it attacks a straw man that need not exist.

To polarize the debate the way that he (and others, most notably Malcolm Gladwell) have done is unnecessary, and it has quite important financial and policy implications for where money should be spent by sports federations and coaches to help improve performance.  Their books and emphasis are not without merit, certainly - they have emphasized how important it is that we recognize that not all young aspirant athletes develop equally, and that we may need to consider how coaching is provided to more children to prevent some from falling through the cracks.  But sports science already knew this.  What these books have done is spawn a theory that now says that practice is sufficient for expert performance, which it clearly is not.

The work of Elferink-Gemser, who presented today after Syed, confirmed this, because she has been studying the progress of young sportspeople for 10 years, and has found large differences between children in terms of how they respond to training sessions and coaching.  But more important, she finds that it is possible to predict which children will become professional within the first few years of them entering the sports academy.  In other words, by the time children are 15 or 16, there are already differences between those who will become "great" and those who are merely "good".  It has little to do with accumulating the "magical 10,000 hours".   The mere fact that these young athletes have such different responses to training tells you that you can't generalize potential performance to a group, and that the outcome of training will also differ between individuals.

The three 'failings' of the 10,000 hour, "practice is sufficient" model

I think there are three key points about this 10,000 hour concept:

Firstly, if you can find ONE case of an exception, then you have disproved the "rule".  That is, if you can find a guy who trains 10,000 hours but doesn't succeed, then you have shown that it's not sufficient.  Or, if you can find a guy who trains only 5,000 hours, but who does succeed, then you have shown that it is not necessary.

And the truth is that both of these cases exist, everywhere.  Baker has shown it in triathlon, it has been found in chess (so it's not only "physiological" sports where innate ability seems to matter), and it has been found in football, wrestling, field hockey, skeleton.  Every single sport has examples of athletes who have shot to the top within a few years of starting the sport, and it is littered with athletes who fail despite doing 20,000 hours.  Today I spoke with a woman whose husband taught music for a school for gifted musicians in New York, and they discover children who within months of starting are playing at near-professional expert levels.  Now, unless those children have managed to get 10,000 hours of training in in one hour (by discovering how to slow down time), they have achieved expertise well before the theoretical minimum.

There's no question that talent, or innate ability, or genetics, play a role.

The second point is that there is no good evidence at all to suggest that 10,000 hours is required for expert performance.  The study that is always cited is a violin study, which found that expert violinists had accumulated an AVERAGE of 10,000 hours by the time they went to music school, whereas those who were merely good had done 8,000 hours.  Two problems.  First, you can't infer cause from this kind of retrospective study.  Who is to say that the talented, genetically gifted violinists didn't train more BECAUSE they had more talent from the age of 8?  Perhaps their innate ability was the catalyst to get them more practice (mom sends them for lessons, and they enjoy it).  And secondly, the study showed absolutely no indication of ranges or variance.  So we don't know whether there are some people who became experts with less training, and nor do we know whether some failed despite doing their 10,000 hours, because the author did not show that data.  I hope I don't have to emphasize that if either of these people exist, then the theory is wrong.

Which brings me to the third point about this theory - it is entirely unfalsifiable.  To the "evangelists" who proclaim that anyone can become an expert if they just practice enough, it's too easy and too convenient to simply dismiss the exceptions because they clearly didn't practice in the right way.  So if someone has done 25,000 hours and has not succeeded, then they simply say "He obviously didn't practice the right way".  Or if someone becomes an expert in only 3,000 hours (which happens, all the time), they say "He must have compressed his 10,000 hours into a third of the time".

So it's a completely unfalsifiable theory.  It cannot be proven, and it cannot be disproven.  Therefore, it does not belong in science.

What the science does say - "responders" and "non-responders"

What does belong in science are studies that look at how different individuals have been shown to adapt to training.  And sure enough, those studies exist, though Gladwell and Syed would never admit to them.  The Heritage study, for example, took hundreds of unrelated people and gave them standardized training programmes, and then measured the responses.

The result?  A complete spectrum, ranging from those who show absolutely no response to training, all the way to those who improve by more than 40% as a result of training.  And as expected by the scientific theory, the difference between these people can very reliably be linked to genetic factors.  Specifically, there are Single Nucleotide Polymorphisms (SNPs) which account for half this training resopnse.  Individuals who have 9 or fewer of the identified 21 SNPs are the "low-responders", whereas people who have 19 or more of these SNPs are "high responders".

The answer therefore is that it's not about having different genes, it could also be about having different variants of the same gene, the result being that you and I show completely different responses to training.  And you have to ask yourself, if you are a coach, would you rather have an individual who is a "high responder" or a "low responder"?  And more importantly, if you have $100,000 to invest in a sport, where do you spend it to find a champion?  On talent ID, to find those "high responders", or do you believe that anyone can succeed if you just spend the money to help them all do 10,000 hours of training?  In terms of policy, it's clear that the science, at least for this physiological variable, points you in the direction of finding the right people to spend the money on.  And that means understanding the value of genetic factors to performance.

And just to dispel the idea that skill-based activities benefit more from training, when you look at studies in chess, you find that there is a massive difference in the time taken to reach Master level - some do it in 3,000 hours, some have been at it for 25,000 hours and counting.  In darts, 15 years of practice (almost 15,000 hours) only accounts for 28% of the variability in performance.  In otherwords, 72% of the difference in performance between two players cannot be explained by the hours spent training.  In darts...

In sport, countless studies show that elite athletes get to the top within 6,000 hours of starting their sport, and the success of Talent ID programmes proves that talent transfer (something that is impossible if the 10,000 hour theory is correct) exists. 

Conclusion - training is the realization of genetic potential

The bottom line is that a theory of deliberate practice gives us one important message - if you want to succeed, practice.  Coaches around the world breathe a sigh of relief, you're not redundant.  But this is so obvious, I guess the reminder is always good though.

But the application of this theory, and the dismissal of genes that it somehow seems associated with, is a huge oversimplication and wrong, at least for sports.  Syed today argued about school performance, and about how teachers should downplay the idea that some children are more "talented" with numbers or better at mathematics than others.  And that's fine, because whatever helps people improve is great.  But if we're in the business of finding Olympic champions, then this theory has no place in its polarized form.

Not only this, but it could be extremely damaging.  If you take it literally, and you buy into a 10,000 hour concept, then you'll be obliged to start training a child at the age of about 10, because you need them to become world-class in their early-20s.  All good and well, except the evidence shows quite clearly that the earlier you start intensive training, the LESS likely you are to succeed.  And so there are all kinds of implications for how we manage children's sport participation.

The ultimate conclusion, in my opinion (and as always, I welcome your views), is that training is nothing more than the realization of genetic potential.  Without both, you will not become an Olympic champion (in a competitive sport, that is).  Training will improve everyone, and so everyone should be encouraged to train.  But genetic factors determine where we start, how we respond to training (trainability), how much training we can tolerate before burnout or injury (because let's face it, chess players rarely get injuries that force 6-week layoffs, like stress fractures), and finally, where the "performance ceiling" exists.

Training will get you to your ceiling, you'll realize your genetic potential.  But will it win you a medal?  Only if you chose your parents right!

Ross

P.S.  For a more detailed discussion of these issues, please do read the previous two articles I wrote on the subject:
  1. A look at the 10,000 hour concept.  What does it say, and why it fails to pass the test of validity
  2. The evidence for how genes influence elite sporting performance


Tidak ada komentar:

Posting Komentar