Let’s start with that short answer, lest many people stop at the headline and first paragraph and take away that having a left and right knee that are perfectly symmetrical explain why Usain Bolt is a six-time Olympic Gold medalist.
If I told you that LeBron James is the best player in basketball because he is 203cm tall, you should call me out on the basis that there are dozens of men at least as tall as James in the NBA, and many thousands of tall men who are not in the NBA, and who are not as good at the sport as he is. One trait does not make a champion, but the media is never slow to pounce on these findings to solve complex problems. That’s why this headline: “Could symmetrical knees be the secret to Usain Bolt’s success? A new study seems to think so“, is both intriguing and infuriating.
So let’s repeat – no, having symmetrical knees does not make Usain Bolt faster than the rest of the world. In fact, it is exactly this kind of over-simplified, extreme thinking that I wrote about in my New Year’s Resolution post. I’d hoped there’d be less of it in 2014, but turns out we didn’t quite make it to the end of week 1 before this overstep occurred again. So this is a post not only on symmetry and sprinting, but an illustration of how we get the application of science to performance so badly wrong.
Now, let’s move onto the interesting discussion.
[ribbon toplink=true]The finding[/ribbon]
The study in question, published in PLOS One towards the end of last year, found that sprint performance in a group of Jamaicans was significantly correlated with the symmetry of their knees when they were children.
By symmetry, they are talking about the degree to which the left and right knees are the same. The research is part of a long-term study looking at a number of different traits, including ear height, finger length, wrist and elbow width, knee and ankle width, and foot length. The researchers tested a large group of Jamaican children in 1996, and again in 2006, and then had a subset of that group perform a 90m and 180m sprint trial in 2010. The sprint performances were evaluated as a function of symmetry, and it was found that the more symmetrical the knees, the faster the sprint performance. A weak relationship, to be clear, but a significant one nonetheless. Ergo, the conclusion that “high knee symmetry in childhood is linked to an ability to sprint fast in adult Jamaicans as well as a readiness to sprint”.
That conclusion is accurate, based on that data. What happens next – the application of the data to explain why Jamaica win so many sprint medals by the media – is not.
[ribbon toplink=true]The significance of symmetry[/ribbon]
The idea behind symmetry is fairly obvious – if your left and right knee are exactly the same width, then they are perfectly symmetrical. Same for the ankle, or length of the foot, finger, or ear. The reasoning behind why this should matter to sport is less clear to me, although there is a fairly large body of research on the subject, including findings that symmetry of facial features is a fairly good predictor of how many sexual partners a person will have in their lifetime, and that women perceive men with symmetrical bodies and faces to be more attractive than men with higher asymmetry!
I won’t profess to be an expert on symmetry and why it matters, but the relevance, for the researchers anyway, is that the degree of symmetry is assumed to be an important component of “fitness” (in the evolutionary sense) and developmental stability. For the purposes of sports performance, I guess it’s easier to conceptualize from the other direction, namely that a person who is highly asymmetrical (imagine a large left vs right leg length discrepancy as an extreme case) would be considered LESS ‘fit’ or physically capable than a symmetrical person. So, symmetry = fitness.
[ribbon toplink=true]Jamaican vs UK children: Asymmetry comparison[/ribbon]
Now, consider this data from the Jamaicans when they were children. This data comes from a paper published in 1999, where John Manning and his team compare a group of Jamaican children to children from the UK. The comparison is important, because it drives our interpretation of the sprint performance finding of 2013.
The table below shows the relative asymmetry of Jamaican and UK children for 11 traits, but they’re only paired for seven of them. I’ve highlighted the asymmetry score of the Jamaican’s knees in green. That score, incidentally, is calculated as (left-right)/[0.5x(left + right)], and so you can work out that if the left and right dimensions are identical, the asymmetry score equals zero. If however the left knee width were 113mm and the right 114mm, for instance, then the asymmetry score is 0.009.
Now, the point I would make about that data is that the AVERAGE Jamaican asymmetry score is lower than the AVERAGE UK child score for all but one of these traits (see composite asymmetry at the bottom). However, you can also deduce a high likelihood that there will be overlap between the groups, and this means that there will almost certainly be English children with an asymmetry score of zero (that is, perfectly symmetrical).
Unfortunately, we don’t know this, because data is rarely presented this way in scientific papers, which is a real problem for the ultimate conclusion. Also unfortunately, they don’t provide the asymmetry score for knee width in UK children, but I think its fair to suggest that the close proximity of the averages, along with the relatively large variability, means that once again, there will be children in both the Jamaican and English population who are perfectly symmetrical and have a score of 0 using this method.
[ribbon toplink=true]The difference between prevalence and presence of an advantage[/ribbon]
Why is that important? Because if you are going to suggest that improved sprinting performance in one population is the result of more symmetry, then you must be sure that the populations are completely different. You cannot, as the media did, single out Usain Bolt as a champion because of his symmetry when there could very well be individuals in other populations who are equally symmetrical (never mind that we don’t actually have Bolt’s dimensions, so it’s a leap anyway).
Now, this does not mean the data has no meaning – it’s clear that the Jamaican children differ from UK children. And so the typical Jamaican might be expected to be more symmetrical. Or put differently, if you took 100 Jamaicans and 100 English children in the UK, you would likely find more children who are perfectly symmetrical in the Jamaican population. Now, IF (and it’s a big if) symmetry does really matter to performance, then you could argue that the probability of finding a Jamaican with the potential to be a world-class sprinter is higher as a result of the PREVALENCE of an advantageous characteristic. But it’t not because of the PRESENCE of the characteristic.
This is a subtle, but important difference. It’s the difference between saying that Jamaicans (as individuals) are better sprinters because they are symmetrical (which is the media angle) and saying that Jamaica (as a collective) is more likely to produce sprinters because the probability of an individual within that population having the necessary characteristics is higher.
Look at it again from the other direction – if the population of Nation X tended to have individuals who were highly asymmetrical, you would not invest as much time, people or money into discovering a champion sprinter in that group. But it doesn’t mean there are none, just that the prevalence is lower. And prevalence, in a real world, limited by resource availability, drives behaviour (you don’t drop a billion dollars on a gold mine when you aren’t sure you’ll discover gold, in other words).
[ribbon toplink=true]The genetic basis for distance running – another prevalence vs presence difference[/ribbon]
Last year this time, I wrote a paper with two colleagues, Jordan Santos and Malcolm Collins, on the genetic basis for elite distance running, and we attempted to argue this same point, which seems to elude even some highly published scientists. That point is that the difference between Kenyan distance runners and those from the rest of the world is not necessarily that there is a unique gene present in the Kenyans (it may well be, incidentally, but it hasn’t been discovered yet), but rather than within the Kenyan population, there are more “candidate” elite distance runners because the prevalence of the necessary characteristics in each individual is higher.
Here, we’re talking about innumerable characteristics, things like running economy, metabolic capacity, tendon elasticity, cardiovascular capacity, mechanical factors related to leg length and tendon length, and many others we probably have yet to discover. To use a basketball analogy again, if you look for NBA stars in the pygmy population, with an average height of under 150cm, you are likely to be wasting your time. But if you do happen to discover that one outlier, who gets to 190cm, it doesn’t mean that pygmies are suddenly a viable population for NBA scouts to recruit from, it just means that height (the performance characteristic, in this case) is not unique to the rest of the world’s populations.
[ribbon toplink=true]The prudent conclusion[/ribbon]
So, similarly, symmetrical knees are not likely to be unique to Jamaicans, and thus it is over-stepping the data to suggest that symmetrical knees explain the Jamaican sprint advantage. What you can suggest (as a hypothesis that then requires further testing) is that symmetry of the knees is a contributing factor to sprint performance (as the Manning study finds, albeit very weakly), and that the prevalence of symmetry is different between populations (data not yet reported). Thus, all other factors being equal (which they never are), some populations may produce more sprinters than others.
None of this precludes the need to find, develop and then manage the talent and training pathways, and none of it means that anyone other than Jamaicans should not bother. But it does mean there are questions of leverage, and “cost-benefit”, which have to be respected in the real world. But just as there are Italian, American, British and South African runners with the necessary characteristics for endurance success, the rest of the world CAN still discover world class athletes.
The problem is a numbers game, one of probability and the realization that they may have to work that much harder to discover it. Then on top of this, you overlay generations of success, the resultant heritage and culture of a specific sport in a given population, and add also the economic drivers that exist in these groups, and you have this extreme dominance. However, it’s wrong to say that the success is ALL down to physiology (including symmetrical knees), just as it is equally wrong to suggest that the success is not driven by genetic factors, because it is. Both matter.