On “Windows of Trainability”
Posted by Dean Holden at December 4th, 2014
by Craig Taylor, 22 March 2010
I’ve been told I need to blog more, so instead of the infrequent notices and pictures I’ve been posting lately, here’s a meatier post that I’ve been meaning to write for a while – although I suspect there are only 3 or 4 coaches who might be interested, and everyone else is likely running for the blog links already. Onwards.
I first read the Sport Canada Long Term Athlete Development Model (aka LTAD) approximately 6 years ago. The LTAD examines sport structure and sport/athlete development in Canada, and makes recommendations for how we can better develop our athletes for international success and/or lifelong sport enjoyment. The LTAD model was developed right here in Canada (originating in BC), and has been adopted by many countries around the globe, most notably the UK. In fact, every sport in Canada – and many sport federations in other countries – have used the LTAD as a framework from which to craft their own sport-specific LTADs. Google “LTAD” to get an idea of how far and wide the model has travelled.
When I first read the LTAD, I thought it was a great resource, and I think that’s still how it’s generally received by people. To be sure, there are several excellent aspects to the LTAD (ie. the FUNdamentals stage, emphasizing physical literacy, avoiding early specialization, recommending that top coaches work with developing athletes, etc.), but there are also a few things, presented as fact, that simply don’t stand up to any sort of critical investigation. Falling squarely in the latter category is the concept of “Windows of Trainability”.
As with anything presented as fact, I think it’s a good idea to ask a simple question: “Is that true?” For two or three years after first reading the LTAD, I took the content to be fact. I even taught those facts to other coaches. It wasn’t until I was enrolled at the NCI, where I was required to prepare a LTAD project, that I took a more critical look at the LTAD model. I came across this article from the UK – still one of the only critiques of LTAD to be found online – which raised some concerns with the model and suggested that readers “drill down to the actual claims in science and follow the references.” So that’s the approach I took.
First, a quick summary:
The LTAD begins by identifying shortcomings and consequences of the Canadian Sport System, which include:
-coaches largely neglect the critical periods of accelerated adaptation to training.
-failure to reach optimal performance levels in international competitions.
-poor movement abilities.
-lack of proper fitness.
-poor skill development.
-fluctuating national performance due to lack of TID and a developmental pathway.
-athletes failing to reach their genetic potential and optimal performance level.
As part of the solution to these and other problems, the LTAD model recommends that young athletes be trained and organized according to their biological maturity, rather than their chronological maturity. Biological maturity can be assessed several ways, but the least invasive method is by frequently measuring and charting the height of a young athlete. The adolescent (pubertal) growth spurt, or ‘peak heigh velocity’ (PHV), is a good marker for puberty and the myriad biological, physiological and psychological changes which ensue. So far so good; it makes sense to consider the needs of early, normal, and late maturing athletes (ie. biological age) rather than always simply grouping them chronologically.
More importantly however, the LTAD goes on to identify “Windows of Trainability”. These are specific times in the life of a young athlete, some aligned with chronological age and others aligned with biological age (via PHV), when an athlete is uniquely sensitive to specific types of training and capable of enhanced adaptation. These are the ‘critical periods’ which are being ‘neglected’ by coaches, resulting in the failure to develop athletes appropriately.
These Windows of Trainability are listed here (bottom of page), and summarized in the chart below, where the x-axis is developmental age and the y-axis is annual growth rate. The specific “windows” are identified by the boxes.
The reader is therefore left with the understanding that there are specific identifiable episodes in the life of a young athlete when specific types of training result in enhanced or accelerated adaptations. Furthermore, neglecting these opportunities limits the long term success of the athlete. For a coach working with youth/junior athletes, this looks like the holy grail of development coaching, but Is that True? Dr. David Collins, the former performance director of UK Athletics says no, arguing that “there’s no evidence for [the LTAD]”. For anyone who takes the time to review the literature, it’s hard to disagree with Dr. Collins.
The truth is: if you were to read the references cited in any version of the LTAD (or any one of several official LTAD presentations I’ve attended), you would not find one single peer-reviewed paper which provides evidence for these “Windows”. I’ll repeat that: None of the articles cited in any version of the LTAD, past or present, provides empirical support for the concept of Windows of Trainability. To be honest, there is one article you might not be able to read because I’ve been unable to find an english translation (Harsanyi, L.”A 10-18 eves atletak felkeszitesenek modellje.” Budapest: Utanpotlas-neveles, No.10, 1983.), but aside from that one exception, you could read every book and article and not find a shred of evidence.
So what exactly is the evidence used to support this concept?
There seem to be two main types of support used for Windows of Trainability in the LTAD. The first are articles by Balyi & Way (including the chart above), who also happen to be two of the principle authors of the LTAD. “Drilling down” through earlier documents reveals an interesting pattern; each new version of the LTAD cites an earlier article by Balyi & Way (or Bayli) which claims the existence of Windows of Trainability. To be fair, the concept has been proposed by researchers and coaches for at least 30-40 years, so this isn’t a new idea. Yet nowhere in this chain of citations can a reader find any original research conducted by Balyi & Way. Furthermore, none of the articles have been subjected to the rigour of peer-review in a scientific journal. The result is a series of opinion pieces repeatedly cited as fact and bolstered by the claim that “this document (the LTAD) is fully based on and supported by the coaching and exercise science literature.”
The second type of support is what I will term ‘implied’ or ‘associated’ – there are several references in the bibliography with titles that sound like they support the concept of Windows of Trainability. Although some (perhaps half) are peer-revivewd, most aren’t actually cited in the text of the document, and none actually provide proof of the concept.
One such example is Vorontsov’s (1999) “Patterns of Growth for Some Characteristics of Physical Development: Functional and Motor Abilities in Boy Swimmers 11 – 18 Years“, which sounds impressive, but makes no attempt to prove that Windows of Trainability exist.
Perhaps the best example however is Viru et al.’s (1999) “Critical Periods in the Development of Performance Capacity During Childhood and Adolescence“. From the title alone, it would sound like an important piece of supporting evidence for the concept of Windows of Trainability. In fact, I’ve seen this specific article used in LTAD presentations as proof of the concept. But anyone who bothers to read the article will find only that the authors suggest in passing that training during times of accelerated growth might result in enhanced adaptation, and that further research is required to investigate the issue; hardly rock-solid proof.
Fortunately, the same group of researchers made good on their recommendation and conducted research to investigate the issue in female athletes. In a cross-sectional study of more than 1544 girls aged 10-17 (642 active in track & field, 902 inactive), they found that “sport training at the age of 10-13 did not bring about a more intensive development of motor performance compared with that of nonactive girls of the same age” (Loko et al, 2003). That finding is completely counter to the LTAD model which states that there are critical windows for speed and strength development for girls in this age range. Although it was published 7 years ago, this study has yet to be included in the LTAD or related presentations.
Perhaps more interesting are two studies which investigated the effects of training in adolescent identical twin boys. By studying identical twins with one brother training and one brother remaining inactive, these studies control for genetics (since we know that genetics influence training response (Bouchard et al., 1999)) and maturation (since identical twins are likely to progress through puberty at the same rate), allowing the investigators to isolate the training effect.
More than 30 years ago, Weber et al. (1976) studied 12 pairs of genetically identical twin boys; four 10 year-old pairs, four 13 year-old pairs, four 16 year-old pairs. One twin from each pairing completed 7 workouts a week for 10 weeks, while his brother acted as the sedentary control.
Not surprisingly, the trained brothers in the 10 year-old and 16 year-old groups demonstrated better fitness than their sedentary brothers after 10 weeks of training. Among the 13 year-olds however, there was no difference between the trained and untrained brothers after 10 weeks – their VO2max and other parameters improved at the same rate. During peak growth, when the LTAD says that training targeted to improve aerobic capacity and power should have the greatest effect, there was no enhanced adaptation. In fact, not only was there no accelerated training adaptation, there was no adaptation at all, beyond that which was due to growth. The maximal oxygen uptake of the sedentary twin improved just as much (15.95% +/- 8.95%) as his brother who trained 7 days a week for 10 weeks (see chart below).
Based on their findings, the authors concluded, “the old hypothesis that more may be gained by introducing extra exercise at the time when the rate of growth is greatest is not tenable” (Weber, 1976).
Weber et al’s (1976) work did earn some criticism however, primarily that the active twins only trained approximately 2-3 hours per week, and that chronological age, rather than biological maturity was used to classify the groups. The first criticism is easily addressed by noting that the training load was enough for both the 10 year-old and 16 year-old trained subjects to demonstrate a training effect compared to the sedentary controls. But what about the question of biological maturity? Were the 13 years olds really in that PHV “Window of Trainability” identified in the LTAD?
Fortunately, Danis et al. (2003) considered this issue in a study of 9 pairs of identical male twins, aged 11 to 14. The biological age of the twins was assessed using Tanner Stages (the gold standard), with all pairs being found to be pre-pubertal or pubertal. According to the LTAD model, the pre-pubertal subjects would demonstrate a modest increase in VO2 due to training, but the pubertal subjects would demonstrate an enhanced or accelerated training adaptation.
One twin was trained for 6 months, 3 times a week, while his brother remained inactive. At the end of the study, training resulted in a higher VO2max amongst the trained pre-pubertal twins compared to their sedentary brothers. But once again, there was no different in VO2max (L/min) amongst the pubertal twin pairs, despite 6 months of training for the trained group (Note: relative VO2max (ml/kg/min) was actually higher in the trained twin due to a reduction in body fat, but training did not improve metabolic or cardiovascular function beyond the improvements due to growth observed in the control group). Echoing the findings published 27 years earlier, Danis et al. (2003) concluded that “growth at puberty, regardless of structural and functional acceleration, leaves no place for the effective influence of training on variables of maximal effort” (see proposed model below).
So where does this leave the concept of “Windows of Trainability”? To date, I have not seen one peer-reviewed study demonstrating an accelerated or enhanced adaptation to training (speed, stamina, power, strength) amongst males or females during one of the so-called “Windows of Trainability”. Even the references in the LTAD which are intended to support the concept fail to provide any proof.
Even more damaging however are the peer-reviewed studies, three of which are discussed in this post, which demonstrate that rather than enhancing adaptation during these windows of trainability, targeted training during these episodes has no effect on parameters such as aerobic power, aerobic capacity, speed, and power. Far from being Windows of Trainability, these are actually periods of the least trainability. During these phases, adolescents get stronger/faster/fitter simply by growing.
If there are studies out there which actually support the concept, I’d love to see them. I’m not against the idea if it can actually be demonstrated to exist, but I do take issue with continuing to push an idea when all of the available evidence is against it. Presenting this concept as fact to coaches, parents, athletes and administrators does not serve the development of athletes.
But let’s bring this post back full-circle; don’t take my word for it, find out for yourself. Read the studies and come to your own conclusion – it’s the best way to become educated. If anyone has made it this far, and they want to keep going, e-mail me and I’ll send you full text PDFs of the studies in question. Review the literature and determine for yourself where the truth lies – you’ll be better educated and more informed for your effort, and maybe you’ll expand your understanding of athlete development in the process.
In the meantime, if you coach developing athletes, you’ve got one less thing to worry about because you don’t have to sequence training to match critical development windows. I would still recommend however that you read the generic LTAD (www.ltad.ca) and your sport specific LTAD if your sport has one. There are several great ideas in these documents, as long as you remember that the concept of Windows of Trainability isn’t one of them.
Bouchard, C., An, P., Rice, T., Skinner, J., Wilmore, J., Gagnon, J., Perusse, L., Leon, A., and D. C. Rao (1999). Familial aggregation of VO2max response to exercise training: Results from the HERITAGE Family Study. Journal of Applied Physiology, 87(3), 1003-1008.
Danis, A., Kyriazis, Y., and V. Klissouras. (2003). The effect of training in male prepubertal and pubertal monozygotic twins. European Journal of Applied Physiology, 89, 309-318.
Loko, J., Aule, R., Sikkut, J., Ereline, J., & A. Viru. (2003). Age differences in growth and physical abilities in trained and untrained girls 10-17 years of age. American Journal of Human Biology, 15, 72-77.
Viru, A., Loko, J., Maarike, H., Volver, A., Laaneots., L, and M. Viru. (1999). Critical periods in the development of performance capacity during childhood and adolescence. Physical Education and Sport Pedagogy, 4(1), 75-199.
Vorontsov, A.R. (1999). Patterns of growth for some characteristics of physical development: Functional and motor abilities in boy swimmers 11 – 18 years. In: Biomechanics and Medicine in Swimming VIII. Eds. Keskinen, K.L., Komi, P.V. and Hollander, A.P. Jyvaskyla, Gunners.
Weber, G., Kartodihardjo, W., and V. Klissouras. (1976). Growth and physical training with reference to heredity. Journal of Applied Physiology, 40(2), 211-215.
Category: 10K hours / 10 years, athleticism, coaching culture, communication, curiosity, education, evaluation, exercise, Fitness / Training, genetics, growth & development, learning, LTAD, metrics / measures, parents, planning / periodization, recommended reading / books, research, skill acquisition, Skills, statistics, talent, teaching