Toolbox: Talent Identification Crystal Balls
The intent of this week’s Toolbox is two-fold. As always, the primary goal is to explore an interesting scientific question. In this case, can you predict ultimate success in cycling based on test scores? Secondly, it is to honour Pez-friend Dr. Aldo Sassi and to wish him the best in his health battles.
It’s About Time
It may be annoying as heck for us cycling lifers to constantly have our legs ridden off by a rotating procession of fast cadet and junior racers, but it’s also the best possible thing for our sport to continue to grow and thrive. I well remember back in my days in Vancouver as a senior Cat 3 getting repeatedly dropped race after race by a young Roland Green as a cadet rider. Now in St. Catharines, our club has a specific Youth Cycling Team dedicated to nurturing individual development and teamwork for our young racers.
Junior racers are great because they bring so much energy and enthusiasm to our sport. However, as each junior sprints off up the road and leaves me in the dust, my thoughts repeatedly turn to “just what is that magic ingredient that turns a fast junior into an actual pro?” Is there some way of predicting the future for some of these riders in my club? Does being fast now automatically mean that they’ll take the next step up? What might be holding them back? And what can we do as sport scientists or senior members of cycling clubs to help their progression?
Most importantly for sport scientists and cycling federations, is there some predictor that can be used for talent identification?
Talent identification isn’t a popularity contest like American Idol or the other nauseating reality shows out there. It is the process of finding young athletes with a natural talent and inclination for a sport and then supporting them in their progression through the ranks. It can be taken to nefarious extremes, as was the case in the former Eastern Bloc and presently in China, where young children are often taken from their families and simply assigned to sports and sports academies.
It can also be done right, allowing a country to use limited financial or infrastructure resources to maximum efficiency. The best example in international sports has certainly been the amazing success of Australia, a country of 20 million that dominated the Sydney Olympics and has continued to do so in many sports in the decade since. Within cycling, both Australia and Great Britain have become dominant forces in the sport despite their relatively small population and the weak cycling history among the general population.
So what’s going on to identify the diamonds in the rough? Can it even be done or is it the proverbial needle in the haystack?
Just Let the Kids Play?
In one of his essays about talent identification, best-selling author Malcolm Gladwell (who incidentally was classmates with my wife in small-town Elmira, Ontario) wrote about how many top college quarterbacks flame out once they make it to the NFL. In essence, he concluded that playing the same position at two levels are so fundamentally different that the only way of determining who would be a great NFL quarterback is simply to let them try out as NFL quarterbacks. Back in 2006, I wrote about the lack of predictive power between riders who make it to the Junior World Champs and who end up as top professionals. So on a basic level, it appears that performance at one level does not predict performance at the next levels above.
If not performance then, what about lab measures? As endurance athletes, we are obsessed with the concept of maximal oxygen uptake (VO2max), a measurement of how much oxygen your body can process and utilize in a minute. In theory, the higher your VO2max, the more highly developed your aerobic metabolism and your capacity for energy production for sustained periods of time.
Complementing VO2max, another fundamental laboratory test is some measure of your threshold. What this threshold (anaerobic, lactate, ventilatory, functional threshold power, critical power, etc.) means both physiologically and practically remains highly contentious, but the notion that you can quantify your performance based on your sustained power remains an important idea for most coaches and scientists.
So is cycling simply a matter of one-upmanship of whoever has the highest test measures wins?
Menaspa et al. 2010
Dr. Sassi, in his role as director of the Mapei Sport Centre in Italy, has both tested and guided innumerable cycling champions, from the late Franco Ballerini through to newly-crowned Giro champion Ivan Basso and World Champion Cadel Evans. In the April 2010 issue of Medicine and Science in Sports and Exercise, he co-authored a study retrospectively analyzing the Centre’s massive database of tests to see whether there was any relationship between test scores as juniors and current success.
This study is really quite unique, because most studies of talent identification are “cross-sectional” ones. That is, you might take a group of pro riders and a matched group (e.g. age, years in the sport) of elite but non-professionals, do a number of tests on them, and then assume that whatever was different was the cause of their different level of performance.
Rather, the power of this study is that it’s “longitudinal” in that the same athletes are followed for a prolonged period of time, allowing a better look at the overall role of specific determinants. It’s still not perfect “cause and effect” but it’s definitely a preferable approach despite its logistical difficulty.
• 309 cyclist data were used (17 y, 178 cm, 66 kg average), tested between 1996 and 2002. Where an athlete had multiple tests available, the highest values were used.
• Data of their eventual performance level was calculated as of December 2008, when the cyclists would be 24-30 y old. This long time frame of up to 12 years permitted the subjects to progress from juniors through to their prime years as competitive cyclists.
• The vast majority of athletes, not surprisingly, were from Italian junior teams, although there was approximately 10% from elsewhere in Europe, Asia, and Oceania. Obviously, the athletes were already at a decent level when they were tested. Therefore, we’re not talking about what it takes for a couch potato to become a world-beater.
• At the time of testing, 72 were on their respective national teams.
• 28 cyclists became professionals, at a mean age of 21.4 y.
• Professionals were defined as having had a pro contract for at least 3 years. As most neo-pro contracts were for two years, having this cutoff ensured that the cyclists had enough ability to have earned a second contract. None have ever had a doping infraction.
• The pros were further categorized as “professionals” (n=10) or “ProTour” (n=18).
• The pros were also categorized according to specialty as “climbers” (n=8), “flat riders” (n=11), and “sprinters” (n=9).
• The ProTour riders were definitely world-beaters, winning Olympics, Worlds, Giro d’Italia, and Grand Tour stages along with seven Classics. Indeed, 6 of the ProTour riders accounted for 241 victories overall.
I won’t go into the details, but the authors crunched all the data using complex statistics to look at the predictive ability of VO2peak, ventilatory threshold (VT), and respiratory compensation thresholds (RCP). The latter two are measures roughly analogous to what we would commonly call the lactate threshold and the onset of blood lactate accumulation – see the article links above for more information. Specifically, did these test measure predict who would fall into what category of rider?
First, let’s look at comparing the junior test data itself:
• As juniors, the national team riders tended to be slightly bigger than the non-nats, and this may have contributed to their higher absolute VO2peak and absolute VO2 at both VT and RCP. In other words, the national riders had higher test scores, but this may have been only because they were bigger and therefore had more muscle mass.
• The pro flat riders were bigger than non pros, and also had higher absolute VO2peak along with VO2 at VT and RCP.
• The pro climbers had lower body weight than non pros, higher relative VO2peak, and VO2 at VT and RCP. So both this and the above comparison fits the stereotype of what it takes to be a rouleur or climber.
Now let’s look at whether the junior test data can predict future success:
• The short answer is NO! None of the test variables could predict whether a junior became a professional.
• There was a very small of predicting who became a ProTour rider based on relative VO2peak. Stress the very small part. Indeed, overall, the statistical model would ultimately predict all of the actual professional cyclists as ultimately being non-professionals.
• Junior riders on the national team had a slightly better chance for becoming ProTour riders than non-national team members.
The Lessons of Youth
The authors summarized their study thus, “traditional measures of aerobic fitness… can differentiate the competitive level of the young cyclists.” In other words, it is possible to say that junior national team riders are distinctly better in these tests than riders who are not on the national team. This cross-sectional analysis may appear useful on one level, because it can help in deciding team selection in a group of racers who, at that stage of their careers, may have the same overall experience and exposure to races, making it difficult to clearly differentiate between athletes.
But it certainly is not that clear cut, as when you look at their entire career, these same measures “are not able to predict the competitive level that junior cyclists can reach in adulthood.” So the lesson here seems to be that fitness tests may seem to be useful in the short term, they are minimally useful in the big picture.
OK, but what lessons can we learn from the differences we see between juniors on the national teams and those who aren’t, especially since there seems to be a bit of predictive power between being on the national team and ultimately riding the ProTour? Here, the vagaries of sports can mask any real discrimatory ability of test measures. For example, the national team riders tended to be bigger and have higher absolute VO2peak values, suggesting that they are the big rouleur types. However, considering that a typical junior race at the Worlds has less climbing (1557 m) than the pros (4179 m), it’s not surprising that bigger rouleurs might be favoured in selection over climbers.
Then why would the riders who were on the Junior national teams have a slightly higher odds of making it to the ProTour? Here again, external factors may trump physiological capacity. Perhaps getting onto the national team opened a lot of doors that remained shut to the other cyclists. For example, maybe this was the catalyst that got them more/better coaching, financial support from sponsors, ability to ride bigger/harder races, and gave them more visibility to pro team managers. All of these little factors may help to skew the ultimate path for one group of riders compared to the other.
So what is the big lesson from this? The biggest lesson is that test data is great for monitoring training and fitness, but it cannot by itself tell us who will be the next Armstrong, Evans, or Cancellara.
Rather than being frustrating, I see this as an incredibly powerful and optimistic result. It tells me that all the young athletes out there just might become our next giant of the road, and that our role should therefore be to provide as much grass-roots support to our juniors and young athletes as possible. Also, rather than quickly assessing someone’s potential early in their career, we should take as much time as possible to assess their future.
So with that in mind, here’s my plea to all of you Pez readers, and especially the Masters riders out there. Do something this year to get more youth into the sport, or else find some way to mentor or support a cadet or junior rider in your club.
Have fun and ride safe!
Menaspa P, Sassi A, Impellizzeri FM. Aerobic fitness variables do not predict the professional career of young cyclists. Medicine and Science in Sports and Exercise 42:805-812, 2010.
Stephen Cheung is a Canada Research Chair at Brock University, and has published over 50 scientific articles and book chapters dealing with the effects of thermal and hypoxic stress on human physiology and performance. He has just published the book Advanced Environmental Exercise Physiology dealing with environments ranging from heat and cold through to hydration, altitude training, air pollution, and chronobiology. Stephen’s currently writing “Cutting Edge Cycling,” a book on the science of cycling, and can be reached for comments at [email protected] .