In the relationship between sport science and elite performance, tales from the road are just as important as the latest experimental findings from the lab. The need for understanding exactly what makes elite athletes tick can provide valuable clues about the potential contributors to elite performance. As my colleague, Dr. Ben Sporer of Canadian Sport Centre – Pacific likes to say, ”science drives practice while practice leads science.” That is, often the best clues or ideas about how to optimize performance comes from seeing what top performers are already doing, and then taking those ideas into the lab, tweaking and experimenting on them, and then re-applying that information back to the road.
One of the real boons in advancing this lab-road relationship has been the introduction and popularization of heart rate and power monitors amongst both top cyclists and recreational and amateur cyclists. This is especially the case for the ability to download and analyze training data. No longer are training data restricted to potentially inaccurate memory or scattered written training logs, but all sorts of important training metrics (heart rate, power, speed, distance, cadence, elevation, etc.) can all be recorded continuously and in incredible detail.
With Polar, SRM, PowerTap, QuarQ, Garmin, and other high-tech monitoring companies wanting to showcase their wares, we have also seen a proliferation of pro rider data being broadcast live during races or being available for download. For most fans, this can serve as a “gee whiz” factor, seeing the wattage of racers up Alpe d’Huez. For scientists and discerning coaches, however, these data are a treasure trove for understanding elite performance. For example, we have previously explored real-life pro race data on the cadence during flat and climbing, and also the power outputs of T-Mobile during a 5-day stage race.
Beyond just the racing data, what is even more valuable would be access to all of the training files of pro cyclists. In essence, the race data gives us a snapshot in time of what the racer can do, but doesn’t tell us how they got there. For example, how much time was spent in different training zones or activities? How hard were the intervals? How much change was there in functional threshold power (FTP) over the year? Taken to a wider perspective, if there was access to data from a broad range of elite cyclists, we can begin to see whether any patterns emerge that predict success. For example, what are the differences in training patterns or power levels between world-beaters versus national-class cyclists?
Nimmerichter et al. 2011
No prizes if you guessed that we’re going to spend this Toolbox talking about a study analyzing the training patterns of a group of world-beaters through elite national-class cyclists! A UK and Austrian research group a complete season’s worth of training and racing data from 11 national/world level cyclists and then set about dissecting the data. Note that this is an observational study, where the scientists simply observed what happened, rather than an experimental study where they sought to manipulate or guide the training in any way.
On one level, this study is a no-brainer, and it is a bit surprising that nobody has published this in the open scientific literature before. On the other hand, there are a number of logistical hurdles to obtain top-flight scientific data. This included ensuring that subjects were willing to divulge their training data to begin with, as many athletes view that as giving away their secrets to their competitors.
The study involved 1 female and 10 male cyclists, all of whom are elite (national/world level) athletes very familiar with power meters and competing for at least six years. Other details on the experimental design:
• Data was tracked from the start of December through the end of October, for an 11 month complete season.
• All cyclists used a SRM monitor, which was calibrated by the experimenters at the beginning of the study. The subjects followed manufacturer calibration protocols prior to each training session. All subjects have used a power meter for at least two years.
• All subjects, in general, periodized their training from December to April with a “two-peak” plan, with one peak in May-June and another from August-October timeframe.
• Training sessions not using the SRM (e.g. most subjects only had the SRM on one bike) and which did not have both power and HR data were removed from analysis. In total, 96% of all available data sets (1800+) were included in the analysis. This encompassed 60% of all training time, and 69% of all cycling time.
• FTP was defined as the peak average power over a 20 min field test, which was performed ~4 times over the course of the year.
• Data was processed using TrainingPeaks WKO+ software and then analyzed.
Seven power zones were defined: Zone 1 < 50% FTP, Zone 2 50-70% FTP, Zone 3 71-85% FTP, Zone 4 86-105% FTP, Zone 5 106-125% FTP, Zone 6 126-170% FTP, Zone 7 >170% FTP.
Types of training workouts were also categorized by the subjects according to pre-arranged categories: “recovery”, “basic aerobic endurance”, “aerobic capacity”, “anaerobic threshold”, “maximal oxygen uptake”, “strength”, “maximal power”, “competition”, and “non-cycling activities.”
Digging into the Data
From such a store of data, an almost infinite number of analyzes are possible (even though many of course are not relevant). The key with such an epidemiological perspective is to have pre-planned analyses with clear rationale for each analysis, rather than just doing a “fishing” expedition where you analyze everything, then make up a reason afterwards for why it may be of importance.
With that in mind, here are some interesting observations and some of my interpretation of what they might mean:
• FTP increased over the season, from 4.7 W/kg to 4.8, 5.0, and ultimately 5.1 W/kg. No big surprise here. Athletes are able to progressively improve their fitness and performance over the course of a season. Note here that their fitness (FTP) was higher during the second peak, suggesting that a well-designed program should have you fitter and faster after adequately recovering and rebuilding from a first peak.
• Number of training sessions (268 +/- 60) and total training time (689 +/- 191 h) were strongly correlated with the subject’s overall classification (e.g. national level versus World Cup winners). This suggests that there simply is no shortcut to the top, and that cycling is a sport where effort in equals effort out. Besides being smart with your available training time, the other ingredient of success seems to be simply more training time. So when considering making the next big step up in your performance (e.g. upgrading a category), quality can equal quantity.
• Total training time was strongly correlated with FTP and VO2max. This observation strongly supports the above correlation with overall rider ranking.
• The amount of time spent in “aerobic endurance” workouts and at the Zone 2 (Endurance) power zones also strongly correlated with both FTP and VO2max. This is a very interesting finding and runs counter to most our pre-conceptions that better fitness comes through better/harder/more intervals. Rather, this points to the importance of developing that big “aerobic engine” as the foundation for better fitness. Indeed, many of us, despite our limited training time, probably do as much higher intensity work as elite riders with double our training volume. So if you can arrange to have a week or two of increased training volume, it may be best to focus on endurance efforts rather than more intervals or high-intensity work.
• The above finding is about total time in aerobic endurance work. However, the overall distribution of training time at different zones were similar across all 11 subjects, averaging ~73 in Zones 1-2, 22% in Zones 3-4, and 5% in Zones 5-7. Note again the preponderance of relatively “easy” endurance work even in elite/world class cyclists. This suggests that the mix of training is about the same across these elite athletes, and again that “quality equals quantity” even at these elite levels.
• FTP was most strongly correlated with total training time spent doing “strength” workouts, which consisted of low cadence (40-60 rpm) high gear efforts for 2-20 min.
• A bit of a “chicken and egg” finding, but the better riders were doing intervals at higher wattages during the “strength” workouts.
• In analyzing the coefficient of variation of power output during training (i.e. how variable was power over the course of training), a strong inverse correlation was observed (i.e., the less variable the power output over the course of workouts, the higher the FTP across subjects). This suggests that the better riders were able to regulate and control their workouts better.
This is quite an interesting study, because it’s one of the first times that we have been able to take an in-depth look at both the racing and especially the training data for a group of elite cyclists over the course of an entire season. The main thing I take away from this study is the power of data analysis. If you have either a downloadable power or heart rate monitor, you should make the effort to look at your files over time to learn how you function best over the course of one or ideally multiple seasons. More than a gee-whiz gizmo, that monitor is the key to understanding your own physiology and responses to training.
Ride safe and have fun!
Nimmerichter A, Eston G, Bachl N, Williams C. Longitudinal monitoring of power output and heart rate profiles in elite cyclists. J Sport Sci 29(8): 831-9, 2011.
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 firstname.lastname@example.org .