I love data, I love numbers, I love the development of systems and models, and at the moment I am in my element crunching through ten (10) years of time trial data from the Triathlon Australia Junior program. Now this data set this includes performance records for athletes from short high intensity anaerobic 30 second efforts through longer aerobic efforts, and the element that has piqued my interest is the decay (or fatigue) rate of performance over time is an interest to me both as a measure for performance and an important element for the construction of individual training plans.

As a coach and sport science analyst I am interested in the link between the development of fatigue and exercise performance. For a working definition we think of Fatigue as an ongoing dynamic process during exercise involving central and peripheral factors that limit the power-producing capabilities of an athlete. Fatigue is distinct from task failure, which is defined as the point at which fatigue develops to the point at which it, or its factors, cause inhibition, resulting in a reduction of the desired exercise performance.

The link between fatigue and performance has historically been regarded as elusive; however, in recent years, compelling evidence has indicated that it is embedded within the concept of critical power (CP). The concept of power profiling is not new, and was packaged up well by Coggan and


Allen in their book Training and Racing with a Power Meter, however their modelling built on the work previously done by Monod and Scherrer who introduced the Critical Power (CP) model in in 1965, to describe the power/exhaustion-time relationship. But these are not the only models; there have been a number of formulations of the relationship, such as the power law models dating back to 1906 and the exponential decay function developed by Wilkie in 1960. However the power duration model remains the formulae of choice as it comes closest to achieving both a “good fit” to the underlying data and a reasonable approximation of the underlying physiology.


When the time to the limit of tolerance is plotted against particular constant speeds or power outputs, the relationship is not linear (straight line) but is rather curvilinear (curved path), with the ability to sustain exercise deteriorating more sharply at higher compared with lower power outputs – we can also utilise velocity if we want a variable that we can easily apply to swimming or running. For the math boffins out there this relationship is described as being hyperbolic, and dates back to 1925 when the British physiologist A. V. Hill plotted the relationship between mean speed and sustainable time using world record performance times across a variety of distances for both male and female running and swimming events.

But what does all of this mean for us? Irrespective of the underlying mechanics, the characteristics of the relationship between power output (or speed) and time to task failure are intuitively obvious:

  • maximal neuromuscular power is limited to a few seconds at most;
  • endurance athletes have a high capacity to sustain exercise if the power (or speed) requirements are relatively low;
  • between these two extremes, the sustainable power output declines in a characteristic curvilinear pattern.



Research has identified four intensity domains that represent the physiological response to various exercise outputs:

  1. Moderate – power output below lactate threshold (LT).
  2. Heavy – power outputs between LT and CP.
  3. Severe – power outputs above CP that can be sustained until VO2max is attained.
  4. Extreme – power output resulting in task failure before VO2max is attained.

Within each intensity domain the course of the VO2 response differs, the moderate domain is the sole zone in which a steady state is attained with two to three minutes of exercise onset. Steady state is delayed in the heavy domain by the emergence of a VO2 slow component, increasing the O2 cost of exercise. There is no steady state possible in the severe domain as VO2 climbs toward VO2max, and task failure will occur in the extreme domain before VO2 reaches VO2max. These differences within each domain are mirrored in the muscle metabolic response, suggesting a distinct fatigue mechanism will also characterise each domain.

In cycling and triathlon, where steady state effort and its subsequent quantification is of primary importance, we have long used Critical Power (CP) (Monod) and Functional Threshold Power (FTP) (Coggan) to provide athletes with a reliable measure to calculate their exertion / intensity over time, both within racing and training. We have also utilised the power profile based on watts / kilogram to develop an understanding of how an athlete measures up to world best, the area’s they need to improve and / or the types of racing they are best suited to. We base much of our own coaching practice upon this type of profiling and a foundation upon understanding its application to developing and improving applicable intensity domains, and will often utilise the AIS power profiling test set to capture many of these key metrics within a single session.

For those who want to read more on the AIS test these articles are a good place to start.



The following outlines the power profile of four (4) elite riders within the Cycling Canada National Program, utilising the AIS Power Profile Test in 2014.

Cycling Canada National Testing Protocols (Version 2014.D)

The following table outlines, the same data as watts/kg for those of you who want to compare.

  6s 15s 30s 1min 4min 10min
Senior (F) 13.7 10.7 8.3 6.6 4.8 4.2
Junior (F) 13.6 11.4 8.9 6.3 4.6 4.1
Track U23 (M) 14.7 11.9 9.6 7.5 5.6 4.8
Road U23 (M) 14.6 11.5 9.2 7.4 5.5 4.9

Cycling Canada National Testing Protocols (Version 2014.D)

If we look at the performance decay from 6 seconds to 15 seconds, performance fell by between 16% and 22%, when we stretch this target out to 30 seconds, 1 minute, 4 minutes and 10 minutes, performance fell by an average of 36%, 51%, 64% and 68% respectively. If we think about this decay in performance and compare it to the curvilinear relationship above in the velocity/power chart, we can see this decay in performance output.

The following chart is a representation of the power profile of various cycling disciplines, all from elite or high performance programs. The Professional road cyclist is far and away above the other riders represented, however the high performance BMX racer has an interesting profile, with significantly higher short duration power outputs compared to his longer aerobic power, due to the nature of the BMX events.

Within WKO4 Coggan et al have developed the power duration curve and this is what I use with many of my athlete’s in conjunction with the AIS testing protocol and the odd FTP test when appropriate. I think it is very important that the coach have an understanding of the concepts here, and the application of the appropriate protocols to aid the development of the athlete to meet the needs of competition.

The Sufferfest have recently released the 4DP protocol as methodology to develop and dictate training loads (sound a bit familiar?), and the single test, Full Frontal, to determine the rider’s power profile. I think this is interesting, and I am more talking about the development of the software to alter training loads automatically for differently power numbers. Prior to this with software such as Perf Pro, Trainer Road and Zwift – all of which I use – you have been somewhat limited to basing sessions on the FTP metric, and we know that this is not necessarily appropriate for those more intense short power efforts. Especially if we are designing a session for multiple athletes and the resulting multiple differing power profiles. It would be interesting to have a look at the test and the modelling.

The final note I want to leave you with is that testing and profiling is not the everything, we do not compete in a lab or in the garage (unless we only race on Zwift), but on the road in the sun and wind, dealing with nutrition and other racers. But with testing and power profiling we can gain a better understanding of an athlete’s strengths and weaknesses and develop a program to help them meet the rigors of competition. A plan that takes into account the type of racing they undertake, and address their individual profile. And now I back to where I started, crunching numbers, in the attempt to pull meaning from ten years of data from National junior performance testing. I do love this stuff.


Cycling Canada National Testing Protocols (Version 2014.D)

Gonzalez-Tablas, A., Martin-Santana, E., Torres, M., Designing a Cost-Effective Power Profile Test for Talent Identification Programs, (2016), J. Sci. Cycling. Vol. 5(2).

Novak, A. R., Dascombe, B. J., Physiological and performance characteristics of road, mountain bike and BMX cyclists, (2014) J. Sci. Cycling. Vol. 3(3). 9-16

Pinot, J., Grappe, F., Determination of Maximal Aerobic Power on the field in cycling, J. Sci. Cycling,(2014) Vol. 3(1). 26-32

Quod, M.j., Martin, D.T., Martin, J.C., Laursen, P.B., The Power Profile Predicts Road Cycling MMP, Int. J. Sports Med 2010; 31: 397-402

Leave a Reply

Your email address will not be published. Required fields are marked *