Usually, high-intensity interval training is prescribed as a percentage of VO2max., FTP/CP/Anaerobic Threshold. However, these are, primarily, aerobic parameters. These tend to work for everything below VO2max. (each one with its own nuances, caveats, limits and positive aspects), but, when higher intensities need to be trained, more individualization is needed. Two athletes can have the same VO2max., the same FTP/CP, but can have different anaerobic capabilities. One of them may be able to do 5x30" intervals at 600w, but the other one may only be able to do them at 500w, due to a lower anaerobic capacity.

In order to individualize that training, the APR is a good tool. APR is defined as the difference between the maximal sprint power and the maximal aerobic power. So, performing two tests, we can estimate the APR. The maximal sprint power would equate to the peak sprinting power, while the maximal aerobic power or power at VO2max. (pVO2max.) would equate to ~the peak 6min power (or the lowest power that elicits VO2max. in an incremental test). Then, by using the following exponential formula, we can estimate what power the athlete can produce for any given duration (between 5s and 6min), therefore, optimizing training prescription:

POt = POaer + (POsp – POaer) * e(-k*t) 

where t is the duration for which we want to estimate power (POt), POaer is the maximal aerobic power, POsp is the athlete’s peak power, and k is the exponent showing the decrease in power over time (k=0.026).

However, in very high-level cyclists, using 3' peak power output in place of maximal aerobic power usually works better to estimate the power for any given duration (between 5s and 3'), as the exponent of the formula was developed with runners initially. Sanders has worked with world tour cyclists trying to improve the original APR model, suggesting that using the 3' peak power, instead of the maximal aerobic power, and an exponent of 0.0277 works better. So, in the improved model, POaer would equate to the maximal 3’ power achieved in an all-out 3’ test, and k would equate to 0.0277.

In the photo attached there is an example of the APR model and how it can be used. The graph shows the power-duration curve for 5s to 180s. The model predicts 565w for 45s, 493w for 60s, and 445w for 75s. Let's say that we want to prepare this athlete for a race that is flat but ends with a short steep climb lasting 45-75s. We want to develop this athlete's power for that duration of time, improving his APR. In order to do so, we could plan 45s interval at his predicted 60s power, doing an x number of 45s intervals at ~495w with good recovery to work on the anaerobic/glycolytic system instead of the aerobic system. 

To sum up, in order to plan HIIT sessions some reference of the athlete's anaerobic capacity needs to be taken into account beyond aerobic parameters. APR is a good tool for that, it can give a framework within which we can plan HIIT sessions, and slightly tweak it to adapt it to the athlete. As with all the other parameters, it is another tool in our toolbox, and by using it properly it can be helpful. However, we shouldn't forget everything else and just use APR. It is important to listen to the athlete, understand his characteristics, and analyze how he responds to the impulse. Maybe his predicted best power for 60s is 500w, but he can only repeat 95% of that intensity 4 times within a session, while another athlete could repeat it 6 times. The APR model gives a good tool for coaches and self-coached athletes to individualize HIIT sessions, especially HIIT sessions working on the anaerobic system.


Sanders D and Heijboer M. "The Anaerobic Power Reserve and its Applicability in Professional Road Cycling". J Sports Sci, 2018.

Sanders D, Heijboer M, Akubat I, Meijer K, and Hesselink MK. "Predicting High-Power Performance in Professional Cyclists". Int J Sports Physiol Perform 12: 410-413, 2017.

Weyand PG and Bundle MW. "Energetics of high-speed running: integrating classical theory and contemporary observations".Am J Physiol Regul Integr Comp Physiol288: 956-965, 2005.

Weyand PG, Lin JE, and Bundle MW. "Sprint performance-duration relationships are set by the fractional duration of external force application". Am J Physiol Regul Integr Comp Physiol290: 758-765, 2006.

Get in Contact

If you’d like more information on how Spokes can tailor a training programme to your exact needs, why not check out our products and services. Need more info or would you like to speak to one of our coaches? Get in contact.


More posts by