Training Levels for a Power Based Traing System: Andrew Coggan
Training and Workouts

Training Levels for a Power Based Traing System: Andrew Coggan

As promised, here is the schema I’ve put together. In developing it, I’ve drawn from a number of sources, including Peter Janssen’s book ‘Lactate Threshold Training‘, Joe Friel’s ‘The Cyclist’s Training Bible’, the British Cycling Federation’s training guidelines (developed by Peter Keen), in addition to my own background in exercise physiology and experience of training and racing with a PowerTap the last couple of years. I would also like to recognize all the people who responded to my initial request for power data, as that has helped me to verify/refine the system. I’ll begin by describing the various ‘levels’ in the system first, then discuss some of the details…

Level 1: Active recovery
Average power: <55% of 40k TT average power
Average heart rate: <68% of 40k TT average heart rate
Perceived exertion: <2
Description: “Easy spinning” or “light pedal pressure”, i.e., very low level exercise, too low in and of itself to induce significant physiological adaptations. Minimal sensation of leg effort/fatigue. Requires no concentration to maintain pace, and continuous conversation possible. Typically used for active recovery after strenuous training days (or races), between interval efforts, or for socializing.

Level 2: Endurance
Average power: 56-75% of 40k TT average power
Average heart rate: 69-83% of 40k TT average heart rate
Perceived exertion: 2-3
Description: “All day” pace, or classic long slow distance (LSD) training (note that the “slow” refers to the very high intensity, interval-centered training programs that were popular at the time the term was coined in the 1970’s). Sensation of leg effort/fatigue generally low, but may periodically to higher levels (e.g., when climbing). Concentration generally required to maintain effort only during very long rides. Breathing is more regular than at level 1, but continuous conversation is still possible. Frequent (daily) training sessions of moderate duration (i.e., 1-2 h) at level 2 are possible (provided dietary carbohydrate intake is adequate), but complete recovery from longer workouts may take more than 24 hours.

Level 3: Tempo
Power: 76-90% of 40k TT average power
Heart rate: 84-94% of 40k TT average heart rate
Perceived exertion: 3-4
Description: Typical intensity of fartlek workout, ‘spirited’ group ride, or briskly moving paceline. More frequent/greater sensation of leg effort/fatigue than at level 2. Requires concentration to maintain alone, especially at upper end of range, to prevent effort from falling back to level 2. Breathing deeper and more rhythmic than level 2, such that any conversation must be somewhat or very halting, but not as difficult as at level 4. Recovery from level 3 training sessions more difficult than after level 2 workouts, but consecutive days of level 3 training still possible if duration isn’t excessive.

Level 4: Threshold
Average power: 91-105% of 40k TT average power
Average heart rate: 95-105% of 40k TT average heart rate (may not be achieved during initial phases of effort(s))
Perceived exertion: 4-5
Description: Just below to just above TT effort, taking into account duration, current fitness, environment, etc. Essentially continuous sensation of moderate or even greater leg effort/fatigue. Continuous conversation difficult at best, due to depth/frequency of breathing. Effort sufficiently high that continuous cycling at this level is mentally very taxing – therefore typically performed in training as multiple ‘repeats’, ‘modules’, or ‘blocks’ of 10-30 min duration. While consecutive days of training at level 4 is sometimes possible, in general such workouts should only be performed when sufficiently rested/recovered from prior training so as to be able to maintain intensity.

Level 5: Aerobic power
Average power: 106-120% of 40k TT average power
Average heart rate: >106% of 40k TT average heart rate (may not be achieved due to slowness of heart rate response and/or ceiling imposed by maximum heart rate)
Perceived exertion: 6-7
Description: Typical intensity of longer (3-8 min) intervals intended to raise VO2max. Strong to severe sensations of leg effort/fatigue, such that completion of more than 30-40 min total training time is difficult at best. Conversation not possible due to often ‘ragged’ breathing. Should only be attempted when adequately recovered from prior training – consecutive days of level 5 work generally not desirable even if possible.

Level 6: Anaerobic capacity
Average power: >121% of 40k TT average power
Average heart rate: N/a
Perceived exertion: >7
Description: Short (

Level 7: Neuromuscular power
Average power: N/a
Average heart rate: N/a
Perceived exertion: * (maximal)
Description: Very short, very high intensity efforts (e.g., jumps, standing starts, short sprints).that generally place greater stress on the musculoskeletal rather than metabolic systems. Power useful as guide, but only in reference to prior similar efforts, not TT pace.


1. Choice of 40k TT power as basis: Average power during a 40k TT provides a logical basis for a training system because it correlates very highly with power at lactate threshold, the most important physiological determinant of endurance cycling performance (since it integrates VO2max, the percentage of VO2max that can be sustained, and cycling efficiency). (Indeed, beyond the first few seconds of exercise the entire power-duration performance curve can be described quite closely using just two mathematical parameters, representing anaerobic capacity and lactate threshold respectively.) While shorter efforts might be more convenient, 40k was chosen because it is a standard distance and because power during a 40k is only slightly less than that generated during shorter TTs. In theory, one could derive specific correction factors to be used with data during shorter TTs (e.g., power during a ~20 min TT will be ~1.05 times that of a 40k) in order to fit such data into the system, but given individual variation in the exact shape of the power-duration curve, day-to-day variability in performance, and the breadth of the specified power levels, this may only convey a false sense of precision. Somewhat along the same lines, one could base a system on laboratory-derived measures, such as lactate threshold itself, but relatively few people have access to such measurements (as opposed to simply going out and measuring their own power during a TT). Conversely, one could dispense with using one single ‘anchor’ measurement, and simply reference all workouts back to the maximum power that an individual can generate for that duration (i.e., Friel’s ‘critical power paradigm’). However, such an approach requires much more testing than simply using TT power, while (in my opinion) providing little, if any, advantage in actual practice.

2. Number of levels: A compromise had to be made between defining more levels, to better reflect the continuum of physiological responses, and defining fewer levels, for simplicity. The seven levels specified were considered the minimum needed to adequately describe the different types of training required/used to meet the demands of competitive cycling. Even with seven levels, though, the range within each is somewhat broad. However, this should not be a major disadvantage, for several reasons. First, there is obviously an inverse relationship between power output and the duration that power can be sustained. Thus, it is axiomatic that shorter training sessions or efforts will be conducted at the higher end of a given range, whereas longer sessions or efforts will fall towards the middle or lower end of a given range. Second, since power is a more precise indicator of exercise intensity than, e.g., heart rate, workouts should still be adequately controlled despite the seemingly large range in power within each level. Finally, as with all systems training exercise prescriptions should be individualized, in this case taking into account the power the athlete has generated in previous similar or identical workouts…the primary reference, therefore, is not to the system itself, but to the athlete’s own unique (and current) ability. In this regard, the classification scheme described above should be viewed primarily as an overall framework, not a detailed plan.

3. Heart rate guidelines: The suggested heart rate ranges must be considered as imprecise, because of individual differences in the positive y-intercept of the power-heart rate relationship. That is, even when power is zero, heart rate is not, with differences between individual in this ‘zero power’ (not resting) heart rate significantly influencing the percentage of 40k TT heart rate corresponding to any given power output. Because of this, I do not believe it is really useful to try to derive power ranges from heart rate ranges (as Friel’s initial attempt to do so readily shows). (Expressing heart rate as a percentage of the range from that at zero power (derived by back-extrapolation of the linear power-heart rate relationship) to that at 40k TT power – akin to the Karvonen formula for heart rate reserve – corrects for this individual effect and allows you to more precisely specify the levels based on heart rate. However, I rejected this approach as simply being too complex, especially given that this is a power-based system.) Nonetheless, I have derived guideline for heart rate (as well as perceived exertion) from power data, such that can be used along with power to help guide training.

4. Perceived exertion guidelines: The values given are from Borg’s 10 point category-ratio scale (reproduced below), not the original 20 point scale that is probably more familiar to most people. This choice was made because the category-ratio scale explicitly recognizes the non-linear response of many physiological variables (e.g., blood and muscle lactate), and thus provides a better indicator of overall effort.
0 = Nothing at all
0.5 = Extremely weak (just noticeable)
1 = Very weak
2 = Weak (light)
3 = Moderate
4 = Somewhat strong
5 = Strong (heavy)
7 = Very strong
10 = Extremely strong
* = Maximal

Since perceived exertion increases over time even at a constant exercise intensity (power), the suggested values or ranges obviously refer to perceived effort as determined relatively early in a training session/series of intervals.

5. Other limitations: While the system is based on the average power during a workout or interval effort, consideration must also be given to the distribution of power within a ride. For example, average power during mass start races typically falls within the range defined as level 3 (‘tempo’), but races are usually more stressful due to the greater variability (and therefore higher peaks) in power. Similarly, due to soft-pedaling/coasting down hills, the same average power achieved during a hilly (or even mountainous) ride will not reflect the same stress as the same average power achieved during a completely flat workout. In part, the variability in power is taken into account in defining the various levels, especially levels 2 and 3 (training at the higher levels is likely to be much more structured, thus tending to limit variations in power). Nonetheless, a workout consisting of, for example, 30 min at level 1 (as warm-up), 60 min at level 3, and another 30 min at level 1 (as warm down) would best be described as a tempo training session, even though the overall average power might fall within level 2 (‘endurance’).

One last caveat: Defining various training ‘levels’ is obviously only the first step in developing a training plan, as what really matters is the distribution of training time or effort devoted to each of the various levels. Discussion of such matters, however, is probably best left to the future, so all I will say at this point is this: 1) I believe that training should be highly individualized, to account for each athlete’s unique abilities, goals, and state of development (e.g., age, training background), and 2) compared to some, I tend to place more value in training at levels 2, 3, and 4 – indeed, what many consider to be ‘junk training’. In that regard, my philosophy apparently parallels that of Peter Keen, or at least how his ideas are seemingly reflected in the BCF guidelines…