As a former UCI professional cyclist, performance coach, and sports scientist, I’ve learned that a single number doesn’t define a rider.
Functional Threshold Power (FTP) may give some suggestions of ability on paper, but an athlete’s ability to repeat efforts, recover between surges, or sustain power late in a race often tells a more complete story.
This is where using different models, like FTP and Critical Power (CP), can provide a fuller picture, giving coaches a better understanding of the power an athlete can actually produce, repeat, and sustain in real-world racing.
Two different models, two different meanings
Despite often being used interchangeably, FTP and CP are not the same. They emerge from different theoretical models and are derived from different methods. Understanding these differences is important as the chosen model directly influences training prescription, fatigue interpretation, pacing strategy, and performance modeling.
For coaches, the real question is not just whether FTP and CP differ, but how those differences affect training zones, athlete profiling, training strategy, race-specific decision-making, and, most importantly, how to use both for more effective training.
What is FTP?
FTP was popularised by Dr. Andrew Coggan and is commonly defined as the highest power output an athlete can steadily sustain for 40-80 minutes while maintaining a relatively stable physiological effort.
In practice, FTP is rarely measured directly through a maximal 60-minute effort. Instead, it’s commonly estimated from shorter tests, most notably 95% of the mean maximal power during a 20-minute time trial.
The original rationale behind FTP was pragmatic rather than mechanistic. With power meters becoming more available, coaches required a field-based metric that was easy to test, repeatable, and useful for prescribing training zones and tracking performance.
Research has shown that FTP-style testing demonstrates acceptable reliability in trained cyclists (McGrath et al., 2019), although its validity at the individual level remains debated (Borszcz et al., 2018).
What is Critical Power?
Critical Power originates from the hyperbolic power-duration relationship first described by Monod and Scherrer (1965). CP represents the asymptote of the power-duration curve and theoretically defines the highest metabolic rate at which physiological homeostasis can still be maintained.
CP is based on the idea that the harder an athlete rides, the less time they can hold that effort. Within this relationship, CP represents the point at which effort shifts from hard but relatively sustainable to an intensity at which fatigue begins to accumulate much more quickly.
Above CP, fatigue develops predictably, and exhaustion occurs following depletion of W′. Below CP, physiological variables can stabilise, and exercise tolerance becomes markedly prolonged. Modern CP testing typically uses multiple maximal efforts between 2-15 minutes, allowing both CP and W′ to be modelled simultaneously (Muniz-Pumares et al., 2019).
The physiological difference
The key distinction is that FTP is primarily a performance-derived construct, whereas CP is a mathematically modeled construct used to determine the physiological boundary between heavy and severe intensity domains.
Several investigations have directly compared FTP and CP and found poor agreement between the two metrics and that they should not be used interchangeably (Karsten et al., 2021). In many athletes, CP is higher than FTP, and the metabolic responses at each intensity are not identical. Research has also questioned whether the FTP test can predict hour-long performances (Wong et al., 2022).
A practical example highlights why this matters. Consider two cyclists who both produce a 20-minute power of 300 watts. Athlete A has a large anaerobic contribution, allowing them to achieve high short-term power, but they struggle to sustain power output during longer steady-state efforts. Athlete B, in contrast, has a stronger aerobic engine and can maintain a more stable power output over prolonged durations. Despite these physiological differences, both riders may end up with the same FTP value, even though their performance profiles, strengths, and training needs are very different.
If both athletes are prescribed the same FTP-based threshold work, Athlete A may be pushed too close to their limit, while Athlete B may be under-prescribed in certain sessions. A CP model, because it accounts for both CP and W′, can help coaches better understand whether an athlete’s performance is driven more by sustainable aerobic power or by finite work capacity above CP.
Advantages and limitations of FTP
FTP remains popular because of its simplicity, accessibility, and integration within major training platforms like TrainingPeaks.
However, FTP has several limitations. The traditional 95% correction from a 20-minute effort assumes athletes have similar fatigue characteristics, which is often not the case. Athletes with a larger anaerobic contribution may overestimate sustainable metabolic steady state, whereas highly aerobic athletes may underestimate it.
FTP values are also sensitive to the testing protocol used, meaning different methods may produce substantially different outputs for individuals (Borszcz et al., 2018).
Advantages and limitations of Critical Power
CP aligns more closely with exercise-intensity domains, lactate dynamics, and VO2 kinetics, giving it stronger mechanistic validity (Jones et al., 2010). The inclusion of W′ also makes CP particularly useful for modelling stochastic race demands such as road racing, MTB, and cyclocross.
Nevertheless, CP testing can be more demanding, requiring multiple maximal efforts as opposed to the single FTP test. Additionally, different mathematical models may also produce slightly different CP values, while the physiological interpretation of W′ remains debated (Clark and MacDermid, 2021).
Why this matters for training zones
Because FTP and CP may produce different values for the same athlete, the choice of model can meaningfully affect training-zone prescription. If CP is higher than FTP, workouts based on CP may place threshold and VO2max intervals at higher absolute intensities than an FTP-based model.
For some athletes, this may be appropriate. For others, it may unintentionally shift sessions into a more fatiguing intensity domain.
Coaches should avoid assuming these values are interchangeable and this is where the individualisation of the CP model comes in.
If using the CP model, you also have the W’ component. Meaning, rather than prescribing supra-threshold work by a percentage of FTP, you can instead individualise the workload of each athlete to target a certain percentage of depletion, improving training prescription and monitoring.
Coaching Tip: Use modelled power for a more complete athlete profile
FTP and CP are useful reference points, but both have limitations when interpreted as single numbers. FTP is often estimated from one field test, while traditional CP models typically rely on a small number of maximal efforts across select durations to estimate the relationship between critical power and W′.
TrainingPeaks’ modelled power system, available in Analyze 360, takes this concept further by using maximal efforts from across the full power-duration curve. Rather than relying on one test result or a handful of durations, the model uses a broader range of data to continually update the athlete’s profile.
At the center of this approach is modelled FTP (mFTP), which can be thought of as a modern descendant of both FTP and CP. Like FTP, it provides a practical estimate of sustainable threshold power. Like CP, it is derived from the power-duration relationship rather than a single fixed-duration test. The difference is that mFTP is informed by the shape of the full curve, making it a more complete and responsive way to evaluate threshold.
TrainingPeaks’ modelled power also includes functional reserve capacity (FRC) and Pmax. Together, mFTP, FRC, and Pmax represent sustainable power, work capacity above threshold, and maximal power production, respectively. This allows coaches to understand not only how much power an athlete can sustain, but also how they produce power across different durations and energy systems.
The Peak Power Chart with mFTP can be found in TrainingPeaks’ Analyze 360. Learn more about how to use it here.
Used well, the power-duration model gives a more complete picture than FTP alone. It helps answer more useful coaching questions, such as: Does the athlete’s profile match the demands of their target event? Is their limiter sustainable threshold power, anaerobic work capacity, sprint power, or the ability to maintain output over time?
FTP is still useful, but it is only one part of the picture. The full power-duration curve provides a richer view of performance and should be used to guide more individualised training decisions.
Practical applications for coaches
FTP may be preferable when working with large groups of athletes, recreational athletes, or in environments where simplicity and adherence are priorities. It remains a useful anchor for setting training zones, prescribing workouts, and tracking progress over time.
CP and W′ may be preferable in settings where coaches need a more detailed understanding of severe-domain physiology, pacing, fatigue resistance, and race demands. Rather than looking only at a single threshold value, CP and W′ allow coaches to consider both sustainable power and the finite amount of work an athlete can perform above that boundary.
From FTP and CP to modelled power
However, the modern coaching conversation does not stop at FTP or CP alone. CP is typically interpreted alongside W′, which represents the finite amount of work an athlete can perform above CP. In TrainingPeaks’ modeled power system, mFTP and FRC serve a similar practical purpose: mFTP estimates sustainable power, while FRC represents an athlete’s capacity to perform work above that threshold.
This matters because racing rarely rewards one isolated number. FTP can provide a useful guide for steady-state pacing, such as long climbs, time trials, or controlled endurance events. But when races involve repeated surges, short climbs, technical accelerations, or attacks, coaches need to understand not only what an athlete can sustain, but also how much work they can produce above that sustainable level.
Modelled power uses the full power-duration curve to evaluate the athlete’s broader profile. For example, two athletes may have similar FTP values but very different FRC capacities. One may be better suited to repeated attacks or punchy terrain, while the other may perform better in long, even-paced efforts.
Ultimately, FTP, CP, W′, mFTP, and FRC should not be viewed as absolute truths. They are models that simplify complex physiology into practical coaching tools.
Used well, FTP remains a simple and accessible anchor for training prescription, CP and W′ offer a more detailed view of the power-duration relationship and severe-domain fatigue, and mFTP and FRC provide a more dynamic view of the athlete’s full power profile. Used carelessly, any of these metrics can create a false sense of precision.
References
Barsumyan, A. et al. (2025). Enhanced durability predicts success in amateur road cycling: Evidence of power output declines. Retrieved from https://doi.org/10.3389/fspor.2025.1530162
Borszcz, F.K. et al. (2018, October). Functional threshold power in cyclists: Validity of the concept and physiological responses. Retrieved from https://doi.org/10.1055/s-0044-101546
Clark, B. & Macdermid, P.W. (2021, May 30). A comparative analysis of critical power models in elite road cyclists. Retrieved from https://doi.org/10.1016/j.crphys.2021.05.001
Jones, A.M. et al. (2010, October). Critical power: Implications for determination of V˙O2max and exercise tolerance. Retrieved from https://doi.org/10.1249/MSS.0b013e3181d9cf7f
Jones, A.M. et al. (2019). The maximal metabolic steady state: redefining the “gold standard.” Retrieved from https://doi.org/10.14814/phy2.14098
Karsten, B. et al. (2021, January 22). Relationship between the critical power test and a 20-min functional threshold power test in cycling. Retrieved from https://doi.org/10.3389/fphys.2020.613151
McGrath, E. et al. (2019, November 1). Is the FTP Test a reliable, reproducible and functional assessment tool in highly-trained athletes? Retrieved from https://doi.org/10.70252/RQOO7391
Monod, H. & Scherrer, J. (1965). The work capacity of a synergic muscular group. Retrieved from https://doi.org/10.1080/00140136508930810
Muniz-Pumares, D. et al. (2019, February). Methodological approaches and related challenges associated with the determination of critical power and curvature constant. Retrieved from https://doi.org/10.1519/JSC.0000000000002977
Wong, S. et al. (2022, December). Functional threshold power is not a valid marker of the maximal metabolic steady state. Retrieved from https://doi.org/10.1080/02640414.2023.2176045
