ArticlePDF AvailableLiterature Review

Abstract and Figures

Emerging trends in technological innovations, data analysis and practical applications have facilitated the measurement of cycling power output in the field, leading to improvements in training prescription, performance testing and race analysis. This review aimed to critically reflect on power profiling strategies in association with the power-duration relationship in cycling, to provide an updated view for applied researchers and practitioners. The authors elaborate on measuring power output followed by an outline of the methodological approaches to power profiling. Moreover, the deriving a power-duration relationship section presents existing concepts of power-duration models alongside exercise intensity domains. Combining laboratory and field testing discusses how traditional laboratory and field testing can be combined to inform and individualize the power profiling approach. Deriving the parameters of power-duration modelling suggests how these measures can be obtained from laboratory and field testing, including criteria for ensuring a high ecological validity (e.g. rider specialization, race demands). It is recommended that field testing should always be conducted in accordance with pre-established guidelines from the existing literature (e.g. set number of prediction trials, inter-trial recovery, road gradient and data analysis). It is also recommended to avoid single effort prediction trials, such as functional threshold power. Power-duration parameter estimates can be derived from the 2 parameter linear or non-linear critical power model: P ( t ) = W ′/ t + CP ( W ′—work capacity above CP; t —time). Structured field testing should be included to obtain an accurate fingerprint of a cyclist’s power profile.
This content is subject to copyright. Terms and conditions apply.
Vol.:(0123456789)
1 3
European Journal of Applied Physiology (2022) 122:301–316
https://doi.org/10.1007/s00421-021-04833-y
INVITED REVIEW
Power profiling andthepower‑duration relationship incycling:
anarrative review
PeterLeo1 · JamesSpragg2· TimPodlogar3,4· JustinS.Lawley1· IñigoMujika5,6
Received: 2 June 2021 / Accepted: 14 October 2021 / Published online: 27 October 2021
© The Author(s) 2021
Abstract
Emerging trends in technological innovations, data analysis and practical applications have facilitated the measurement of
cycling power output in the field, leading to improvements in training prescription, performance testing and race analysis.
This review aimed to critically reflect on power profiling strategies in association with the power-duration relationship in
cycling, to provide an updated view for applied researchers and practitioners. The authors elaborate on measuring power
output followed by an outline of the methodological approaches to power profiling. Moreover, the deriving a power-duration
relationship section presents existing concepts of power-duration models alongside exercise intensity domains. Combining
laboratory and field testing discusses how traditional laboratory and field testing can be combined to inform and individual-
ize the power profiling approach. Deriving the parameters of power-duration modelling suggests how these measures can be
obtained from laboratory and field testing, including criteria for ensuring a high ecological validity (e.g. rider specialization,
race demands). It is recommended that field testing should always be conducted in accordance with pre-established guide-
lines from the existing literature (e.g. set number of prediction trials, inter-trial recovery, road gradient and data analysis). It
is also recommended to avoid single effort prediction trials, such as functional threshold power. Power-duration parameter
estimates can be derived from the 2 parameter linear or non-linear critical power model: P(t) = W/t + CP (W—work capacity
above CP; t—time). Structured field testing should be included to obtain an accurate fingerprint of a cyclist’s power profile.
Keywords Racing· Training· Analysis· Performance· Prediction· Power output
Abbreviations
%
̇
V
O2max Fractional utilization of the maximum oxygen
uptake
2-P CP Two-parameter critical power model
3-P CP Three-parameter critical power model
APR Anaerobic power reserve
ATP Adenosine tri phosphate
BMX Bicycle motocross
CT Critical torque
CP Critical power
CPTTF Time to task failure at critical power
e Basis of the natural logarithm (e = 2.178)
EVA Exposure variation analysis
FPCA Functional principal component analysis
FTP Functional threshold power
GET Gas exchange threshold
GXT Laboratory incremental graded exercise test
k The rate of the exponential decline in power
output
LT Lactate threshold
MAP Maximum aerobic power
Communicated by Michael Lindinger.
* Peter Leo
peter.leo@uibk.ac.at
1 Division ofPerformance Physiology & Prevention,
Department ofSport Science, University Innsbruck,
Innsbruck, Austria
2 Health Physical Activity Lifestyle Sport Research
Centre (HPALS), University ofCape Town, CapeTown,
SouthAfrica
3 Faculty ofHealth Sciences, University ofPrimorska, Izola,
Slovenia
4 Department ofAutomatics, Biocybernetics andRobotics,
Jožef Stefan Institute, Ljubljana, Slovenia
5 Department ofPhysiology, Faculty ofMedicine andNursing,
University oftheBasque Country, Leioa, BasqueCountry,
Spain
6 Exercise Science Laboratory, School ofKinesiology, Faculty
ofMedicine, Universidad Finis Terrae, Santiago, Chile
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
302 European Journal of Applied Physiology (2022) 122:301–316
1 3
MLSS Maximum lactate steady state
MMP Maximal mean power output
OmPD Omni power duration model
P&T Peronnet and Thibault model
P(t) Power output
Pmax Peak power over 1s
SEE Standard error of the estimate
TT Cycling time trial
TTF Time to task failure
̇
V
O2 Oxygen uptake
̇
V
O2max Maximum oxygen uptake
W Work capacity above critical power
WEP Work above end test power
Introduction
Since the invention of the first mobile power meter for
cycling in the late 1980s training and racing with this tool
has become standard practice in multiple cycling disciplines
including road, track, mountain bike, cyclo-cross, bicycle
motocross (BMX) and triathlon. Mechanical power output
measured by strain gauges, most commonly mounted in the
bike’s crank spindle, crank arm or pedal spindle and con-
nected to a head unit mounted in the handlebar allows power
output data to be accurately recorded in field conditions in
real time (Maier etal. 2017). This enables an in-depth analy-
sis of a cyclist’s mechanical power output during training
and/or competition, and the assessment of an athlete’s endur-
ance capacity outside of a laboratory setting (Passfield etal.
2017).
These aforementioned technological innovations have
allowed applied scientific research to be undertaken in
cycling, including real-time measurements of internal (e.g.
heart rate) and external (e.g. power output) workloads (van
Erp and de Koning 2019; Mujika 2017; Muriel etal. 2021;
Padilla etal. 2000; Padilla etal. 2008). This in turn allows
the demands of racing to be described (Ebert etal. 2005,
2006; van Erp etal. 2021b; Menaspà etal. 2015; Menaspà
etal. 2013; Vogt etal. 2007b), training/racing performance
analysis to be conducted (Leo etal. 2021c; Lucia etal. 2001;
Mujika and Padilla 2001; Pinot and Grappe 2011) and train-
ing prescription to be quantified (Leo etal. 2020; Sanders
etal. 2020; Sanders and Heijboer 2019a).
Power profiling in cycling is most commonly defined as
the assessment of field derived power outputs, i.e. values
obtained during training and racing (Coggan 2003; Leo etal.
2020). Power profiling can be used for the tracking of longi-
tudinal changes in performance and race analysis (Leo etal.
2021b). There is a growing interest in the theoretical and
practical implications of power profiling. However, to date,
there is no consensus on what constitutes the best practice
for power profiling, especially given that there are numerous
methodological issues and approaches. Therefore, the aim
of this narrative review is to present and discuss existing
practices and methods, their implementation, interpreta-
tion, and practical applications, provide recommendations
to unify power profiling approaches for both practice and
research, and suggest future directions for research.
Measuring power output
Before analysing power output data, it is important to under-
stand how power output is measured during cycling and any
associated methodological issues. In cycling, when a force
is created by the muscles and applied perpendicular to the
bicycle crank arm, one crank arm revolution creates two
angular impulses (one per leg); this results in forward drive.
Optimal force production, and as a result optimal forward
drive, is a complex interplay of innervation, muscle recruit-
ment patterns, the contractile function of muscle as well
as an elastic tendon–muscle interaction and metabolic pro-
cesses occurring in these tissues. The properties of force
generation are often described using physics expressions
such as mean torque or mean power output; the former
describing the force and the latter the amount of work pro-
duced in a given time (Winter etal. 2016). Power output is
often expressed as a steady-state value (e.g. 100W), but this
value is a product of many impulses over a given period of
time or a given proportion of the pedal stroke. Some have
argued that ‘mean power output’ is therefore a more accurate
descriptor (Winter etal. 2016). Notwithstanding the validity
of this argument, for the purposes of this review the authors
will employ the customarily used term ‘power output
throughout. However, it should be noted that power output
does not include the energy used to accelerate the cyclist’s
limbs nor forces applied in non-propulsive directions.
Mechanical (or external) power output can either be meas-
ured by strain gauges or calculated mathematically (Maier
etal. 2017; Martin etal. 1998). Depending on the position of
the strain gauge (e.g., pedal spindle, crank, bottom bracket),
the recorded power output is expected to deviate slightly as
some energy is lost via drivetrain inefficiencies (Coyle etal.
1991; Maier etal. 2017; Martin etal. 1998). This highlights
that power output values derived from different strain gauge
positions may not be comparable. Likewise, different power
meter brands and models have different levels of trueness
and precision. Maier etal (2017) found that while on average
commercially available power meters record at a trueness of
0.9 ± 3.2% some units will deviate by more than 5%. The
authors also reported that some power meter brands have
significantly greater precision than others.
On average Maier etal. (2017) found that the small-
est worthwhile change for the accuracy of commercially
available power meters was 1.1–2.8%. This implies that
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
303European Journal of Applied Physiology (2022) 122:301–316
1 3
any performance improvements of less than 1.1% cannot
be accurately quantified by commercially available power
meter devices. However, this value may differ from brand to
brand and model to model. Validation studies have been con-
ducted for most commercially available power meters, but
there is no agreed-upon gold standard to which power meters
should be compared. Therefore, researchers and practitioners
should take note of the comparative measure when assess-
ing the validity of any power output measuring device. We
draw the reader’s attention to the aforementioned study by
Maier and colleagues (2017) for a broader discussion of the
methodological issues surrounding power meter validation.
To ensure high data quality the authors strongly recommend
accurate calibration according to the manufacturer’s recom-
mendations prior to the collection of any power meter data.
Additionally, dynamic (Gardner etal. 2004), static (Wooles
etal. 2005), and day-to-day calibration, known as ‘zero-
offsetting’ are all recommended before data derived from
power meters are used for power profiling purposes.
Methodological approaches topower
proling
Numerous methodologies have been applied in the field
of power profiling. The most basic of these is simply the
reporting of average power output values for a given race or
event (Ebert etal. 2005; Vogt etal. 2007a, b). While this is
the starting point in understanding the demands of a given
event, it fails to fully utilise the full potential of power pro-
filing. Another disadvantage is that unless data are derived
from cyclists with differing performance levels within an
event, this approach does not provide any information on the
demands of peak performance, instead it merely describes
the demands of participation.
A more advanced approach is to describe the power out-
put by time at a given intensity. This approach is normally
described as ‘binning’. Binning is where each power output
value is categorized into a bin; each bin represents a range
of intensities (for example 100–200W). The resulting cat-
egorization of each output value can then be expressed as
either total cumulative time in each bin or as a percentage of
total time. (Abbiss etal. 2010; Ebert etal. 2006; Leo etal.
2021b; Metcalfe etal. 2017). Typically, but not always, the
bins are defined by normalizing the power output to body
mass (for example 4–5W kg−1). However, the suitability of
this approach can be questioned; for example, in some events
aerodynamic drag is a far more important factor than body
mass (Pringle etal. 2011). Besides scaling power output
relative to the frontal area (Padilla etal. 1999), to the best
of the authors’ knowledge no studies have been published
where the bins represent ranges of power output values nor-
malised to aerodynamic drag (W CdA−1).
Binning has advantages in that it can describe the range
of intensities that are required to compete or perform in a
given event. Typically, cycling events are not completed
at a fixed power output; instead, power output is stochas-
tic in nature, even in individual time trials (Gordon 2005).
Whilst binning allows the total time at different intensities
to be described, there are weaknesses with this approach.
Firstly, the choice of the range of intensities for a given bin
will influence the results. Often arbitrary bins are chosen,
based on a given power output normalized to body mass,
for example 5.9–7.9W kg−1. If the range of intensities is
too wide the granularity of the power output data cannot be
captured. Another problem is that binning gives no insight
into the length of individual efforts. The cumulative time
in each power output bin may represent one long effort or
multiple short efforts. Finally, if arbitrary bins are used then
it may be that the range of intensities covered by a single
bin includes power outputs that are both sustainable and
unsustainable from a physiological point of view. A solu-
tion to this problem is to use physiological thresholds to
define the bins (Abbiss etal. 2010; Passfield etal. 2013).
For example, the submaximal physiological thresholds that
define the exercise intensity domains could be determined
during laboratory testing and used to define the bins. While
this approach does give a greater insight into the physiology
of a given event for individual athletes, problems occur when
data from multiple athletes are amalgamated, as the bins,
while representing consistent physiological responses, do
not necessarily represent the same absolute or relative power
output for all athletes.
As previously mentioned, one of the main problems with
binning is that duration of individual efforts are not rep-
resented within the data. However, there is a small body
of work that uses exposure variation analysis (EVA) to try
and overcome this limitation (Abbiss etal. 2010; Passfield
etal. 2013). This approach uses a two-bin system; one set of
bins is used in the traditional manner to describe the inten-
sity. Bins can be associated with either arbitrary values or
physiological thresholds. The second set of bins is used to
describe the duration of individual efforts. Here arbitrary
durations are used, for example 0–5s, 5–10s or > 1min.
The intensity bins are plotted on the x-axis, the duration of
individual efforts is plotted on the z-axis and the percentage
of total race time is plotted on the y-axis (see sample data
in Fig.1).
Whilst the exact power output of individual efforts is
still not displayed, EVA is a very powerful tool to show the
pacing strategy and stochastic nature of power output in a
given event. This approach may be especially powerful to
describe events where lots of short submaximal sprints are
interspersed by periods of recovery, for example cyclo-cross
or Olympic cross country mountain biking. EVA is an effec-
tive way to describe the duration of efforts and recovery
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
304 European Journal of Applied Physiology (2022) 122:301–316
1 3
bouts. This information can be valuable for coaches and
practitioners when prescribing interval training sessions to
replicate the demands of an event.
A major limitation of the approaches discussed thus far is
that they fail to describe power outputs for individual efforts.
To do this the mean maximal power output (MMP) approach
can be used (van Erp and Sanders 2020; Puchowicz etal.
2020; Quod etal. 2010; Vogt etal. 2007b). MMP values
represent the highest average power that was recorded for a
given (arbitrary) duration, during an event. For example, the
highest average power output recorded over a 5min duration
in a race would be the 5min MMP. Such MMP data are very
valuable because they make it possible to identify the power
output and duration that a cyclist is required to produce to
be competitive in an event. For example, MMP data analysis
shows that a top male general classification contender in a
grand tour is required to produce 5.8W kg−1 for 20min on
key mountain climbs (van Erp etal. 2020a, b). For coaches
and practitioners this is very valuable information.
There are however some fundamental issues with MMP
data. Firstly, it is not known if the recorded MMP values
were derived from a maximal effort. This contrasts with
values derived from formal testing where the maximality
of an effort can be verified. For example, in a traditional
laboratory incremental graded exercise test (GXT) a given
perception of effort and respiratory exchange ratio need to be
obtained for the test to be considered maximal in nature and
therefore a valid maximum oxygen uptake (
̇
V
O2max) value to
be obtained (Jones etal. 2016). It is hypothesised that almost
none of the MMP values derived from races are maximal
in nature. If a rider were to produce a maximal effort at
any point other than at the finish of a race, it may compro-
mise their ability to subsequently follow their competitors
in bunch events or compromise their pacing strategy in indi-
vidual events (Leo etal. 2021b, c). Secondly, MMP data
from a specific (arbitrary) duration could be the result of
the bracketing of a subsection of a longer effort, or a shorter
duration effort and a subsequent recovery (Leo etal. 2021a,
b). For example, it is very unlikely that a 5-min MMP value
derived from a race represents a maximal effort of exactly
5min in duration. As a result, there is a high probability of
an inherent underestimation of maximal power output when
using MMP values alone. MMP data are only indicative of
what a cyclist did, not what the cyclist is capable of.
Another issue with MMP data in research is that there is
no agreed-upon set of (arbitrary) durations that are being
applied. This means that when trying to compare data from
various studies coaches and practitioners cannot perform
like for like comparisons. This situation has improved some-
what as research groups have started to incorporate a wider
range of MMP durations from ~ 5 to ~ 1800s. This allows a
power-duration curve to be developed using the MMP val-
ues, allowing for some comparisons between studies. A final
issue with MMP analysis is that it may not actually define
‘race winning efforts’. Recent work by Leo and colleagues
(2021b) and van Erp and colleagues (2021a; b) showed that
the power output that cyclists produce falls throughout an
event; and that MMP values are not predictive of race per-
formance. Instead, it is the power output that riders produce
at key moments in the race that is predictive of performance.
For example, in the case of a sprinter in road cycling it is
the power that they can produce in the final moments of the
race that is important, but this is not necessarily the same
as their 10s MMP. This means that MMP analysis may
be missing the very efforts that it is trying to identify. To
better identify these race-winning efforts an approach has
been taken in research whereby the event is broken down
into segments and MMP values in each segment have been
reported (van Erp etal. 2021a; Leo etal. 2021b; Sanders and
van Erp 2021). To date, these segments have been defined
Fig. 1 EVA—exposure varia-
tion analysis in the final hour of
a race in six U23 cyclists (N = 6)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
305European Journal of Applied Physiology (2022) 122:301–316
1 3
via accumulated work, either absolute values or normal-
ized to body mass, for example MMP values after 2.500kJ
of work. However, this approach, which has thus far only
been applied in road cycling has introduced some further
limitations. Road cycling is a team sport in which riders
perform individual tasks such as sheltering a team leader or
collecting nutrition from a following car. It is not the goal of
every rider to try and win the race. Therefore, the reported
decrease in MMP values, as accumulated work increases,
may partially be a product of the fact that some riders have
simply finished their tasks and are therefore no longer pro-
ducing maximal efforts.
To alleviate the problem of arbitrary MMP durations not
matching actual effort durations, some studies have selected
specific sections of the event and identified power output
exclusively in that section (Jobson etal. 2008; Leo etal.
2021c; Padilla etal. 2008). For example, Leo and colleagues
(2021a; b) looked at MMP values exclusively on classified
climbs. This approach, while potentially beneficial in certain
circumstances, does require researchers to identify the key
moments in races for analysis. While this may be possible
for some events, such as a road race stage that starts out flat
and concludes with a mountain top finish, it is not always
possible to accurately identify the key moment in a race. A
possible solution to this is to seek the input of athletes when
identifying the key periods in the race. Whilst an attrac-
tive proposition, to the best of the authors’ knowledge this
approach has not been used in published research.
As mentioned before, the major issue with MMP analysis
is the uncertainty surrounding whether an effort was maxi-
mal in nature, and whether the MMP duration is equal to the
effort duration. To counter this problem, the authors recom-
mend using power output values derived from formal test-
ing to provide a comparative measure to MMP values. This
approach has particular benefits for coaches and practition-
ers as comparisons between MMP data and formal testing
data can be used to monitor changes in the power profile;
and if a rider records a MMP value which exceeds the pre-
diction from formal testing a new formal performance test
can be scheduled. This is particularly useful when analysing
performance in timed events where the in-competition power
output and event duration can be compared to the theoreti-
cal power-duration relationship. This example highlights
the importance of developing a power-duration relationship
rather than simply using standard duration performance
tests, as the likelihood of the test and competition durations
being identical is low. For methodological issues surround-
ing the development of theoretical power duration relation-
ships please see the section ‘Deriving a power-duration
relationship’ below.
Unfortunately, this approach (i.e. comparing MMP
against a pre-established theoretical power-duration relation-
ship derived from prior formal testing) was only undertaken
by a few research groups (Leo etal. 2020; Leo etal. 2021b;
Nimmerichter etal. 2020; Quod etal. 2010). However, all
research has shown good to very good agreement between
power output values from formal testing and MMP values.
Of particular interest is work by Leo and colleagues (2020,
2021a) that shows the formal testing values are only predic-
tive of race performance for a 6-month period before formal
re-testing is required.
Methodological issues
Thus far we have discussed methodological approaches in
power profiling, however, there are also methodological
issues that are pertinent to all approaches. Recorded power
output values can be highly influenced by the topography of
the event (Padilla etal. 2000, 2008; Sanders and Heijboer
2019a), differences between single day and multi-day stage
racing (van Erp and de Koning 2019; van Erp and Sanders
2020; Lucía etal. 2003) and race category (Sanders and
van Erp 2021). In professional road cycling race category
was found to influence power output: higher power outputs
over shorter durations (< 2min) were reported in lower-
ranked races, and higher power outputs over longer durations
(> 10min) were observed in races with higher difficulty.
Another important consideration when performing power
profiling are environmental factors. Altitude, temperature,
and humidity can all influence the power output athletes can
produce. Therefore, from a research perspective the authors
recommend that the environment and race conditions should
be reported whenever possible.
Recent research has also shown that power profiling anal-
ysis conducted exclusively on either training or racing data
produces different results in the same participants (Leo etal.
2020). This is an important factor and further highlights
the need to provide adequate information on the context in
which any power profiling data were collected.
Finally, in competition settings, alongside the aforemen-
tioned issues surrounding team roles there is an influence of
other team-mates and competitors on power output due to
drafting, which lowers the power output requirement for a
given speed (Ouvrard etal. 2018, van Druenen and Blocken
2021). Research has also suggested that competition may
influence the pacing strategy adopted by cyclists (Bossi etal.
2018).
Deriving apower‑duration relationship
When power output is plotted against time to task failure
(TTF) a consistent power-duration relationship emerges
(Burnley and Jones 2018). The first researchers to math-
ematically describe this relationship were Monod and
Scherrer (1965) who analysed muscle fatigue during static
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
306 European Journal of Applied Physiology (2022) 122:301–316
1 3
and dynamic work (knee extension exercise) and created a
mathematical model describing the hyperbolic relationship
between completed work and TTF. Due to the strong scien-
tific evidence over decades (Burnley and Jones 2018; Jones
etal. 2010; Poole etal. 2016) the power-duration relation-
ship can be considered to represent an integrative approach
to the limits of tolerable exercise in humans.
From a physiological perspective the power-duration
relationship is comprised of four distinct exercise intensity
domains; namely, moderate, heavy, severe, and extreme
(Burnley and Jones 2007), which are characterised by dis-
tinct whole-body physiological responses (Jamnick etal.
2020; Vanhatalo etal. 2016; Whipp 1996). While a complete
physiological background on the systemic and mechanistic
bases of the power-duration relationship would be beyond
the scope of this narrative review, interested readers are
referred to the following review articles: Burnley and Jones
(2018), Jones etal. (2010), Poole etal. (2016), Poole etal.
(2021), Vanhatalo etal. (2016).
Various models are available to coaches and practitioners
to model the power-duration relationship for use in power
profiling (Sreedhara etal. 2019). However, most models only
cover a specific section of the power-duration relationship
(see Fig.2).
Modelling power output intheextreme exercise
intensity domain
Previous research (Bundle etal. 2003; Bundle and Wey-
and 2012; Weyand etal. 2006) has demonstrated that the
anaerobic power reserve (APR) is capable of predicting
short duration (< 3min) power outputs within the extreme
exercise intensity domain, where
̇
V
O2max may not be attained
before task failure occurs. The APR approach was initially
developed in laboratory settings where the maximum aero-
bic power (MAP) recorded during a GXT and the maximal
power an athlete can produce over one pedal revolution or
over one second (Pmax) are used as parameter inputs. How-
ever, Sanders etal. (2017, 2019b) developed a field testing
method where 3min MMP can be used as a surrogate for
MAP. In this approach the time constant (k), which can be
defined as the rate of the exponential decline in power out-
put (i.e. the reciprocal of the corresponding time constant:
k = 1/τ), can be varied between values of 0.024–0.027 to best
Fig. 2 An illustration of the spectrum of physiological responses
across the power-duration relationship using arbitrary power output
values. Pmax 1s peak power, W work above critical power, CP criti-
cal power, LT lactate threshold, GET gas exchange threshold, APR
anaerobic power reserve model, 2-P CP two-parameter critical power
model, 3-P CP three-parameter critical power model, P&T Peronnet
and Thibault Model, OmPD omni power duration model
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
307European Journal of Applied Physiology (2022) 122:301–316
1 3
fit the MMP data. This allows for an individualisation of the
power-duration relationship modelling, which may provide
a better fit (Sanders and Heijboer 2019b) [see sample data
in Fig.3 and Table1 (Eq.1)].
Alongside the APR model, power output in the extreme
exercise intensity domain can also be predicted using the
three-parameter critical power (3-P CP) (Morton 1996), the
Peronnet and Thibault model (P&T) (1989) and Puchow-
icz’s omni power duration model (OmPD) (Puchowicz etal.
2020). It should be noted that in the P&T model, Pmax is pro-
vided as a parameter estimate, whereas in the APR model,
3-P CP model and the OmPD model Pmax is required as an
input parameter. These different modelling approaches con-
siderably influence power output predictions in the extreme
exercise intensity domain (see Fig.4).
Modelling power output inthesevere exercise
intensity domain
Multiple approaches based on the CP concept have been
proposed to predict power outputs within the severe exercise
intensity domain. Although all CP models are equivalent
from a mathematical perspective (i.e. they can be derived
mathematically from one another) they produce different sta-
tistical parameter estimates for CP and work above CP (W)
(Jones etal. 2010; Muniz-Pumares etal. 2019), and there-
fore slightly different predictions within the severe exercise
intensity domain; particularly at the extremes of the domain.
The 3-P CP model (Morton 1996) aimed to overcome these
limitations for short duration power outputs toward the upper
end of the severe and into the extreme exercise intensity
domain by incorporating Pmax as a model parameter, but it
still overestimates power outputs in the moderate exercise
intensity domain (see Fig.4).
Modelling power output belowthecritical power
The CP represents the theoretical asymptote of the power-
duration curve, suggesting that a given power output is infi-
nitely sustainable. However, this is clearly not the case for
real-world performances where exercise at the CP is limited
to 20–40min (Poole etal. 2016). For this reason, previ-
ous research (Peronnet and Thibault 1989; Puchowicz etal.
0306090120 150 180
0
500
600
900
1200
1500
Time (s)
PowerOutput (W)
Fig. 3 Sample data for the anaerobic power reserve model, black
dots—record power output over 5, 10, 15, 30, 60, 90, 120 and 150s
durations; horizontal black dashed line:—anaerobic power reserve;
green, blue and red dashed lines representing the power duration
curve with the rate constant (k) of the exponential decline in power
output (k = 0.024, k = 0.026, k = 0.027) according to Sanders and Hei-
jboer (2019b)
Table 1 Power-duration models corresponding to the respective exercise intensity domains
Equation1: P(t) power output, P(3-min) 3min field test, P(max) 1s peak power, e base of the natural logarithm (2.718), k the rate constant of the
exponential decline in power output, t time in seconds
Equation2: t time in seconds, Wʹ work above critical power, P power output, CP critical power, P(max) 1s peak power
Equation3: P(t) power output, Wʹ work above critical power, CP critical power, t time in seconds
Equation4: Pmap(t) power output at maximum aerobic power, MAPTTF time to task failure at maximum aerobic power, t time in seconds, A rep-
resents a fixed constant for the decline in power output over time, Ln natural logarithm to the base of e (2.718)
Equation5: P(t) power output, Wʹ work above critical power, CP critical power, t time in seconds, CPTTF time to task failure at critical power, A
represents a fixed constant for the decline in power output over time, Ln natural logarithm to the base of e (2.718)
Exercise intensity domains Model Equation
extreme Anaerobic power reserve
P(t)
=P
(
3
min)
+
(
P
(max)
P
(
3
min
))
×e(k×t
)
(1)
extreme and severe 3-parameter critical power model
t
=
W
P
CP
+
W
CP
Pmax
(2)
Severe 2-parameter critical power model
P
(t)=
W
t
+
CP
(3)
extreme, severe and heavy Peronnet and Thibault model
Pmap
(t)=MAP A×Ln
(
t
MAPTTF )
;t>MAP
TTF
(4)
Omni power duration model
P
(t)=W
t
×
(
1et×
P
max
CP
W
)
+CP;tCP
TTF
(t)=W
t×
1et×
max
W
+CP A×Ln
t
CP
;t>CP
(5)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
308 European Journal of Applied Physiology (2022) 122:301–316
1 3
2020) has suggested an exponential decay term below the
CP to predict power outputs in the heavy exercise intensity
domain (see Fig.2 and Eqs.4 and 5). However, these decay
terms are not necessarily routed in the underlying physiol-
ogy of fatigue in the heavy and moderate exercise intensity
domains (see Black etal. (2017), Clark etal. (2019) and
Amann (2011) for overviews of possible fatigue mecha-
nisms at these intensities). They do however represent the
best models to date for estimating exercise tolerance below
the CP (see equations in Table1).
Choosing amodelling approach
The authors recommend that coaches and practitioners refer
to the physiological demands of a given discipline or train-
ing modality to guide their choice. They should then select
the model that best predicts the power-duration relationship
across the range of intensities in which athletes will train and
race. For example, the two-parameter CP model (Moritani
etal. 1981; Whipp etal. 1982) overestimates both short-
and long-duration power outputs outside the severe exer-
cise intensity domain (see Fig.4), thus potentially limiting
its utility. To give some practical examples; power outputs
in the team sprint falls exclusively in the extreme exercise
intensity domain, whereas power outputs in the individual
pursuit falls within both the extreme and severe exercise
intensity domains (Gardner etal. 2005). In road cycling a
large proportion of the power output falls within the heavy
and moderate exercise intensity domains (van Erp and de
Koning 2019); however, power outputs in the extreme and
severe exercise intensity domains are more important in
predicting race performance (Menaspà etal. 2017). Longer
duration (ultra) endurance events, for example, ironman dis-
tance triathlons (Laursen 2011) or the ‘Race Across Amer-
ica’ (Hulton etal. 2010) fall within the moderate exercise
intensity domain, as do extensive training sessions in cycling
or triathlon (van Erp etal. 2020b; Laursen 2011). A differ-
ent modelling approach may be required for each of these
examples.
Interestingly some of the aforementioned models are able
to predict exercise tolerance in multiple exercise intensity
domains. Whilst there is a considerable body of evidence
indicating that the physiological responses in each exercise
intensity domain is unique (Burnley and Jones 2007), it
should be noted that most research is derived from exer-
cise intensities that are not in close proximity to the thresh-
olds that define a given exercise intensity domain. Work
by Pethick and colleagues (2020) looking at responses in
proximity to the critical torque (CT) during isolated knee
extension exercise, a proxy for CP, showed that above the
CT participants displayed physiological responses consist-
ent with the severe exercise intensity domain. Likewise,
slightly below the CT physiological responses associated
with the heavy exercise intensity domain were recorded.
Another pertinent example is that research has shown that
although the
̇
V
O2 slow component is a defining characteris-
tic of the heavy exercise intensity domain, a variant of the
slow component, albeit smaller in magnitude, also occurs in
the moderate exercise intensity domain (Davies and Thomp-
son 1986). Whilst a proportion of the change in
̇
V
O2 uptake
may be due to a shift in substrate utilisation, this change
wouldn’t account for the entire increase in
̇
V
O2, suggesting
altered or additional muscle recruitment (Burnley and Jones
2018). Together, these findings suggest that rather than each
exercise intensity domain inducing distinct physiological
responses, there is instead a spectrum of responses across the
power-duration relationship (see Fig.2). Indeed, this would
explain why the power-duration curve is smooth in nature
and doesn’t contain ‘turn-points’ as would be expected if
the thresholds between exercise intensity domains were
indeed ‘hard’ in nature. It may also explain why some of
the aforementioned models are able to predict exercise toler-
ance across intensities in multiple exercise intensity domains
(Fig.2 and Table1).
Combining laboratory andeld testing
Both laboratory and field testing have been used in isola-
tion and in conjunction with each other to investigate physi-
ological and performance capacity in cycling (Gardner etal.
0
200
300
600
900
1200
1500
1800
Power Output (W)
0
5s
30s
1-min
2-min
15-min
40-min
1-h
2-h
4-h
MMPOmPD P&T
2-PCP3-PCP
Fig. 4 Various power duration modelling approaches applied to
the same MMP data. MMP Mean Maximum Power, OmPD Omni
Power Duration model, P&T Peronnet and Thibault model, 2-P CP
two-parameter critical power model, 3-P CP three-parameter critical
power model; horizontal dashed line—critical power asymptote; ver-
tical dashed lines represent the approximate transitions between the
exercise intensity domains (extreme, severe, heavy and moderate)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
309European Journal of Applied Physiology (2022) 122:301–316
1 3
2007; Jobson etal. 2009; Jones and Vanhatalo 2017; Lucia
etal. 2001; Paton and Hopkins 2001).
In cycling, the most commonly reported measures from
laboratory testing include peak power output from sprint-
ing or graded incremental exercise tests,
̇
V
O2max, %
̇
V
O2max,
MAP, fractional utilization of MAP, first and second lactate
or ventilatory thresholds, maximum lactate steady state and
cycling efficiency (Laurent etal. 2007; Lucia etal. 2000;
Mujika and Padilla 2001). Although good agreement exists
between some of these laboratory measures and cycling per-
formance, none of the aforementioned physiological vari-
ables can be used to create a power-duration relationship
as recommended by the authors for the purposes of power
profiling.
As demonstrated before, a critical component of the
power-duration relationship is the border between the
heavy and severe exercise intensity domains; power outputs
at which a steady state can and cannot be achieved (Poole
etal. 2016; Poole etal. 1988). The physiological bound-
ary between these domains has been most associated with
endurance performance (Burnley and Jones 2007; Poole
etal. 1988). For a long time, the maximum lactate steady
state (MLSS) was considered as the gold standard for this
boundary (Billat etal. 2003; Keir etal. 2015; Kilding and
Jones 2005). However, recent work (Galán-Rioja etal. 2020;
Jamnick etal. 2020; Jones etal. 2019; Nixon etal. 2021) has
suggested that CP better estimates the maximal metabolic
steady state, the highest power output where a steady state
in the oxygen uptake (
̇
V
O2) response can still be observed,
despite increasing blood lactate values (Bräuer and Smekal
2020). There is still some debate as to which method (if any)
is superior for differentiating between metabolic steady state
and non-steady state exercise, and whether both MLSS and
CP can actually be used interchangeably (Jones etal. 2019;
Keir etal. 2015; Nixon etal. 2021; Poole etal. 1988).
In applied settings, it has been suggested that an alter-
native approach, namely the functional threshold power
(FTP), can be used as a surrogate for the maximal meta-
bolic steady state: (Mackey and Horner 2021). FTP was
first described as the cycling power output that can be sus-
tained for one hour in a “quasi physiological steady-state”
(Bassett etal. 1999; Coggan 2003; Mackey and Horner
2021). FTP is therefore a surrogate of the 60min MMP.
It has been proposed that FTP can also be predicted either
by taking 95% of the power output in a 20-min maximal
field test (Borszcz etal. 2018; Morgan etal. 2019; Valen-
zuela etal. 2018) or by taking 90% of the power output
in a 8-min maximal field test (Sanders etal. 2020); the
former being commonly used (Valenzuela etal. 2018). In
contrast to CP and MLSS, where multiple determination
trials are required, FTP can be predicted from a single trial
and is, therefore, less time consuming. This time efficient
approach may explain why the concept has been widely
adopted in cycling (Mackey and Horner 2021). However,
whilst CP and MLSS can be considered as estimates of the
maximal metabolic steady state (Keir etal. 2015; Poole
etal. 1988), this cannot be confirmed for FTP (Morgan
etal. 2019). Whilst both MLSS and FTP are single-param-
eter estimates, the CP concept can be used to predict TTF
for a range of power values within the severe exercise
intensity domain and provides an estimate of the border
between metabolic steady state and non-steady state exer-
cise. The same cannot be said for either MLSS or FTP,
which can only predict a single point on the power-dura-
tion relationship, or a border between exercise intensity
domains, but not TTF for a range of power output values.
Physiologically speaking, CP has been shown to represent
the highest power output at which there is no progressive
derangement in the muscle metabolite milieu (Burnley and
Jones 2018); however, instead of a ‘hard’ border, the CP
represents a phase transition between the heavy and severe
exercise intensity domains (Pethick etal. 2020). Mitchell
and colleagues (2018) also reported a strong relationship
between CP and muscle capillary density, underpinning the
aerobic component of CP. Similarly, Vanhatalo etal. (2016)
demonstrated that CP was strongly associated with the per-
centage of highly oxidative type I muscle fibres. Above CP,
in the severe exercise intensity domain a non-metabolic
steady state occurs, characterized by a reduction in intra-
muscular creatine phosphate stores, continuously increasing
concentrations of inorganic phosphate, hydrogen ions and
blood lactate, which are all associated with a reduced con-
tractile function of the working muscle (Allen etal. 2008;
Burnley and Jones 2007, 2018; Jones etal. 2010; Poole etal.
2016, 1988).
Although a strong relationship exists between FTP and
CP estimates (Denham etal. 2020; Karsten etal. 2020; Mor-
gan etal. 2019, Mackey and Horner 2021), and FTP and
MLSS (Borszcz etal. 2019), the cited studies have demon-
strated that the limits of agreement between parameters are
too large for them to be used interchangeably. This ques-
tions the relevance of FTP (Borszcz etal. 2018; Karsten
etal. 2020; Morgan etal. 2019; Valenzuela etal. 2018).
Furthermore, Borszcz and colleagues (2018) demonstrated
that the 95% of 20min power output overestimates 60min
power output, and recommended that 20min power output
alone should be used for training prescription and perfor-
mance monitoring, rather than trying to make estimates
of 60min power output (i.e. FTP). After all, both 20 and
60min power output are arbitrary in nature. However, whilst
FTP might represent an arbitrary value, rather than a physi-
ological threshold, it may still have practical utility in terms
of informing the training process (Valenzuela etal. 2018).
However, to the best of the authors’ knowledge no studies
exist that compare performance outcomes when prescribing
training based on different concepts, i.e. FTP, CP and MLSS.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
310 European Journal of Applied Physiology (2022) 122:301–316
1 3
That said, for the reasons outlined above the authors
consider CP the most useful concept in terms of deriving a
power-duration relationship, and therefore recommend the
use of the CP concept in the field of power profiling.
Deriving theparameters ofpower‑duration
modelling
There is currently no consensus on how best to derive the
parameters that are needed to model the power-duration rela-
tionship; namely Pmax, CP and W. Likewise, there is consid-
erable debate on which mathematical model should be used
to derive CP and W (Maturana etal. 2018; Muniz-Pumares
etal. 2019; Nimmerichter etal. 2020).
Traditionally, performing three to five prediction trials
between 2 and 15min of duration (Karsten etal. 2015; Mat-
urana etal. 2018; Muniz-Pumares etal. 2019) allows CP and
W to be derived through weighted least square or geometric
mean linear and nonlinear regression analysis (Vinetti etal.
2017; Vinetti etal. 2020). Prediction trials shorter than 2min
do not ensure the attainment of
̇
V
O2max (i.e. they fall outside
the severe intensity domain) (Hill and Smith 1994; Maturana
etal. 2018; Muniz-Pumares etal. 2019; Nimmerichter etal.
2020), while prediction trials longer than 15min are not
recommended due to the influence of glycogen depletion and
psychological factors (i.e. motivation) (Karsten etal. 2015;
Maturana etal. 2018). To avoid any skewness during the
mathematical modelling and reduce errors in the calculation
of CP and W the shortest prediction trial should last between
2 and 5min and the longest prediction trial between 12 and
15min (Karsten etal. 2015; Maturana etal. 2018; Muniz-
Pumares etal. 2019). Inter-trial recovery between prediction
trials should be set to a minimum of 30min during a single
visit or 24h during multiple days (Karsten etal. 2017). The
benefit of multiple days if that any fatigue induced by the
initial prediction trial does not affect the subsequent one, but
possible error due to day-to-day variation in power output
is introduced.
Once the performance trials have been completed the
respective power output and trial duration values can be
used to derive CP and W. Computing CP and W estimates
from a nonlinear two- or three-parameter models requires
access to statistical software to perform a weighted least
square or geometric mean regression analysis (Vinetti
etal. 2017, 2020). To simplify this process for coaches
and practitioners there are two options available to lin-
earize the hyperbolic power-duration relationship (see
Fig.5). Practitioners can either use a) the linear work time
CP model (see Eq.3 and Fig.5c or b) the linear power
inverse of time CP model (see Eq.3 and Fig.5b), where
CP and W can be derived as the slope and intercept of
the linear relationship (Clarke and Skiba 2013; Sreedhara
etal. 2019). All mathematical models from Fig.5 provide
a high accuracy for the model fit, but there is a possibility
that the power-duration parameter estimates (CP and W)
diverge somewhat depending on which fitting method is
used (Muniz-Pumares etal. 2019). As a result, Hill (1993)
suggested that the best fit mathematical model could be
more objectively selected, where the model producing the
lowest standard error of the estimate (SEE) should be the
0200 400600 800
0
300
400
500
600
Time (s)
Poweroutput(W)
0.0000.002 0.0040.006 0.0080.010
0
300
400
500
600
1/Time
Poweroutput(W)
0200 400600 800
0
100000
200000
300000
Time (s)
Work (J)
a
b
c
Fig. 5 Graphical illustration of the power-duration relationship for the
hyperbolic (a), inverse of time (b) and linear work time (c). Model
adopted from Clarke and Skiba (2013)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
311European Journal of Applied Physiology (2022) 122:301–316
1 3
preferred way to derive CP (Hill 1993; Muniz-Pumares
etal. 2019).
CP and W parameter estimates can also be derived using
only two prediction trials (Parker Simpson and Kordi 2017).
While this can be seen as a time-efficient testing protocol,
the limitation of this approach is that the linear relation-
ship always results in a perfect fit (R2 = 1.0). In addition, no
parameters for the goodness of fit (i.e., SEE) can be derived.
Therefore, it is recommended to use at least three prediction
trials to ensure a low standard error for CP (2–5%) and W
(< 10%) (Black etal. 2016; Dekerle etal. 2015). Performing
three prediction trials and using a two-parameter CP model
to fit the data results in one degree of freedom. For instance,
a standard error of 5W for a cyclist with a CP of 385W
would then need to be multiplied by 12.7 to calculate the
95% confidence limits (± 64W) in both directions. Adding a
fourth prediction trial would reduce the CP standard error to
3W and the 95% confidence limits (± 38W) thus improving
the CP predictive ability.
The 3-min all out test has also been proposed as a more
time efficient way to derive CP and W (Vanhatalo etal.
2007, 2008). The principal assumption in this test is that W
or more accurately WEP (work above end test power) as it is
known in this test, is fully depleted within the first 150s and
therefore during the last 30s only CP (end test power) can
be sustained. Despite showing good reliability and validity
compared with traditional CP testing in some circumstances
(Wright etal. 2017), other research in elite cyclists shows
significantly higher CP estimates are derived from the 3-min
all out test than traditional protocols (McClave etal. 2011)
which can lead to overestimation of performance capacity
in the severe exercise intensity domain (Nicolò etal. 2017).
This finding brings into question whether the 3-min all out
test can be used in the field of power profiling.
In some power-duration models (see Table1) Pmax is
an additional input parameter when modelling the power-
duration relationship. Extensive research (Douglas etal.
2021; Driss and Vandewalle 2013; McCartney etal. 1983,
1985) was conducted on the assessment and mechanisms of
Pmax in cycling (Sargeant etal. 1981). Assessing Pmax in the
laboratory or field settings requires a thoughtful reflection
on testing protocols. Recent research used the highest 1s
power output within 4s, 10s and 15s sprints to derive Pmax
(Driss and Vandewalle 2013; Ferguson etal. 2021; Gardner
etal. 2007; Sanders and Heijboer 2019b). If efforts longer
than 10s are used Pmax could be negatively influenced as the
cyclist may apply a pacing strategy (Driss and Vandewalle
2013; Gardner etal. 2007). Practitioners should also be
aware of a “learning effect” during all-out sprint efforts, and
it is therefore recommended that adequate familiarization is
undertaken prior to formal testing of Pmax. Additional impor-
tant factors to consider when testing Pmax in a laboratory
setting are; the torque factor setting (Forbes etal. 2014) and
whether the expected Pmax is within the range of validity of
the power measuring device. For example, a commercially
available smart trainer is only valid up to 700W, which is
much lower than the expected Pmax for some populations
(Zadow etal. 2016).
Ecological validity
Cadence, body position as well as topography, i.e. level
ground or uphill conditions, have also been shown to influ-
ence model parameter estimates due to different biomechani-
cal recruitment patterns (Bertucci etal. 2005; Kordi etal.
2019; Nimmerichter etal. 2012). Therefore, rider specializa-
tion (for example climber vs. time trial specialist) and race
demands (uphill vs. flat, on-road vs. off-road, etc.) need to
be considered in the selection of testing environments (Nim-
merichter etal. 2012). The testing conditions should mirror
the conditions in which athletes are expected to perform.
For example, it is recommended that time trial specialists
perform prediction trials on a time trial bike on level ground,
while climbing specialists conduct testing in uphill condi-
tions on a road bike.
Previous research has also investigated whether time
trials or TTF trials should be favoured as prediction trials
(Coakley and Passfield 2018; Karsten etal. 2018). Tradi-
tionally, TTF trials have been based on a fixed percentage
(i.e. 80–105%) of the power output in a GXT. The main
limitation with this approach being that inter-individual dif-
ferences could influence the trial duration (Jamnick etal.
2020). In contrast, maximum effort time trialling requires a
high level of pacing ability and may therefore only be suit-
able for use with experienced cyclists (Karsten etal. 2018).
However, time trials are inherently easier to perform in field
settings, as Simpson and Kordi (2017) have shown a particu-
larly time-effective protocol using time trials in elite athletes
can produce valid CP and W´ estimates. However, in less
trained participants higher power output values have been
reported in TTF trials resulting in higher CP and W´ estima-
tions (Coakley and Passfield 2018).
As mentioned above, environmental factors should be
considered whenever performing any formal testing. Testing
conditions during formal testing should therefore aim to mir-
ror as closely as possible the competition settings to ensure
environmental validity. To illustrate this point, CP has been
shown to decline significantly as altitude increases, while
W only decreased above 4.000m of altitude (Townsend
etal. 2017); heat and humidity have been shown to influ-
ence power outputs in formal testing (Racinais etal. 2015).
Previous research also investigated the influence of
cadence on time trial performance and power-duration
parameter estimates. While CP estimates were higher at
cadences 60 vs. 100 revolutions per minute in recreation-
ally trained individuals (Barker etal. 2006; Carnevale and
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
312 European Journal of Applied Physiology (2022) 122:301–316
1 3
Gaesser 1991), no statistically significant differences in
physiological determinants (gross efficiency, energy turno-
ver) were reported at cadences between 80 vs. 100 revolu-
tions per minute in elite cyclists during cycling time trials
(Foss and Hallén 2005). Although higher power outputs can
be achieved at lower cadences, elite cyclists tend to prefer
higher cadences around ~ 90 revolutions per minute despite
reductions in cycling efficiency.
Agreement betweenmodelled
power‑duration relationship andMMP
values
Good agreement between CP estimates derived from for-
mal testing and MMP values has been reported (Leo etal.
2020, 2021a; b; Nimmerichter etal. 2020; Quod etal. 2010).
While a good agreement between CP derived from formal
testing and racing has been shown, the same cannot be
confirmed for W. Both Leo etal. (2020) and Karsten etal.
(2015) reported low agreement between W´ derived from
formal testing and MMP data. This low agreement may be
due to cyclists not performing maximal efforts in race situ-
ations apart from very specific circumstances (i.e., during
time trials or at the finish of races). If cyclists were to fully
deplete W´ in any other circumstance (i.e. uphill mountain
finish, lead out or tine trial), there is a chance that they may
subsequently not be able to match the power requirement to
follow the peloton. These scenarios have direct implications
on the recorded MMP values thereafter, as they are not as
high as the MMP values recorded earlier in the race (Leo
etal. 2021b). Thus, these efforts are not being captured via
basic MMP analysis per se.
Good agreement has been reported between power out-
puts predicted by the APR model and race-derived MMP
data for short duration power outputs (< 2min) in profes-
sional male cyclists (Sanders etal. 2017; Sanders and Heij-
boer 2019b). However, only limited research exists to verify
if this approach could also be applied to other populations.
Future directions
Although many approaches concerning power profiling have
been developed in the literature, it remains unclear which
approach provides the greatest insight. Arguably, the most
convenient way for practitioners to create a power profile
would be to retrospectively use field derived MMP data
from training and racing over pre-defined durations (Ebert
etal. 2005; Menaspà etal. 2017; Sanders and van Erp 2021;
Vogt etal. 2007b). Although this kind of data may provide
valuable insights into racing demands in highly trained
cyclists, little information can be retrieved in terms of the
power-duration relationship due to the arbitrary selection of
MMP values.
Deriving a comparative measure allows longitudinal
analysis: for example, if a rider records a MMP value in
racing which exceeds the prediction from formal testing,
practitioners can use that information to monitor changes
in the power profile. However, deriving W´ from racing or
field testing has shown poor predictive ability (Karsten etal.
2015; Leo etal. 2021a) questioning the practical utility of
W for power profiling purposes. When creating a theoreti-
cal power-duration curve from formal testing, care should
be taken that the appropriate models are used. For example,
application of the CP concept outside the severe exercise
intensity domain involves an overestimation in short MMP
(< 2min) ability and long duration MMP (> 40min) sus-
tainability. For this reason, the APR model provides a use-
ful concept to predict the power-duration relationship in the
extreme exercise intensity domain.
While the power-duration relationship in the severe exer-
cise intensity domain has been well investigated based on
the CP concept (Jones etal. 2010; Poole etal. 2016), limited
research exists on deriving the power-duration relationship
in the moderate and heavy exercise intensity domains (Black
etal. 2017). Hence Puchowicz etal. (2020) and Peronnet
and Thibault (1989) proposed mathematical models with an
aerobic decay term, but limited research exists to assess if
these concepts have a high predictive ability for the power-
duration relationship in the moderate and heavy exercise
intensity domains in relation to the muscle bioenergetic sys-
tem (Korzeniewski 2019; Korzeniewski and Rossiter 2020,
2021; Vanhatalo etal. 2016).
Recent work (van Erp etal. 2021b; Leo etal. 2021b) has
shown a reduction in MMP values as prior work increases.
However, future research is needed to better understand the
mechanisms which lead to alterations in the power-dura-
tion due to fatigue, especially the influence of the exercise
intensity and if work in different exercise intensity domains
induce the same degree of downward shift in the power-
duration curve. This is important as improved performance
capacity, i.e. smaller alterations in the power-duration rela-
tionship, has been positively related to race success (van Erp
etal. 2021b; Leo etal. 2021b).
In the era of big data science a novel approach intro-
duced by Puchowicz (2018) on the Golden Cheetah open
data project (Liversedge 2020) could provide novel insights
into power profiling. Functional principal component anal-
ysis (FPCA) enables an in-depth view of the components
of variability in MMP data between cyclists via eigenfunc-
tions which classify riders on their phenotype (sprinter vs.
climber) and performance level. Currently, however, the use
of FPCA for the purposes of power profiling still requires
adequate scientific validation before any potential findings
can be applied by coaches and practitioners.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
313European Journal of Applied Physiology (2022) 122:301–316
1 3
Practical recommendations inapplied
settings
Based on the current literature and the authors’ experience
conducting power profiling in applied settings, the fol-
lowing recommendations can be made as a starting point
for coaches and practitioners: to derive the parameters
to model a power-duration curve a formal test protocol
should include one sprint effort (i.e. ~ 10–15s) and at least
three maximum efforts between 2 and 15min (Karsten
etal. 2015; Leo etal. 2021a; Muniz-Pumares etal. 2019;
Sanders and Heijboer 2019b). These efforts can be com-
pleted in a single testing session, though it is recom-
mended to divide field testing into two sessions over two
consecutive days. The order of efforts should preferably be
randomized for scientific research or follow the cyclist’s or
coach’s individual preference in applied settings. Inter-trial
recovery between efforts should be set to a minimum of
30min of active recovery (< 2 rating of perceived exer-
tion) (Karsten etal. 2017). CP and W should be derived by
the non-linear two-parameter CP model (Muniz-Pumares
etal. 2019), while Pmax should be referred to the 1s peak
power during the ~ 10–15s sprint effort (Sanders and Hei-
jboer 2019b). This protocol will allow coaches and practi-
tioners to derive valid Pmax, CP and W estimates. Coaches
can then choose the best modelling approach based on the
exercise intensity domain(s) that are important for race
analysis and training prescription in a given discipline.
Power meters should be verified for accurate and reli-
able measurement and a zero-offset or re-calibration
according to the manufacturer’s recommendations is
recommended.
The authors do not recommend using single effort field
tests (i.e. 8min or 20min TT) to derive the FTP estimate
because it lacks physiological background and only rep-
resents a single point on the power-duration curve. Nor
do they recommend the use of the 3- min all-out test as
this may lead to an overestimation of the power-duration
relationship in the severe exercise intensity domain.
To increase the ecological validity of power profiling
we recommend a careful selection of the power-duration
modelling approach, based on biomechanical and physi-
ological principles. Standardized laboratory and field test-
ing should be conducted in line with performance analysis
from training and racing to increase the practical utility of
performance prediction and training related consequences.
In addition, any formal testing should consider the envi-
ronmental and topographical conditions in which the power
profile information is to be applied in. Therefore, the dura-
tion of the effort, gradient, inter-trial recovery, rider type
specialization (climbers vs. flat specialist) and race demands
(climb vs. time trial) should be replicated as best possible.
Collectively, power profiling provides an advanced
opportunity for performance modelling based on power
output data from training and racing in combination with
traditional laboratory and field-testing methods to maximize
cycling performance.
Acknowledgements We would like to thank Dr. Richard Ferguson and
Mag. Dieter Simon for their theoretical input and assistance in real-
izing this project.
Author contributions PL, JS and IM had the idea of the article. PL, JS
and TP performed the literature search and writing and and graphical
illustration. JS, TP, JL and IM critically revised the manuscript.
Funding Open access funding provided by University of Innsbruck
and Medical University of Innsbruck. No funding was received for the
preparation of this manuscript.
Availability of data and material Not applicable.
Code availability Not applicable.
Declarations
Conflict of interest Author P.L., J.S., T.P., J.L. and I.M. declare that
they have no conflict of interests.
Ethics approval Not applicable.
Consent to participate Not applicable.
Consent for publication All authors agree to the publication of the
submitted manuscript.
Open Access This article is licensed under a Creative Commons Attri-
bution 4.0 International License, which permits use, sharing, adapta-
tion, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons licence, and indicate if changes
were made. The images or other third party material in this article are
included in the article's Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in
the article's Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will
need to obtain permission directly from the copyright holder. To view a
copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
References
Abbiss CR etal (2010) Examining pacing profiles in elite female
road cyclists using exposure variation analysis. Br J Sports Med
44(6):437–442
Allen DG, Lamb GD, Westerblad H (2008) Skeletal muscle fatigue:
cellular mechanisms. Physiol Rev 88(1):287–332
Amann M (2011) Central and peripheral fatigue: interaction dur-
ing cycling exercise in humans. Med Sci Sports Exerc
43(11):2039–2045
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
314 European Journal of Applied Physiology (2022) 122:301–316
1 3
Barker T, Poole DC, Noble ML, Barstow TJ (2006) Human critical
power-oxygen uptake relationship at different pedalling frequen-
cies. Exper Physiol 91(3):621–632
Bassett DR etal (1999) Comparing cycling world hour records, 1967–
1996: modeling with empirical data. Med Sci Sports Exerc
31(11):1665–1676
Bertucci W etal (2005) Effects on the crank torque profile when chang-
ing pedalling cadence in level ground and uphill road cycling. J
Biomech 38(5):1003–1010
Billat VL etal (2003) The concept of maximal lactate steady state:
a bridge between biochemistry, physiology and sport science.
Sports Med 33(6):407–426
Black MI etal (2016) The constant work rate critical power protocol
overestimates ramp incremental exercise performance. Eur J
Appl Physiol 33(6):407–426
Black MI etal (2017) Muscle metabolic and neuromuscular deter-
minants of fatigue during cycling in different exercise intensity
domains. J Appl Physiol 122(3):446–459
Borszcz FK etal (2018) Functional threshold power in cyclists:
validity of the concept and physiological responses. Int J
Sports Med 39(10):737–742
Borszcz FK, Tramontin AF, Costa VP (2019) Is the functional thresh-
old power interchangeable with the maximal lactate steady
state in trained cyclists? Int J Sports Med 14(8):1029–1035
Bossi AH etal (2018) Pacing strategy and tactical positioning during
cyclo-cross races. Int J Sports Physiol Perform 13(4):452–458
Bräuer EK, Smekal G (2020) VO2 steady state at and just above
maximum lactate steady state intensity. Int J Sports Med
41(9):574–581
Bundle MW, Weyand PG (2012) Sprint exercise performance: does
metabolic power matter? Exerc Sport Sci Rev 40(3):174–182
Bundle MW, Hoyt RW, Weyand PG (2003) High-speed running per-
formance: a new approach to assessment and prediction. J Appl
Physiol 95(5):1955–1962
Burnley M, Jones AM (2007) Oxygen uptake kinetics as a determi-
nant of sports performance. Eur J Sports Sci 7(2):63–79
Burnley M, Jones AM (2018) Power–duration relationship: physiol-
ogy, fatigue, and the limits of human performance. Eur J Sports
Sci 18(1):1–12
Carnevale TJ, Gaesser GA (1991) Effects of pedaling speed on the
power-duration relationship for high-intensity exercise. Med
Sci Sports Exerc 23(2):242–246
Clark IE etal (2019) Dynamics of the power-duration relationship
during prolonged endurance exercise and influence of carbo-
hydrate ingestion. J Appl Physiol 127(3):726–736
Clarke DC, Skiba PF (2013) Rationale and resources for teaching the
mathematical modeling of athletic training and performance.
Am J Physiol 37(2):134–152
Coakley SL, Passfield L (2018) Cycling performance is superior for
time-to-exhaustion versus time-trial in endurance laboratory
tests. J Sports Sci 36(11):1228–1234
Coggan A (2003) Level II coaching manual training and racing with
a power meter: an introduction
Coyle EF etal (1991) Physiological and biomechanical factors
associated with elite endurance cycling performance. Med Sci
Sports Exerc 23(1):93–107
Davies CTM, Thompson MW (1986) Physiological responses to
prolonged exercise in ultramarathon athletes. J Appl Physiol
61(2):611–617
Dekerle J etal (2015) Exercise tolerance can be enhanced through a
change in work rate within the severe intensity domain: work
above critical power is not constant. PLoS ONE 10(9):1–15
Denham J, Scott-Hamilton J, Hagstrom AD, Gray AJ (2020) Cycling
power outputs predict functional threshold power and maxi-
mum oxygen uptake. J Strength Cond 34(12):3489–3497
Douglas J, Ross A, Martin JC (2021) Maximal muscular power: les-
sons from sprint cycling. Sports Med Open 7(1):48–63
Driss T, Vandewalle H (2013) The measurement of maximal (anaer-
obic) power output on a cycle ergometer: a critical review.
BioMed Res Int 2013:1–40
Ebert TR etal (2005) Power output during women’s world cup road
cycle racing. Eur J Appl Physiol 95(5):29–36
Ebert TR, Martin DT, Stephens B, Withers RT (2006) Power out-
put during a professional men’s road-cycling tour. Int J Sports
Physiol Perform 1(4):324–335
Ferguson HA, Harnish C, Chase JG (2021) Using field based data
to model sprint track cycling performance. Sports Med - Open
7(1):1–12
Forbes SC, Kennedy MD, Boule NB, Bell G (2014) Determination
of the optimal load setting for arm crank anaerobic testing in
men and women. Int J Sports Med 35(10):835–839
Foss Ø, Hallén J (2005) Cadence and performance in elite cyclists.
Eur J Appl Physiol 93(4):453–462
Galán-Rioja MÁ, González-Mohíno F, Poole DC, González-Ra
JM (2020) Relative proximity of critical power and metabolic/
ventilatory thresholds: systematic review and meta-analy-
sis. Sports Med 50(10):1771–1783. https:// doi. org/ 10. 1007/
s40279- 020- 01314-8
Gardner AS etal (2004) Accuracy of SRM and power tap power
monitoring systems for bicycling. Med Sci Sports Exerc
36(7):1252–1258
Gardner SA etal (2005) Power output demands of elite track sprint
cycling. Int J Sports Perform Analys 5(3):149–154
Gardner AS etal (2007) Maximal torque-and power-pedaling rate rela-
tionships for elite sprint cyclists in laboratory and field tests. Eur
J Appl Physiol 101(3):287–292
Gordon S (2005) Optimising distribution of power during a cycling
time trial. Sports Eng 8(2):81–90
Hill D (1993) The critical power concept. Sports Med 16(4):237–254
Hill DW, Smith JC (1994) A method to ensure the accuracy of esti-
mates of anaerobic capacity derived using the critical power
concept. J Sports Med Phys Fit 34(1):23–37
Hulton AT etal (2010) Energy expenditure in the race across America
(RAAM). Int J Sports Med 31(7):463–467
Jamnick NA etal (2020) An examination and critique of cur-
rent methods to determine exercise intensity. Sports Med
50(10):1729–1756
Jobson SA, Woodside J, Passfield L, Nevill AM (2008) Allomet-
ric scaling of uphill cycling performance. Int J Sports Med
29(09):753–757
Jobson SA etal (2009) The analysis and utilization of cycling training
data. Sports Med 39(10):833–844
Jones AM, Vanhatalo A (2017) The ‘critical power’ concept: applica-
tions to sports performance with a focus on intermittent high-
intensity exercise. Sports Med 47(s1):65–78
Jones AM etal (2010) Critical power: implications for determina-
tion of VO2max and exercise tolerance. Med Sci Sports Exerc
42(10):1876–1890
Jones AM etal (2016) Sport and exercise physiology testing guidelines:
the British Association of Sport and Exercise Sciences Guide.
Routledge
Jones AM etal (2019) The maximal metabolic steady state: redefining
the ‘gold standard.’ Physio Rep 7(10):e14098
Karsten B etal (2015) Validity and reliability of critical power field
testing. Eur J Appl Physiol 115(1):197–204
Karsten B etal (2017) Comparison of inter-trial recovery times for the
determination of critical power and W’ in cycling. J Sports Sci
35(14):1420–1425
Karsten B etal (2018) Time trials versus time-to-exhaustion tests:
effects on critical power, W’, and oxygen-uptake kinetics. Int J
Sports Physiol Perform 13(2):183–188
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
315European Journal of Applied Physiology (2022) 122:301–316
1 3
Karsten B etal (2020) Relationship between the critical power test
and a 20-min functional threshold power test in cycling. Front
Physiol 11:1877
Keir DA etal (2015) Exercise intensity thresholds: identifying the
boundaries of sustainable performance. Med Sci Sports Exerc
47(9):1932–1940
Kilding AE, Jones AM (2005) Validity of a single-visit protocol to
estimate the maximum lactate steady state. Med Sci Sports Exerc
37(10):1734–1740
Kordi M, Fullerton C, Passfield L, Parker Simpson L (2019) Influence
of upright versus time trial cycling position on determination of
critical power and W in TRAINED CYCLISTS. Eur J Sports
Sci 19(2):192–198
Korzeniewski B (2019) Pi-induced muscle fatigue leads to near-
hyperbolic power-duration dependence. Eur J Appl Physiol
119(10):2201–2213
Korzeniewski B, Rossiter HB (2020) Exceeding a ‘critical’ muscle Pi:
implications for VO2 and metabolite slow components, muscle
fatigue and the power-duration relationship. Eur J Appl Physiol
120(7):1609–1619
Korzeniewski B, Rossiter HB (2021) Factors determining training-
induced changes in V̇O2max, critical power, and V̇O2 on-kinetics
in skeletal muscle. J Appl Physiol 130(2):498–507
Laursen PB (2011) Long distance triathlon: demands, preparation and
performance. 6(2): 247–63
Laurent CM, Meyers MC, Robinson CA, Green JM (2007) Cross-val-
idation of the 20-versus 30-s wingate anaerobic test. Eur J Appl
Physiol 100(6):645–651
Leo P etal (2020) Training characteristics and power profile of pro-
fessional U23 cyclists throughout a competitive season. Sports
8(12):167
Leo P, Spragg J, Simon D, Lawley J etal (2021a) Climbing perfor-
mance in U23 and professional cyclists during a multi-stage race.
Int J Sports Med (in print)
Leo P, Spragg J, Simon D, Mujika I etal (2021b) Power profiling,
workload characteristics and race performance of U23 and pro-
fessional cyclists during the multistage race tour of the alps. Int
J Sports Physiol Perform (in press)
Leo P, Spragg J, Menz V etal (2021c) Power profiling in U23 profes-
sional cyclists during a competitive season. Int J Sports Physiol
Perform 16(6):881–889
Liversedge M (2020) GoldenCheetah OpenData Project.” https:// osf.
io/ 6hfpz/. 16 Feb 2021
Lucia A, Hoyos J, Chicharro JL (2000) Physiological response to pro-
fessional road cycling: climbers vs. time trialists. Int J Sports
Med 21(7):505–512
Lucia A, Hoyos J, Chicharro JL (2001) Physiology of professional road
cycling. Sports Med 31(5):325–337
Lucía A etal (2003) Tour de France versus Vuelta a España: which is
harder? Med Sci Sports Exerc 35(5):872–880
Mackey J, Horner K (2021) What is known about the FTP20 test
related to cycling? A scoping review. J Sports Sci 1(aop):1–11
Maier T etal (2017) Accuracy of cycling power meters against a
mathematical model of treadmill cycling. Int J Sports Med
38(6):454–461
Martin JC etal (1998) Validation of a mathematical model for road
cycling power. J Appl Biomech 14(3):276–291
Maturana FM etal (2018) Critical power: how different protocols and
models affect its determination. J Sci Med Sport 21(7):742–747
McCartney N, Heigenhauser GJ, Jones NL (1983) Power output and
fatigue of human muscle in maximal cycling exercise. J Appl
Physiol 55(1):218–224
McCartney N, Obminski G, Heigenhauser GJ (1985) Torque-veloc-
ity relationship in isokinetic cycling exercise. J Appl Physiol
58(5):1459–1462
McClave SA, LeBlanc M, Hawkins SA (2011) Sustainability of critical
power determined by a 3-minute all-out test in elite cyclists. J
Strength Cond Res 25(11):3093–3098
Menaspà P etal (2015) Physical demands of sprinting in professional
road cycling. Int J Sport Nutr Exerc Metab 36(13):1058–1062
Menaspà P, Quod M, Martin DT (2013) Physiological demands of
road sprinting in professional and U23 cycling. A pilot study. J
Sci Cyc 2(2):35–39
Menaspà P, Sias M, Bates G, La Torre A (2017) Demands of world cup
competitions in elite women’s road cycling. Int J Sports Physiol
Perform 12(10):1293–1296
Metcalfe AJ etal (2017) Within-season distribution of external train-
ing and racing workload in professional male road cyclists. Int J
Sports Physiol Perform 12(2):2–142
Mitchell E, Martin N, Bailey SJ, Ferguson RA (2018) Critical power
is positively related to skeletal muscle capillarity and type I
muscle fibers in endurance-trained individuals. J Appl Physiol
125(3):737–745
Monod H, Scherrer J (1965) The work capacity of a synergic muscular
group. Ergonomics 8:329–338
Morgan PT etal (2019) Road cycle TT performance: relationship to
the power-duration model and association with FTP. J Sports
Sci 37(8):902–910
Moritani T, Ata AN, Devries HA, Muro M (1981) Critical power as
a measure of physical work capacity and anaerobic threshold.
Ergon 24(5):339–350
Morton RH (1996) A 3-parameter critical power model. Ergon
39(4):611–619
Mujika I (2017) Quantification of training and competition loads in
endurance sports: methods and applications. Int J Sports Physiol
Perform 12(2):229
Mujika I, Padilla S (2001) Physiological and performance char-
acteristics of male professional road cyclists. Sports Med
31(7):479–487
Muniz-Pumares D, Karsten B, Triska C, Glaister M (2019) Methodo-
logical approaches and related challenges associated with the
determination of critical power and curvature constant. J Strength
Cond 33(2):584–596
Muriel X etal (2021) Physical demands and performance indicators in
male professional cyclists during a grand tour: worldtour versus
proteam category. Int J Sports Physiol Perform 1(aop):1–9
Nicolò A, Bazzucchi I, Sacchetti M (2017) Parameters of the 3-min-
ute all-out test: overestimation of competitive-cyclist time-trial
performance in the severe-intensity domain. Int J Sports Physiol
Perform 12(5):655–661
Nimmerichter A, Eston R, Bachl N, Williams C (2012) Effects of low
and high cadence interval training on power output in flat and
uphill cycling time-trials. Eur J Appl Physiol 112(1):69–78
Nimmerichter A etal (2020) Field-derived power-duration variables
to predict cycling time-trial performance. Int J Sports Physiol
Perform 10(1):1–8
Nixon RJ, Kranen SH, Vanhatalo A, Jones AM (2021) Steady-state
VO2 above MLSS: evidence that critical speed better represents
maximal metabolic steady state in well-trained runners. Eur J
Appl Physiol. https:// doi. org/ 10. 1007/ s00421- 021- 04780-8
Ouvrard T etal (2018) Mechanisms of performance improvements due
to a leading teammate during uphill cycling. Int J Sports Med
13(9):1215–1222
Padilla S, Mujika I, Cuesta G, Goiriena JJ (1999) Level ground and
uphill cycling ability in professional road cycling. Med Sci
Sports Exerc 31(6):878–885
Padilla S, Mujika I, Orbañanos J, Angulo F (2000) Exercise intensity
during competition time trials in professional road cycling. Med
Sci Sports Exerc 32(4):850–860
Padilla S etal (2008) Exercise intensity and load during uphill cycling
in professional 3-week races. Eur J Appl Physiol 102(4):431–438
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
316 European Journal of Applied Physiology (2022) 122:301–316
1 3
Parker Simpson L, Kordi M (2017) Comparison of critical power and
W derived from 2 or 3 maximal tests. Int J Sports Physiol Per-
form 12(6):825–830
Passfield L, Dietz KC, Hopker JG, Jobson SA (2013) Objective
time-binning in exposure variation analysis. J Manag Math
24(3):269–282
Passfield L etal (2017) Knowledge is power: issues of measuring train-
ing and performance in cycling. J Sports Sci 35(14):1426–1434
Paton CD, Hopkins WG (2001) Tests of cycling performance. Sports
Med 31(7):489–496
Peronnet F, Thibault G (1989) Mathematical analysis of running perfor-
mance and world running records. J Appl Physiol 67(1):453–465
Pethick J, Winter SL, Burnley M (2020) Physiological evidence that
the critical torque is a phase transition, not a threshold. Med Sci
Sports Exerc 52(11):2390–2401
Pinot J, Grappe F (2011) The record power profile to assess perfor-
mance in elite cyclists. Int J Sports Med 32(11):839–844
Poole DC, Ward SA, Gardner GW, Whipp BJ (1988) Metabolic and
respiratory profile of the upper limit for prolonged exercise in
man. Ergon 31(9):1265–1279
Poole DC etal (2016) Critical power: an important fatigue threshold
in exercise physiology. Med Sci Sports Exerc 48(11):2320–2334
Poole DC, Rossiter HB, Brooks GA, Bruce Gladden L (2021) The
anaerobic threshold: 50+ years of controversy. J Physiol
599(3):737–767
Pringle JS etal (2011) Critical power and aerodynamic drag accurately
predict road time-trial performance in British champion cyclists.
Med Sci Sports Exerc 43(5):160–161
Puchowicz MJ (2018) Functional PCA of the Golden Cheetah Power
Duration Data. http:// veloc linic. com/ funct ional- pca- of- the-
golden- cheet ah- power- durat ion- data/
Puchowicz MJ, Baker J, Clarke DC (2020) Development and field vali-
dation of an omni-domain power-duration model. J Sports Sci
38(7):801–813
Quod MJ, Martin DT, Martin JC, Laursen PB (2010) The power profile
predicts road cycling MMP. Int J Sports Med 31(6):397–401
Racinais S, Périard JD, Karlsen A, Nybo L (2015) Effect of heat and
heat acclimatization on cycling time trial performance and pac-
ing. Med Sci Sports Exerc 47(3):601
Sanders D, van Erp T (2021) The physical demands and power pro-
file of professional men’s cycling races: an updated review. Int J
Sports Physiol Perform 16(1):3–12
Sanders D, Heijboer M (2019a) Physical demands and power profile
of different stage types within a cycling grand tour. Eur J Sports
Sci 19(6):736–744
Sanders D, Heijboer M (2019b) The anaerobic power reserve and
its applicability in professional road cycling. J Sports Sci
37(6):621–629
Sanders D etal (2017) Predicting high-power performance in profes-
sional cyclists. Int J Sports Physiol Perform 12(3):410–413
Sanders D, Taylor RJ, Myers T, Akubat I (2020) A field-based cycling
test to assess predictors of endurance performance and establish-
ing training zones. J Strength Cond 34(12):3482–3488
Sargeant AJ, Hoinville E, Young A (1981) Maximum leg force and
power output during short-term dynamic exercise. J Appl Physiol
51(5):1175–1182
Sreedhara VSM, Mocko GM, Hutchison RE (2019) A survey of math-
ematical models of human performance using power and energy.
Sports Med Open 54(1):1–13
Townsend NE etal (2017) Prediction of critical power and W’ in
hypoxia: application to work-balance modelling. Front Physiol
23(8):180
Valenzuela PL etal (2018) Is the functional threshold power a valid
surrogate of the lactate threshold? Int J Sports Physiol Perform
13:1–6
van Druenen T, Blocken B (2021) Aerodynamic analysis of uphill
drafting in cycling. Sports Eng 24(1):1–11
van Erp T, de Koning JJ (2019) Intensity and load characteristics of
professional road cycling: differences between men’s and wom-
en’s races. Int J Sports Physiol Perform 14(3):296–302
van Erp T, Sanders D (2020) Demands of professional cycling races:
influence of race category and result. Eur J Sports Sci 16(1):1–12
van Erp T, Hoozemans M, Foster C, De Koning JJ (2020a) Case report:
load, intensity, and performance characteristics in multiple grand
tours. Med Sci Sports Exerc 52(4):868–875
van Erp T, Sanders D, De Koning JJ (2020b) Training characteristics of
male and female professional road cyclists: a 4-year retrospective
analysis. Int J Sports Physiol Perform 5(1):1–7
van Erp T, Kittel M, Lamberts RP (2021a) Demands of the tour de
France: a case study of a world-class sprinter (Part I). Int J Sports
Physiol Perform 9:1363–1370 (in press)
van Erp T, Sanders D, Lamberts RP (2021b) Maintaining power output
with accumulating levels of work done is a key determinant for
success in professional cycling. Med Sci Sports Exerc 53:1903–
1910 (in press)
Vanhatalo A, Doust JH, Burnley M (2007) Determination of critical
power using a 3-min all-out cycling test. Med Sci Sports Exerc
39(3):548–555
Vanhatalo A, Doust JH, Burnley M (2008) A 3-min all-out cycling test
is sensitive to a change in critical power. Med Sci Sports Exerc
40(9):1693–1699
Vanhatalo A etal (2016) The mechanistic bases of the power-time
relationship: muscle metabolic responses and relationships to
muscle fibre type. J Physiol 594(15):4407–4423
Vinetti G etal (2017) Effects of recovery interval duration on the
parameters of the critical power model for incremental exercise.
Eur J Appl Physiol 117(9):1859–1867
Vinetti G, Taboni A, Ferretti G (2020) A regression method for the
power-duration relationship when both variables are subject to
error. Eur J Appl Physiol 120(4):765–770
Vogt S, Schumacher YO, Blum A etal (2007a) Cycling power output
produced during flat and mountain stages in the Giro d’Italia: a
case study. J Sports Sci 25(12):1299–1305
Vogt S, Schumacher YO, Roecker K etal (2007b) Power output during
the tour de France. Int J Sports Med 28(9):756–761
Weyand PG, Lin JE, Bundle MW (2006) Sprint Performance-duration
relationships are set by the fractional duration of external force
application. Am J Physiol 290(3):58–65
Whipp B etal (1982) A constant which determines the duration
of tolerance to high intensity work. Fed N Am Soc Exp Biol
41(5):1591
Whipp BJ (1996) Domains of Aerobic function and their limiting
parameters. Springer, Boston
Winter EM etal (2016) Misuse of ‘power’ and other mechanical terms
in sport and exercise science research. J Strength Cond Res
30(1):292–300
Wooles AL, Robinson AJ, Keen PS (2005) A static method for obtain-
ing a calibration factor for SRM bicycle power cranks. Sports
Eng 8(3):137–144
Wright J, Bruce-Low S, Jobson SA (2017) The reliability and validity
of the 3-min all-out cycling critical power test. Int J Sports Med
38(6):462–467
Zadow EK etal (2016) Validity of power settings of the wahoo KICKR
power trainer. Int J Sports Perform Anal 11(8):1115–1117
Publisher's Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
... The logical contradictions observed in our simulations occur specifically in cases falling outside the model assumptions for an interval duration of fewer than 3 min. While these limitations are well acknowledged for the continuous hyperbolic model (Leo et al. 2021), we provide mathematical evidence that they propagate to the Skiba intermittent exercise model and affect the prescription and analysis of intermittent efforts in the severe and extreme intensity domain. ...
... The W' reconstitution requires further work and individual adjustmentsChorley and Lamb 2020;Chorley et al. 2019). At the foundation of the Skiba HIIT model, the continuous exercise hyperbolic model relies on a specific set of assumptions(Chorley and Lamb 2020) restricting its validity to limited intensity and duration windows(Leo et al. 2021). We suspect that the limitations of the hyperbolic model propagate to the Skiba HIIT model and confine its validity to certain intensity windows. ...
... The curves of the graphs in Figures 2 and 3 (especially for Figures 2A,C and 3A,C) are not necessarilysmooth, which will be discussed in the next section. However, looking at the general trends of each curve, we observe that impossible sessions are characterized by a small number of short intervals at high target intensity, in what is typically referred to as the severe and extreme intensity domains(Leo et al. 2021), with long recovery periods in between the efforts. For the same relative power output, interval duration, total time spent at the target intensity or rest duration, the profile corresponding with the highest level of endurance (Time-Trialist) is subject to a higher percentage of impossible sessions than the athlete with a high anaerobic capacity(Sprinter). ...
Thesis
Full-text available
In many physical activities, exercise is not continuous, but intermittent: it involves a sequence of exercise fractions at varying intensities, some higher than others. In planned training, this type of exercise is found in the form of high intensity interval training (HIIT), which is an effective and time-efficient approach that has been popular in high-performance sports over the last century, and in clinical settings for the past twenty years. Models are available to predict performance during continuous exercise (without intensity variation) over various durations. The ecological validity of some continuous exercise models has been reported. However, this is not the case for intermittent exercise, which has several parameters that can be modified, leading to a large variation in individual responses. The purpose of this master’s thesis is to compare the major models of intermittent exercise and determine their strengths and weaknesses, the constructs on which they are based, and their applicability to various physical activities. The master’s thesis also reviews the evolution of continuous exercise models to better understand the elements that need to be considered to improve the validity of intermittent exercise modelling. Due to the lack of quality data to compare a set of HIIT sessions of the same degree of difficulty, the thesis presents a study that uses simulations to identify the main limitations of the intermittent exercise models included in commercial applications, i.e., the Coggan and Skiba models. The study reveals the limitations of these models in prescribing sessions with a low number of repetitions performed at supramaximal intensity, interspersed with long recovery periods. The main intermittent exercise models have limitations that restrict their widespread use. In order for intermittent exercise modelling to evolve into more valid models that improve understanding of the physiological phenomena involved, it is crucial that the models be tested against a robust set of comparable intermittent exercise data. The thesis draws a detailed portrait of the continuous and intermittent exercise models, accounts for their evolution over time, and provides elements to guide future exercise modelling. Finally, the thesis identifies the limits of the current intermittent exercise models, makes recommendations to sports practitioners to promote their good use, and proposes a modification to the Coggan model that reduces its limitations. Key words: modelling, critical power, anaerobic reserve, training load Dans plusieurs activités physiques, l’exercice n’est pas continu, mais intermittent : il comprend un enchaînement de fractions d’exercice à des intensités variées, certaines plus élevées que d’autres. Dans l’entraînement planifié, on retrouve ce type d’exercice sous la forme de l’entraînement par intervalles (EPI), qui est une approche efficace et économe en temps, très populaire dans les milieux sportifs depuis plus d’un siècle, et dans les milieux cliniques depuis plus d’une vingtaine d’années. Des modèles sont disponibles permettant de prédire les performances lors de l’exercice continu (sans variation d’intensité) sur des durées variées. La validité écologique de certains modèles de l’exercice continu a été rapportée, montrant leur capacité à s’appliquer aux situations observées sur le terrain. Ce n’est toutefois pas le cas pour l’exercice intermittent, qui comporte plusieurs paramètres pouvant être modifiés, et menant à une grande variation des réponses individuelles. L’objectif du mémoire est de comparer les principaux modèles de l’exercice intermittent et déterminer leurs forces et leurs faiblesses, les construits sur lesquels ils sont fondés, et leur applicabilité dans diverses activités physiques. Il s’agit aussi de revoir l’évolution des modèles de l’exercice continu pour mieux comprendre les éléments à considérer pour améliorer la validité de la modélisation de l’exercice intermittent. Face au manque de données de qualité permettant de comparer un ensemble de séances d’EPI de même degré de difficulté, le mémoire présente une étude qui procède par simulations pour identifier les principales limites des modèles de l’exercice intermittent inclus dans des applications commerciales, soit les modèles de Coggan et de Skiba. L’étude révèle les limites de ces modèles quant à la prescription de séances comprenant un faible nombre de répétitions effectuées à intensité supramaximale, entrecoupées de longues périodes de récupération. Les principaux modèles de l’exercice intermittent présentent des limites restreignant leur utilisation généralisée. Pour que la modélisation de l’exercice intermittent évolue vers des modèles plus valides, permettant d’améliorer la compréhension des phénomènes physiologiques en jeu, il est crucial de confronter les modèles à un ensemble robuste de données comparables de l’exercice intermittent. Le mémoire dresse un portrait détaillé des modèles de l’exercice continu et intermittent, fait état de leur évolution au fil du temps, et propose des éléments pour guider la suite des travaux de modélisation. Enfin, le mémoire identifie les limites des modèles de courants de l’exercice intermittent, présente des recommandations aux intervenants sportifs pour favoriser la bonne utilisation de ceux-ci, en plus de fournir une modification du modèle de Coggan qui diminue les limites de celui-ci. Mots-clés : modélisation, puissance critique, réserve anaérobie, charge d’entraînement
... For the purpose of this study, the highest mean PO from a 5-second sprint interval represents maximum sprint power (MSP), whereas the average PO achieved during 2x20 minutes best average time trial efforts represents FTP. The methods applied to define MSP and FTP were based on power profiling with the mean maximum PO approach 41 and popular training sets, which is a feasible and practical solution commonly applied in triathlon and cycling. The participants were able to monitor the PO in real-time with a bike or multisport computers placed on the bicycle's bar. ...
... While guidelines for cardiopulmonary exercise testing and measuring traditional parameters associated with performance exist 42,43 , their applicability in high-performance environments may be limited. There is currently no universally agreed-upon set of best practices for performance testing or power profiling in elite athletes 41,44 , and the measurement protocols Interestingly, higher BMI was associated with higher FTP obtained indoors compared to outdoors, both absolutely and relative to body mass. Body composition and BMI are well known to influence athletic performance 48,49 . ...
Article
BACKGROUND: Cyclists and triathletes test, train, and race both indoors and outdoors. However, the differences between indoor and outdoor performance remain understudied. This study aimed to analyze the relationship between indoor and outdoor performance in cycling. We investigated the influence of training environment history (indoor vs. outdoor) and application of e-sport platform on the aforementioned relationship. METHODS: Forty-three well-trained triathletes performed indoor and outdoor field tests to establish maximum sprint power (MSP) and functional threshold power (FTP). The main effects for FTP and MSP were assessed by repeated-measures ANOVA to analyze the differences between power output obtained indoors and outdoors. Multiple covariates, including training environment history and application of ZWIFT e-sport platform, were applied. Multiple linear regression was performed to investigate outdoor FTP prediction based on indoor testing. RESULTS: No significant differences were found for MSP and FTP obtained indoors and outdoors. However, the high individual variability in the disparity between indoor and outdoor power output (PO) in triathletes was observed. There was a statistically significant interaction between FTP and training environment history (P<0.001 for FTP expressed in W*kg-1 and W) and FTP and BMI (P=0.042 and P=0.034 for FTP expressed in W*kg-1 and W, respectively). The prediction formulas to establish outdoor FTP based on indoor testing had high accuracy (R2 0.80 and 0.68 for full and simple model, respectively). CONCLUSIONS: The study underlines the crucial role of the appropriate testing environment, corresponding to the training environment and racing demands. A high individual variability in the disparity between indoor and outdoor PO at FTP is associated with training environment history and BMI. No such interactions were found for MSP.
... Without technological assistance, this cumbersome task leaves little room for active coaching and bolstering the coach-athlete relationship. 3 Compared to other sports such as cycling, 4 swimming has been trailing in the direct measurement of performancedetermining factors in the daily training practice due to the complexity of measuring in an aquatic environment and the absence of a connection with a mechanical device from which relevant performance data can be readily obtained, such as in cycling. However, recent sensor developments have provided the opportunity to collect a large amount of data simultaneously. ...
... 6,62 The light gray range is defined as ± 0.6 s for LT, ± 1 stroke for SC, ± 1 str·min −1 and ± 3% of HR of the white zone. The white and light gray zone of the IE can be calculated based on the other parameters following equation (4). Any variable outside of the white or light gray zone is then in the dark gray zone. ...
Article
Full-text available
Swim coaches routinely monitor the performance and performance determining variables of their athletes to optimize training programs in an individual-specific manner with the ultimate aim to swim faster and win races. To interpret the collected data, they require a suitable, and practically useful, conceptual framework, which can be found in the power balance of swimming. For coaches a heuristic model for training and performance optimization based on the power balance has been proposed. We build on this model and illustrate how it can be employed in the training practice using data of an exemplary sub-elite swimmer (700 FINA points), which was collected continuously during two training sessions. Variables that are used in daily swim training practice, such as heart rate (HR), stroke count (SC), stroke rate (SR), and lap time (LT), were measured. By combining external load variables (e.g. LT) and internal load variables (e.g. HR) with technical variables (e.g. SR), the degree to which the swimmer complies with the training program can be determined. The measured values of these variables are presented using a traffic light feedback system indicating the degree of compliance. The traffic light system enables coaches to adjust the program if deemed necessary. It is thus shown how the model and commonly measured variables can assist swim coaches in the design and evaluation of training sessions in their pursuit of personal performance improvement leading to greater athletic success.
... Several authors have reported improvements in various markers of performance. After the addition of conventional strength training, increases in type IIA fiber proportions and decreases in type IIX; improvements in pedaling efficacy and cycling economy after fatigue; and increases in peak power output (PO), time trial performance, and PO at fixed blood lactate concentrations are some of the highlighted changes [6][7][8]. However, a previous study has also reported no effect of strength training when added to cycling training [9]. ...
... Some of these possible benefits could be of interest to road cyclists given that most of them translate into a better exercise economy [1,25]. Given that an improvement in exercise economy may not necessarily translate into an increase in mean maximal POs, the current study could have overlooked this hypothetical benefit, which should be explored in the future [7]. Despite this possibility, most cyclists have limited time to train and need to limit cycling time in order to incorporate some kind of strength training into their routine. ...
Article
Full-text available
Conventional strength training and core exercises are commonly prescribed to improve cycling performance. Although previous studies have explored the utility of strength training in various cycling populations, this intervention has never been compared to core exercises. Thirty-six trained road cyclists were divided into three groups of 12 participants that performed either no strength training, conventional strength training, or core exercises, in all cases together with their regular cycling training during a 12-week period. Peak power outputs (POs) across different durations (five seconds, 60 seconds, five minutes, and 20 minutes) were recorded before and after the intervention. The results of the present study showed higher increases in relative PO with conventional strength training when compared to core training and no strength training for all measured durations: five-second Δ = 1.25 W/kg vs 0.47 W/kg and -0.17 W/kg; 60-second (Δ = 0.51 W/kg vs0.13 W/kg and 0.02 W/kg; five-minute Δ = 0.22 W/kg vs 0.06 W/kg and 0.05 W/kg; and 20-minute Δ = 0.22W/kg vs 0.07 W/kg and 0.06 W/kg. According to the data obtained in this study, conventional strength training is superior to core exercises, and no strength training was performed by trained road cyclists. Accordingly, it is recommended that this population incorporates strength training during their regular weekly workouts.
... This technique has been applied to open-source projects such as the GoldenCheetah Open Data project [6], which provides training information from users of the GoldenCheetah training platform. Containing anonymized data from thousands of amateur athletes using the platform, it has proven a valuable source of information to perform physiological studies on a much broader scale [7]. ...
Preprint
Full-text available
The field of cycling analytics has only recently started to develop due to limited access to open data sources. Accordingly, research and data sources are very divergent, with large differences in information used across studies. To improve this, and facilitate further research in the field, we propose the publication of a data set which links thousands of professional race results from the period 2017-2023 to detailed geographic information about the courses, an essential aspect in road cycling analytics. Initial use cases are proposed, showcasing the usefulness in linking these two data sources.
... Functional threshold power (FTP), critical power (CP), W prime (W'), and anaerobic power reserve were calculated based on the best MMP values of the coinciding cycling season. 26 Power distribution during the different stage types (TT, FLAT, SMT, and MT) was analyzed based on 5 different power zones. These zones, as defined by Allen et al, 27 were zone 1 ≤55% of FTP, zone 2 56% to 75% FTP, zone 3 76% to 90% FTP, zone 4 91% to 105% FTP, and zone 5 ≥106% FTP. ...
Article
Purpose : To describe and compare the race characteristics, demands, and durability profile of a male and a female Grand Tour winner. Methods : Overall and stage-type-specific (ie, time trials, flat, semimountainous, and mountain) demands and race characteristics during 2 Grand Tours were determined and compared between the female and male cyclists. In addition, relative power output distribution and pacing, percentage of functional threshold power (FTP), and changes in maximal mean power outputs (MMPs) with increasing levels of kilojoules burned were determined. Results : Although many differences were found between course and absolute racing demands between the male (FTP: 413 W; critical power: 417 W) and female (FTP: 297 W; critical power: 297 W) cyclists, similar power distributions and pacing strategies were found if data were expressed relatively. However, the female cyclist rode a higher percentage of her FTP during the first 2 quarters of flat stages (14.7%–15.1%) and the last quarter of mountain stages (9.8%) than the male cyclist. Decrements in MMPs were only observed after burning 30 kJ·kg ⁻¹ in the female and 45 kJ·kg ⁻¹ in the male Grand Tour winner. Conclusions : Both the male and female Grand Tour winners produced very high 20- to 60-minute MMPs, whereas decrements in MMPs were only observed after having burned 75% (female) and 80% (male) of total kilojoules burned during a stage. These are the latest and lowest in MMPs reported in the scientific literature and highlight the importance of durability in combination with excellent climbing and time-trial skills, which are needed to be able to win a Grand Tour.
... The PO-duration relationship was assessed through the aforementioned field tests, as proposed elsewhere. 11,12 All the tests were performed with participants' own bikes on the same uphill climb (distance, 5.8 km; average inclination, 5.2%). Temperature was measured with a portable meteorological station (Kestrel Meter 5500) and remained similar between testing sessions (15-17°C). ...
Article
Purpose : To assess the effect of 2 work-matched efforts of different intensities on subsequent performance in well-trained cyclists. Methods : The present study followed a randomized controlled crossover design. Twelve competitive junior cyclists volunteered to participate (age, 17 [1] y; maximum oxygen uptake, 71.0 [4.7] mL·kg ⁻¹ ·min ⁻¹ ). The power–duration relationship was assessed through 2-minute, 5-minute, and 12-minute field tests under fresh conditions (control). On subsequent days and following a randomized order, participants repeated the aforementioned tests after 2 training sessions matched for mechanical work (∼15 kJ/kg) of different intensities (ie, a moderate-intensity continuous-training [60%–70% of critical power; CP] session or a session including high-intensity intervals [3-min repetition bouts at 110%–120% of the CP interspersed by 3-min rest periods]). Results : A significantly lower power output was found in the 2-minute test after the high-intensity training session compared not only with the control condition (−8%, P < .001) but also with the moderate-intensity continuous-training session (−7%, P = .003), with no significant differences between the latter conditions. No significant differences between conditions were found for the remaining tests. As a consequence, the high-intensity training session resulted in significantly lower W′ values compared to both the control condition (−27%, P = .001) and the moderate-intensity continuous-training session (−26%, P = .012), with no differences between the 2 latter conditions and with no differences for CP. Conclusion : A session including high-intensity intermittent efforts induces a greater fatigue, particularly in short-duration efforts and W′, than a work-matched continuous-training session of moderate intensity.
... At the time of publication, no research had yet been conducted investigating the power-profile of H3/H4 handcyclists. Power profiling can be defined as the assessment of field derived power outputs and is commonly used for training prescription, performance testing, and race analysis in AB cycling(Leo et al., 2022). As such, powerprofiling during a 15-km TT may have provided interesting insights into the typical TT power profile of handcyclists.Given the novel findings of this study, it was subsequently submitted, and accepted for publication in the International Journal of Sports Physiology and Performance (impact factor: 3.652) ...
Thesis
Full-text available
Handcycling has become a popular recreational and competitive form of Para-cycling. Like their able-bodied cycling counterparts, competitive handcyclists continue to explore ways by which to gain a performance edge. Whilst our collective understanding as to the influence of handbike design and configuration, handcycling technique, and the physiological determinants of competitive handcycling has improved over the past 20-years, there still remains several gaps in our scientific knowledge as to the most effective approach by which to optimise handcycling performance capabilities. Therefore, the following PhD thesis summarises five thematically linked publications, and two related conference presentations which aimed to investigate the development and implementation of appropriate training interventions designed to enhance the performance of recumbent handcyclists across the spectrum of competitive events including road races, time-trials, and ultra-endurance challenges. Publication 1 (Chapter 3) represents a pilot research project which aimed to investigate the effectiveness of concurrent strength and endurance training on handcycling performance. Whilst demonstrating that concurrent training was more effective than endurance training alone this study generated several pertinent research questions. These included what are the physiological determinants of real-world handcycling performance? What is the relationship between upper-body strength and handcycling performance? and would a long-term concurrent training intervention elicit greater improvements in performance capabilities? To address these questions Publication 2 (Chapter 4) and Publication 3 (Chapter 5) build upon the published literature and identify the physiological determinants of handcycling performance. However, for the first time in the literature these studies also investigate the relationship between upper-body strength measures, anaerobic capacity and identified determinants of handcycling performance. Based upon these findings Publication 4 (Chapter 6) reports upon the effectiveness of a 30-week concurrent training program based upon a block periodisation model. Furthermore, this study reports the performance profile of an elite handcyclist during a 1407-km ultra-endurance handcycling challenge. Building upon this body of work, Publication 5 (Chapter 7) represents a holistic narrative review led by the author and written in conjunction with a group of international researchers in the field of handcycling. This piece aims to translate handcycling specific research and provide useful insights to riders, coaches and sports scientists as to the history of handcycling, functional classification levels, handbike configuration, the physiological determinants of handcycling performance, and the best approach by which to develop handcycling performance capabilities. In summary, the body of work presented within this PhD thesis has added to our collective knowledge in regard to understanding the physiological determinants of handcycling performance including the importance of quantifying anaerobic capacity and upper-body strength. Furthermore, from an applied perspective the work presented demonstrates that concurrent strength and endurance training based upon a block periodisation model appears to be an effective approach by which to develop both TT and ultra-endurance handcycling performance. Taken collectively this knowledge adds to the existing body of literature and will positively impact upon the ability of riders, coaches, and sport scientists to optimise recumbent handcycling performance capabilities. Future studies should aim to use classifiable handcyclists with the intent of translating their findings to the wider handcycling community with the goal of not only enhancing handcycling performance but also improving the functional capabilities of a valued but often under represented section of society.
Article
Full-text available
The metabolic boundary separating the heavy-intensity and severe-intensity exercise domains is of scientific and practical interest but there is controversy concerning whether the maximal lactate steady state (MLSS) or critical power (synonymous with critical speed, CS) better represents this boundary. We measured the running speeds at MLSS and CS and investigated their ability to discriminate speeds at which V˙O2\dot{V}{\text{O}}_{2} V ˙ O 2 was stable over time from speeds at which a steady-state V˙O2\dot{V}{\text{O}}_{2} V ˙ O 2 could not be established. Ten well-trained male distance runners completed 9–12 constant-speed treadmill tests, including 3–5 runs of up to 30-min duration for the assessment of MLSS and at least 4 runs performed to the limit of tolerance for assessment of CS. The running speeds at CS and MLSS were significantly different (16.4 ± 1.3 vs. 15.2 ± 0.9 km/h, respectively; P < 0.001). Blood lactate concentration was higher and increased with time at a speed 0.5 km/h higher than MLSS compared to MLSS ( P < 0.01); however, pulmonary V˙O2\dot{V}{\text{O}}_{2} V ˙ O 2 did not change significantly between 10 and 30 min at either MLSS or MLSS + 0.5 km/h. In contrast, V˙O2\dot{V}{\text{O}}_{2} V ˙ O 2 increased significantly over time and reached V˙O2max\dot{V}{\text{O}}_{2\,\,\max } V ˙ O 2 max at end-exercise at a speed ~ 0.4 km/h above CS ( P < 0.05) but remained stable at a speed ~ 0.5 km/h below CS. The stability of V˙O2\dot{V}{\text{O}}_{2} V ˙ O 2 at a speed exceeding MLSS suggests that MLSS underestimates the maximal metabolic steady state. These results indicate that CS more closely represents the maximal metabolic steady state when the latter is appropriately defined according to the ability to stabilise pulmonary V˙O2\dot{V}{\text{O}}_{2} V ˙ O 2 .
Article
Full-text available
Functional Threshold Power (FTP) in cycling is increasingly used in exercise prescription, particularly with the rise in use of home trainers and virtual exercise platforms. FTP testing does not require biological sampling and is considered a more practical test than others. This scoping review investigated what is known about the 20-minute FTP (FTP²⁰) test. A three-step search strategy was used to identify studies in relevant databases (PubMed, CINAHL, SportDiscus, Google Scholar, Web of Science) and grey literature. Data were extracted and common themes identified which allowed for descriptive analysis and thematic summary. Fifteen studies were included. The primary focus fitted broadly into four themes: reliability, association with other physiological markers, other power-related concepts and performance prediction. The FTP²⁰ test was reported as a reliable test. Studies investigating the relationship of FTP²⁰ with other physiological markers and power-related concepts reported large limits of agreement suggesting parameters cannot be used interchangeably. Some findings indicate that FTP²⁰ may be useful in performance prediction. The majority of studies involved trained male cyclists. Overall, existing literature on the FTP²⁰ test is limited. Further investigation is needed to provide physiological justification for FTP²⁰ and inform use in exercise prescription in a range of populations.
Article
Full-text available
Maximal muscular power production is of fundamental importance to human functional capacity and feats of performance. Here, we present a synthesis of literature pertaining to physiological systems that limit maximal muscular power during cyclic actions characteristic of locomotor behaviours, and how they adapt to training. Maximal, cyclic muscular power is known to be the main determinant of sprint cycling performance, and therefore we present this synthesis in the context of sprint cycling. Cyclical power is interactively constrained by force-velocity properties (i.e. maximum force and maximum shortening velocity), activation-relaxation kinetics and muscle coordination across the continuum of cycle frequencies, with the relative influence of each factor being frequency dependent. Muscle cross-sectional area and fibre composition appear to be the most prominent properties influencing maximal muscular power and the power-frequency relationship. Due to the role of muscle fibre composition in determining maximum shortening velocity and activation-relaxation kinetics, it remains unclear how improvable these properties are with training. Increases in maximal muscular power may therefore arise primarily from improvements in maximum force production and neuromuscular coordination via appropriate training. Because maximal efforts may need to be sustained for~15-60 s within sprint cycling competition, the ability to attenuate fatigue-related power loss is also critical to performance. Within this context, the fatigued state is characterised by impairments in force-velocity properties and activation-relaxation kinetics. A suppression and leftward shift of the power-frequency relationship is subsequently observed. It is not clear if rates of power loss can be improved with training, even in the presence adaptations associated with fatigue-resistance. Increasing maximum power may be most efficacious for improving sustained power during brief maximal efforts, although the inclusion of sprint interval training likely remains beneficial. Therefore, evidence from sprint cycling indicates that brief maximal muscular power production under cyclical conditions can be readily improved via appropriate training, with direct implications for sprint cycling as well as other athletic and health-related pursuits. Maximal muscle power production under cyclical conditions is interactively constrained by force-velocity properties, activation-relaxation kinetics and muscle coordination across the continuum of possible movement frequencies. Fatigue alters the power-frequency relationship, with a higher degree of power loss at higher movement frequencies. Maximal muscular power production can be readily increased with appropriate strength and power training; it remains less clear if rates of power loss during brief maximal sustained efforts can be improved with training.
Article
Full-text available
Some teams aiming for victory in a mountain stage in cycling take control in the uphill sections of the stage. While drafting, the team imposes a high speed at the front of the peloton defending their team leader from opponent’s attacks. Drafting is a well-known strategy on flat or descending sections and has been studied before in this context. However, there are no systematic and extensive studies in the scientific literature on the aerodynamic effect of uphill drafting. Some studies even suggested that for gradients above 7.2% the speeds drop to 17 km/h and the air resistance can be neglected. In this paper, uphill drafting is analyzed and quantified by means of drag reductions and power reductions obtained by computational fluid dynamics simulations validated with wind tunnel measurements. It is shown that even for gradients above 7.2%, drafting can yield substantial benefits. Drafting allows cyclists to save over 7% of power on a slope of 7.5% at a speed of 6 m/s. At a speed of 8 m/s, this reduction can exceed 16%. Sensitivity analyses indicate that significant power savings can be achieved, also with varying bicycle, cyclist, road and environmental characteristics.
Article
Full-text available
Cycling performance models are used to study rider and sport characteristics to better understand performance determinants and optimise competition outcomes. Performance requirements cover the demands of competition a cyclist may encounter, whilst rider attributes are physical, technical and psychological characteristics contributing to performance. Several current models of endurance-cycling enhance understanding of performance in road cycling and track endurance, relying on a supply and demand perspective. However, they have yet to be developed for sprint-cycling, with current athlete preparation, instead relying on measures of peak-power, speed and strength to assess performance and guide training. Peak-power models do not adequately explain the demands of actual competition in events over 15-60 s, let alone, in World-Championship sprint cycling events comprising several rounds to medal finals. Whilst there are no descriptive studies of track-sprint cycling events, we present data from physiological interventions using track cycling and repeated sprint exercise research in multiple sports, to elucidate the demands of performance requiring several maximal sprints over a competition. This review will show physiological and power meter data, illustrating the role of all energy pathways in sprint performance. This understanding highlights the need to focus on the capacity required for a given race and over an event, and therefore the recovery needed for each subsequent race, within and between races, and how optimal pacing can be used to enhance performance. We propose a shift in sprint-cyclist preparation away from training just for peak power, to a more comprehensive model of the actual event demands.
Article
Purpose: To compare the physical demands and performance indicators of male professional cyclists of 2 different categories (Union Cycliste Internationale WorldTour [WT] and ProTeam [PT]) during a cycling grand tour. Methods: A WT team (n = 8, 31.4 [5.4] y) and a PT team (n = 7, 26.9 [3.3] y) that completed "La Vuelta 2020" volunteered to participate. Participants' power output (PO) was registered, and measures of physical demand and physiological performance (kilojoules spent, training stress score, time spent at different PO bands/zones, and mean maximal PO [MMP] for different exertion durations) were computed. Results: WT achieved a higher final individual position than PT (31 [interquartile range = 33] vs 71 [59], P = .004). WT cyclists showed higher mean PO and kilojoule values than their PT peers and spent more time at high-intensity PO values (>5.25 W·kg-1) and zones (91%-120% of individualized functional threshold power) (Ps < .05). Although no differences were found for MMP values in the overall analysis (P > .05), subanalyses revealed that the between-groups gap increased through the race, with WT cyclists reaching higher MMP values for ≥5-minute efforts in the second and third weeks (Ps < .05). Conclusions: Despite the multifactorial nature of cycling performance, WT cyclists spend more time at high intensities and show higher kilojoules and mean PO than their PT referents during a grand tour. Although the highest MMP values attained during the whole race might not differentiate between WT and PT cyclists, the former achieve higher MMP values as the race progresses.
Article
The aim of this study was to analyze climbing performance across two editions of a professional multistage race, and assess the influence of climb category, prior workload, and intensity measures on climbing performance in U23 and professional cyclists. Nine U23 cyclists (age 20.8 ± 0.9 years) and 8 professional cyclists (28.1 ± 3.2 years) participated in this study. Data were divided into four types: overall race performance, climb category, climbing performance metrics (power output, ascent velocity, speed), and workload and intensity measures. Differences in performance metrics and workload and intensity measures between groups were investigated. Power output , ascent velocity, speed were higher in professionals than U23 cyclists for Cat 1 and Cat 2 (p ≤ 0.001-0.016). Workload and intensity measures (Work total , Work total •km-1 , Elevation gain , eTRIMP and eTRIMP•km-1) were higher in U23 compared to professionals (p = 0.002-0.014). Climbing performance metrics were significantly predicted by prior workload and intensity measures for Cat 1 and 2 (R 2 = 0.27-0.89, p ≤ 0.001-0.030) but not Cat 3. These findings reveal that climbing performance in professional road cycling is influenced by climb categorization as well as prior workload and intensity measures. Combined, these findings suggest that Cat 1 and 2 climbing performance could be predicted from workload and intensity measures.
Article
Purpose: The aim of this study was to compare the power profile, internal and external workloads, and racing performance between U23 and professional cyclists and between varying rider types across 2 editions of a professional multistage race. Methods: Nine U23 cyclists from a Union Cycliste Internationale "Continental Team" (age 20.8 [0.9] y; body mass 71.2 [6.3] kg) and 8 professional cyclists (28.1 [3.2] y; 63.0 [4.6] kg) participated in this study. Rider types were defined as all-rounders, general classification (GC) riders, and domestiques. Data were collected during 2 editions of a 5-day professional multistage race and split into the following 4 categories: power profile, external and internal workloads, and race performance. Results: The professional group, including domestiques and GC riders, recorded higher relative power profile values after certain amounts of total work (1000-3000 kJ) than the U23 group or all-rounders (P ≤ .001-.049). No significant differences were found for external workload measures between U23 and professional cyclists, nor among rider types. Internal workloads were higher in U23 cyclists and all-rounders (P ≤ .001-.043) compared with professionals, domestiques, and GC riders, respectively. The power profile significantly predicted percentage general classification and Union Cycliste Internationale points (R2 = .90-.99), whereas external and internal workloads did not. Conclusion: These findings reveal that the power profile represents a practical tool to discriminate between professionals and U23 cyclists as well as rider types. The power profile after 1000 to 3000 kJ of total work could be used by practitioners to evaluate the readiness of U23 cyclists to move into the professional ranks, as well as differentiate between rider types.