ArticlePDF Available

Relationship of critical velocity to marathon running performance

Authors:

Abstract and Figures

The purpose of this investigation was to evaluate the critical velocity (CV) test for prediction of marathon running performance. Twelve subjects [mean age (SD) = 29 (4) years; mean body mass = 63 (13) kg] were tested for CV and completed the 1994 New York City Marathon. The CV (m · s−1) was determined from times to exhaustion at four treadmill running velocities. In addition, peak oxygen consumption (\(\) O 2 peak; ml · kg−1 · min−1) and ventilatory threshold (Thvent) were determined from an incremental treadmill test. The Thvent was calculated using bi-segmental linear regression and was expressed as the velocity (m · s−1) at Thvent. Separate simple linear regression analyses showed that marathon time [MT; mean (SD) = 231.9 (27.4) min] correlated more highly with CV [MT = 445.3 – 50.3 (CV); r 2 = 0.76, SEE = 14.1 min] than either \(\)O2peak [MT = 390.7 – 2.7 (\(\)O2peak); r 2 = 0.51, SEE = 20.1 min] or Thvent [MT = 353.5 – 30.1 (Thvent) r 2 = 0.28, SEE = 27.4 min]. A stepwise regression analysis resulted in CV (entered first) and Thvent being included in the prediction equation [MT = 443.5 – 78.9 (CV) + 34.3 (Thvent), R 2 = 0.88, SEE = 10.7 min], while \(\)O2peak was not included. These preliminary data indicate that the CV test may be an attractive field test for assessing marathon performance capabilities.
Content may be subject to copyright.
ORIGINAL ARTICLE
Shelly-lynn Florence · Joseph P. Weir
Relationship of critical velocity to marathon running performance
Accepted: 5 September 1996
Abstract The purpose of this investigation was to eval-
uate the critical velocity (CV) test for prediction of
marathon running performance. Twelve subjects [mean
age (SD) = 29 (4) years; mean body mass = 63 (13) kg]
were tested for CV and completed the 1994 New York
City Marathon. The CV (m · s
–1
) was determined from
times to exhaustion at four treadmill running velocities.
In addition, peak oxygen consumption (V
˙
O
2peak
;
ml · kg
–1
· min
–1
) and ventilatory threshold (Th
vent
) were
determined from an incremental treadmill test. The
Th
vent
was calculated using bi-segmental linear regres-
sion and was expressed as the velocity (m · s
–1
)atTh
vent
.
Separate simple linear regression analyses showed that
marathon time [MT; mean (SD) = 231.9 (27.4) min]
correlated more highly with CV [MT = 445.3 50.3
(CV); r
2
= 0.76, SEE = 14.1 min] than either V
˙
O
2peak
[MT = 390.7 2.7 (V
˙
O
2peak
); r
2
= 0.51, SEE = 20.1 min]
or Th
vent
[MT = 353.5 30.1 (Th
vent
) r
2
= 0.28, SEE =
27.4 min]. A stepwise regression analysis resulted in CV
(entered first) and Th
vent
being included in the prediction
equation [MT = 443.5 78.9 (CV) + 34.3 (Th
vent
), R
2
=
0.88, SEE = 10.7 min], while V
˙
O
2peak
was not included.
These preliminary data indicate that the CV test may be
an attractive field test for assessing marathon perfor-
mance capabilities.
Key words Marathon · Ventilatory threshold · Critical
velocity
Introduction
A variety of testing procedures have been employed to
assess endurance exercise capability, including maximal
oxygen consumption (V
˙
O
2max
), ventilatory threshold
(Th
vent
), and various lactate measurement protocols.
While moderate to good correlations have been reported
among these procedures and athletic events such as the
marathon (Fohrenback et al. 1987; Rhodes and
McKenzie 1984; Sjodin and Jacobs 1981; Zoladz et al.
1993), they are not widely accessible for the majority of
athletes. A test that could predict endurance capabilities
with comparable accuracy to that of metabolic testing
would be a valuable aid in optimizing training of en-
durance athletes.
The critical power test, originally developed using
synergic muscle groups (Monod and Scherrer 1965) and
later modified for use with cycle ergometry (Moritani
et al. 1981), allows for the simultaneous determination
of aerobic and anaerobic capabilities (Hill 1993). In this
procedure, three to five work bouts at a power loading
of sufficient intensity to elicit exhaustion within 1–10 min
(Housh et al. 1990) are performed. For each of the work
bouts, the total work performed (work limit; WL) is
regressed against the corresponding time to exhaustion
(time limit; TL). The WL-TL relationship for a given
subject has been reported to be highly linear (Moritani
et al. 1981; Nebelsick-Gullet et al. 1988) and, as such,
can be described by the equation: WL = a + b (TL),
where a and b are the y-intercept and slope, respectively.
The y-intercept has been termed anaerobic work capa-
city (Monod and Scherrer 1965; Moritani et al. 1981)
and theoretically represents stored intramuscular energy
reserves. Similarly, the slope has been termed critical
power and is theoretically representative of the power
loading that corresponds to the maximal work load that
can be maintained indefinitely (Monod and Scherrer
1965; Moritani et al. 1981).
Critical velocity (CV), a treadmill analog of critical
power (Housh et al. 1991b; Pepper et al. 1992), provides
Eur J Appl Physiol (1997) 75: 274 278 Springer-Verlag 1997
S. Florence · J.P. Weir
Applied Physiology Laboratory,
Department of Movement Sciences and Education,
Teachers College, Columbia University,
New York, N.Y., USA.
J.P. Weir (&)
Program In Physical Therapy,
University of Osteopathic Medicine and Health Sciences,
3200 Grand Ave, Des Moines,
IA 50312, USA
a running test that may be an attractive alternative to
metabolic testing for predicting endurance performance
in events such as the marathon. Since calculation of
power output during activities such as running is pro-
blematic (Hughson et al. 1984), modification of the cri-
tical power test for treadmill exercise involves
substitution of velocity for power and distance for work
(Hughson et al. 1984). The regression of the distance run
(distance limit; DL) versus the TL at several exhaustive
running velocities on the treadmill results in the gen-
eralized equation: DL = a + b(TL), where a is considered
to be the anaerobic running capacity (ARC) and the
slope (b) is termed CV (Housh et al. 1992). In practice,
the CV test involves performing a series of fatiguing runs
on a treadmill. At each velocity, the TL is recorded and
the corresponding DL is calculated. From these data,
the regression of the DL–TL relationship is calculated,
and CV and ARC are determined. Figure 1 shows the
DL–TL relationship for a representative subject. Similar
to critical power, CV theoretically represents the velocity
at which an individual could run indefinitely (Housh
et al. 1991b; Pepper et al. 1992), and CV should be
highly correlated with prolonged endurance perfor-
mance capability such as that required in the marathon.
Therefore, the purpose of this investigation was to ex-
amine the ability of the CV test to predict marathon
running performance. In addition, the predictive ability
of the CV test was compared to that of traditional me-
tabolic indices of endurance capability such as peak
oxygen consumption (V
˙
O
2peak
) and Th
vent
.
Methods
Subjects
Six males and six females [mean age (SD) = 29 (4) years; mean mass
63 (13) kg] who were training for the 1994 New York City Mara-
thon volunteered to be subjects for this investigation. The subjects
had been involved in running for a mean of 7.8 (5.2) years (range =
6 months to 18 years) and all had completed at least one marathon
[mean = 3.4 (1.9)]. All procedures were approved by the Institu-
tional Review Board and written informed consent was obtained
from the subjects prior to all testing.
Experimental protocol
The testing started 3 weeks prior to the marathon and continued
for 2 weeks. This time frame allowed for adequate recovery from
the testing prior to the race. The subjects came to the laboratory
twice, with each visit separated by a minimum of 48 h. During the
first visit, the subjects performed an incremental treadmill test for
the determination of V
˙
O
2peak
and Th
vent
. The protocol for the test
was patterned after that used by Rhodes and McKenzie (1984).
Briefly, following a 5-min warm-up at 1.6 m · s
–1
, the velocity was
increased to 2.2 m · s
–1
for 2 min and was subsequently increased by
0.2 m · s
–1
every minute until volitional exhaustion. Heart rate was
determined from ECG recordings of leads I–III and expired gas
samples were measured using a calibrated metabolic measurement
cart (MMC Horizon, Sensor Medics, Anaheim, Calif., USA) which
provides 15-s averaged data. The V
˙
O
2peak
was considered the
highest V
˙
O
2
attained during the incremental test. The Th
vent
was
determined using a computerized two-line segment linear regression
program patterned after the procedure of Orr et al. (1982). The
plots of minute ventilation (V
˙
E
) and CO
2
output (V
˙
CO
2
) versus
time as well as V
˙
CO
2
versus V
˙
O
2
(V slope) were analyzed with the
computer program. Visual inspection of the two plots was used to
further delineate the Th
vent
from the differences between the two
computerized analyses.
During the second visit, CV was determined using a protocol
similar to that described previously (Housh et al. 1991b; Pepper
et al. 1992). Briefly, the subjects performed a series of four ran-
domly ordered treadmill runs at velocities ranging from 3.6 to 6.0
m · s
–1
. At each velocity, the subjects warmed up for 5 min at 1.6
m · s
–1
at a zero per cent incline. Following the warm-up, the sub-
jects straddled the treadmill belt while the velocity was adjusted.
Once the velocity was set, the subjects began running while still
holding the handrails. When the subjects adjusted to the treadmill
velocity and released the handrails, the timing for the run was
started. Timing for the run was stopped when the subjects again
grabbed the handrails at the point of volitional exhaustion.
Throughout each run, the subjects were verbally encouraged to run
for as long as possible. A minimum of 20 min of rest was given bet-
ween runs, which allowed the heart to fall below 100 beats · min
–1
.
The subjects were allowed to drink water ad libitum during the rest
period, but food was not consumed. For each run, DL was cal-
culated as the product of the treadmill velocity and the TL. Based
on the four runs, CV and ARC were determined from the slope and
y)intercept of the DL-TL relationship, respectively.
Data analysis
Univariate relationships between CV, V
˙
O
2peak
, and Th
vent
and
marathon time (MT) were evaluated using a simple linear regres-
sion. A stepwise multiple regression analysis was performed using
CV, V
˙
O
2peak
, and Th
vent
as predictor variables. In addition, dif-
ferences between CV, velocity at Th
vent
, and mean marathon ve-
locity (MS) were evaluated with a one-way repeated measures
analysis of variance with the Huynh-Feldt correction (Keppel
1982). An alpha level of 0.05 was considered to be significant.
Results
Table 1 provides descriptive data for the individual
subjects. All 12 subjects finished the marathon. The
mean MT was 231.9 (27.4) min (range = 192–261 min)
and the temperature and humidity at the start of the race
Fig. 1 The relationship between distance limit (DL in m) and the time
limit (TL in s) for a representative subject [DL = 129.3 + 4.8 m· s
–1
(TL); r
2
= 0.99]. The slope of this relationship (4.8 m· s
–1
) is termed the
critical velocity and the y)intercept (129.3 m) is termed the anaerobic
running capacity
275
were 22.2°C and 82%, respectively. There were occa-
sional showers and wind.
The separate simple linear regression analyses for the
three predictor variables showed that CV correlated
more highly with MT [MT = 445.3 50.3 (CV); r
2
=
0.76, SEE = 14.1 min; see Fig. 2] than either V
˙
O
2peak
[MT = 390.7 – 2.73 (V
˙
O
2peak
); r
2
= 0.51, SEE = 20.1 min;
see Fig. 3] or Th
vent
[MT = 353.5 ) 30.1 (Th
vent
);
r
2
= 0.28, SEE = 27.4 min; see Fig. 4]. The stepwise
multiple regression analysis resulted in CV (entered first)
and Th
vent
, being included in the prediction equation
[MT = 443.5 ) 78.9 (CV) + 34.3 (Th
vent
) R
2
= 0.88, SEE
= 10.7 min] while V
˙
O
2peak
was not included. The one-
way analysis was significant (F = 120.5, P < 0.0001),
and Tukey post hoc comparisons showed that CV [4.43
(0.48) m · s
–1
was significantly higher than the speed at
Th
vent
[4.04 (0.48) m · s
–1
] and that both CV and Th
vent
were significantly higher than the marathon speed [3.07
(0.35) m · s
–1
].
Discussion
While strong relationships have been reported between
various laboratory procedures and MT (Fohrenback
et al. 1987; Rhodes and McKenzie 1984; Sjodin and
Jacobs 1981; Zoladz et al. 1993), the CV test is appealing
in that measurement requires only a treadmill and a
stopwatch. Thus, this procedure may allow for testing of
a larger number of athletes than is currently possible
with more traditional laboratory procedures, especially
if the protocol can be eventually modified to allow for
testing without a treadmill (i.e., running-track testing).
The results of this investigation showed that CV corre-
lates highly with MT in this relatively small hetero-
geneous sample (r
2
= 0.76). The predictive ability of CV
was higher than that of Th
vent
(r
2
= 0.28) or V
˙
O
2peak
(r
2
=
0.51). These results are encouraging with respect to the
potential utility of the critical velocity test as a tool for
assessing endurance performance capabilities in athletes.
Theoretically, CV represents the fastest running ve-
locity that can be maintained indefinitely (Pepper et al.
1992). In practice, CV over-predicts this velocity and
subjects are typically only able to maintain a running
pace at a speed equal to CV for less than 30 min (Housh
et al. 1991b; Pepper et al. 1992). Indeed, the mean
Table 1 Individual subject characteristics. (M Male, F Female)
Subject Sex Age (years) Height (cm) Mass (kg)
1 M 34 178 73.0
2 F 26 168 53.1
3 F 34 157 59.4
4 M 34 178 73.5
5 F 33 168 51.7
6 M 24 180 66.7
7 M 25 188 70.3
8 F 31 155 42.2
9 F 25 160 47.2
10 F 28 170 58.5
11 M 29 183 84.8
12 M 27 180 75.3
Fig. 2 The relationship between marathon time (min) and critical
velocity (m · s
–1
)
Fig. 3 The relationship between marathon time (min) and peak O
2
consumption (V
˙
O
2peak
)(ml·kg
–1
· min
–1
)
Fig. 4 The relationship between marathon time (min) and ventilatory
threshold (m · s
–1
)
276
marathon velocity for the subjects in this investigation
(3.07 m · s
–1
) was significantly less than the speed at CV
(4.43 m · s
–1
). Nonetheless, the strong correlation be-
tween CV and MT suggests that the CV parameter en-
compasses information that is reflective of an
individual’s ability to perform prolonged endurance
exercise. Critical power, derived from cycle exercise, has
been shown to be correlated with V
˙
O
2max
(Moritani et al.
1981), onset of blood lactate accumulation (4-mmol le-
vel; Housh et al. 1991a), Th
vent
(Moritani et al. 1981),
and an electromyographic fatigue threshold (deVries
et al. 1982), and is sensitive to changes in aerobic fitness
following training (Gaesser and Wilson 1988; Jenkins
and Quigley 1992). Thus, critical power and, by exten-
sion, critical velocity are believed to reflect aerobic me-
tabolic capabilities (Hill 1993). It is not surprising, then,
that CV should be significantly correlated with MT,
while the over prediction of actual MT by the absolute
CV is consistent with previous research regarding the
maximal duration of exercise at CV (Housh et al. 1991b;
Pepper et al. 1992).
In contrast to previous work (Rhodes and McKenzie,
1984), we did not find a strong correlation between MT
and running speed at Th
vent
. Rhodes and McKenzie
(1984) reported a correlation of 0.94 between actual MT
and predicted MT, based on the breakpoint in the excess
CO
2
curve (‘‘anaerobic threshold’’). The correlation
between MT and Th
vent
observed in this investigation
was –0.53. The lower correlation reported in this study is
likely to be due to at least two factors. First, differences
in subject characteristics with respect to endurance
capabilities are likely to have affected the relationship
between Th
vent
and MT in two studies, especially con-
sidering that the treadmill protocol for Th
vent
determi-
nation in the two studies was essentially the same.
Specifically, the subjects in the study of Rhodes and
McKenzie (1984) were more accomplished marathon
runners [mean MT = 172 min; see Table 2 of Rhodes
and McKenzie (1984)] than the subjects in this in-
vestigation (mean MT = 231 min) and were more highly
conditioned in terms of V
˙
O
2max
and Th
vent
. These dif-
ferences in performance capability and training status
are likely to influence the ability of individuals to
maintain a pace at or near the ‘‘anaerobic threshold’’,
which Rhodes and McKenzie (1984) suggest is an opti-
mal pace for trained marathon runners. This is sup-
ported by the significant difference in mean marathon
speed (3.07 m · s
–1
) and speed at Th
vent
(4.04 m · s
–1
) for
the subjects in this study. Thus, while other studies have
reported that high-level marathon runners are able to
run at a velocity near the velocity above which changes
in blood lactate, acid–base, ventilatory, and gas ex-
change indices occur (Zoladz et al. 1993), the less ex-
perienced and conditioned subjects in this study ran at a
mean marathon velocity that was considerably below the
velocity at Th
vent
.
Second, differences in course characteristics may in-
fluence the relationship between Th
vent
and MT. The New
York City Marathon has more hills than the relatively flat
course reported by Rhodes and McKenzie (1984). This
would probably require the runners to have a more
variable race pace and energy expenditure. This increased
variability may decrease the utility of the speed at Th
vent
as a predictor of marathon running performance.
Interestingly, the multiple regression analysis resulted
in both CV and Th
vent
being included in the model while
V
˙
O
2peak
was excluded. While interpretation of inclusion/
exclusion of variables stepped into regression equations
must be done with caution (Pedhazur 1982), especially
with the sample size of this study, these data indicate
that CV and Th
vent
provide unique information for MT
preditciton, and the information associated with V
˙
O
2peak
significantly overlaps with that of CV. This is surprising
since both theoretical and empirical evidence regarding
critical power has indicated that critical power is re-
flective of the ability to supply energy without depen-
dence on anaerobic processes, which would presumably
be correlated with indices of ‘‘anaerobic threshold’’ as
opposed to maximal aerobic power.
In summary, the results of this investigation showed a
high correlation between CV and MT. These results
suggest that the CV test may be useful for determining
performance capabilities in prolonged endurance events
such as the marathon.
References
Fohrenback R, Mader A, Hollman W (1987) Determination of
endurance capacity and prediction of exercise intensities for
training and competition in marathon runners. Int J Sports
Med 8: 11–18
Gaesser GA, Wilson LA (1988) Effects of continuous and interval
training on the parameters of the power-endurance time re-
lationship for high-intensity exercise. Int J Sports Med 9: 417–
421
Hill DW (1993) The critical power concept. A review. Sports Med
16: 237–254
Housh DJ, Housh TJ, Bauge SM (1990) A methodological con-
sideration for determination of critical power and anaerobic
work capacity. Res Q Exerc Sport 61: 406–409
Housh TJ, deVries HA, Housh DJ, Tichy MW, Smyth KD, Tichy
AM (1991a) The relationship between critical power and the
onset of blood lactate accumulation. J Sports Med Phys Fitness
31: 31–36
Housh TJ, Johnson GO, McDowell SL, Housh DJ, Pepper M
(1991b) Physiological responses at the fatigue threshold. Int J
Sports Med 12: 305–308
Housh TJ, Johnson GO, McDowell SL, Housh DJ, Pepper M
(1992) The relationship between anaerobic running capacity
and peak plasma lactate. J Sports Med Phys Fitness 32: 117–
122
Hughson RL, Cook CJ, Staudt LE (1984) A high velocity treadmill
test to assess endurance running potential. Int J Sports Med 5:
23–25
Jenkins DG, Quigley BM (1992) Endurance training enhances
critical power. Med Sci Sports Exerc 24: 1283–1289
Keppel G (1982) Design and analysis. A researcher’s handbook.
Prentice Hall, Engelwood Cliffs, N.J., p 472
Monod H, Scherrer J (1965) The work capacity of a synergic
muscular group. Ergonomics 8: 329–338
Moritani T, Nagata A, deVries HA, Muro M (1981) Critical power
as a measure of physical work capacity and anaerobic thresh-
old. Ergonomics 24: 339–350
277
Nebelsick-Gullet LJ, Housh TJ, Johnson GO, Bauge SM (1988) A
comparison between methods of measuring anaerobic work
capacity. Ergonomics 31: 1413–1419
Orr GW, Green HJ, Hughson RL, Bennett GW (1982) A computer
linear regression model to determine ventilatory anaerobic
threshold. J Appl Physiol 52: 1349–1352
Pedhazur EJ (1982) Multiple regression in behavioral research.
Explanation and prediction. Holt, Rinehart, and Winston, Fort
Worth, Tex.
Pepper ML, Housh TJ, Johnson GO (1992) The accuracy of the
critical velocity test for predicting time to exhaustion during
treadmill running. Int J Sports Med 13: 121–124
Rhodes EC, McKenzie DC (1984) Predicting marathon time from
anaerobic threshold measurements. Phys Sportsmed 12: 95–98
Sjodin B, Jacobs I (1981) Onset of blood lactate accumulation and
marathon running performance. Int J Sports Med 2: 23–26
deVries H, Moritani T, Nagata A, Magnussen K (1982) The rela-
tion between critical power and neuromuscular fatigue as esti-
mated from electromyographic data. Ergonomics 25: 783–791
Zoladz JA, Sargeant AJ, Emmerich J, Stoklosa J, Zychowsk A
(1993) Changes in acid-base status of marathon runners during
an incremental field test. Eur J Appl Physiol 67: 71–76
278
... Physiological determinants of long-distance running performance include maximal oxygen uptake (V _ O 2max ), lactate threshold (or gas exchange threshold) and running economy (Midgley et al. 2007). Critical speed also represents an important parameter of aerobic function, being significantly correlated with 5-km to 42-km running events (Kolbe et al. 1995;Florence and Weir 1997;Nimmerichter et al. 2017;Jones and Vanhatalo 2017;Bergstrom et al. 2017), suggesting that CS encompasses information that is reflective of an individual's ability to perform prolonged endurance exercise. ...
... It has been previously reported that the CS is significantly correlated with 5-km , 10-km and 21-km (Kolbe et al. 1995), and 42-km running events (Florence and Weir 1997;Jones and Vanhatalo 2017;Bergstrom et al. 2017). Indeed, it has been estimated that elite marathon runners sustain 96% of their CS during competition (Jones and Vanhatalo 2017). ...
... Indeed, it has been estimated that elite marathon runners sustain 96% of their CS during competition (Jones and Vanhatalo 2017). These previous studies determined CS through conventional methods, which require the completion of 3 to 5 trials to exhaustion on a treadmill on the same or separate days (Kolbe et al. 1995;Florence and Weir 1997;Nimmerichter et al. 2017), or timed runs on a track, also on the same or separate days (Hill 1993;Jones et al. 2019). The need for several tests, added to its exhaustive nature may prevent the use of the CS concept as a powerful tool for fitness diagnostics, prescription and evaluation of training. ...
Article
It has been shown that the critical speed (CS) predicted from a perceptually self-regulated 10-min submaximal treadmill test (T10) is reliable and closely matches the CS estimated from conventional methods. To assess the relationship between the T10 and 5-km and 10-km running performances, 36 recreational runners (mean SD: age: 32.2 ± 6.2 years, height: 173.2 ± 7.3 cm, weight: 70.9 ± 8.8 kg, maximal oxygen uptake (V̇O 2max ): 53.3 ± 6.1 mL·kg ⁻¹ ·min ⁻¹ ) performed a ramp incremental test and 2 T10 tests (the first as a familiarization trial). Results showed that the T10 CS (3.9 ± 0.44 m·s ⁻¹ ) was significantly correlated with runners’ last 6 months best performances in 5 km (20.3 ± 2.7 min; r = –0.90) and 10 km (42.7 ± 5.7 min; r = –0.91), the V̇O 2max (r = 0.75), the speed associated with the gas exchange threshold (vGET: 3.38 ± 0.36 m·s ⁻¹ ; r = 0.76), the speed associated with the second ventilatory threshold (vVT 2 : 4.15 ± 0.49 m·s ⁻¹ ; r = 0.84), and the speed associated with the V̇O 2max (vV̇O 2max : 4.78 ± 0.54 m·s ⁻¹ ; r = 0.87). Moreover, 79% and 83% of the variance in 5-km and 10-km performances could be explained solely by the CS predicted from the T10. Results evidenced the strong relationship and practical performance relevance of the T10 CS test. Novelty: Critical speed derived from a 10-min submaximal treadmill test (T10) is significantly correlated with 5-km and 10-km running performances. The T10 critical speed test may represent a useful tool for assessing running performance capabilities.
... The demarcating intensity between the two domains has been described as critical power (CP), critical speed (CS), maximal lactate steady state (MLSS), or the second ventilatory threshold (VT2) (24). Although all three represent physiological landmarks occurring at a similar exercise intensity, the current weight of evidence points toward the CP/CS model offering the most comprehensive explanation of performance over various exercise durations (25)(26)(27)(28)(29). It has also been suggested that the CP/CS best represents the threshold between steady and nonsteady exercise (24,30,31); however, the arguments surrounding this topic are outside the scope of this viewpoint (32,33). ...
... durations and exercise modalities than either V _ O 2max or Wmax (28), or indeed any other measure of the heavy/severe exercise-intensity domain border (37). Indeed, the CP concept has been applied to predict performance across exercise durations from single repetition maximum (29) to marathon performance (27). ...
... The first mathematical model was the linear total distance (Lin-TD). The Lin-TD model consists of linear regression analysis between the covered distance (D) and the running duration (t) in 6-, 9-and 12-minute running tests [5,[18][19][20][21][22][23][24]: ...
... Florence and Weir [23] found a high correlation between marathon time and CV values. The marathon running is an aerobic activity and it dominantly uses the aerobic way of energy production. ...
Article
Full-text available
Introduction: The aim of this study was to examine relationships between Yo-Yo intermittent recovery level 1 test (YYIR1T) performance and critical velocity determined by test protocols consisting of runs with a change of direction and straight runs. Material and Methods: Twelve young soccer players voluntarily participated in study (age: 17.07±0.24 years, training experience: 8.42±2.50 years, height: 178.58±5.76 cm, weight: 70.67±6.14 kg, body mass index: 22.16±1.59). To determine critical velocity, 6-, 9- and 12-minute runs were performed with maximum effort on a straight-line running track and a running track with a change of direction. The critical velocity was determined by two linear regression models. YYIR1T was performed to determine players’ aerobic endurance. The critical velocity value of the two test tracks was compared by the paired samples T-test. The correlation between test parameters was determined by Spearman’s correlation coefficient. Results: A significant difference between anaerobic distance capacity and the mean running speed of the two test tracks (p<0.05) was found. There was a significant correlation between critical velocity in the straight-line test track and the YYIR1T performance (p<0.05). Conclusions: Consequently, it might be said that runs with a change of direction might increase energy expenditure and anaerobic distance capacity was highly affected by those compared to critical velocity.
... The asymptote of the hyperbola is known as critical speed (CS) and the curvature constant (D') represents the finite amount of exercise that can be performed faster than CS. Most athletes can run at their CS for approximately 20-45 min, and CS and D' have been used as predictors of fitness (Florence and Weir 1997); see also (Emig and Peltonen 2020) for related ideas. ...
... We evaluate these four models based on their ability to predict marathon performance, which is a common approach for evaluating other types of fitness estimates (Florence and Weir 1997). ...
Article
Full-text available
Every year millions of people, from all walks of life, spend months training to run a traditional marathon. For some it is about becoming fit enough to complete the gruelling 26.2 mile (42.2 km) distance. For others, it is about improving their fitness, to achieve a new personal-best finish-time. In this paper, we argue that the complexities of training for a marathon, combined with the availability of real-time activity data, provide a unique and worthwhile opportunity for machine learning and for recommender systems techniques to support runners as they train, race, and recover. We present a number of case studies—a mix of original research plus some recent results—to highlight what can be achieved using the type of activity data that is routinely collected by the current generation of mobile fitness apps, smart watches, and wearable sensors.
... Performance during the marathon is determined by a variety of factors, including the physiological and anthropometric characteristics, and training of the subject (Doherty et al., 2020). A significant body of research has been developed to identify reliable correlation to predict the marathon performance time (MPT); for instance, some studies were based on the result of incremental treadmill test (Florence and Weir, 1997), ventilatory threshold (Florence and Weir, 1997), maximal aerobic power (Hagan et al., 1981 and1987), skinfold assessment of body fat (Barandun et al., 2012, Tanda andKnechtle, 2013), and training indices (Slovic, 1977, Hagan et al., 1981and 1987, Tanda, 2011, Barandun et al., 2012, Tanda and Knecthle, 2013. Marathon time predictors are also based on previous race performance (e.g., https://www.runnersworld.com/uk/training/a761681/rws-race-time-predictor/); ...
... Performance during the marathon is determined by a variety of factors, including the physiological and anthropometric characteristics, and training of the subject (Doherty et al., 2020). A significant body of research has been developed to identify reliable correlation to predict the marathon performance time (MPT); for instance, some studies were based on the result of incremental treadmill test (Florence and Weir, 1997), ventilatory threshold (Florence and Weir, 1997), maximal aerobic power (Hagan et al., 1981 and1987), skinfold assessment of body fat (Barandun et al., 2012, Tanda andKnechtle, 2013), and training indices (Slovic, 1977, Hagan et al., 1981and 1987, Tanda, 2011, Barandun et al., 2012, Tanda and Knecthle, 2013. Marathon time predictors are also based on previous race performance (e.g., https://www.runnersworld.com/uk/training/a761681/rws-race-time-predictor/); ...
Article
Full-text available
The aim of this work is to provide further validation of a predictive formula for marathon time performance (MPT) published in 2011. The predictive formula has been correlated with new sample points derived mainly from publicly available data on Strava. The new marathon data points confirm the predictive correlation between mean weekly distance run, mean training pace and marathon performance. The RMSE of 5.4 min for MPT in the 2:47−3:36 (hour:min) range is statistically significant. The extension of the correlation validity for MPT below 2:47 (hour:min) is possible but results are affected by a larger RMSE (9.5 min for MPT). Therefore, the predictive formula for the MPT can be used by coaches and athletes to adjust training programmes and to adopt optimal pace strategies during the marathon race.
... Nevertheless, there are some methodological considerations which might affect an accurate determination of CP (e.g., day-to-day variability, protocol, methodology, mathematical modeling) (Muniz-Pumares et al., 2019). The work rate at CP is closely associated with performance in endurance events (Kranenburg and Smith, 1996;Florence and Weir, 1997;Joyner and Coyle, 2008;Nimmerichter et al., 2017), and moreover, training above or below this MMSS leads to differences in physiological adaptations and specific performance outcomes (Vanhatalo et al., 2011;Iannetta et al., 2018). Due to the increasing availability of affordable on-the-bike power meters though, field-based methods of threshold assessment have been validated (Karsten et al., 2015;Triska et al., 2015). ...
Article
Full-text available
To investigate the agreement between critical power (CP) and functional threshold power (FTP), 17 trained cyclists and triathletes (mean ± SD: age 31 ± 9 years, body mass 80 ± 10 kg, maximal aerobic power 350 ± 56 W, peak oxygen consumption 51 ± 10 mL·min-1·kg-1) performed a maximal incremental ramp test, a single-visit CP test and a 20-min time trial (TT) test in randomised order on three different days. CP was determined using a time-trial (TT) protocol of three durations (12, 7 and 3 min) interspersed by 30 min passive rest. FTP was calculated as 95% of 20-min mean power achieved during the TT. Differences between means were examined using magnitude-based inferences and a paired-samples t-test. Effect sizes are reported as Cohen’s d. Agreement between CP and FTP was assessed using the 95% limits of agreement (LoA) method and Pearson correlation coefficient. There was a 91.7 % probability that CP (256 ± 50 W) was higher than FTP (249 ± 44 W). Indeed, CP was significantly higher compared to FTP (P = 0.041) which was associated with a trivial effect size (d = 0.04). The mean bias between CP and FTP was 7 ± 13 W and LoA were -19 to 33 W. Even though strong correlations exist between CP and FTP (r = 0.969; P < 0.001), the chance of meaningful differences in terms of performance (1% smallest worthwhile change), were greater than 90%. With relatively large ranges for LoA between variables, these values generally should not be used interchangeably. Caution should consequently be exercised when choosing between FTP and CP for the purposes of performance analysis.
Article
Full-text available
The elegant concept of a hyperbolic relationship between power, velocity, or torque and time to exhaustion has rightfully captivated the imagination and inspired extensive research for over half a century. Theoretically, the relationship’s asymptote along the time axis (critical power, velocity, or torque) indicates the exercise intensity that could be maintained for extended durations, or the “heavy–severe exercise boundary”. Much more than a critical mass of the extensive accumulated evidence, however, has persistently shown the determined intensity of critical power and its variants as being too high to maintain for extended periods. The extensive scientific research devoted to the topic has almost exclusively centered around its relationships with various endurance parameters and performances, as well as the identification of procedural problems and how to mitigate them. The prevalent underlying premise has been that the observed discrepancies are mainly due to experimental ‘noise’ and procedural inconsistencies. Consequently, little or no effort has been directed at other perspectives such as trying to elucidate physiological reasons that possibly underly and account for those discrepancies. This review, therefore, will attempt to offer a new such perspective and point out the discrepancies’ likely root causes.
Article
Full-text available
Purpose: The prediction of running anaerobic sprint test and 800 m performance by parameters of critical velocity was examined in this study. Material: The participants of study were consisted of thirteen amateur soccer players (n=13, age=22.69±5.29 years, weight=72.46±6.32 kg, height=176.92±6.73 cm). The 800 and 2400 m running tests were performed for determination of critical velocity and anaerobic distance capacity. The critical velocity and anaerobic distance capacity were determined by three mathematical models (linear total distance, linear velocity, non-linear two parameter model). The repeated sprint and sprint endurance ability was determined by running anaerobic sprint test and 800 m running test. The simple and multiple linear regression analysis was used for prediction of dependent variables (running anaerobic sprint test and 800 m running performance) by independent variables (critical velocity and anaerobic distance capacity) of study. The correlation between variables was determined by Pearson correlation coefficient. Results: It was found that anaerobic distance capacity was a significant predictor of running anaerobic sprint test and 800 m running performance (p<0.05). However, it was determined that critical velocity predicted significantly only time parameters of running anaerobic sprint test and 800 m test (p<0.05). Also, the parameters of 800 m test (except for average velocity) were significantly predicted by running anaerobic sprint test parameters (p<0.05). Conclusions: It may be concluded that anaerobic distance capacity is an indicator of repeated sprint and speed endurance ability in soccer and may be used in improvement of sprint endurance performance.
Chapter
In this work we generate user profiles from the raw activity data of over 12000 marathon runners. We demonstrate that these user profiles capture accurate representations of the fitness and training of a runner, and show that they are comparable to current methods used to predict marathon performance – many of which require many years of prior experience or expensive laboratory testing. We also briefly investigate how these user profiles can be used to help marathon runners in their training and race preparation when combined with current recommender systems approaches.
Article
Full-text available
Physiological variables such as maximal oxygen uptake (VO2max), velocity at maximal oxygen uptake (vVO2max), running economy (RE) and changes in lactate levels are considered the main factors determining performance in long-distance races. The aim of this review was to present the mathematical models available in the literature to estimate performance in the 5000 m, 10,000 m, half-marathon and marathon events. Eighty-eight articles were identified, selections were made based on the inclusion criteria and the full text of the articles were obtained. The articles were reviewed and categorized according to demographic, anthropometric, exercise physiology and field test variables were also included by athletic specialty. A total of 58 studies were included, from 1983 to the present, distributed in the following categories: 12 in the 5000 m, 13 in the 10,000 m, 12 in the half-marathon and 21 in the marathon. A total of 136 independent variables associated with performance in long-distance races were considered, 43.4% of which pertained to variables derived from the evaluation of aerobic metabolism, 26.5% to variables associated with training load and 20.6% to anthropometric variables, body composition and somatotype components. The most closely associated variables in the prediction models for the half and full marathon specialties were the variables obtained from the laboratory tests (VO2max, vVO2max), training variables (training pace, training load) and anthropometric variables (fat mass, skinfolds). A large gap exists in predicting time in long-distance races, based on field tests. Physiological effort assessments are almost exclusive to shorter specialties (5000 m and 10,000 m). The predictor variables of the half-marathon are mainly anthropometric, but with moderate coefficients of determination. The variables of note in the marathon category are fundamentally those associated with training and those derived from physiological evaluation and anthropometric parameters.
Article
In brief: To examine the relationship between actual performance times and predicted marathon times calculated from running velocity at the anaerobic threshold (AT), 18 male subjects ran on a treadmill to volitional fatigue. The AT was determined by examining the excess CO2 elimination curve. The subjects' average velocity at the AT was 9.22 mph. The average predicted time was 2:53:49, while the average marathon time was 2:52:06. There was a highly significant correlation between the predicted and actual marathon times (r =.94, p <.01). This relationship suggests that running velocity at the AT may be critical in determining efficient running speed during marathons.
Article
A new conception of dynamic or static muscular work tests is presented. The authors define the critical power of a muscular work from the notions of maximum work and maximum time of work. The work capacity is then considered in the case of dynamic work, and of continuous or intermittent static work. From the data presented it is possible to define the maximum amount of work that can be performed in a given time as well as the conditions of work performed without fatigue. (French & German summaries) (22 ref.) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
The present investigation was conducted to determine whether critical power (CP) assesses the ability to perform continuous aerobic exercise and to determine whether training-induced changes in aerobic endurance are reflected by changes in the slope, but not the y-intercept of the CP function. Twelve healthy, active, but untrained male students (mean age +/- SD = 19.1 +/- 0.8 yr) undertook 8 wk of cycle ergometer endurance training (30-40 min a day, three times a week) at an intensity corresponding to their CP. Six control subjects of similar age and initial training status refrained from regular exercise for the same period. Before and immediately following the training period, each of the 18 participants completed three cycle ergometer tests to determine their CP function, an incremental exercise task to establish their maximal oxygen uptake (VO2max), and 40 min of continuous cycle ergometry at or near their calculated CP. CP was significantly correlated with endurance time at 270 W (r = 0.65, P < 0.05) and with the mean power that could be maintained for 40 min (r = 0.87-0.95, P < 0.01), but overestimated the latter by less than 6%. In response to endurance training, CP increased from a mean of 196 +/- 40.9 W to 255 +/- 28.4 W (31%) (ANCOVA, P < 0.01), while the mean power output maintained for 40 min of exercise increased from 190 +/- 34.5 W to 242 +/- 34.9 W (28%). VO2max increased from 49.2 +/- 7.8 ml.kg-1.min-1 to 53.4 +/- 6.4 ml.kg-1.min-1 (8.5%) (P < 0.01).(ABSTRACT TRUNCATED AT 250 WORDS)
Article
The purpose of this study was to examine the relationship between anaerobic running capacity and peak plasma lactate. Twelve adult males (mean age +/- SD = 21.9 +/- 1.2 yrs) performed a critical velocity test from which anaerobic running capacity was determined and a maximal treadmill running test from which peak plasma lactate was determined from post-exercise blood samples taken at one minute intervals. The results indicated that anaerobic running capacity (mean +/- SD = 0.18 +/- 0.04 km) was not significantly (r = -0.06, p greater than 0.05) correlated with peak plasma lactate (9.3 +/- 1.8 mM). These findings do not support anaerobic running capacity as an indirect indicator of anaerobic capabilities.
Article
The purpose of this investigation was to determine the accuracy of the critical velocity (CV) test for predicting time to exhaustion (time limit = TL) during treadmill running. Ten adult males (mean +/- SD of age = 23 +/- 2 years) volunteered to perform a maximal treadmill test, a CV test, and five exhaustive treadmill runs at 70%, 85%, 100%, 115% and 130% of CV for the determination of actual TL. Related t-tests revealed significant (p less than 0.05) differences between the predicted and actual TL values for velocities equal to 100 and 130% of CV. The correlations between predicted and actual TL values for velocities above CV ranged from r = 0.957 to 0.980 (SEE = 0.28-0.82 minutes). At 100% of CV, the subjects maintained the running pace for an average of 16.43 +/- 6.08 minutes (range = 9.96-31.90 minutes) while, at 85% of CV, 8 of the 10 subjects were able to maintain the running pace for 60 minutes. These findings did not support the validity of the CV test for predicting the actual TL during treadmill running and indicated that, in 20% of the cases, CV overestimated the running velocity that could be maintained for 60 minutes by greater than 15%.
Article
The purpose of this investigation was to determine the contribution of the onset of blood lactate accumulation (OBLA), the heartrate-workload slope (HR-WL) and the efficiency of electrical activity (EEA = slope of IEMG vs. workload) of the leg extensor muscles to Critical Power (CP). Twelve adult males (mean age +/- SD = 24.5 +/- 2.8 yrs) volunteered as subjects for this study. Zero-order correlations indicated that OBLA was significantly (p less than 0.05) related to CP (r = 0.616) and EEA (r = -0.577). Stepwise multiple regression resulted in a one variable model with OBLA the only significant (p less than 0.05) predictor of CP. Furthermore, a related t-test resulted in a significant difference between the means of the power out-put at CP (mean +/- SD = 230.0 +/- 22.1 watts) and OBLA (179.6 +/- 31.8 watts). The results of this study indicated that the two threshold parameters, CP and OBLA, were significantly related and therefore it is likely that the physiological factors responsible for OBLA also influence CP. However, the significant mean differences indicated that the mechanisms which underly CP and OBLA were not identical. Furthermore, the HR-WL slope (mean +/- SD = 0.343 +/- 0.071 beats per watt) and EEA (0.969 +/- 0.572 microvolts per watt) were not potent predictors of CP.
Article
The purpose of this investigation was to determine the oxygen consumption, heart rate and plasma lactate responses at the fatigue threshold (FT) and estimate the length of time the FT could be maintained. Ten adult males (mean age +/- SD = 21.1 +/- 1.3 yrs) volunteered to perform a maximal treadmill test and FT test. During the maximal test, VO2 heart rate and plasma lactate measurements were taken. The results of the investigation indicated that the FT (14.0 +/- 1.2 km.hr-1, 197 +/- 8 bpm; 47.5 +/- 5.7 ml/kg.min-1, 5.4 +/- 1.3 mM) was very close to a maximal effort (VO2max = 14.4 +/- 1.2 km.hr-1, 203 +/- 10 bpm; 49.5 +/- 6.1 ml/kg.min-1, 7.4 +/- 2.1 mM) and could be maintained for only 0.16 to 0.28 hrs. These findings do not support the validity of the FT as a measure of the maximal running velocity that can be continued for an extended period of time without exhaustion.