Training & Testing 209

Velocity- and Power-Load Relationships of the Bench

Pull vs. Bench Press Exercises

Authors L. Sánchez-Medina1, J. J. González-Badillo2, C. E. Pérez3, J. G. Pallarés4

1
Affiliations Studies, Research and Sports Medicine Centre, Instituto Navarro de Deporte y Juventud (INDJ), Pamplona, Spain
2
Faculty of Sport, Pablo de Olavide University, Seville, Spain
3
Sports Medicine Centre, University of Murcia, Spain

This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
4
Exercise Physiology Laboratory, University of Castilla La Mancha, Toledo, Spain

Key words Abstract obtained prediction equations. Important differ-

●
▶ resistance training
▼ ences in the load that maximizes power output
●
▶ muscle strength
This study compared the velocity- and power- (Pmax) and the power profiles of both exercises
●
▶ 1RM prediction
load relationships of the antagonistic upper-body were found according to the outcome variable
●
▶ exercise testing

●
▶ isoinertial assessment exercises of prone bench pull (PBP) and bench
●
▶ maximal power output press (BP). 75 resistance-trained athletes per-
formed a progressive loading test in each exer-
cise up to the one-repetition maximum (1RM) in
random order. Velocity and power output across
the 30–100 % 1RM were significantly higher for
PBP, whereas 1RM strength was greater for BP.
A very close relationship was observed between
relative load and mean propulsive velocity for
both BP (R2 = 0.97) and PBP (R2 = 0.94) which ena-
bles us to estimate %1RM from velocity using the
used: mean (MP), peak (PP) or mean propulsive
power (MPP). When MP was considered, the Pmax
load was higher (56 % BP, 70 % PBP) than when
PP (37 % BP, 41 % PBP) or MPP (37 % BP, 46 % PBP)
were used. For each variable there was a broad
range of loads at which power output was not
significantly different. The differing velocity- and
power-load relationships between PBP and BP
seem attributable to the distinct muscle archi-
tecture and moment arm levers involved in these
exercises.

Introduction its predictive ability on sport-specific perform-

accepted after revision
June 21, 2013
Bibliography
DOI http://dx.doi.org/
10.1055/s-0033-1351252
Published online:
July 30, 2013
Int J Sports Med 2014; 35:
209–216 © Georg Thieme
Verlag KG Stuttgart · New York
ISSN 0172-4622
Correspondence
Dr. Luis Sánchez-Medina
Studies, Research and Sports
Medicine Centre
Instituto Navarro de Deporte y
Juventud (INDJ)
C/Sangüesa 34
Pamplona
Spain
31005
Tel.: + 34/948/292 623
Fax: + 34/948/292 636
lsmedina@gmail.com
▼ ance [21, 22, 27, 29] or have examined the effect
The kinematics and kinetics associated with
resistance training (RT) are believed to be critical
stimuli for neuromuscular adaptations to occur
[4]. A detailed description of the load-velocity
and load-power relationships for the most
widely used RT exercises will enable strength
and conditioning coaches and sport scientists:
i) to individually prescribe the loads that opti-
mize strength and power gains while improving
training efficiency, and ii) to periodically assess
the neuromuscular adaptations induced by
training or detraining against the relative loads
(% 1RM) that are most related to sport-specific
performance.
Although less popular than the bench press (BP),
the prone bench pull (PBP) is a multi-joint RT
exercise commonly used in sport disciplines
which require upper-body pulling such as can-
oeing [10–12], rowing [17, 18, 21, 22], sailing
[26, 27], combat sports [7], cross country skiing
[34] and swimming [29]. Even though some
studies have described the utilization of the PBP
exercise as part of physical conditioning pro-
grams at the elite level [10–12, 17], have analyzed
of alternating agonist (BP) and antagonist (PBP)
exercises on strength and power development
during complex training [2, 30, 31], the velocity-
and power-load relationships of the opposing
upper-body actions of pushing and pulling have
received only minor attention.
To the best of our knowledge, there is only one
study [25] that has compared the kinematics and
kinetics of the BP and PBP. Pearson et al. [25]
described the load-velocity and power output
profiles of these exercises in a sample of 12 elite-
level male sailors. However, at least with regard
to the BP exercise, their results are somewhat
conflicting with the findings of several studies
conducted by our research group [14, 32] and
others [9, 19]. Of particular concern is the fact
that the mean concentric velocities reported by
Pearson et al. [25] for each relative load ( %1RM)
are ~45–50 % lower than those of the aforemen-
tioned studies [9, 14, 19, 32]. These velocities are
unexpectedly low, especially if one considers that
the sample was composed of very strong and
experienced sailors [25]. Since mechanical power
is the product of movement velocity and applied

Sánchez-Medina L et al. Velocity- and Power-Load Relationships … Int J Sports Med 2014; 35: 209–216
210 Training & Testing
force, large differences in velocity will influence the obtained
power output values, making it difficult to obtain a clear picture
of the kinetics and kinematics of a given exercise. In particular,
this can affect the load that maximizes the mechanical power
output (Pmax load), which is a variable often taken as a reference
for prescribing resistance exercise or assessing neuromuscular
performance [6, 38]. It was therefore deemed necessary to fur-
ther explore the load-velocity and load-power relationships of
the PBP vs. the BP in a large sample of highly experienced ath-
letes from different sports.
Research attention [14, 32] has recently been placed on monitor-
ing movement velocity during resistance exercise. A very close
relationship between relative load and mean velocity was
described for the BP exercise, a novel finding which has impor-
tant practical applications for the prescription and monitoring
of training load in RT [13, 14, 32]. It seems appropriate to con-
firm whether this relationship between movement velocity and
load ( % 1RM) also exists in the PBP exercise as well as to exam-
ine the possible differences with the BP.
In other respects, the importance of considering the contribu-
tion of the propulsive phase (i. e., that portion of the concentric
phase during which the measured acceleration is greater than
acceleration due to gravity) when assessing strength and muscle
power in isoinertial conditions has recently been highlighted
[13, 33]. Referring the mean mechanical values only to the pro-
pulsive phase when assessing the velocity and power with
which a load is lifted in a concentric action prevents underesti-
mating an individual’s neuromuscular ability, especially when
lifting light and medium loads.
Therefore the main purpose of the present study was to compare
the load-velocity and load-power relationships of these 2 antag-
onistic upper-body exercises, BP and PBP. A secondary aim was
to analyze the relationship between movement velocity and
relative load ( % 1RM) in the PBP in order to assess the possibility
of using velocity data to estimate loading intensity. Finally, we
aimed to study the relative contribution of the propulsive and
braking phases [33] across different percentages of the one-rep-
etition maximum (1RM) as well as to determine the Pmax load ( %
1RM) in both exercises using 3 different outcome measures:
mean concentric power (MP), mean propulsive power (MPP) and
peak power (PP).
Materials and Methods
▼ formed on the same Smith machine.
Subjects
75 men (age = 26.7 ± 5.4 years, body mass = 76.0 ± 8.8 kg, height =
176.9 ± 6.8 cm, body fat = 12.3 ± 3.7 %) volunteered to take part in
this study. Their 1RM strength for the BP and PBP were
90.3 ± 16.3 kg and 80.2 ± 11.8 kg (i. e., 1.19 ± 0.21 and 1.06 ± 0.15
normalized per kg of body mass), respectively. Participants were
junior and senior national-level athletes in their respective sport
(wrestling, canoeing, rowing or judo) with a strength-training
background ranging from 4 to beyond 8 years. In the 12 months
preceding this study, subjects had been performing 2–3 strength
training sessions per week and were capable of performing both
exercises with excellent technique. None of the subjects was
taking drugs, medications or dietary supplements known to
influence physical performance. The study met the ethical
standards of this journal [15] and was approved by the Research
Ethics Committee of Pablo de Olavide University. After being
Sánchez-Medina L et al. Velocity- and Power-Load Relationships … Int J Sports Med 2014; 35: 209–216
informed of the purpose and experimental procedures, subjects
signed a written informed consent form prior to participation.
Testing procedures
For each subject, testing was conducted over 3 sessions sepa-
rated by 48 h. The first session was used for body composition
assessment, personal data and health history questionnaire
administration, and familiarization with the BP and PBP testing
protocols. Subjects arrived at the laboratory in the morning, in a
well-rested condition and fasted state. After being interviewed
and having their body composition determined, they were
instructed on how to properly perform the lifts and carried out
some practice sets with light and medium loads.
After a resting day, 2 testing sessions (one for each exercise)
were conducted in random order. In each session, the individual
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
load-velocity and load-power output relationships as well as
1RM strength were determined by means of a progressive load-
ing test as explained later. Both testing sessions were conducted
at the same time of day (17:00 h) to control for circadian rhythms
effects on neuromuscular performance [24]. Warm-up consisted
of 5 min of stationary cycling at a self-selected easy pace, 5 min
of static stretching and upper-body joint mobilization exercises,
followed by 2 sets of 5 repetitions for each exercise with fixed
loads of 20 and 40 kg.
A description of the BP testing protocol has recently been
reported in detail elsewhere [33]; start and finish positions for
this exercise are shown in ● ▶ Fig. 1a, b. In the PBP, subjects were
instructed to lie prone and place their chin on the padded edge
of a high bench. The pulling phase began with both elbows in full
extension, while the barbell was grasped with hands shoulder-
width apart or slightly wider (4–5 cm). The participants were
instructed to pull with maximum effort until the barbell struck
the underside of the bench, after which it was again lowered to
the starting position (● ▶ Fig. 1c, d); they were not allowed to use
their legs to hold onto the bench. There was a distance of 8 cm
between the underside of the bench and the subjects’ chest. Sub-
jects were required to always perform the concentric action of
both exercises in an explosive manner, at maximal voluntary
velocity. A momentary pause, which lasted approximately 1.5 s,
was interposed between the eccentric and concentric phases of
each exercise to minimize the contribution of the rebound effect
and allow for more reproducible, consistent measurements.
Only the concentric actions (pushing for BP and pulling for PBP)
were analyzed in the present study. Both exercises were per-
For both exercises, initial load was set at 20 kg for all subjects
and was progressively increased in increments of 10 kg until
the attained mean propulsive velocity (MPV) was lower than
0.5 m · s − 1 and 0.7 m · s − 1 for BP and PBP respectively. Thereafter,
load was individually adjusted using smaller increments (5
down to 2.5 kg). The heaviest load that each subject could prop-
erly lift while completing full range of motion and without any
external help was considered to be his 1RM. In the PBP, the bar-
bell was required to touch the underside of the bench at the end
of the concentric pulling phase. Trained spotters were present
when high loads were lifted to ensure safety. 3 attempts were
executed for light ( < 50 % RM), 2 for medium (50–80 % RM) and
only 1 for the heaviest ( > 80 % RM) loads. Inter-set rest intervals
were 3 min for the light and medium loads and 5 min for the
heaviest loads. Only the best repetition at each load, according to
the criteria of fastest MPV [33], was considered for subsequent
analysis. In a previous pilot study performed with part of these

subjects (n = 12), the intra-subject mean coefficients of variation
(CV) for the whole load-velocity relationships were 2.4 % and
3.0 % for BP and PBP, respectively.
Measurement equipment and data acquisition
Height was measured to the nearest 0.5 cm during a maximal
inhalation using a wall-mounted stadiometer (Seca 202, Seca
Ltd., Hamburg, Germany). Body weight was determined and fat
percentage estimated using an 8-contact electrode segmental
body composition analyzer (Tanita BC-418, Tanita Corp., Tokyo,
Japan) with the subjects in a morning fasting state and only
wearing underclothes.
A Smith machine (Multipower Fitness Line, Peroga, Murcia,
Spain) was used for all testing sessions (● ▶ Fig. 1). This machine
allows only vertical displacement of the barbell along a fixed
pathway, and its guide rods and bearings are specially designed
to ensure a smooth operation, with very low friction force
between the barbell and the support rails. The Smith machine
did not have any kind of counterweight mechanism, acting iden-
tical to free-weights (i. e., isoinertial loading). The weight of the
barbell, including the guidance system, totalled 20.0 kg. Extra
load was added by sliding equal weight discs onto both ends of
the barbell.
Sánchez-Medina L et al. Velocity- and Power-Load Relationships … Int J Sports Med 2014; 35: 209–216
Training & Testing 211
Fig. 1 Start and finish positions of the concentric
phase for the Bench Press a, b and Prone Bench
Pull c, d resistance training exercises performed in
a Smith machine.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
A dynamic measurement system (T-Force System, Ergotech,
Murcia, Spain) automatically calculated the relevant kinematic
parameters of every repetition, provided auditory velocity
feedback and stored data on disk for analysis. This system con-
sists of a linear velocity transducer interfaced to a personal
computer by means of a 14-bit resolution analog-to-digital
data acquisition board and custom software. Instantaneous
velocity was sampled at a frequency of 1 000 Hz and subse-
quently smoothed with a 4th order low-pass Butterworth filter
with a cut-off frequency of 10 Hz. A digital filter with no phase
shift was then applied to the data. Reliability (ICC = 1.00,
CV = 0.57 %) of this system has been recently reported else-
where [32]. Mean propulsive velocity (MPV) and power (MPP)
were calculated as the average velocity and power output val-
ues, respectively, measured only through the propulsive phase,
defined as that portion of the concentric action during which
the measured acceleration (a) is greater than acceleration due
to gravity (g), i. e., a ≥ − 9.81 m · s − 2. The final braking phase, on
the other hand, corresponds to the remaining part of the con-
centric action, during which a < g [33]. Peak power (PP) was the
highest instantaneous power output value registered at a par-
ticular instant (1 ms) during the concentric phase. Since the
effect of friction force was negligible in pilot testing, it was not
212 Training & Testing

taken into account in the calculations. The constant downward

PBP-BP (m · s − 1)
MPV Difference
force exerted by the cable (~5 N) was not taken into considera-
tion since it was minimal compared to the weights being lifted.

Statistical analyses

0.21
0.23
0.24
0.25
0.27
0.28
0.29
0.30
0.31
0.32
0.33
0.34
0.34
0.35
0.36
Standard statistical methods were used for the calculation of
means, standard deviations (SD), standard error of the estimate

Braking Phase ( %)
(SEE), confidence intervals (CI) and Pearson product-moment
correlation coefficients (r). Relationships between load ( % 1RM)
and velocity and power output were studied by fitting second-

Table 1 Mean propulsive velocity attained with each % 1RM, and relative contribution of the propulsive and braking phases to the total concentric duration in the BP and PBP exercises (n = 75).
order polynomials to data. A 2-way (exercise × load) ANOVA was

15
14
13
12
11
11
10
9
8
7
6
5
4
3
2
used to detect differences between exercises as well as differ-
ences between power output at different loads. Scheffé post-hoc

Propulsive Phase ( %)
test was used to identify the source of any significant differ-
ences. Significance was accepted at the p ≤ 0.05 level. Analyses

This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Prone Bench Pull (PBP)
were performed using SPSS software version 19.0 (SPSS, Chi-
cago, IL).

85
86
87
88
89
89
90
91
92
93
94
95
96
97
98
Results
▼

95 % CI (m · s − 1)
1RM strength

1.47–1.52
1.40–1.45
1.33–1.37
1.25–1.30
1.18–1.23
1.11–1.15
1.04–1.08
0.97–1.01
0.90–0.94
0.83–0.87
0.77–0.80
0.70–0.74
0.64–0.67
0.57–0.60
0.51–0.54
1RM strength was significantly different (p < 0.001) between
exercises: 90.3 ± 16.3 kg for BP and 80.2 ± 11.8 kg for PBP. The
maximum strength ratio between BP and PBP was 1.12 ± 0.11.
Mean concentric displacement was significantly different
(p < 0.001) for BP (51.3 ± 5.6 cm) and PBP (48.1 ± 3.6 cm).
MPV (m · s − 1)
1.50 ± 0.11*
1.42 ± 0.11*
1.35 ± 0.10*
1.28 ± 0.10*
1.20 ± 0.10*
1.13 ± 0.10*
1.06 ± 0.09*
0.99 ± 0.09*
0.92 ± 0.09*
0.85 ± 0.08*
0.79 ± 0.08*
0.72 ± 0.07*
0.65 ± 0.07*
0.59 ± 0.06*
0.52 ± 0.06*
Relationship between relative load and velocity
After plotting mean propulsive velocity (MPV) against % 1RM
and fitting a second-order polynomial to all data points, a very
close relationship between these 2 variables was found for the
Braking Phase ( %)

BP (R2 = 0.97) and PBP (R2 = 0.94) exercises (● ▶ Fig. 2). Individual

curve fits for each test gave an R2 of 0.995 ± 0.004 (range: 0.978–
0.999; CV = 0.43 %) for BP and R2 of 0.994 ± 0.005 (range: 0.973–

MPV = Mean Propulsive Velocity; CI = Confidence Interval; * Significantly different to the BP exercise (p < 0.001)
0.999; CV = 0.47 %) for PBP. The mean MPV attained with each
24
21
19
17
15
12
10
8
6
3
1
0
0
0
0
percentage of 1RM was obtained from these polynomial fits,
from 30 % 1RM onwards, in 5 % increments (● ▶ Table 1). MPV
Propulsive Phase ( %)

attained with each %1RM was significantly higher (p < 0.001) for
PBP than BP. The actual mean velocity attained with the 1RM
Bench Press (BP)

2.0
y= 3.106.10–5 x2 – 0.01824 x + 2.043
Mean Propulsive Velocity (m.s–1)

1.8
100
100
100
100
76
79
81
83
85
88
90
92
94
97
99

2
R = 0.94, SEE= 0.089, N=532
1.6
1.4
95 % CI (m · s − 1)

1.2
1.0
1.27–1.30
1.18–1.22
1.09–1.13
1.00–1.04
0.92–0.95
0.84–0.87
0.76–0.79
0.68–0.71
0.60–0.63
0.52–0.55
0.45–0.47
0.37–0.39
0.30–0.32
0.23–0.25
0.16–0.18

0.8
0.6
0.4
y = 3.849.10–5 x2 – 0.0209 x + 1.886
0.2
R2 = 0.97, SEE= 0.076, N= 626
MPV (m · s − 1)

0.0
1.29 ± 0.08
1.20 ± 0.08
1.11 ± 0.07
1.02 ± 0.07
0.94 ± 0.07
0.85 ± 0.07
0.77 ± 0.07
0.69 ± 0.06
0.61 ± 0.06
0.53 ± 0.06
0.46 ± 0.05
0.38 ± 0.05
0.31 ± 0.05
0.24 ± 0.04
0.17 ± 0.04

10 20 30 40 50 60 70 80 90 100
Load (%1RM)
BP PBP
Load ( %1RM)

Fig. 2 Relationship between relative load ( % 1RM) and mean propulsive

velocity for the BP and PBP exercises directly obtained from raw data pairs
derived from the 75 progressive isoinertial loading tests performed in
100
30
35
40
45
50
55
60
65
70
75
80
85
90
95

each exercise.

Sánchez-Medina L et al. Velocity- and Power-Load Relationships … Int J Sports Med 2014; 35: 209–216
 Training & Testing 213

100% RM a 30 % RM
0.75 Propulsive Braking
1500 85 % 15 %
Vpeak
0.60 Vpeak 0.51 m.s–1 1000
500

Power Output (W)

0.72 m.s–1 25%
Velocity (m.s–1)

75%
0.45 0
–500 20 40 60 80 100
MVPBP =0.47 m.s–1
0.30 –1000
–1500
0.15 –2000
BP: MP= 373 W; MPP= 590 W
MVBP =0.20 m.s–1 –2500
PBP: MP= 464 W; MPP= 849 W
0.00 –3000
0 20 40 60 80 100 b 60 % RM
Concentric Phase Duration (%) Propulsive Braking

This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
PBP BP 1500 90 % 10 %
1000
Fig. 3 Example of the actual velocity-time curves for the concentric 500

Power Output (W)

phase of each exercise when lifting a maximal 1RM load for a representa- 88% 12%
0
tive subject. Vpeak = Peak velocity; MV = Mean velocity; BP = Bench Press; 20 40 60 80 100
PBP = Prone Bench Pull. –500
–1000
–1500
load (V1RM) was very different between exercises (0.16 ± 0.04 for –2000
BP: MP= 441 W; MPP= 507 W
BP and 0.52 ± 0.05 for PBP; p < 0.001). An example of the actual –2500
PBP: MP= 599 W; MPP= 754 W
velocity-time curves for a representative subject when lifting his –3000
1RM is provided in ●▶ Fig. 3. c 90 % RM
Propulsive Braking
1500 95 %
Predicting load ( % 1RM) from velocity data 5%
1000
A prediction equation to estimate relative load (Load, % 1RM)
500 100 % 0%
from mean propulsive velocity data (MPV, in m · s-1) could be
Power Output (W)

obtained for both exercises: 0

20 40 60 80 100
BP Load = 11.2988 MPV2–78.05 MPV + 113.04 –500
(R2 = 0.972; SEE = 4.39 % 1RM) –1000
PBP Load = 13.2596 MPV2–93.867 MPV + 144.38 –1500
(R2 = 0.945; SEE = 5.90 % 1RM) –2000
In the case that mean concentric velocity (MV) is used, the BP: MP= MPP= 364 W
–2500
resulting equations were: PBP: MP= 556 W; MPP= 594 W
–3000
BP Load = 11.4196 MV2–81.904 MV + 114.03
Concentric Phase Duration (%)
(R2 = 0.969; SEE = 4.59 % 1RM)
PBP BP
PBP Load = 18.5797 MV2–104.182 MV + 147.94
(R2 = 0.955; SEE = 5.31 % 1RM)
Fig. 4 Comparison of power output-time curves for each exercise when
lifting a: a light load (30 % 1RM); b medium load (60 % 1RM); and c high
Contribution of the propulsive and braking phases to load (90 % 1RM) for a representative subject. The percent contribution of
different loading conditions the propulsive and braking phases to the whole concentric duration of the
●
▶ Table 1 shows the contribution of the propulsive and brak- lift is shown. MP = Mean Power; MPP = Mean Propulsive Power; BP = Bench
ing phases [33] to the total concentric time from 30 to 100 % Press; PBP = Prone Bench Pull.
1RM, according to calculations made using the high correla-
tion that exists between load ( % 1RM) and relative contribu-
tion of the propulsive phase to the total concentric duration The load that maximizes the mechanical power output
of the lift in both exercises (BP: r = 0.90, p < 0.001; PBP r = 0.84, (Pmax load)
p < 0.001) from individual data obtained in the 75 BP and PBP The load that maximized the mechanical power output was
tests. An example of the power-time curves detailing the con- found to be dependent on the outcome variable used (MP, MPP or
tribution of the propulsive and braking phases to the total PP) and was statistically different between exercises (●▶ Fig. 5). In

concentric duration when light (30 % 1RM), medium (60 % order to calculate the Pmax load, a second-order polynomial was
1RM) and high (90 % 1RM) loads are lifted is shown for a rep- fitted to individual load-power data points. When using MP
resentative subject in ● ▶ Fig. 4. The diff erent diminishing con- (●▶ Fig. 5a), power output was maximized at a load of 56 ± 2 %

tribution of the braking phase and the resulting power output 1RM for BP and 70 ± 4 % 1RM for PBP, although no statistically
(MP and MPP) values as load increases can be compared significant differences in power output were found for loads
between exercises. between 40 % and 70 % 1RM for BP and for loads between 50 %
and 90 % for PBP. When using MPP (● ▶ Fig. 5b), power output was

maximized at 37 ± 8 % 1RM for BP and 46 ± 10 % 1RM for PBP,

Sánchez-Medina L et al. Velocity- and Power-Load Relationships … Int J Sports Med 2014; 35: 209–216
214 Training & Testing

BP and PBP (p < 0.001). The absolute Pmax values directly obtained
a Mean Power
for each variable (i. e., MP, MPP and PP) for the BP (400 ± 80 W for
700 Pmax 70% 1RM
† MP, 494 ± 109 W for MPP and 808 ± 182 W for PP) and the PBP
600 (495 ± 81 W for MP, 630 ± 119 W for MPP and 917 ± 175 W for PP)
tests were all significantly different from one another (p < 0.001).
500
Power Output (W)

400
* Discussion
300
▼
200 † The main finding of the present study was that the load-velocity
Pmax 56 % 1RM and load-power relationships were significantly different
100
between the BP and PBP exercises. Velocity and power output
0 values for any load (30–100 % 1RM) were always higher for the
20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 PBP compared to the BP. The close relationships observed
b Mean Propulsive Power between load ( % 1RM) and mean velocity in both exercises ena-

This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
† ble us to estimate the magnitude of the loads being lifted from
800
velocity measurements. In addition, the obtained results for the
Pmax 46 % 1RM
700 PBP exercise confirm previous published research [33] regarding
600
the importance of considering the contribution of the propulsive
Power Output (W)

and braking phases when assessing muscle strength and power

500
in isoinertial conditions. Finally, the differences observed when
400 determining the Pmax load according to the chosen outcome vari-
300
able (MP, MPP or PP) together with the broad range of relative
† loads at which mechanical power is not statistically significantly
200
Pmax 37% 1RM different seem to minimize the importance of identifying a single
100 load for maximizing power output during resistance training.
0
The load-velocity relationship has been described for isoinertial
20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 resistance exercise performed with maximal voluntary effort
[5, 19]. However, the role played by movement velocity has often
c Peak Power
† been overlooked in the everyday practice of strength training
1200
[28]. Current results, together with previous data for the BP and
1100
Pmax 41 % 1RM squat exercises [14, 32] emphasize the practical importance of
1000
900 considering movement velocity for monitoring training load in
Power Output (W)

800 resistance exercise. The extremely close relationship observed

700 between relative load and MPV for BP (R2 = 0.97) and PBP
600 (R2 = 0.94) in the present study (● ▶ Fig. 2) makes it possible to
500 determine with great precision which % 1RM is being used pro-
400 †
Pmax 37 % 1RM
vided that the first repetition of a set is performed with maximal
300
voluntary velocity. It also allows us to determine the real inten-
200
100 sity of effort being applied by an athlete when using any load
0 from 30 % to 95 % of 1RM. Furthermore, if repetition velocity is
20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 habitually monitored it is possible to determine whether the
Load (%1RM) proposed load (kg) for a given training session truly represents
PBP BP the real effort ( % 1RM) that was intended. Movement velocity is
important because both the neuromuscular demands and the
Fig. 5 Load-power relationships for the BP and PBP exercises according
training effect itself largely depend on the velocity at which
to 3 different outcome measures of power output: a Mean Power; b Mean
loads are lifted.
Propulsive Power; and c Peak Power. Statistically significant differences in
power output between exercises: * p < 0.01; all other cases p < 0.001. The
As shown in ● ▶ Table 1, the MPV attained across the entire load-

† symbol denotes the range of loads at which the obtained power output
is not significantly different than Pmax = load at which power output is
maximized; BP = Bench Press; PBP = Prone Bench Pull.
although no statistically significant differences in power output
were found for loads between 20 % and 60 % 1RM for BP and
between 20 % and 70 % for PBP. Finally, when using PP (●
▶ Fig. 5c),
ing range is significantly higher for PBP than BP, with differences
between exercises for a given % 1RM becoming larger as the load
approaches the 1RM. Of particular interest is the very different
V1RM (0.17 m · s − 1 for BP vs. 0.52 m · s − 1 for PBP). Both the magni-
tude of the velocity values and the shape of the curves are
remarkably different between exercises, as can be appreciated in
the respective velocity-time curves for a representative subject
in ●▶ Fig. 3. This fact can further be used to establish a practical

power was maximized at 37 ± 10 % 1RM for BP and 41 ± 14 % 1RM
for PBP, although no statistically significant differences were
found for loads between 20 % and 65 % 1RM for BP and between
20 % and 75 % for PBP.
Pmax ( % 1RM) calculated attending to MPP or PP outcome varia-
bles was significantly lower to that obtained using MP for both
and objective criteria in terms of mean velocity for a repetition
to be considered a true 1RM in each training exercise.
The greater velocities and power output values observed in the
PBP compared to the BP may be attributed to: i) the differing
skeletal muscle architecture; and ii) the distinct mechanical
levers involved in the opposing actions of pushing and pulling. It

Sánchez-Medina L et al. Velocity- and Power-Load Relationships … Int J Sports Med 2014; 35: 209–216

is well known that a muscle’s architecture determines its func-
tion [23]. The primary movers in the PBP (i. e., latissimus dorsi,
biceps brachii, brachialis) have greater fiber lengths and a longi-
tudinal fiber arrangement, making them more suited to develop
faster shortening velocities, whereas the musculature involved
in the BP (i. e., pectoralis major, triceps brachii, anterior deltoid)
is characterized by shorter fiber lengths, greater pennation
angles and therefore greater force generation capability
[20, 23, 25]. The muscle moment arms, as a result of mechanical
advantages or disadvantages at the specific joint angles, also
seem to differ considerably between the PBP and BP exercises.
For instance, when performing the BP exercise against maximal
and near-maximal loads a sticking region occurs during the con-
centric phase of the movement, a phenomenon which has been
extensively studied [8, 35, 36]. This sticking region is thought to
coincide with a poor mechanical force position where the
lengths and moment arms of the muscles involved are such that
their capacity to exert force is reduced [8, 36]. The sticking
region can be observed in the velocity-time curve for the BP in
the example of ● ▶ Fig. 3 as a region of very low velocity (approx-
imately from 40 to 60 % of concentric duration). However, such a
critical region does not seem to exist in the PBP and, interest-
ingly, the velocity profiles of both exercises (● ▶ Fig. 3) are just the
opposite at this particular position (~50 % of the concentric
duration), where the BP attains a minimal instantaneous veloc-
ity (0.11 m · s − 1), the PBP reaches its peak velocity (0.72 m · s − 1).
The mean velocities observed in the present study for the BP
(●
▶ Table 1) are in agreement with previous published research
[9, 14, 19, 32], but differ from the surprisingly low velocities
reported by Pearson et al. [25]. In the current study much higher
mean velocity values for the whole load-velocity curve were
observed compared to the study conducted by Pearson et al.
[25], with differences ranging between 42.9 % and 51.6 % for BP
and between 9.6 % and 29.3 % higher mean velocity for PBP for a
given loading intensity ( % 1RM). In addition to a larger sample
size (n = 75 vs. n = 12) and greater variety of sport athletes (i. e.,
wrestlers, judoists, flat-water kayakers, oarsmen) analyzed in
the present study, we believe that there are 2 main methodo-
logical differences that could explain the noticeable differences
in velocity between studies. First, the concentric BP in the study
of Pearson et al. [25] was initiated from mechanical stops posi-
tioned ~3 cm off the subjects’ chest, whereas in the present
study the barbell directly rested on each athlete’s chest and was
pushed from that starting position, thus allowing a more com-
plete range of motion and greater initial inertia to overcome the
poor mechanical force position typical of this exercise [8, 36].
Secondly, and most importantly, is the type of Smith machine
used in the study of Pearson et al. [25], which incorporated a
counter-balance weight mechanism. It has recently been shown
that the use of a counter-balance mechanism results in signifi-
cantly lower concentric velocities and accelerations and thus an
underestimation of performance capability [3, 37]. Such a sys-
tem favors slow movements and hinders performance of explo-
sive movements [3, 37]. This seems to be the reason why our
observed velocities (● ▶ Table 1) differ to a greater extent for light
loads than for high loads from those reported by Pearson et al. [25].
Referring mean mechanical values to the propulsive phase
rather than to the entire concentric action, when assessing mus-
cle strength and power in isoinertial conditions prevents under-
estimating an individual’s true neuromuscular ability, especially
when lifting light and medium loads [33]. As shown in ● ▶ Fig. 4,
the magnitude of the braking phase depends on the relative load
Sánchez-Medina L et al. Velocity- and Power-Load Relationships … Int J Sports Med 2014; 35: 209–216
Training & Testing 215
to be lifted. The lighter the load, and higher the movement
velocity, the greater the duration of this braking phase. It is
worth noting that in the BP when the load is sufficiently high
( > 80 % 1RM), the braking phase disappears. By contrast, in the
PBP the braking phase does not completely disappear even at
100 % 1RM, where it remains a minimal ~2 % of the total concen-
tric duration. In fact, this final braking phase in the PBP, with
decelerations that can be as large as − 100 m · s − 2 when lifting
very light loads (15–20 % 1RM), is due to the sudden stop the
barbell experiences when it hits the underside of the bench
(●▶ Fig. 1). These differences between the BP and PBP in the con-
tribution of the propulsive and braking phases to the total con-
centric duration may be attributable to the aforementioned
distinct muscle architecture and moment arm levers involved in
the upper-body pulling and pushing muscle actions. All of these
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
findings indicate that subjects with a higher 1RM will lift the
same absolute loads at faster velocities and will thus experience
a longer braking phase that negatively influences the computa-
tion of the mean concentric value since force, and hence power
output, is negative during the braking phase [33].
Unlike movement velocity, which has just been demonstrated to
show a direct and very close relationship to relative loading
intensity ( %1RM), mechanical power output is a compound vari-
able that can be the result of many combinations of force and
velocity. In terms of power output, it seems evident that the
greater velocity-generating capability of the musculature used in
the PBP compensates the ~11 % lower 1RM strength values meas-
ured in this exercise (80.2 kg) compared to the BP (90.3 kg), espe-
cially as relative load increases [25].
Even though the results of this study strongly support previous
research showing that intermediate loads maximize the mean
mechanical power output in the main RT exercises [1, 5, 6, 16, 33],
it also becomes apparent that the different outcome measures
that can be chosen to analyze power output greatly influence the
determination of the Pmax load ( % 1RM). In the present study, we
compared the effect of selecting mean concentric (MP), mean
propulsive (MPP) or peak power (PP) variables on Pmax determi-
nation. When considering MPP or PP values (● ▶ Fig. 5b, c), the
Pmax load was found at significantly lower loads (p < 0.001) than
when MP was used (● ▶ Fig. 5a) in both exercises. Interestingly,
the Pmax load, although different between BP and PBP, was
almost identical attending to the MPP and PP variables for a
given exercise (~37 % 1RM in the BP for both MPP and PP; 46 % for
MPP and 41 % for PP in the PBP), with the load-power curves
showing a very similar shape (● ▶ Fig. 5b, c). Furthermore, the
range of loads that maximized power output attending to the
MPP and PP variables was even wider than that found for MP
(i. e., no statistically significant differences were found for loads
between 20 % and ~65 % 1RM for either the BP or PBP exercises)
(●▶ Fig. 5). When MP was taken into consideration, the P
max load
was higher (56 % for BP and 70 % for PBP) and there also existed a
large range of loads at which power output was not statistically
significantly different. The fact that there exists such a broad
range of relative loads yielding very similar power output values
makes us wonder whether the issue of identifying an ‘optimal’,
single, Pmax load may have perhaps been overemphasized in pre-
vious research.
Since it is recognized that mechanical stimuli play a key role in
mediating the adaptations to strength training [4], improving
our understanding of the kinematics and kinetics of the most
widely used RT exercises may contribute to optimizing isoiner-
tial testing and training methods. The present study has shown
216 Training & Testing
that there are some mechanical characteristics distinctive of
each exercise that should be taken into account when perform-
ing isoinertial strength and power assessments in those sport
disciplines that typically use upper-body pushing and pulling
actions. Thus, our main findings were that:
i) The load-velocity and load-power relationships are remarka-
bly different between the BP and PBP, probably due to the dis-
tinct muscle architecture and mechanical moment arm levers
involved in each exercise, with significantly higher velocities
and power outputs for upper-body pulling (PBP) compared to
pushing (BP) for a given loading magnitude ( % 1RM). In par-
ticular, the velocity attained against the 1RM load (V1RM) is
distinctive of each exercise and can provide an objective cri-
terion for a repetition to be considered a true maximum.
ii) The mean velocity attained with a given absolute load can be
used as a very good estimate of the magnitude of the load ( %
1RM) in both exercises. Prediction equations are provided for
this purpose. This finding has important practical applica-
tions for RT prescription and training load monitoring and
opens up a new paradigm (velocity-based resistance training)
in which actual movement velocity is used as a marker to
gauge exercise intensity. Measuring repetition velocity dur-
ing daily training (e. g. by means of linear position or velocity
transducers or accelerometer-based units, which are increas-
ingly used in high-level sport) will enable us to accurately
identify and quantify the actual training stimulus undertaken
and therefore to be in a much better position to establish
causal connections between stimulus and adaptation.
iii) There are important differences in the power output profiles
and Pmax load which depend not only on the exercise (BP vs.
PBP) but also on the chosen outcome measure (mean, peak or
mean propulsive power). For each of these distinct measures,
however, there is a broad range of relative loads that maxi-
mize power output, which seems to minimize the importance
of determining an ‘optimal’ Pmax load for a given exercise.
References
1 Baker D, Nance S, Moore M. The load that maximizes the average
mechanical power output during explosive bench press throws in
highly trained athletes. J Strength Cond Res 2001; 15: 20–24
2 Baker D, Newton RU. Acute effect on power output of alternating an
agonist and antagonist muscle exercise during complex training. J
Strength Cond Res 2005; 19: 202–205
3 Buddhadev HH, Vingren JL, Duplanty AA, Hill DW. Mechanisms underly-
ing the reduced performance measures from using equipment with a
counterbalance weight system. J Strength Cond Res 2012; 26: 641–647
4 Crewther B, Cronin J, Keogh J. Possible stimuli for strength and power
adaptation. Acute mechanical responses. Sports Med 2005; 35: 967–989
5 Cronin JB, McNair PJ, Marshall RN. Force-velocity analysis of strength-
training techniques and load: implications for training strategy and
research. J Strength Cond Res 2003; 17: 148–155
6 Cronin J, Sleivert G. Challenges in understanding the influence of maxi-
mal power training on improving athletic performance. Sports Med
2005; 35: 213–234
7 Drid P, Maksimovic N, Matic R, Obradovic B, Milosevic Z, Ostojic SM.
Fitness profiles of elite female judokas of the Serbian national team.
Med Sport 2009; 62: 251–263
8 Elliott BC, Wilson GJ, Kerr GK. A biomechanical analysis of the sticking
region in the bench press. Med Sci Sports Exerc 1989; 21: 450–462
9 Frost DM, Cronin JB, Newton RU. A comparison of the kinematics,
kinetics and muscle activity between pneumatic and free weight
resistance. Eur J Appl Physiol 2008; 104: 937–956
10 García-Pallarés J, Izquierdo M. Strategies to optimize concurrent train-
ing of strength and aerobic fitness for rowing and canoeing. Sports
Med 2011; 41: 329–343
11 García-Pallarés J, Sánchez-Medina L, Carrasco L, Díaz A, Izquierdo M.
Endurance and neuromuscular changes in world-class level kayakers
during a periodized training cycle. Eur J Appl Physiol 2009; 106: 629–638
Sánchez-Medina L et al. Velocity- and Power-Load Relationships … Int J Sports Med 2014; 35: 209–216
12 García-Pallarés J, Sánchez-Medina L, Pérez CE, Izquierdo-Gabarren M,
Izquierdo M. Physiological effects of tapering and detraining in world-
class kayakers. Med Sci Sports Exerc 2010; 42: 1209–1214
13 González-Badillo JJ, Marques MC, Sánchez-Medina L. The importance of
movement velocity as a measure to control resistance training inten-
sity. J Hum Kinet 2011; Special Issue: 15–19
14 González-Badillo JJ, Sánchez-Medina L. Movement velocity as a meas-
ure of loading intensity in resistance training. Int J Sports Med 2010;
31: 347–352
15 Harriss DJ, Atkinson G. Update – Ethical standards in sport and exercise
science research. Int J Sports Med 2011; 32: 819–821
16 Izquierdo M, Häkkinen K, González-Badillo JJ, Ibáñez J, Gorostiaga EM.
Effects of long-term training specificity on maximal strength and
power of the upper and lower extremities in athletes from different
sports. Eur J Appl Physiol 2002; 87: 264–271
17 Izquierdo-Gabarren M, González De Txabarri Expósito R, De Villarreal
ES, Izquierdo M. Physiological factors to predict on traditional rowing
performance. Eur J Appl Physiol 2010; 108: 83–92
18 Izquierdo-Gabarren M, González De Txabarri Expósito R, García-Pal-
larés J, Sánchez-Medina L, De Villarreal ES, Izquierdo M. Concurrent
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
endurance and strength training not to failure optimizes performance
gains. Med Sci Sports Exerc 2010; 42: 1191–1199
19 Jidovtseff B, Croisier JL, Scimar N, Demoulin C, Maquet D, Crielaard
JM. The ability of isoinertial assessment to monitor specific training
effects. J Sports Med Phys Fitness 2008; 48: 55–64
20 Koryak YA. Functional and clinical significance of the architecture of
human skeletal muscles. Human Physiol 2008; 34: 482–492
21 Lawton TW, Cronin JB, McGuigan MR. Strength testing and training of
rowers. A review. Sports Med 2011; 41: 413–432
22 Lawton TW, Cronin JB, McGuigan MR. Strength, power and muscu-
lar endurance exercise and elite rowing ergometer performance. J
Strength Cond Res in press doi:10.1519/JSC.0b013e3182772f27
23 Lieber RL, Fridén J. Functional and clinical significance of skeletal mus-
cle architecture. Muscle Nerve 2000; 23: 1647–1666
24 Mora-Rodríguez R, García-Pallarés J, López-Samanes A, Ortega JF,
Fernández-Elías VE. Caffeine ingestion reverses the circadian rhythm
effects on neuromuscular performance in highly resistance-trained
men. PLoS One 2012; 7: e33807
25 Pearson SN, Cronin JB, Hume PA, Slyfield D. Kinematics and kinetics of
the bench-press and bench-pull exercises in a strength-trained sport-
ing population. Sports Biomech 2009; 8: 245–254
26 Pearson SN, Cronin JB, Hume PA, Slyfield D. Effects of a power-focussed
resistance training intervention on backward grinding performance in
America’s Cup sailing. Sports Biomech 2009; 8: 334–344
27 Pearson SN, Hume PA, Cronin JB, Slyfield D. Strength and power deter-
minants of grinding performance in America’s Cup sailors. J Strength
Cond Res 2009; 23: 1883–1889
28 Pereira MI, Gomes PS. Movement velocity in resistance training. Sports
Med 2003; 33: 427–438
29 Platzer HP, Raschner C, Patterson C. Relationship between physical fit-
ness and freestyle swimming performance. Dtsch Z Sportmed 2009;
60: 321–326
30 Robbins DW, Young WB, Behm DG. The effect of an upper-body agonist-
antagonist resistance training protocol on volume load and efficiency.
J Strength Cond Res 2010; 24: 2632–2640
31 Robbins DW, Young WB, Behm DG, Payne WR. Effects of agonist-antag-
onist complex resistance training on upper body strength and power
development. J Sports Sci 2009; 27: 1617–1625
32 Sánchez-Medina L, González-Badillo JJ. Velocity loss as an indicator
of neuromuscular fatigue during resistance training. Med Sci Sports
Exerc 2011; 43: 1725–1734
33 Sánchez-Medina L, Pérez CE, González-Badillo JJ. Importance of the
propulsive phase in strength assessment. Int J Sports Med 2010; 31:
123–129
34 Stöggl T, Müller E, Ainegren M, Holmberg HC. General strength and
kinetics: fundamental to sprinting faster in cross country skiing?
Scand J Med Sci Sports 2011; 21: 791–803
35 Van den Tillaar R, Ettema G. The sticking period in a maximum bench
press. J Sports Sci 2010; 28: 529–535
36 Van den Tillaar R, Saeterbakken AH, Ettema G. Is the occurrence of the
sticking region the result of diminishing potentiation in bench press?
J Sports Sci 2012; 30: 591–599
37 Vingren JL, Buddhadev HH, Hill DW. Smith machine counterbalance
system affects measures of maximal bench press throw performance.
J Strength Cond Res 2011; 25: 1951–1956
38 Wilson GJ, Newton RU, Murphy AJ, Humphries BJ. The optimal training
load for the development of dynamic athletic performance. Med Sci
Sports Exerc 1993; 25: 1279–1286
Copyright of International Journal of Sports Medicine is the property of Georg Thieme Verlag
Stuttgart and its content may not be copied or emailed to multiple sites or posted to a listserv
without the copyright holder's express written permission. However, users may print,
download, or email articles for individual use.