Journal of Sports Sciences, December 2009; 27(14): 1617–1625

Effects of agonist–antagonist complex resistance training on upper

body strength and power development

DANIEL W. ROBBINS1, WARREN B. YOUNG1, DAVID G. BEHM2, &

WARREN R. PAYNE1
1
School of Human Movement and Sport Sciences, University of Ballarat, Ballarat, Victoria, Australia and 2School of Human
Kinetics and Recreation, Memorial University of Newfoundland, St. John’s, Newfoundland, Canada

(Accepted 25 September 2009)

Abstract
The objective of this study was to examine the chronic effects on strength and power of performing complex versus
traditional set training over eight weeks. Fifteen trained males were assessed for throw height, peak velocity, and peak power
in the bench press throw and one-repetition maximum (1-RM) in the bench press and bench pull exercises, before and after
the eight-week programme. The traditional set group performed the pulling before the pushing exercise sets, whereas the
complex set group alternated pulling and pushing sets. The complex set training sessions were completed in approximately
half the time. Electromyographic (EMG) activity was monitored during both test sessions in an attempt to determine if it was
affected as a result of the training programme. Although there were no differences in the dependent variables between the
two conditions, bench pull and bench press 1-RM increased significantly under the complex set condition and peak power
increased significantly under the traditional set condition. Effect size statistics suggested that the complex set was more time-
efficient than the traditional set condition with respect to development of 1-RM bench pull and bench press, peak velocity
and peak power. The EMG activity was not affected. Complex set training would appear to be an effective method of
exercise with respect to efficiency and strength development.

Keywords: Complex set, bench press throw, bench pull, bench press, complex training

Some research suggests that antagonist pre-

Introduction
Resistance training is an effective method for
developing muscular strength and power (Hakkinen
& Komi, 1981). Complex training is a resistance-
training modality that commonly involves the cou-
pling of biomechanically similar exercises, performed
in an alternating manner. A less common complex
training design pairs biomechanically dissimilar
exercises, targeting agonist and antagonist muscle
groups. Among a number of possible agonist–
antagonist combinations are pairs of heavy tradi-
tional weight-training exercises (e.g. bench pull and
bench press) and pairs of heavy traditional weight-
training and ballistic exercises (e.g. bench pull
and bench press throws). For the purposes of this
research, a complex set refers to agonist–antagonist
pairs of either two heavy resistance exercises or a
heavy resistance and ballistic exercise performed over
repeated trials, in an alternating manner, with rest
intervals between sets.
loading may result in acute performance enhance-
ment of the agonist musculature. Baker and Newton
(2005) reported that power output in the bench
press throw was significantly greater when preceded
by a set of ballistic bench pulls than in a set of bench
press throws with no intervention. These researchers
suggested that the bench pull exercise altered (i.e.
shortened) the braking phase of the triphasic pattern
of the antagonist musculature, thereby allowing for
longer total agonist burst, during the bench press
throws. A mechanistic component (i.e. electromyo-
graphy, EMG) was not incorporated into the
research to support their conclusions.
Research reporting performance enhancement as
a result of antagonist loading is contradicted by
research that has failed to observe augmentation in a
performance measure preceded by antagonist load-
ing. Maynard and Ebben (2003) found a decrease in
isokinetic knee flexion and extension performance
measures (peak torque, rate to peak torque, and peak

Correspondence: D. W. Robbins, School of Human and Sport Sciences, University of Ballarat, PO Box 663, Ballarat, VIC 3353, Australia.
E-mail: drobbins@uvic.ca
ISSN 0264-0414 print/ISSN 1466-447X online Ó 2009 Taylor & Francis
DOI: 10.1080/02640410903365677
1618 D. W. Robbins et al.
power) after pre-loading of the antagonist muscu-
lature. They suggested that the observed increases in
EMG activity (co-contraction) of the antagonist
musculature may have been responsible for the
attenuation in performance measures. It is unclear
if there is a differential response in the upper body
compared with the lower body. It may be that the
level of co-activation is greater in the knee flexors and
extensors than in the chest and back muscle groups.
Greater co-activation in the antagonist musculature
may manifest itself as fatigue and affect that muscle
group adversely when acting as an agonist.
Unlike the two studies discussed above, which
were primarily interested in the augmentation of
the agonist musculature when preceded by anta-
gonist loading, a study combining two heavy
resistance training exercises examined complex
sets in the context of efficiency (Robbins, Young,
Behm, Payne & Klimstra, in press). Specifically,
these authors examined the effects on volume load in
the bench pull and bench press exercises, over three
sets of complex versus a traditional set training
protocol. They observed that the maintenance of
volume load was similar with the complex set and
traditional set training but the former was achieved
in approximately half the time. Complex sets were
determined to be approximately twice as efficient
(output/input, where input is time) as traditional
sets. Electromyographic data were not different
under the two conditions, indicating neuromuscular
fatigue was no greater with complex than traditional
set training.
In the absence of performance enhancement, or
even performance attenuation (depending on the
degree of attenuation), it would appear that
complex set training may be deemed to be time-
efficient. Resistance training schemes that do not
compromise efficacy, or increase efficiency, could
be advantageous not to only athletes, but also to
the general population in terms of improved health
and a decrease in the risk of chronic disease and
disability (Warburton, Nicol, & Bredin, 2006).
Peer-reviewed research into the acute effects of
complex set training is limited and there has been
no research into the chronic effects. There is some
evidence that complex sets may be an efficient
training scheme. Therefore, the purpose of this study
was to assess the efficacy and efficiency of agonist–
antagonist complex sets over the course of an 8-week
training period.
Methods
Design
A randomized, counterbalanced two-group (tra-
ditional and complex) pre- and post-test design
was used to investigate the effects on strength
(one-repetition maximum bench pull and bench
press) and power (throw height, peak velocity,
and peak power) of eight weeks of complex set
versus traditional set training. Under the tradi-
tional set condition, the pulling exercise sets were
completed before performance of the pushing
exercise sets, whereas in the complex set condi-
tion, the pulling and pushing exercise sets were
alternated.
Participants
Fifteen trained males with at least one year’s strength
training experience with pushing and pulling ex-
ercises volunteered to participate in the study. All
participants had experience with complex set-type
training. Participants were randomly assigned to
either the complex set training group (n ¼ 8) or the
traditional set training group (n ¼ 7). The partici-
pants’ descriptive data are displayed in Table I. The
study was approved by the University Human
Research Ethics Committee. All participants were
briefed on the test protocols, the equipment, and the
nature of the study before signing an informed
consent form.
Methodology and procedures
Depending on the training session, loads ranging
between 3- and 6-repetition maximum (RM) were
prescribed for sets of bench pull and bench press
in both protocols and were performed to failure,
which was considered to have been reached when
another repetition using proper technique could
not be performed (Wathen, 1994). High-intensity
loads (e.g. 3- to 6-RM) have been recommended
with respect to strength development (Berger,
1962; Weiss, Coney, & Clark, 1999). Depending
on the training session, 1–4 sets of 3–6 throws at
40% of bench press 1-RM were prescribed for
bench press throw in both protocols. It has been
suggested that over the course of a training, cycle
lighter loads (e.g. 40% of 1-RM) would likely
lead to greater enhancement of power than
heavier loads (i.e. 460%) (Cronin & Crewther,
2004). It has been recommended that when using
Table I. Characteristics of the participants (mean + s).
Characteristic Complex Traditional
Age (years) 25.9 + 5.1 22.4 + 4.9
Height (m) 1.78 + 0.05 1.92 + 0.06
Mass (kg) 84.8 + 20.4 89.9 + 8.0
Strength training 3.6 + 1.7 4.3 + 2.4
experience (years)
 Chronic agonist–antagonist complex set 1619

4 min
4 min
4 min
Rest*
loads designed to achieve maximum power out-
put, fewer repetitions (e.g. 3–6) should be used
(Baker, Nance, & Moore, 2001). In both proto-

40% 1-RM
cols, a 4-min rest interval was instituted between

3-RM
3-RM
Load
like exercise sets. Both the complex set and

Week 4
traditional set programme used a combination of
bench press/bench press throw and bench pull.
That is, bench pull was always alternated with

Reps

3
3
3
either bench press (strength emphasis) or bench
press throw (power emphasis). Participants in

Sets
both the complex set and traditional set groups

6
4
2
were required to perform two training sessions
per week, separated by a minimum of 48 h. The

4 min
4 min
4 min
Rest*
Table II. Phase 1 (weeks 1–4) of the training programme with the emphasis on back and chest strength.
total time required to complete the training
sessions, and the order in which the exercises
were performed, differed between the two proto-

40% 1-RM
cols. The traditional set protocol involved per-

4-RM
4-RM
Load
forming sets of bench pull followed by sets of

Week 3
bench press/bench press throw, with a 4-min rest
interval between all sets (Figure 1). The complex
set protocol performed the same exercises but in

Reps

4
4
4
an alternating manner (Figure 1). The rest
interval between like exercise sets was similar to

Sets
that used in the traditional set protocol (4-min)

6
4
2
and the rest interval between unlike exercise sets
was 2 min. The second exercise (bench press/

4 min
4 min
4 min
Rest*
bench press throw) was performed in such a
manner that the mid-point of the execution of the
second exercise set was 2 min after the beginning

40% 1-RM
of the execution of the first exercise set. Rest

5-RM
5-RM
Load
Week 2

intervals of 2–5 min between resistance training

sets have been recommended when training for
strength and power (American College of Sports
Reps

Medicine, 2002; Baechle, Earle, & Wathen,

5
5
5

2000). The complex set training sessions took

approximately half the time required to complete
Sets

the traditional set training sessions.

5
4
1

To assist in the explanation of any observed

differences in 1-RM (bench pull and bench
4 min
4 min
4 min
Rest*

press), throw height, peak velocity, and peak

power, the EMG responses of four muscles
(pectoralis major, anterior deltoid, latissimus
40% 1-RM

*Rest between like sets. RM ¼ repetition maximum.

dorsi, and trapezius) were monitored during

6-RM
6-RM
Load
Week 1

pre- and post-training programme test sessions.

Specifically, mean amplitude of the root mean
square and the median frequency were recorded.
Reps

Participants were requested to refrain from any

6
6
6

additional upper body resistance training during

the 8-week period.
Sets

4
3
1
Bench press throw
Bench press
Bench pull
Exercise

Figure 1. Time lines for traditional (TS) and complex (CS)

training sessions. A ¼ pulling exercise and B ¼ pushing exercise.
1620 D. W. Robbins et al.

4 min
4 min
4 min
Rest*
Training programmes
A combination of bench press and bench press throw

40% 1-RM
was paired with bench pull such that there was a

3-RM
3-RM
Load
strength phase of 4 weeks in which the emphasis was

Week 4
on bench press (Table II), followed by a power phase
in which the emphasis was on bench press throw
(Table III). Bench pull was used to exercise the back

Reps

3
3
3
musculature over the eight weeks and, as such,
strength was the primary focus. To enhance com-

Sets
pliance, lower body programmes for days on which

6
2
4
the upper body was not trained were provided, but

Table III. Phase 2 (weeks 5–8) of the training programme with the emphasis on back strength and chest power.
were not compulsory. All participants indicated that

4 min
4 min
4 min
Rest*
they completed the lower body programmes. How-
ever, the lower body sessions were not supervised.
The 1-RMs determined in the pre-test sessions were

40% 1-RM
used to calculate the prescribed repetition maximum

4-RM
4-RM
Load
loads during the 8-week training programme. All

Week 3
upper body training sessions were supervised
throughout the 8-week training period, and volume
(repetitions) and intensity (load) were recorded for

Reps

4
4
4
each training set and session. The loads prescribed
were adjusted to ensure progressive overload of the

Sets
targeted muscle groups. That is, the loads were re-

6
2
4
evaluated at the end of each week and adjusted
accordingly in an attempt to ensure true repetition

4 min
4 min
4 min
Rest*
maximum loading, as prescribed.

40% 1-RM
Test procedures

5-RM
5-RM
Load
Before the start of the pre-programme test sessions,
Week 2

a reliability study involving 10 individuals who

later participated in the present study was used to
Reps

determine the test–retest (separated by one week)

5
5
5

intra-class correlation coefficient and percent total

error.
Sets

Participants underwent a familiarization session

5
1
4

in which they were instructed on exercise techni-

que. Testing was performed on two separate days,
4 min
4 min
4 min
Rest*

separated by a minimum of 48 h, before and after

the training programme. One-repetition maximum
bench press was determined on the first day, and
40% 1-RM

*Rest between like sets. RM ¼ repetition maximum.

throw height, peak velocity, peak power, and 1-RM

6-RM
6-RM
Load
Week 1

bench pull were determined on the second day.

Testing of throw height, peak velocity, and peak
power was conducted prior to 1-RM bench pull
Reps

testing. Bench press throw was considered to be a

6
6
6

non-fatiguing exercise and, as such, to have no

implications for the subsequent 1-RM bench pull
Sets

test. To assess strength, bench pull and bench

4
1
3

press 1-RM were determined using the following

procedure. Participants performed a set of 5–10
Bench press throw

repetitions using 40–60% of expected maximum,

followed 1 min later by a set of 3–5 repetitions
Bench press
Bench pull

using 60–80% of expected maximum. After a 2-

Exercise

min rest interval, 1-RM attempts were made with

2-min rest intervals between attempts. If an

attempt was successful using correct technique,
further attempts were made using increasing
increments of weight. The last successful attempt
was recorded as the participant’s 1-RM in that lift.
This procedure was adopted from Stone and
O’Bryant (1987) with one change: rather than 1-
min rest intervals between attempts, 2-min rest
intervals were used to ensure recovery between
attempts. To assess power, throw height, peak
velocity, and peak power were calculated over the
concentric portion for each of four throws in a set
of bench press throw. The totals of the four peak
values for each of throw height, peak velocity, and
peak power were calculated. Participants used 40%
of the predetermined (i.e. familiarization session)
bench press 1-RM as the load in both pre- and
post-training programme test sessions. As a load of
40% of 1-RM was used to develop power over the
eight weeks, a similar load was used for testing.
Before bench press throw testing, three sets of
bench press throw using 40% of 1-RM at a self-
determined 50, 75, and 100% of maximal effort
were completed with a 4-min rest interval between
sets.
The bench pull tests were performed on an
adjustable high bench (Apex B45 adjustable flat
bench), positioned on Step1005 platforms. Partici-
pants were instructed to lie prone on the bench and
grasp an Olympic bar placed on the floor, with a
pronated grip. The bench was adjusted so that the
participant’s arms were straight in this position. A
repetition was deemed to have been completed by
moving the bar from the floor until it touched the
bottom of the bench. Participants were instructed to
keep their head, upper body, and legs flat to the
bench. When performing the bench press, partici-
pants lay supine on a flat bench with feet flat on floor
and head, shoulders, and buttocks flat to the bench.
A repetition was deemed to have been completed
when the bar was moved from the chest to a position
of full elbow extension. When performing bench
press throw, participants lay supine on a flat bench,
in a Smith Machine, which allowed the bar to move
only in the vertical plane. The participants’ feet were
flat on the floor, and head, shoulders, and buttocks
flat to the bench. The starting position of the bar was
touching the chest at the nipples. Participants were
instructed to attempt to throw the bar in the vertical
plane as high as possible, releasing the bar at elbow
extension. Hand placement and tempo were self-
determined for all exercises. An attempt was made to
hold testing at the same time, on the same day of the
week, before and after the 8-week programme. All
participants were asked to refrain from any upper
body training in the 48 h prior to the test sessions.
Post-programme testing was completed 5–10 days
after the final training session.
Chronic agonist–antagonist complex set 1621
Instrumentation
Ballistic measurement system. Throw height (cm), peak
velocity (m  s71), and peak power (W) were
calculated using a position transducer (PT5A linear
position transducer, Fitness Technology, Adelaide,
Australia). The system comprises a cable-extension
potentiometer (distance transducer), USB data
collection interface, and custom software (Ballistic
Measurement System, Fitness Technology, Australia)
to accurately measure vertical movement. The
Ballistic Measurement System was secured to the
barbell in the Smith machine. The barbell is limited
to movement in the vertical plane. The Ballistic
Measurement System software calculated throw
height, peak velocity, and peak power of each bench
press throw.
Electromyography. The EMG data were collected
using surface electrodes (Delsys DE-2.1 sensors),
with an inter-electrode distance of 1 cm using an
active differential preamplifier configuration (Delsys
DE 2.1, Boston, MA). These electrodes were con-
nected to an analog-to-digital converter (Bagnoli
Myomonitor III wireless system, Delsys Inc.,
Boston, MA) and acquired with the assistance of
proprietary software (EMGworks Acquisition 3.5,
Delsys Inc., Boston, MA). The EMG signals were
amplified by 1000 with a frequency band-pass of
20–450 Hz (common mode rejection ratio of 92 dB)
and recorded at 1000 Hz (Bagnoli Myomonitor III
wireless system, Delsys Inc., Boston, MA). The
mean amplitude of the root mean square and the
median frequency analysis was performed using
custom software (National Instruments LabVIEW
8). The averaging window for the root mean square
was 100 ms and all reported values were the mean
root mean square over a predetermined sampling
window from the onset to the offset of each con-
traction. Median frequency was calculated by finding
the frequency that halved the integrated power
spectrum of the EMG signal over a predetermined
sampling window from the onset to the offset of each
contraction. Data were collected throughout the
entire repetition (bench pull and bench press) or
set (bench press throw). Bench pull and bench press
1-RM EMG data were converted to ratios, with the
latissimus dorsi signal being the denominator for the
bench pull and the numerator for the bench press
ratios. These ratios were compared before and after
the 8-week programme. Bench press throw EMG
data collected from the first contraction were
compared with EMG data from the final contraction,
and compared pre- and post-programme.
The EMG signal was acquired from the pectoralis
major, anterior deltoid, latissimus dorsi, and trape-
zius muscles located on the right side of each
1622 D. W. Robbins et al.
participant, using surface electrodes with an inter-
electrode distance of 1 cm. The pectoralis major
electrode was placed at the mid-point between the
acromion process and the xiphoid process. The
anterior deltoid electrode was placed on the mid-
belly, 3–4 cm beneath the anterior margin of the
acromion process. The latissimus dorsi electrode was
placed laterally to the inferior angle of the scapula.
The trapezius electrode was placed midway between
the scapula spine and spinous process at the same
level. A ground electrode (flexible 1-cm disposable
Ag-AgCl surface EMG electrodes, Thought Tech-
nologies Ltd., Montreal, Canada) was placed on the
right elbow. Before electrode placement, the area of
skin was thoroughly prepared with abrasive paper
and isopropyl alcohol swabs to improve conductivity
of the EMG signal.
Statistical analyses
One-repetition maximum bench pull and bench
press and set totals for throw height, peak velocity,
and peak power were calculated before and after
the 8-week programme and analysed using a two-
way analysis of variance (ANOVA) (2 6 2), with
repeated-measures and paired t-tests to determine
whether there were significant main effects or
interactions for the type of training (traditional and
complex set) and time (pre and post). A two-way
ANOVA (2 6 5) was used to determine whether
there were significant main effects or interactions for
the type of training (traditional and complex set) and
relative change in performance measure (bench pull,
bench press, throw height, peak velocity, and peak
power). Analysis of the data to determine if any
significant differences existed within or between the
two training protocols was performed to investigate
the influence of complex sets on the development
of strength and power. Due to the relatively small
sample sizes, effect size calculations were conducted
and Cohen (1988) effect size thresholds were imple-
mented. Specifically, effect size thresholds of 0.2–
0.5, 0.5–0.8, and greater than 0.8 were considered to
be small, medium, and large, respectively. Effect
sizes of less than 0.2 were considered insubstantial.
Efficiency (effect/time) calculations were also con-
ducted and subjected to effect size calculations. The
EMG data (root mean square and median fre-
quency) were gathered during 1-RM bench pull
and bench press testing and for the first and fourth
repetition of bench press throw testing, before and
after the programme. Bench pull and bench press
EMG data were analysed using a two-way ANOVA
(2 6 2) (groups; pre/post), whereas bench press
throw EMG data were analysed using a three-way
ANOVA (2 6 2 6 3) (groups; pre/post; rep. 1 and
rep. 4), to determine whether there were significant
main or interaction effects among the factors. Statis-
tical significance was adjusted using the Bonferroni
technique for all tests and set at P  0.01 and
P  0.003 for the performance measures and EMG
activity, respectively. Statistical tests were completed
using SPSS version 15.
Results
The reliability study determined the intra-class
correlation coefficient (and percent total error) for
1-RM bench pull, 1-RM bench press, throw height,
peak velocity, and peak power as 0.94 (3.2%), 0.89
(2.3%), 0.93 (3.7%), 0.99 (1.2%), and 0.99 (1.1%),
respectively. Paired sample t-tests revealed no
significant (P 5 0.001) differences between the two
test occasions for any of the dependent variables.
The test–retest intra-class correlation coefficient of
the EMG measures for the four monitored muscles
ranged from 0.83 to 0.96.
There were no statistically significant differences in
1-RM bench pull and bench press, or bench press
throw height, peak velocity, and peak power between
the two conditions (Table IV). There was a main effect
for time whereby bench pull and bench press 1-RM
increased significantly under the complex set condi-
tion, and peak power increased significantly under
the traditional set condition. Under the traditional
set condition, medium effect sizes were found for all
three (throw height, peak velocity, and peak power)
power measures. Medium to large effect size statistics
suggested the complex set was more time-efficient
than traditional set training, with respect to the
development of 1-RM bench pull and bench press,
peak velocity and peak power. Efficiency calculations
and effect sizes are shown in Table V. There were no
EMG activity main effects or interactions.
Discussion
Complex training involving various combinations
of heavy resistance and ballistic exercises targeting
agonist/antagonist muscle groups has been pre-
scribed as a means of developing strength and
power. Evidence as to the effectiveness of agonist–
antagonist complex training as a means of developing
strength and power has not been identified. In the
present study, changes in 1-RM bench pull and
bench press, throw height, peak velocity, and peak
power were not significantly different between the
complex and traditional set conditions. However, the
strength measures (1-RM bench pull and bench
press) increased significantly under the complex set
condition and peak power increased significantly
under the traditional set condition. Complex set
training appeared to be more time-efficient (training
effect/time) with respect to the development of 1-RM
Table IV. Changes in bench pull and bench press 1-RM, and bench press throw (BPT) height, peak velocity, and peak power over eight weeks of a complex set (n ¼ 8) versus eight weeks of a traditional
set (n ¼ 7) training protocol (mean + s).

Complex set Traditional set

Variable Pre Post Gain %D Effect size Pre Post Gain %D Effect size

Bench pull 1-RM (kg) 92.1 + 14.1 96.7 + 15.9 4.5 + 3.0* 2.2 + 1.1 0.45 (small) 95.9 + 14.1 98.5 + 15.9 2.6 + 3.8 1.2 + 1.7 0.26 (small)
Bench press 1-RM (kg) 100.9 + 27.8 106.0 + 27.6 5.1 + 3.4* 2.4 + 1.8 0.26 (small) 94.6 + 20.5 99.1 + 20.4 4.5 + 3.5 2.3 + 1.9 0.31 (small)
BPT height (cm), 4 throws 97.1 + 18.5 99.9 + 10.3 2.7 + 15.3 5.0 + 16.5 0.21 (small) 86.8 + 17.6 95.4 + 17.5 8.6 + 8.8 10.8 + 10.7 0.70 (medium)
BPT peak velocity (m  s71), 4 throws 6.8 + 0.5 7.1 + 0.4 0.3 + 0.4 4.2 + 6.3 0.73 (medium) 7.0 + 0.5 7.2 + 0.6 0.2 + 0.3 3.0 + 4.0 0.58 (medium)
BPT peak power (W), 4 throws 3002 + 898 3232 + 716 230 + 227 9.7 + 9.2 0.36 (small) 3047 + 552 3321 + 528 274 + 152* 9.4 + 5.4 0.70 (medium)

*Significant difference between pre and post values (P 5 0.01).

Table V. Bench pull and bench press one-repetition maximum (1-RM) and bench press throw (BPT) height, peak velocity, and peak power efficiency calculations (effect/time) for eight weeks of a
complex set (n ¼ 8) versus eight weeks of a traditional set (n ¼ 7) training protocol (mean + s).

Complex Traditional

Variable Absolute training gains Time (h)* Efficiency Absolute training gains Time (h)* Efficiency Effect size

Bench pull 1-RM (kg) 4.5 + 2.97 4.53 1.00 + 0.66 (kg  h71) 2.6 + 3.80 10.13 0.26 + 0.38 (kg  h71) 1.37 (large)
Bench press 1-RM (kg) 5.1 + 3.37 4.53 1.13 + 0.74 (kg  h71) 4.5 + 3.46 10.13 0.45 + 0.34 (kg  h71) 1.18 (large)
BPT height (cm) 2.7 + 15.32 4.53 0.60 + 0.38 (cm  h71) 8.6 + 8.78 10.13 0.85 + 0.87 (cm  h71) 0.37 (small)
BPT peak velocity (m  s71) 0.3 + 0.41 4.53 0.06 + 0.09 (m  s71  h71) 0.2 + 0.30 10.13 0.02 + 0.03 (m  s71  h71) 0.60 (medium)
BPT peak power (W) 230 + 227 4.53 50.7 + 50.1 (W  h71) 274 + 152 10.13 27.1 + 15.0 (W  h71) 0.64 (medium)

*The final set of each training session was begun either 4 min after initiation of the previous set (traditional set) or such that the mid-point of the execution of the final exercise set was 2 min after the
beginning of the execution of the previous exercise set (complex set). Therefore, the total time to complete the sessions varied slightly (e.g. if 12 s were required to complete the final set in week 1 of the
Chronic agonist–antagonist complex set

traditional set protocol, total time to complete the session would be 28.2 min). The time taken to complete the final set was not included in the calculation.
1623
1624 D. W. Robbins et al.
bench pull and bench press, peak velocity and peak
power. These findings support the hypothesis that
complex set training is an efficacious method of
developing strength and is an efficient training
modality. The EMG activity was similar for both
groups and was not affected over the course of the
programme under either condition.
In response to either training protocol, the EMG
signal did not differ between pre- and post-training.
This is consistent with a number of other investiga-
tions examining responses in EMG amplitude to
training (Cannon & Cafarelli, 1987; Garfinkel &
Cafarelli, 1992; McCarthy, Pozniak, & Agre,
2002; Narici et al., 1996; Thorstensson, Karlsson,
Viitasalo, Luhtanen, & Komi, 1976; Weir, Housh, &
Weir, 1994). Acute changes (rep. 1 to rep. 4) in
EMG signal were not observed during bench press
throw testing before or after training, which is
perhaps not surprising, as the test design in the
present study (one set of four throws) was intended
to be non-fatiguing so as to allow participants to
maximize power output in all four throws.
The EMG signal was not monitored during
training sessions under either condition. It is there-
fore difficult to comment on the level of fatigue
resulting from complex versus traditional set training
sessions. However, it is possible that the greater
training density (training/time) under complex
compared with traditional set training may have
been more fatiguing. Although, in an acute setting,
Robbins et al. (in press) observed no greater deficits
in neuromuscular fatigue under a complex than a
traditional set protocol, it is important to note that
these researchers examined complex set training
over three sets only, whereas in the present study
participants performed training sessions involving
four to six sets. It is likely that training sessions of
four to six sets are more fatiguing than sessions of
three sets and this perhaps explains why fatigue
may have been a factor in the current study. If the
complex set sessions were more fatiguing, this would
not necessarily have been reflected in changes in
EMG amplitude pre- to post-programme.
The changes observed in both strength measures
(1-RM bench pull and bench press) were signifi-
cantly greater, pre- to post-programme, under the
complex set condition. The increases in the power
measures were not statistically significant under the
complex set condition. It is possible that complex
set training is better suited to strength than power
training. It has been suggested that fatigue may act
as a stimulus that leads to increases in strength
(Rooney, Herbert, & Balnave, 1994). Rooney et al.
suggested that training protocols that produce fatigue
result in greater motor unit activation than non-
fatiguing protocols, and that the level of motor unit
activation determines the size of the training
response. The same researchers alternatively sug-
gested that fatigue might provide a more appro-
priate setting in which to encourage activation of
synergist and antagonist muscles and thereby
increase the training response. Another possible
explanation provided by these researchers was that
some relationship might exist between events
related to fatigue and events that trigger muscle
adaptation. Although the mechanism(s) is unclear,
the greater training density performed under the
complex set protocol in the present study, as a
result of less total rest throughout the training
sessions, conceivably resulted in greater fatigue and
may have acted as a stimulus. It is possible that
over a longer training cycle, the non-significant
differences in strength outcomes observed under
the complex compared with the traditional set
condition might continue to grow and become
significant.
Due to the nature of power activities, which
require maximal rates of force development, full
neuromuscular recovery has been recommended
(American College of Sports Medicine, 2002). It
has been suggested that longer rest intervals allow for
acute maintenance of power, which may translate
into greater chronic adaptation (Pincivero, Gear,
Moyna, & Robertson, 1999). It is generally accepted
that fatigue is not a stimulus with respect to
power development, which may explain why the
only significant increase in a power measure (i.e.
peak power) was under the arguably less fatiguing
traditional set condition. Furthermore, medium
effect sizes were observed in all three power mea-
sures, compared with small effect sizes in both
strength measures, under the traditional set condi-
tion. It is possible that traditional set training is better
suited to power than strength training.
Although the finding in the present study that
antagonist preloading (complex sets) over eight
weeks did not have a positive effect on bench press
throw performance would seem to conflict with
the observation by Baker and Newton (2005) that
antagonist preloading resulted in a potentiation of
power output in bench press throw, this is perhaps
explained by the nature of preloading stimulus. It is
possible that the very different antagonist preloading
(ballistic bench pulls) incorporated by Baker and
Newton (2005) was not only non-fatiguing but was
performed in such a manner (i.e. explosively) as to
have some physiological effect (i.e. alteration of the
triphasic pattern) resulting in agonist power output
potentiation. The findings of the current study,
and the suggestion that traditional set training may
be better suited to power than strength training,
is limited to traditional set-type modalities involv-
ing antagonist preloading with heavy resistance
exercises.

Although not statistically significant, increases
were observed in all dependent variables. It is
possible that the relatively low prescribed training
volume (i.e. 18–25 repetitions per muscle group, per
session) and frequency (two sessions per week) did
not provide a great enough stimulus over the 8-week
period to produce significant results in all measures,
under both conditions, in the relatively highly trained
participants (i.e. a minimum of one and generally
several years’ experience). It is possible that longer
(i.e. more repetitions) or more frequent (i.e. more
than two times per week) training sessions over the
course of the eight weeks would result in those gains
that were not statistically significant becoming
statistically significant. It is also likely that the
relatively small sample sizes may have hindered the
attainment of statistical significance.
With the exception of throw height, medium to large
effect sizes suggest complex set training was more time-
efficient with respect to the development of the
performance measures. Training modalities able to
save time without compromising efficacy are beneficial
to athletes and the general population. Athletes may be
able to spend more time on technical aspects of their
sport and thereby better prepare for competition.
Reduction in time commitments may entice greater
numbers of the general population to exercise and
realize health benefits. It is possible that complex set
training could help produce a healthier population.
However, before prescribing such modalities to the
general population, other physiological responses (e.g.
blood pressure) to this type of training should be
investigated.
Conclusions
Although we found similar changes in all performance
measures under both the complex and traditional set
conditions, the findings of the present study seem to
suggest that complex set training may be more
efficacious with respect to strength than power devel-
opment, whereas traditional set training may be more
efficacious with respect to power than strength devel-
opment. With the exception of throw height, complex
set training was more time-efficient with respect to the
development of the performance measures. It would
appear that complex set training is an efficacious means
of developing strength and an efficient method of
training both strength and power.
References
American College of Sports Medicine (2002). Position stand:
Progression models in resistance training for healthy adults.
Medicine and Science in Sports and Exercise, 34, 364–380.
Baechle, T. R., Earle, R. W., & Wathen, D. (2000). Resistance
training. In T. R. Baechle & R. W. Earle (Eds.), Essentials of
Chronic agonist–antagonist complex set 1625
strength training and conditioning (pp. 395–425). Champaign, IL:
Human Kinetics.
Baker, D., Nance, S., & Moore, M. (2001). The load that
maximizes the average mechanical power output during
explosive bench press throws in highly trained athletes. Journal
of Strength and Conditioning Research, 15, 20–24.
Baker, D., & Newton, R. U. (2005). Acute effect on power output
of alternating an agonist and antagonist muscle exercise during
complex training. Journal of Strength and Conditioning Research,
19, 202–205.
Berger, R. A. (1962). Effect of varied weight training programs on
strength. Research Quarterly, 33, 169–181.
Cannon, R. J., & Cafarelli, E. (1987). Neuromuscular adaptations
to training. Journal of Applied Physiology, 63, 2396–2402.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences.
Hillsdale, NJ: Erlbaum.
Cronin, J., & Crewther, B. (2004). Training volume and strength
and power development. Journal of Science and Medicine in
Sport, 7, 144–155.
Garfinkel, S., & Cafarelli, E. (1992). Relative changes in maximal
force, EMG, and muscle cross-sectional area after isometric
training. Medicine and Science in Sports and Exercise, 24, 1220–1227.
Hakkinen, K., & Komi, P. V. (1981). Effect of different combined
concentric and eccentric muscle work regimens on maximal
strength development. Journal of Human Movement Studies, 7,
33–41.
Maynard, J., & Ebben, W. P. (2003). The effects of antagonist
prefatigue on agonist torque and electromyography. Journal of
Strength and Conditioning Research, 17, 469–474.
McCarthy, J. P., Pozniak, M. A., & Agre, J. C. (2002).
Neuromuscular adaptations to concurrent strength and endurance
training. Medicine and Science in Sports and Exercise, 34, 511–519.
Narici, M. V., Hoppeler, H., Kayser, B., Landoni, L., Claassen,
H., Gavardi, C., et al. (1996). Human quadriceps cross-
sectional area, torque and neural activation during 6 months
strength training. Acta Physiologica Scandinavica, 157, 175–186.
Pincivero, D. M., Gear, W. S., Moyna, N. M., & Robertson, R. J.
(1999). The effects of rest interval on quadriceps torque and
perceived exertion in healthy males. Journal of Sports Medicine
and Physical Fitness, 39, 294–299.
Robbins, D. W., Young, W. B., Behm, D. G., Payne, W. R.,
Klimstra, M. D. (in press). Physical performance and electro-
myographic responses to an acute bout of paired set strength
training versus traditional strength training. Journal of Strength
and Conditioning Research.
Rooney, K. J., Herbert, R. D., & Balnave, R. J. (1994). Fatigue
contributes to the strength training stimulus. Medicine and
Science in Sports and Exercise, 26, 1160–1164.
Stone, M. H., & O’Bryant, H. S. (1987). Weight training: A
scientific approach. Minneapolis, MN: Bellweather.
Thorstensson, A., Karlsson, J., Viitasalo, J. H. T., Luhtanen, P., &
Komi, P. V. (1976). Effect of strength training on EMG of human
skeletal muscle. Acta Physiologica Scandinavica, 98, 232–236.
Warburton, D. E., Nicol, C. W., & Bredin, S. S. (2006).
Prescribing exercise as preventative therapy. Canadian Medical
Association Journal, 174, 961–974.
Wathen, D. (1994). Strength training and spotting techniques. In
T. R. Baechle (Ed.), Essentials of strength training and condition-
ing (pp. 345–400). Champaign, IL: Human Kinetics.
Weir, J. P., Housh, T. J., & Weir, L. L. (1994). Electromyographic
evaluation of joint angle specificity and cross-training after
isometric training. Journal of Applied Physiology, 77, 197–201.
Weiss, L. W., Coney, H. D., & Clark, F. H. (1999). Different
functional adaptations to short-term low-, moderate-, and high-
repetition weight training. Journal of Strength and Conditioning
Research, 13, 236–241.
Copyright of Journal of Sports Sciences is the property of Taylor & Francis Ltd 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.