Journal of Sports Sciences

ISSN: 0264-0414 (Print) 1466-447X (Online) Journal homepage: http://www.tandfonline.com/loi/rjsp20

Physical demands of game-based training drills in

women’s team handball

Live S. Luteberget, Hege P. Trollerud & Matt Spencer

To cite this article: Live S. Luteberget, Hege P. Trollerud & Matt Spencer (2017): Physical
demands of game-based training drills in women’s team handball, Journal of Sports Sciences, DOI:
10.1080/02640414.2017.1325964

To link to this article: http://dx.doi.org/10.1080/02640414.2017.1325964

Published online: 16 May 2017.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rjsp20

Download by: [Eastern Michigan University] Date: 16 May 2017, At: 07:32
JOURNAL OF SPORTS SCIENCES, 2017
https://doi.org/10.1080/02640414.2017.1325964

Physical demands of game-based training drills in women’s team handball

Live S. Luteberget, Hege P. Trollerud and Matt Spencer
Department of Physical Performance, Norwegian School of Sports Sciences, Oslo, Norway

ABSTRACT ARTICLE HISTORY

Game-based training drills are popular in team sports. This study compared two game-based Accepted 26 April 2017
training conditions and official matches in team handball. Thirty-one women players wore inertial KEYWORDS
measurement units in five training sessions and five official matches. In training, 3vs3 and 6vs6 Female athletes; inertial
game-based training conditions were performed with a 5-min duration. PlayerLoad™ and high- movement units; match
intensity events (HIEs; >2.5 m · s−1) were extracted from the raw data. Data were analysed using analyses; team sports
magnitude-based inferences and reported with effect sizes (ESs). PlayerLoad™ · min−1 from all
positions combined was 11.37 ± 0.49 (mean ± 90% confidence limits) and 9.71 ± 0.3 for the 3vs3
and 6vs6 conditions, respectively. Backs (ES: 1.63), wings (ES: 1.91), and pivots (ES: 1.58) had greater
PlayerLoad™ in 3vs3 than 6vs6. Substantially greater HIE · min−1 in 3vs3 occurred for all positions.
There was substantially greater PlayerLoad™ · min−1 in 3vs3 and 6vs6 than match play for backs,
wings, and pivots. Wings (ES: 1.95), pivots (ES: 0.70), and goalkeeper (ES: 1.13) had substantially
greater HIE · min−1 in 3vs3 than match play. This study shows greater PlayerLoad™ and HIE in 3vs3
than 6vs6. Both game-based training conditions investigated in this study provide an overload in
overall PlayerLoad™; however, additional exercises might be needed to overload HIE, especially for
backs and pivots.

Introduction intensity of game-based training in team sports. There is

an increase in heart rate, lactate concentration, rating of
Successful performance in team handball is dependent on
perceived exertion (RPE) and greater total distance covered
several factors, including technical, tactical, social and phy-
with fewer players (Belka et al., 2016; Foster et al., 2009;
sical characteristics (Luteberget & Spencer, 2016; Michalsik,
Rampinini et al., 2007).
Aagaard, & Madsen, 2012; Michalsik, Madsen, & Aagaard,
Game-based training mimics specific game demands and
2014). The complexity of team handball requires econom-
thus it is assumed that the training provides an effective
ical training regimes to include all the important perfor-
transfer to match play (Aguiar, Botelho, Lago, Maças, &
mance factors. The use of game-based training drills is a
Sampaio, 2012). In soccer, game-based training simulated the
recommended technique that promotes physical perfor-
overall movement patterns of official matches (Gabbett &
mance enhancement and training effectiveness via a com-
Mulvey, 2008). However, there are no studies comparing the
bination of the components of the game (Hill-Haas, Dawson,
intensity of game-based training to the intensity of official
Impellizzeri, & Coutts, 2011). Indeed, game-based training
matches in team handball. Also, for soccer, playing position
drills are commonly used to improve technical and tactical
affects the intensity of game-based training, and different
skills in many team sports, and it has been shown that such
positions display differences in activity compared to match
skills provide an aerobic stimulus comparable with tradi-
play (Dellal et al., 2012). In team handball, there are substantial
tional interval-training methods (Buchheit et al., 2009;
differences in physical demands among the different playing
Iacono, Eliakim, & Meckel, 2015). Because of the incorpora-
positions in match play (Luteberget & Spencer, 2016; Michalsik
tion of these training drills in team sports, several research-
et al., 2014). However, to our knowledge, no studies in team
ers have focused their attention on physical, physiological
handball have investigated the intensity of game-based train-
and technical activities of specific drills. There are a variety
ing in the context of playing positions.
of factors that affect the intensity of game-based training in
Although there is a substantial growth in research related
team sports, such as the field size (Corvino, Tessitore,
to game-based training in team sports, there is a lack of
Minganti, & Sibila, 2014; Kennett, Kempton, & Coutts, 2012;
research specific to team handball. In light of this, the aim of
Rampinini et al., 2007), the number of players involved
the present study was to compare the intensity of game-based
(Foster, Twist, Lamb, & Nicholas, 2009; Rampinini et al.,
training drills with different player numbers (3vs3 and 6vs6). In
2007), rule modifications (Dellal, Logo-Penas, Wong, &
addition, a comparison of these two training drills with official
Chamari, 2011) and coaches’ encouragement (Rampinini
match intensity was important to specify game-based training
et al., 2007). For example, number of players affects the
for each playing position.

CONTACT Live S. Luteberget [email protected] Norwegian School of Sport Sciences, Postboks 4014 Ullevål Stadion, 0806 Oslo, Norway.
© 2017 Informa UK Limited, trading as Taylor & Francis Group
2 L. S. LUTEBERGET ET AL.

Methods
Experimental approach to the problem
In the present study, we investigated PlayerLoad™ and high-
intensity events (HIEs) of game-based training drills lasting
5 min and compared them to periods of equivalent duration
in official national-standard competition in high-level women
team handball players. The study was conducted during the
first half of the handball season (August–December) and con-
sisted of monitoring 10 training sessions where the game-
based training was conducted, in addition to the monitoring
of 10 official matches in national leagues in Norway.

Participants
Thirty-one semi-professional female handball players (age:
22.2 ± 3.3 years, stature: 171.1 ± 6.4 cm, body mass:
68.5 ± 6.5 kg) volunteered and completed the study. The Figure 1. Setup for the two different game-based training conditions. 6vs6 (a)
participants were from two teams: one playing in the sec- includes a total of 14 players, 6 field players and 1 goalkeeper on each team.
ond-highest division (1. division; n = 15) in Norway and one 3vs3 (b) includes a total of 8 players, 3 field players and 1 goalkeeper on each
team. The playing area was held constant with the area of a standard handball
in the highest division (elite division; n = 16). All participants court (20 × 40 m) in both conditions. Note that the area per player refers to
received verbal and written information about the procedures outfield players only; the goalkeeper area (GKA) is kept constant in both
of the study and gave their written consent to participate in conditions.
the study. The Norwegian Social Science Data Services
approved the study.
each team. The 6vs6 condition consisted of 6 field-players on
each team, in addition to a goalkeeper on each team. The
playing area was held constant with the area of a standard
Monitoring of game-based training
handball court (20 × 40 m). The aim of the game-based
All participants were equipped with an inertial measurement training drills was to create a match-like setting; thus, the
unit (IMU; OptimEye S5, Catapult Sports, Australia) for mon- rules of the drills were kept the same as for official matches,
itoring of game-based training. The device was located in a with the exception that the goalkeeper was allowed to keep a
padded pouch on the upper back in a custom-made vest ball by the goal for a rapid replacement of the ball after a
(Catapult Sports, Australia). The device is integrated with a missed shot. Verbal encouragement provided by the coach
tri-axial accelerometer, a gyroscope and a magnetometer, all was allowed, and the coach was instructed to give encourage-
collecting data at 100 Hz. The device was fitted under the ment similar to what they do in official matches. The order in
training jersey before training commenced. The monitoring of which they performed the two drills was alternated between
game-based training was conducted in a total of 10 training sessions. For players to be included in the analysis, they had to
sessions (5 with each team). The players were monitored at complete a minimum of three monitored training sessions
their respective home-court arena and players were instructed where they were active in both 3vs3 and 6vs6.
to have a similar preparation to each of the monitored ses- Playing positions were set on the basis of the position the
sions, in terms of activity the days before and on the same participants played in official matches. The positions in official
day. Each session began with a general warm-up and a hand- matches and in 6vs6 were the same for all participants. In
ball-warm-up. All participants for each session completed 3vs3, the specific positions change because of the increased
game-based training with 3vs3 and 6vs6 (Figure 1), with a area per player. Thus, the players are not necessarily playing in
duration of 5 min of each. Each participant completed one their assigned playing position in the 3vs3 condition.
or two repetitions of the 3vs3 and 6vs6 game-based training
conditions in each session, depending on the total number of
Match monitoring
players available for the session. Because of practical consid-
erations, it was not possible to standardise the length of the All participants wore the same IMU used in the monitoring of
rest period for each participant. The different game-based game-based training. The data collection was monitored live
training drills began with a 5-min interval, meaning that in using the Catapult Sprint (Version 5.1.4, Catapult Sports, 2014)
most cases the participants had a 5-min rest between condi- software. Interchanges were manually tracked using this soft-
tions. However, in some cases, the rest period was doubled ware to ensure that only time spent on the field was included in
because of the large number of participants available for the the analyses. During team time-outs, all players were inactive.
session. Players were instructed to be active to stay warm As interchanges were frequent and could involve several
between repetitions (e.g., light jog), but were not allowed to players, the interchange area was video-recorded and notes
do any strenuous exercises. The 3vs3 condition consisted of 3 were taken. Thus, uncertainties and eventual errors in inter-
field-players on each team, in addition to a goalkeeper on changes could be corrected in the software. Apart from players
 JOURNAL OF SPORTS SCIENCES 3

wearing the IMU during matches, the study did not intervene considered as trivial, small, moderate, large and very large
with any other aspect of the normal match or match prepara- (Hopkins et al., 2009), respectively. The percentage likelihood
tion. For players to be included in the analysis, they had to have of difference between conditions was calculated (Hopkins,
a minimum of 5 min on-field time in three matches. 2006) and considered almost certainly not (<0.5%), very unli-
kely (<0.5%), unlikely (<25%), possibly (25–75%), likely (>75%),
very likely (>95%) or most likely (>99.5%). A percentage like-
Data processing
lihood of difference <75% was considered a substantial mag-
PlayerLoad™, accelerations, changes of direction, and decel- nitude. Threshold chances of 5% for substantial magnitudes
erations were extracted from the raw files using the Catapult were used, meaning a likelihood of >5% in both positive and
Sprint software. Briefly, PlayerLoad™ is an accelerometer- negative directions was considered an unclear difference
based measurement of external physical loading of team (Hopkins et al., 2009).
sport athletes. Player Load™ was defined as instantaneous
rate of change of acceleration divided by a scaling factor,
expressed as the square root of the sum of the squared Results
instantaneous rates of change in acceleration in each of the The mean values of PlayerLoad™ · min−1 for all players combined
three vectors (X, Y and Z axes) and divided by 100 (Boyd, Ball, were 11.37 ± 0.49, 9.71 ± 0.3, 8.73 ± 0.25 and 9.85 ± 0.36
& Aughey, 2011). Acceleration, change of direction and PlayerLoad™ · min−1 for the 3vs3, 6vs6, mean match and MSP
deceleration events were based on accelerometer (magni- conditions, respectively. Data from all positions in the different
tude), gyroscope and magnetometer (direction) data. All conditions are displayed as mean data in Table 1. Mean values for
events >2.5 m · s−1 were included as either an acceleration, HIE · min−1, when combining all playing positions, were
change of direction or deceleration. The sum of all accelera- 4.27 ± 0.20, 3.03 ± 0.17, 3.29 ± 0.22 and 4.13 ± 0.27 HIE · min−1
tion, change of direction or deceleration events was referred for the 3vs3, 6vs6, mean match and MSP conditions, respectively.
to as HIE. All variables of interest were normalised per min of Mean data from each playing position is presented in Table 1.
on-field time to minimise the variability of absolute values There were substantial differences between the 3vs3 and
with varying match length and individual on-field time in the 6vs6 conditions for all playing positions. Backs had greater
matches. Five-minute periods were calculated from the start PlayerLoad™ · min−1 in the 3vs3 condition with a large ES
of each half and continuously throughout the duration of the (100%, ES: 1.63), and greater HIE · min−1 with a small ES
half. Because of stoppages during play, the duration of halves (85%, ES: 0.58). Wings had greater PlayerLoad™ · min−1 in the
varied both in and among games and could be longer than 3vs3 condition with a large ES (100%, ES: 1.91) and a very large
the effective half-time of 30 min. The 5-min period with the ES for HIE · min−1 (100%, ES: 2.32). Greater PlayerLoad™ · min−1
highest value, for each individual player, was extracted and occurred in the 3vs3 condition for pivots, with a large ES (99%,
represented the most strenuous period (MSP) in match play. ES: 1.58) for PlayerLoad™ · min−1 and a moderate ES (90%, ES:
This serves as a “worst-case scenario”. Players had to be on 1.12) for HIE · min−1. There were no clear differences in the
field for the entire 5-min period (100%), and they had to have 3vs3 and 6vs6 condition in PlayerLoad™ · min−1 for GK, how-
at least two 5-min periods in the match to be included in the ever, HIE · min−1 was greater in the 3vs3 condition with a
analyses of MSP. The MSP was extracted individually for each moderate ES (95%, ES: 0.93).
variable, meaning that PlayerLoad™ · min−1 and HIE · min−1 There was a greater PlayerLoad™ · min−1 in the 3vs3 con-
were not necessarily extracted from the same period. There dition than match play for all playing positions. Similarly, with
were no substantial differences between the two teams in
any of the variables, and thus all results are reported from
the two teams combined. Table 1. Mean and upper and lower confidence limits (CLs) for each playing
position in each condition, for both PlayerLoad™ · min−1 and HIE · min−1. n is
the number of observations in the respective condition and position.
Statistical analyses PlayerLoad™ · min−1 HIE · min−1
90% CL 90% CL
Data are presented as mean ± 90% confidence limits (CLs),
both for the raw data and percentage change data. A linear Position Condition n Mean Upper Lower N Mean Upper Lower
mixed-model in IBM© SPSS© Statistics (Version 21, IBM Corp,) Back 3vs3 63 13.24 12.45 14.07 63 4.50 4.03 5.01
6vs6 78 10.54 9.92 11.19 78 3.69 3.17 4.29
was used for analyses. The dependent variables were log- Match mean 39 9.27 8.71 9.87 39 4.08 3.61 4.61
transformed before analysis, as this approach yields variability MSP 25 10.84 10.13 11.60 25 5.36 4.67 6.15
and differences as the percentage of the mean (Hopkins, Wing 3vs3 56 13.87 13.02 14.79 56 4.36 3.89 4.90
6vs6 72 10.62 9.97 11.30 72 1.98 1.69 2.31
Marshall, Batterham, & Hanin, 2009). The fixed effects in the Match mean 32 8.79 8.23 9.38 32 2.24 1.97 2.55
model were playing position (4 levels: back, wing, pivot and MSP 20 10.46 9.73 11.24 20 3.01 2.59 3.50
goalkeeper (GK)), condition (4 levels: 3vs3, 6vs6, match, and Pivot 3vs3 17 13.14 11.64 14.82 13 5.34 4.23 6.73
6vs6 22 10.52 9.36 11.84 17 3.64 2.57 5.14
MSP), position*condition and team ID. The random effects Match mean 13 9.60 8.54 10.79 12 4.21 3.35 5.28
were athlete ID and Game ID. Separate analyses were per- MSP 10 10.86 9.60 12.28 9 5.29 4.14 6.76
formed for the dependent variables PlayerLoad™ · min−1 and GK 3vs3 47 4.90 4.47 5.37 47 2.83 2.42 3.31
6vs6 32 4.57 4.17 5.01 32 2.06 1.63 2.60
HIE · min−1. Differences between the conditions were analysed Match mean 9 5.17 4.67 5.73 9 1.92 1.54 2.40
using Cohen’s effect size (ES) statistics and ±90% CL. ESs of MSP 9 6.01 5.40 6.69 9 2.88 2.31 3.60
<0.20, 0.20 to 0.59, 0.60 to 1.19, 1.2 to 1.99 and ≥2.00 were GK: goalkeeper; MSP: most strenuous period.
4 L. S. LUTEBERGET ET AL.

the exception of GK, the MSP was greater for all playing condition compared to mean match play for any playing
positions (Figure 2). Differences were also found between position (Figure 3(a)) in HIE · min−1. Backs had lower values
the 6vs6 condition and match mean and MSP (Figure 2). in the 3vs3 condition than MSP, while wings had greater
Wings, pivots, and GK had greater HIE · min−1 in 3vs3 than values (Figure 3(b)). All playing positions had lower
matches. Conversely, there were no differences in the 6vs6 HIE · min−1 in 6vs6 than in MSP.

Figure 2. PlayerLoad™ · min−1 mean ± 90% confidence limits for percentage differences from match mean (a) and MSP (b) for the 3vs3 condition and 6vs6
condition. Effect size (ES) between the different game-based training condition and match mean or MSP is indicated by the stated symbols. Only ESs with a
substantial likelihood of difference (>75%) are shown. * = small, ** = moderate, *** = large, **** = very large. GK: goalkeeper; MSP: most strenuous period.

Figure 3. HIE · min−1 mean ± 90% confidence limits for percentage differences from match mean (a) and MSP (b) for the 3vs3 condition and 6vs6 condition. Effect
size (ES) between the different game-based training conditions and match mean or MSP is indicated by the stated symbols. Only ESs with a substantial likelihood of
difference (>75%) are shown. * = small, ** = moderate, *** = large, **** = very large. GK: goalkeeper; HIE: high intensity events; MSP: most strenuous period.
 JOURNAL OF SPORTS SCIENCES 5

Discussion the overall movement patterns of friendly matches of a

domestic, national, and international standard of competition
This study aimed to investigate the intensity
(Casamichana et al., 2012; Gabbett & Mulvey, 2008). However,
(PlayerLoad™ · min−1) and HIE of game-based training drills
there are, to our knowledge, no previous studies comparing
in team handball. The results of this study indicate that the
the intensity of game-based training to the intensity of official
number of players involved in the game-based training drills
matches in team handball. The current study shows a greater
can affect the intensity of the drill with a lower number of
PlayerLoad™ · min−1 for backs, wings and pivots in both game-
players, resulting in an increase in both PlayerLoad™ · min−1
based training conditions when compared to the mean of the
and HIE · min−1. Moreover, to our knowledge, this is the first
match; this is in line with research in other team sports. The
study to compare game-based training drills to official match
PlayerLoad™ values are also comparable to the values in MSP,
intensity in team handball. The data shows that 3vs3 and 6vs6
and thus can be recommended as an effective training regime
game-based training conditions, with a duration of 5 min,
to mimic/overload official matches. The reduced duration of
show greater values compared to mean match intensity in
the game-based training compared to official matches is most
PlayerLoad™ · min−1 for backs, wings and pivots. However,
likely a contributing factor for greater intensity (Corvino et al.,
the same pattern is not present in HIE · min−1 for backs and
2014). Other factors, such as coach encouragement and spare
pivots.
balls (more effective play time), may also contribute to the
The differences between the 3vs3 and 6vs6 conditions in
training intensity. In addition, the fact that game-based train-
the present study are in line with previous research in team
ing is a training task, and thus the outcome of the “match” is
handball and in other team sports. It has been reported
not as important as an official match, may affect the tactical
greater RPE, greater heart rate and a greater total distance
decisions and risk-taking in play, and again affect the intensity.
covered in 3vs3, compared to 4vs4 and 5vs5 (Belka et al., 2016)
Even though the game-based training is shown to mimic/
in youth male team handball players. It has also been shown
overload the overall intensity of official matches, it is also
that a decrease in player number increases the occurrence of
shown that the training is not sufficient to simulate the
technical actions, such as the number of contacts and number
high-intensity repeated-sprint demands of high-standard
of dribbles in futsal (Duarte, Batalha, Folgado, & Sampaio,
matches in women’s soccer (Gabbett & Mulvey, 2008).
2009). When the player number is changed and the pitch
Similarly, in the current study, the HIE · min−1 does not seem
size is held constant, the area per player will be altered. The
to mimic official match demands to the same extent as
same is true if the pitch size is changed, while the player
PlayerLoad™ · min−1. HIE · min−1 values are not substantially
number is held constant. Studies altering the pitch size show
different from match mean data and are lower in the game-
similar results to the current study. Specifically, a greater
based training compared to MSP. The lower HIE · min−1 might
intensity occurred with an enlargement of area per player.
be a consequence of the greater PlayerLoad™ · min−1, as the
For example, a study on game-based training in team handball
greater intensity might hinder the player’s possibility for
found differences in intensity when altering pitch size and
explosive actions and reduce their ability to execute HIE. In
holding the player number constant (Corvino et al., 2014).
addition, motivation for the activity may play a role in the
The greatest square metre per player (approximately 64 m2)
physical output, and thus, HIE could be greater in matches
elicited greater values in total distance covered, RPE and heart
due to this fact. These data suggest that while overloading the
rate (Corvino et al., 2014). In the current study, the 3vs3
overall intensity in team handball players, game-based train-
condition resulted in an area of 108.5 m2 while the 6vs6
ing conditions do not overload the specific HIE, which is an
resulted in an area of 54.3 m2 per player, and again the largest
important factor for performance in team handball
square metre per player elucidated the highest intensity.
(Luteberget & Spencer, 2016; Michalsik et al., 2012). Wing
When changing the number of players, while holding the
players show a greater HIE · min−1 in the 3vs3 condition
area per player constant, heart rate and blood lactate is
compared to both match mean and MSP. In a 3vs3 condition,
reported not to be different between conditions in soccer,
wing players will have a different role than in a 6vs6 condition.
although some differences in the activity profile have been
They will be located more centrally on the court, and thus be
reported (Castellano, Casamichana, & Dellal, 2013; Randers,
more involved in the game, in both offence and defence. Over
Nielsen, Bangsbo, & Krustrup, 2014; Randers et al., 2010).
all, 3vs3 seems to be a good overload condition for wing
Thus, some changes might occur by manipulating player num-
players for both PlayerLoad™ · min−1 and HIE · min−1.
ber alone. However, it is also important to note that the
However, the fact that wing players change their role in the
changes in the current study might not be solely contributed
3vs3 condition will affect the specificity of the drill in relation
to player number per se, as the changes in area per player
to their typical match role. This is also applicable to pivot
might be of equal importance.
players in the 3vs3 condition.
Game-based training is thought to mimic specific physical
In addition to the PlayerLoad™ and HIE investigated in this
and physiological game demands, and thus it is assumed that
study, team handball players are also subjected to isometric
it can be an effective and specific form of training for team
actions. Such actions are not registered by the IMU, and thus
sport athletes (Aguiar et al., 2012). In addition, an overload of
an underestimation of the intensity of players is present in the
physical components is often preferable to increase or main-
current study, both in the game-based training and in match
tain the players’ physical conditioning (Casamichana,
play. This might be more pronounced in pivot players due to
Castellano, & Castagna, 2012; Hill-Haas et al., 2011). In soccer,
their tactical role in both offence and defence. However, this
it has been reported that game-based training can simulate
issue of isometric actions should not have affected the
6 L. S. LUTEBERGET ET AL.

differences between game-based training conditions and References

matches in the present study.
Aguiar, M., Botelho, G., Lago, C., Maças, V., & Sampaio, J. (2012). A review
During match play, GKs display lower PlayerLoad™ · min−1 on the effects of soccer small-sided games. Journal of Human Kinetics,
and HIE · min−1 than the other playing positions (Luteberget & 33, 103-113. doi:10.2478/v10078-012-0049-x
Spencer, 2016). In the present study, GKs had lower Belka, J., Hulka, K., Safar, M., Duskova, L., Weisser, R., & Riedel, V. (2016).
PlayerLoad™ · min−1 values in the game-based training condi- Time-motion analysis and physiological responses of small-sided team
tions than in MSP. Thus, GKs might need a different training handball games in youth male players: Influence of player number.
Acta Gymnica. doi:10.5507/ag.2016.019
setup for intensity overloading than the other playing posi- Boyd, L. J., Ball, K., & Aughey, R. J. (2011). The reliability of minimaxx
tions. The fact that GKs had spare balls close to their goal accelerometers for measuring physical activity in Australian football.
might also affect the intensity for this position, as they do not International Journal of Sports Physiology and Performance, 6, 311–321.
have to move over larger areas to get the ball back in play. In doi:10.1123/ijspp.6.3.311
both 3vs3 and 6vs6, the GKs’ roles/tasks are the same; how- Buchheit, M., Laursen, P. B., Kuhnle, J., Ruch, D., Renaud, C., & Ahmaidi, S.
(2009). Game-based training in young elite handball players.
ever, the results in this study show that the amount of HIE International Journal of Sports Medicine, 30(4), 251–258. doi:10.1055/s-
differs between the two conditions. For tactical reasons, the 0028-1105943
frequency and type of shots could have been different during Casamichana, D., Castellano, J., & Castagna, C. (2012). Comparing the
the game-based training than official match play, especially in physical demands of friendly matches and small-sided games in semi-
the 3vs3 condition. This could in turn affect the amount of HIE professional soccer players. The Journal of Strength and Conditioning
Research, 26(3), 837–843. doi:10.1519/JSC.0b013e31822a61cf
that GKs execute. The methods for monitoring handball Castellano, J., Casamichana, D., & Dellal, A. (2013). Influence of game
players used in this study might not be optimal for measuring format and number of players on heart rate responses and physical
the load of GKs (Luteberget & Spencer, 2016), and different demands in small-sided soccer games. Journal of Strength and
analysis methods might be more useful. However, this needs Conditioning Research /National Strength & Conditioning Association, 27
to be investigated further. (5), 1295–1303. doi:10.1519/JSC.0b013e318267a5d1
Corvino, M., Tessitore, A., Minganti, C., & Sibila, M. (2014). Effect of
court dimensions on players’ external and internal load during
small-sided handball games. Journal of Sports Science and
Conclusions Medicine, 13(2), 297–303.
The results from this study demonstrate that coaches can Dellal, A., Logo-Penas, C., Wong, D. P., & Chamari, K. (2011). Effect of the
number of ball contacts within bouts of 4 vs. 4 small-sided soccer
manipulate physical demands in game-based training drills
games. International Journal of Sports Physiology and Performance, 6
for women team handball players by modifying the number (3), 322–333. doi:10.1123/ijspp.6.3.322
of players involved. The study shows that there are substantial Dellal, A., Owen, A., Wong, D. P., Krustrup, P., van Exsel, M., & Mallo, J.
differences in PlayerLoad™ and HIE when the number of (2012). Technical and physical demands of small vs. large sided games
players in game-based training drills is altered; fewer players in relation to playing position in elite soccer. Human Movement Science,
31(4), 957–969. doi:10.1016/j.humov.2011.08.013
resulted in higher values both in PlayerLoad™ · min−1 and
Duarte, R., Batalha, N., Folgado, H., & Sampaio, J. (2009). Effects of exercise
HIE · min−1 for all outfield positions. In addition, both game- duration and number of players in heart rate responses and technical skills
based training conditions facilitated a greater PlayerLoad™ during futsal small-sided games. The Open Sports Sciences Journal, 2, 37-41.
than match data (both mean and MSP). However, HIE is not Foster, C. D., Twist, C., Lamb, K. L., & Nicholas, C. W. (2009). Heart rate
overloaded to the same extent for backs and pivots. Thus, responses to small sided games among elite junior rugby league players.
Journal of Strength and Conditioning Research /National Strength &
coaches can use game-based training drills, such as the ones
Conditioning Association, 1–6. doi:10.1519/JSC.0b013e3181aeb11a
used in this study, to overload the PlayerLoad™ of match Gabbett, T. J., & Mulvey, M. J. (2008). Time-motion analysis of small-sided
situations. However, to overload the important HIE for backs training games and competition in elite women soccer players. Journal
and pivots, there is a need for additional drills. Further of Strength and Conditioning Research /National Strength & Conditioning
research is required to fully understand all factors that influ- Association, 22(2), 543–552. doi:10.1519/JSC.0b013e3181635597
Hill-Haas, S. V., Dawson, B., Impellizzeri, F. M., & Coutts, A. J. (2011).
ence the intensity and amount of HIE of game-based training
Physiology of small-sided games training in football: A systematic
drills in team handball. In addition, a technical comparison of review. Sports Medicine, 41, 199–220. doi:10.2165/11539740-
game-based training drills and matches is required to under- 000000000-00000
pin the specificity of the drills. Hopkins, W. (2006). A spreadsheet for combining outcomes from several
subject groups. Retrieved from sportsci.org/2006/wghcom.htm
Hopkins, W. G., Marshall, S. W., Batterham, A. M., & Hanin, J. (2009).
Acknowledgements Progressive statistics for studies in sports medicine and exercise
science. Medicine and Science in Sports and Exercise, 41(1), 3–12.
The authors would like to thank the players participating in this study for doi:10.1249/MSS.0b013e31818cb278
their time, effort and enthusiastic participation, and the Norwegian hand- Iacono, A. D., Eliakim, A., & Meckel, Y. (2015). Improving fitness of elite
ball federation for giving us the opportunity to execute this study. The handball players: Small-sided games vs. high-intensity intermittent
authors would also like to thank Will Hopkins for his help and assistance training. Journal of Strength and Conditioning Research /National
with the statistical model used in this study. Strength & Conditioning Association, 29(3), 835–843. doi:10.1519/
JSC.0000000000000686
Kennett, D. C., Kempton, T., & Coutts, A. J. (2012). Factors affecting exercise
Disclosure statement intensity in rugby-specific small-sided games. Journal of Strength and
Conditioning Research, 26(8), 2037–2042. doi:10.1519/
No potential conflict of interest was reported by the authors. JSC.0b013e31823a3b26
 JOURNAL OF SPORTS SCIENCES 7

Luteberget, L. S., & Spencer, M. (2016). High intensity events in interna- small-sided soccer games. Journal of Sports Sciences, 25(6), 659–666.
tional female team handball matches. International Journal of Sports doi:10.1080/02640410600811858
Physiology and Performance. doi:10.1123/ijspp.2015-0641 Randers, M. B., Nielsen, J. J., Bangsbo, J., & Krustrup, P. (2014).
Michalsik, L. B., Aagaard, P., & Madsen, K. (2012). Locomotion character- Physiological response and activity profile in recreational small-sided
istics and match-induced impairments in physical performance in male football: No effect of the number of players. Scandinavian Journal of
elite team handball players. International Journal of Sports Medicine, 34, Medicine & Science in Sports, 24(Suppl 1), 130–137. doi:10.1111/
590–599. doi:10.1055/s-0032-1329989 sms.12232
Michalsik, L. B., Madsen, K., & Aagaard, P. (2014). Match performance and Randers, M. B., Nybo, L., Petersen, J., Nielsen, J. J., Christiansen, L.,
physiological capacity of female elite team handball players. Bendiksen, M., . . . Krustrup, P. (2010). Activity profile and physiological
International Journal of Sports Medicine, 35, 595–607. doi:10.1055/s- response to football training for untrained males and females, elderly
0033-1358713 and youngsters: Influence of the number of players. Scandinavian
Rampinini, E., Impellizzeri, F. M., Castagna, C., Abt, G., Chamari, K., Sassi, A., Journal of Medicine & Science in Sports, 20(Suppl 1), 14–23.
& Marcora, S. M. (2007). Factors influencing physiological responses to doi:10.1111/j.1600-0838.2010.01069.x