See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/359481021

Small-Sided Games in Elite Football: Practical Solutions to Replicate the 4-

min Match-Derived Maximal Intensities

Article in The Journal of Strength and Conditioning Research · March 2022

DOI: 10.1519/JSC.0000000000004249

CITATIONS READS

16 2,449

3 authors:

Andrea Riboli Fabio Esposito

A.C. Milan University of Milan
58 PUBLICATIONS 303 CITATIONS 237 PUBLICATIONS 4,275 CITATIONS

SEE PROFILE SEE PROFILE

Giuseppe Coratella
Università degli Studi di MIlano
130 PUBLICATIONS 2,327 CITATIONS

SEE PROFILE

All content following this page was uploaded by Giuseppe Coratella on 26 March 2022.

The user has requested enhancement of the downloaded file.

 Original Research

Small-Sided Games in Elite Football: Practical

Solutions to Replicate the 4-min Match-Derived
Maximal Intensities
Andrea Riboli, Fabio Esposito, and Giuseppe Coratella
Department of Biomedical Sciences for Health, University of Milan, Milan, Italy
Downloaded from http://journals.lww.com/nsca-jscr by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC4/OAVpDDa8KKGKV0Ymy+78= on 03/26/2022

Abstract
Riboli, A, Esposito, F, and Coratella, G. Small-sided games in elite football: practical solutions to replicate the 4-min match-derived
maximal intensities. J Strength Cond Res XX(X): 000–000, 2022—The purpose of the study was to determine the area per player
(ApP, m2·player21) to replicate the maximal 4-min match-play period (4-minPeak) using small-sided or large-sided games (SSGs)
lasting the same duration in elite football players. Twenty players (age: 26 6 5 years) competing in European championship and
international UEFA competitions were monitored during SSGs with (SSGwith) or without (SSGwithout) goalkeepers, both lasting 4-
min. A total of 1,724 and 561 individual observations for SSGwith and SSGwithout across 99 and 34 different SSGs formats were
undertaken. Whole (90-minavg) and 4-minPeak match demands were calculated. Relative (m·min21) total, high-speed (HSRD), very
high-speed (VHSRD), sprint, and acceleration/deceleration distance were collected. For both SSGwith and SSGwithout, ApP was
small-to-very-largely correlated with TD, HSRD, VHSRD, and sprint, whereas an inverse trivial-to-small correlation with
acceleration/deceleration was found. In SSGwith, the ApP to replicate 4-minPeak was larger (p , 0.001, effect size [ES]: 0.33–5.80)
than that to mimic 90-minavg for total distance, HSRD, VHSRD, and sprint, whereas for acceleration/deceleration the ApP was lower
(p , 0.001, ES: 23.13 to 0.48) for 4-minPeak. In SSGwithout, the ApP to replicate 4-minPeak was larger (p , 0.001, ES: 0.04–3.72)
than that to mimic 90-minavg for HSRD, VHSRD, and sprint. Some positional differences (p , 0.05, ES: 0.22–1.10) in ApP were
found. The ApP to replicate 4-minPeak was higher for SSGwith than SSGwithout (p , 0.001, ES: 0.67–3.58). These findings may help
practitioners to recreate the 4-minPeak to manage ApP during SSGs in elite football.
Key Words: team sports, soccer, performance, training load, global positioning system, locomotor activities

Introduction conditioned for more intense demands of the competition (10,39).

For example, accordingly with the distribution of match activities
In football, coaches and sport scientists need to control the
relative to maximal intensities (36), football-specific drills lasting
training load applied to each player to maximize individual ad-
2-, 3-, 4-, or 5-minutes could be targeted accordingly with the
aptations and improve performance (22,28). To this purpose,
most intense phases of similar durations determined during offi-
player-tracking technologies such as global positioning systems
cial matches (7,39).
are typically used to quantify total distance (TD), high-speed
Small-sided games (SSGs) are largely used to improve physical
running distance (HSRD), very-high-speed running distance
fitness (21,32) recreating technical, tactical, and physical
(VHSRD), sprint, and acceleration/deceleration distance allowing
football-specific abilities (19,34). In practice, continuous
to profile the individual training and match locomotor de-
mands (34). (i.e., 10–30 minutes) or interval (i.e., 2–10 bouts lasting 3- to 5-
The activities recorded during the matches are used to plan the minute) training regimens are used to manage the workload using
training workload (25,28,34). Although the most common time- SSGs (17,19). Interestingly, the literature shows that higher
motion analysis has been focused on the 90-minute match de- HSRD and VHSRD were found using SSGs intermittent vs.
mands (12), the use of a rolling average method, where distance is continuous regimens (20). In addition, 4-min SSGs bout was
divided in set intervals from every time point sampled, was sug- suggested as “the optimal” physical training stimulus to elicit
gested as a more appropriate method when quantifying the run- training adaptations (14). Hence, the intermittent 4-min SSGs
ning intensity periods in football (3,39). Several investigations model could be used to target the training stimuli toward the
from Serie A (39), French Ligue 1 (25), English Championship intense phases of match-play recorded in similar time-windows.
(15), and reserve squad Spanish La Liga (3,27) highlighted that When using SSGs in practice, several independent parameters
different time windows (e.g., 1-, 2-,3-, 4-min, etc.) should be could be manipulated to vary the training-induced adaptations
considered across training routine to allow the players to be (8). For example, SSGs may be played with (SSGwith) or without
(SSGwithout) the presence of the goalkeeper, when the aim is to out-
Address correspondence to Andrea Riboli, riboliandrea@outlook.com. score the opponent team or to maintain ball possession as long as
Supplemental digital content is available for this article. Direct URL citations appear possible, respectively (19,34). Despite some controversial find-
in the printed text and are provided in the HTML and PDF versions of this article on ings (24,26), TD and HSRD were higher in SSGwithout than
the journal’s Web site (http://journals.lww.com/nsca-jscr). SSGwith of similar pitch-size in Serie A soccer players because of a
Journal of Strength and Conditioning Research 00(00)/1–9 lower player tactical organization, thus increasing the SSGs de-
ª 2022 National Strength and Conditioning Association mands (34). In addition, greater pitch size and lower number of

Copyright © 2022 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Area Per Player to Replicate 4-minPeak (2022) 00:00

players may lead to increase in the locomotor activities specially A total of 524 individual samples during official matches with a
at high intensities (18). Since the pitch size and number of players median of 34.2 6 15.4 (range 5 43–2) individual samples were
could be manipulated independently resulting in a very high monitored. The official-match pitch size was 105 3 66 m, with a
number of combinations, the area per player (ApP, expressed as grass surface.
m2·player) was suggested to have possible reference values for
training purposes (19,31,34). The manipulation of the ApP was
reported to induce different physiological responses (5), loco- Subjects
motor activities (34), technical factors (23), and tactical behaviors Twenty elite football players competing in top-class European
(30) in elite football players. Focusing on the locomotor activities, championship and international UEFA competitions (age [6SD]:
increments in ApP were reported to increase TD, HSRD, VHSRD, 26 6 5 years; body mass: 80 6 7 kg; body height: 1.85 6 0.06 m)
and sprint in Serie A football players (25,34). In addition, larger were involved in the present study. All subjects were classified
ApP may be used during SSGwith than SSGwithout to replicate the according to their position: forwards (n 5 3), wide forwards, (n 5
90-minute match-play demands (90-minavg) for TD and HSRD 3), central-midfielders (n 5 4), wide-midfielders (n 5 4), and
distance, whereas no difference was observed in SSGwith vs. central-defenders (n 5 6). The goalkeepers were excluded from
SSGwithout for sprint (34). Moreover, between-position differ- the data collection. The club’s medical staff certified the health
ences were reported highlighting that central defenders need status of each player. An injured player was excluded from data
lower and forwards and central midfielders higher ApP (34). collection for at least 1 month after their return to full training.
However, no previous study has investigated the minimal ApP to The procedures were fully explained to the subjects and the club
replicate the match demands across a specific 4-min time-window staff. The subjects gave their written consent. The Ethics Com-
determined during the official matches. Because the 4-min time mittee of the University of Milan (protocol #102/14) approved
window is representative of the optimal SSGs duration (14), this the study that was performed in accordance with the principles of
may have potential practical implications. Therefore, the aim of the Declaration of Helsinki (1975) for studies involving human
the present study was to calculate the ApP during SSGwith and subjects.
SSGwithout to replicate the 4-minPeak determined during the offi-
cial match using a rolling average method for TD, HSRD,
VHSRD, sprint, and acceleration/deceleration in elite football
Procedures
players. In addition, between-position differences were
investigated. A 10-Hz Global-positioning system unit (K-Sport, Mon-
telabbate, Italy) was used to collect data during training (34).
Each device was turned on at least 15-minutes before each
Methods session to allow for acquisition of the satellite signal (34). To
reduce the interunit differences, each player wore the same unit
Experimental Approach to the Problem for every training session over the whole investigation (34).
The present investigation was conducted during the competition The locomotor activities during the official matches were col-
period across 2 consecutive seasons. The subjects undertook their lected using a computerized semi-automated video-based
traditional weekly training routine, and the data were extrapo- multicamera image system (Stats Perform, Chicago, IL) and
lated from the sessions as proposed by the team manager and the processed by a dedicated software (34). The system has pre-
technical staff. All sessions were performed on grass or artificial- viously been shown to provide valid and reliable measurements
surface pitches preserved by qualified operators and were con- of the match activity in football (12,34). GPS and the video-
ducted at the same time of day to limit the effects of circadian based multi-camera image system was previously determined
variation. A specialized and high-qualified physician staff rec- as interchangeable (34).
ommended and monitored the diet regime of each player before During both training sessions and matches, TD, HSRD
and after every training session. (15–19.9 km·h21), VHSRD (20–24 km·h21), sprint distance
Two different formats of SSGs were analyzed: SSGwith and (.24 km·h21), and acceleration/deceleration distance (.3 m·s21)
SSGwithout. A total of 2,285 individual observations (1,724 and were measured (34). The distance covered across the entire match
561 for SSGwith and SSGwithout, respectively) across 133 different (90-minavg) for each player and metric was collected. To assist in
formats of SSGs (99 and 34 for SSGwith and SSGwithout, re- the development of velocity-based movement indicators, rolling
spectively) were undertaken. Small-sided games with ranges from moving averages were calculated across 4-min time duration (4-
4 vs. 4 to 10 vs. 10 with a ApP from 38–487 m2; SSGwithout ranged minPeak) for each player across each match with the maximum
from 4 vs. 4 to 10 vs. 10 with a ApP from 33.3 to 150 m2. A value collected at each metric (39). The rolling moving average
detailed description of the SSGs characteristics and the average across 4-min time duration was calculated across 1-minute time-
locomotor demands are reported in Supplemental Digital Con- windows analysis determined during official matches throughout
tent (see Tables 1 and 2, http://links.lww.com/JSCR/A318) for a specifically developed spreadsheet. The 4-minPeak and 90-
SSGwith and SSGwithout, respectively. The ApP was calculated minavg for TD, HSRD, VHSRD, sprint, and acceleration/
excluding the goalkeepers in SSGwith. Each SSGwith or SSGwithout deceleration distances were normalized as the relative distance
lasted 4-min in duration. The SSGwith or SSGwithout were per- covered in one minute (m·min21) and inserted into the data
formed under the supervision and motivation of several coaches analysis.
to keep up a high work-rate. For the same reason, a ball was To determine the ApP that replicates the normalized TD,
always available by prompt replacement when it went out-of-play HSRD, VHSRD, sprint, and acceleration/deceleration
(19). In SSGwith, the corners were replaced by a prompt ball-in- (m·min21) recorded during the 90-minavg and/or the 4-
game from the goalkeeper. Both SSGwith or SSGwithout were minPeak, we first recorded these variables during the official
completed after a standardized 20-minute warm-up under the matches. Thereafter, we separately plotted each relationship
guidance of club staff. between ApP and the normalized TD, HSRD, VHSRD, sprint,

Copyright © 2022 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Area Per Player to Replicate 4-minPeak (2022) 00:00 | www.nsca.com

and acceleration/deceleration during SSGs. Then, the 90- presented for SSGwith and SSGwithout. To determine the individual
minavg or 4-minPeak values recorded during the official ApP to replicate the desired locomotor demands, the random
matches were used to intersect each ApP/TD, HSRD, VHSRD, effects parameters should be considered for each player as follow:
sprint, and acceleration/deceleration relationship recorded in y 5 ß0 fixed 1 ß1 fixed·xplayer 1 ß0 random 1 ß1 random·xplayer, where y
SSGwith and SSGwithout to calculate the forecasted ApP that is the locomotor demands during SSGs, x is the ApP, ß0 is the
corresponded to the 90-minavg or 4-minPeak official match intercept, and ß1 is the regression coefficient.
demands, as previously proposed (34,37). Moreover, using the
same procedures, we provided the information required to
determine the equation to calculate the ApP depending on the The Area Per Player in Small-Sided Games With vs. Small-
desired locomotor demands or ApP for TD, HSRD, VHSRD, Sided Games Without to Replicate 4-minPeak
sprint, and acceleration/deceleration. Table 1 shows the ApP to replicate 4-minPeak for SSGwith and
SSGwithout. The magnitude of the between-drills differences (p ,
0.001) is also reported. Some between-position differences (p ,
Statistical Analysis
0.05) in the ApP to replicate locomotor match demands for
SPSS (version 26, Chicago, IL) was used to perform the statistical SSGwith (ES: 0.22–1.10) and SSGwithout (ES: 0.36–1.49) were
analysis. To check the normal distribution of the sampling, a found and reported in Figure 3.
Shapiro-Wilk test was used. A linear mixed model analysis was Comparing with the ApP to replicate 90-minavg, the ApP to
used to calculate the individual relationship between TD, HSRD, replicate 4-minPeak was larger (p , 0.001; ES: 1.03–3.30) in
VHSRD, sprint, acceleration/deceleration, and the ApP during SSGwith for TD, HSRD, VHSRD, and sprint; conversely, larger
SSGwith and SSGwithout. The correlation coefficient between ApP ApP to replicate acceleration/deceleration in 90-minavg than 4-
and TD, HSRD, VHSRD, sprint, and acceleration/deceleration in minPeak was found (p , 0.05; ES: 21.32, CI: 23.12 to 0.08). For
SSGwith and SSGwithout was calculated and interpreted as follows: SSGwithout, the ApP to replicate 4-minPeak was larger (p , 0.001;
r 5 0.00–0.09 trivial, 0.10–0.29 small, 0.30–0.49 moderate, ES: 0.41–1.88) than the ApP required to mimic 90-minavg for
0.50–0.69 large, 0.70–0.89 very large, and 0.90–0.99 nearly HSRD, VHSRD, and sprint.
perfect. Thereafter, a linear mixed model analysis was used to
calculate the difference in the minimal ApP in TD, HSRD,
VHSRD, sprint, and acceleration/deceleration calculated across Discussion
each position. Cohen’s d effect size with 95% confidence intervals
(CI) was used to describe the magnitude of the pairwise differ- The current study determined the ApP to replicate the 4-minPeak
ences and interpreted as follows: ,0.20: trivial; 0.20–0.59: small; using SSGwith and SSGwithout. To increase the locomotor demands
0.60–1.19: moderate; 1.20–1.99: large; and $2.00: very large. for TD, HSRD, VHSRD, and sprint, the ApP should be enlarged
Statistical significance was set at a , 0.05. Unless otherwise during both SSGwith and SSGwithout, whereas acceleration/
stated, all values are presented as mean 6 SD. deceleration is quite similar across different ApP. The ApP to
replicate the 4-minPeak was larger, the higher the speed threshold
considered. Interestingly, the present novel findings modelled a
Results specific ApP to recreate the 4-minPeak using both SSGwith (ApP
;245, ;349, ;367, ;541 and ;121 m2·player21 for TD,
In SSGwith, to assess the individual relationship between ApP and HSRD, VHSRD, sprint, and acceleration/deceleration, re-
locomotor demands, the estimates of covariance parameters jus- spectively) and SSGwithout (i.e., ;90, ;142, ;186, ;366 and
tified the use of linear mixed model for both intercept and slope ;76 m2·player21, respectively). Moreover, some between-
only in TD (p , 0.05); in addition, the use of linear mixed model position differences in the ApP to replicate 4-minPeak were
was justified for slope in HSRD (p 5 0.029) and sprint (p 5 found, albeit no specific trend was observed.
0.012). In SSGwithout, linear mixed model was justified for in- Increments in the ApP lead to higher locomotor demand for
tercept in TD (p 5 0.047) and acceleration/deceleration (p 5 TD, HSRD, VHSRD, and sprint during SSGwith and SSGwithout.
0.012). Otherwise, linear model with ApP and locomotor de- These findings confirm previous results in which a correlation
mands as fixed-effect was used. between ApP and TD, HSRD, and sprint was found during both
SSGwith and SSGwithout in elite Serie A soccer players (34). In line
with these findings, higher HSRD and sprint were reported when
Correlations Between Area Per Player and
incrementing the ApP during 5-, 6- and 9-a-side SSGwith in second
Locomotor Demands
division Portuguese football players (9) or during 5-, 7-, 10-a-side
Figure 1 shows the individual relationship between ApP and lo- SSGwith and SSGwithout in English Premier League (16) football
comotor demands for both SSGwith and SSGwithout. The correla- players. The current results further highlighted that larger pitch
tion between ApP and locomotor demands for the estimates of size is required to increase the locomotor demands especially for
fixed effects for both SSGwith and SSGwithout are reported in the high-speed running. Conversely, acceleration/deceleration
Figure 1. In SSGwith, large-to-very large correlations for TD, remained quite unchanged across the different ApP, as reported in
HSRD, VHSRD, and sprint were found; conversely, acceleration/ Portuguese (9), Italian (34), and English (16) football players.
deceleration showed inversely small correlation with ApP. In Therefore, practitioners should use larger ApP to increase HSRD,
SSGwithout, small-to-large correlations for TD, HSRD, VHSRD, VHSRD, and sprint, whereas acceleration/deceleration seems
and sprint were found; conversely, acceleration/deceleration only slightly affected by the ApP manipulation and quite consis-
showed inversely trivial correlation with ApP. tent across different ApP.
Figure 2 shows the individual relationship between ApP and For the first time here, a specific ApP to recreate the 4-minPeak
the locomotor demands express as percentages than the in- was reported. Compared with SSGwith (ApP ;245, ;349, ;367,
dividual 4-minPeak determined during official matches. Data are ;541 and ;121 m2·player21 for TD, HSRD, VHSRD, sprint,

Copyright © 2022 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Area Per Player to Replicate 4-minPeak (2022) 00:00

Figure 1. The linear mixed model analysis for the individual (gray line) and pooled (black line) relationship between
area per player (m2·player21) and the relative locomotor demands (m·min21) during small-sided games with (SSGwith)
and without (SSGwithout) goalkeepers is shown. The correlation of the estimates of fixed effects between the area per
player and the relative locomotor demands are shown for total distance (panel A and panel F for SSGwith and
SSGwithout, respectively), high-speed running (panel B and panel G, respectively), very high-speed running (panel C
and panel H, respectively), sprint (panel D and panel I, respectively), and acceleration/deceleration (panel E and panel
L, respectively).

Copyright © 2022 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Area Per Player to Replicate 4-minPeak (2022) 00:00 | www.nsca.com

Figure 2. The linear mixed model analysis for the individual (gray line) and pooled (black line) relationship between
area per player (m2·player21) and the relative locomotor demands expressed as percentage than official match 4-
minPeak is shown. Data are presented for small-sided games with (SSGwith) and without (SSGwithout) goalkeepers.
The gray area highlights the 610% than the 4-minPeak demands determined during official matches. Total distance:
panel A and panel F for SSGwith and SSGwithout, respectively; high-speed running: panel B and panel G, re-
spectively; very high-speed running: panel C and panel H, respectively; sprint: panel D and panel I, respectively;
and acceleration/deceleration: panel E and panel L, respectively.

Copyright © 2022 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Area Per Player to Replicate 4-minPeak (2022) 00:00

Table 1
The minimal area per player (m2·player21) to replicate 4-minPeak locomotor match demands using small-sided games with (SSGwith) or
without (SSGwithout) goalkeepers is reported across different metrics.*
SSGwith SSGwithout ES (95% CI)
Total distance 246 6 104† 90 6 22 3.58 (1.67–6.00)
High-speed running 350 6 129† 143 6 34 2.93 (0.80–5.06)
Very high-speed running 367 6 88† 187 6 101 2.17 (0.26–4.08)
Sprint 541 6 145† 366 6 124 1.31 (0.39–2.23)
Acceleration/deceleration 121 6 47† 77 6 26 1.84 (1.06–2.62)
*Data are reported as mean (SD). Effect size (ES) and 95% confidence intervals (CI) between SSGwith and SSGwithout are also reported.
†p , 0.05 vs. SSGwithout.

and acceleration/deceleration, respectively), lower ApP in (4,13). In the training practice, this may limit the stimulus and
SSGwithout was calculated (i.e., ;90, ;142, ;186, ;366 and this extremely great ApP could seem prohibitive within a team
;76 m2·player21). This was probably because of the greater training routine, so that SSGwithout (i.e., that require an ApP
tactical organization during SSGwith that lowered the locomotor similar to the relative pitch-size during official matches) or
demands as previously suggested, thus requiring a greater ApP to supplemental running-based exercises should be used when
increase the intensity of the activities (34). These latter findings aiming to condition the 4-minPeak match demands for sprinting
support previous studies in a similar elite football population activities. As an alternative, during SSGwith, supplemental
(34). In addition, higher high-intensity running was previously rules such as specific timeframes (i.e., 10–20 seconds) in which
observed without goalkeepers in different 3-, 4-, 5-, and 7-a-side possession-play should be maintained or goals should be
SSGs across the same ApP (6,24,26). Partially in contrast with the achieved and/or other specific rules modifications (19) may be
present outcomes, it was reported that SSGwith resulted in higher used to increase the locomotor demands. In contrast,
sprint than in SSGwithout (16). However, the authors investigated SSGwithout could negatively affect the tactical learning envi-
a maximum ApP of 135 m2, and this does not allow an appro- ronments, and this factor should be considered in the training-
priate comparison. Other researchers reported that TD and the load management for football-specific performance de-
time spent in high-intensity running (.17 km·h21) was higher velopment. Moreover, SSGwith in large ApP was reported as
with goalkeepers (11). Although the authors argued that the useful to recreate the tactical variability for attacking explo-
goalkeeper presence might have motivated the players, 2 sub- ration and defending organization (31). Therefore, SSGwith in
sequent reviews (1,19) consistently remarked that the goalkeeper large ApP seem to be a good solution to contextualize the lo-
presence could improve the players’ organization, thus decreasing comotor demands (2), physiological responses (5), and tactical
the SSGs demands. Therefore, because the SSGs intensity is a behaviors (30) regarding maximal match demands, whereas
crucial issue (17,25,34) for high-performance development, the SSGwithout permits to recreate similar locomotor demands us-
present findings remarked that practitioners should use a specific ing smaller ApP when required.
large ApP when increments in locomotor demands are required, The present findings also showed some between-position
greater using SSGwith than SSGwithout. differences in the ApP to replicate 4-minPeak using SSGwith
The maximal match-play intensity was recently contextual- (i.e., without a clear trend), suggesting that an individualized
ized across the training microcycles, suggesting that practi- training-load manipulation may be needed to overload in-
tioners need to condition football players by increasing the dividual peak performance demands. Given that the use of dif-
training workloads up to the maximal demands of the compe- ferent positional SSGs is not feasible in real-life training routine,
tition (29). The distribution of the official match demands with practitioners should consider to overload the locomotor de-
respect to the maximal intensity provides a practical tool to mands using individual supplemental activities when required.
support training load management and training prescriptions Individual positional-drills, running-based exercises, and/or
(36). Therefore, because 4-min is largely used as typical duration specific rules could be considered as supplemental tools to in-
for the SSGs within the typical weekly routine (19,20), the 4- dividualize the training exposure near to individual maximal
minPeak could be an effective reference for intermittent SSG match-derived locomotor demands. Between-position differ-
prescriptions. The current findings also showed for the first ences during SSGwith were already reported in similar pop-
time, a higher ApP to replicate the 4-minPeak than 90-minavg in ulation (25,34). In contrast, because few between-position
SSGwith and SSGwithout, providing practical information for differences were found in SSGwithout, this could allow to simul-
stimulating the maximal demands of the competition. More- taneously condition each player up to the 4-minPeak match-play
over, the present results report that the relative high-intensity demands reducing between-player variability in the locomotor
locomotor demands during most SSGs formats are far lower that demands. This was probably because of lower tactical-
official match demands across the same time-window. In- organization during SSGwithout, in which players are free to
terestingly, only 2 players (i.e., a central-defender and a central- move to keep or regain ball-possession without tactical restric-
midfielders) achieved or overloaded the maximal 4-minpeak us- tions that may have affected the between-position overloading
ing SSGwith, suggesting that an individual approach is manda- or underloading.
torily required during the field practice. Some limitations accompany the present investigation, also
That said, the ApP in SSGwith exceeded the match-derived opening to future research perspectives. First, the internal load
ApP (;340 m2·player21) to replicate the 4-minPeak sprint de- parameters (i.e., heart rate) and the rate of perceived exertion
mands. This may possibly derive from the lower psycho- were not examined, and we acknowledge that should be coupled
physiological stressors during training than official matches with the external load metrics to describe accurately both the

Copyright © 2022 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Area Per Player to Replicate 4-minPeak (2022) 00:00 | www.nsca.com

Figure 3. The minimal area per player (m2·player21) to replicate the 4-minPeak lo-
comotor match demands using small-sided games with (SSGwith, panel A) or
without (SSGwithout, panel B) goalkeepers is reported for each metric across different
positions. Data are reported as mean (SD). TD 5 total distance; HSRD 5 high-
speed running distance; VHSRD 5 very high-speed running distance; SPD 5 sprint
distance; Acc + Dec 5 acceleration/deceleration distance; CD 5 central defenders;
WM 5 wide midfielders; CM 5 central midfielders; WF 5 wide forwards; FW 5
forwards. ap , 0.05 vs. WM. bp , 0.05 vs. CD.

SSGs and match demands (22). However, some technological stimulus, individual exercises based on submaximal (38) and/or
limitations (e.g., the use of portable thoracic bands especially maximal physiological capacities (33,35) should be included.
during official matches) or some contextual limitations (e.g., the Finally, for replication purposes, the interchangeability between
possibility to collect rate of perceived exertion after each SSG global-positioning system and computerized semi-automated
format during daily real-life training routine) can limit the op- video-based multicamera image system needs to be carefully
portunity to monitor consistently the internal and perceived load. checked, especially when recording high-speed or nonlinear
Second, as both SSGs and running-based exercise may be equally movements. However, the present results are based on the trivial
effective to enhance aerobic training status (21), it should be differences in the metrics and adjusted using a calibration equa-
remarked that soccer-specific drills only could insufficiently pre- tion as previously investigated using the same technologies (34).
pare players for the match demands. This may be because of the In conclusion, the present study determined the ApP to repli-
individual capacity that may exceed the actual stimuli received for cate the 4-minPeak using SSGwith and SSGwithout across different
some positions (e.g., central defenders). In case of nonsufficient metrics and positions. For SSGwith and SSGwithout, high-intensity

Copyright © 2022 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Area Per Player to Replicate 4-minPeak (2022) 00:00

activities required greater ApP, with higher values in SSGwith vs. players to pitch area on internal and external load in soccer players. Int J
SSGwithout. Finally, between-position differences in ApP across Sports Physiol Perform 14: 1265–1272, 2019.
6. Castellano J, Casamichana D, Dellal A. Influence of game format and
each metric were observed. number of players on heart rate responses and physical demands in small-
sided soccer games. J Strength Cond Res 27: 1295–1303, 2013.
7. Castellano J, Martin-Garcia A, Casamichana D. Most running demand
Practical Applications passages of match play in youth soccer congestion period. Biol Sport 37:
367–373, 2020.
The present findings have a number of practical applications. 8. Clemente FM, Afonso J, Sarmento H. Small-sided games: An umbrella
review of systematic reviews and meta-analyses. PLoS One 16: 1–22,
In the first instance, the specific ApP can be used in 4-min 2021.
SSGs to replicate, underload, or overload the 4-minPeak, as 9. Clemente FM, Sarmento H, Rabbani A, et al. Variations of external load
also previously suggested for the 90-minavg (34). The present variables between medium- and large-sided soccer games in professional
findings also highlighted that an ApP ;350 m2·player21 players. Res Sports Med 27: 50–59, 2019.
using SSGwith should be used to recreate the 4-minPeak at a 10. Delaney JA, Thornton HR, Pryor JF, et al. Peak running intensity of in-
ternational rugby: Implications for training prescription. Int J Sports
similar extent to match-play demands for high-speed to very- Physiol Perform 12: 1039–1045, 2017.
high-speed activities. Interestingly, this ApP was similar to 11. Dellal A, Chamari K, Pintus A, et al. Heart rate responses during small-
the ApP during official match-play (;350 m2·player21), sided games and short intermittent running training in elite soccer players:
suggesting that a consistency between the ApP during A comparative study. J Strength Cond Res 22: 1449–1457, 2008.
12. Di Salvo V, Baron R, Tschan H, et al. Performance characteristics
SSGwith and during official matches is required to replicate according to playing position in elite soccer. Int J Sports Med 28:
the 4-minPeak. However, a greater pitch size 222–227, 2007.
(.350 m2·player21) or supplementary exercises would be 13. Dixon JG, Jones MV, Turner MJ. The benefits of a challenge approach on
required to replicate 4-minPeak for sprint using SSGwith. match day: Investigating cardiovascular reactivity in professional acad-
When lower ApP is used, practitioners should consider high- emy soccer players. Eur J Sport Sci 20: 375–385, 2020.
14. Fanchini M, Azzalin A, Castagna C, et al. Effect of bout duration on
speed to sprinting supplementations using specific rules exercise intensity and technical performance of small-sided games in
during SSGs, individualized positional drills and/or running- soccer. J Strength Cond Res 25: 453–458, 2011.
based exercises to prepare the players. Another possible 15. Ferraday K, Hills SP, Russell M, et al. A comparison of rolling averages
football-specific solution may be to use the SSGwithout that versus discrete time epochs for assessing the worst-case scenario loco-
motor demands of professional soccer match-play. J Sci Med Sport 23:
required a lower ApP (.186 m2·player21) to replicate 4- 764–769, 2020.
minPeak for VHSRD to sprinting activities. However, the 16. Gaudino P, Alberti G, Iaia FM. Estimated metabolic and mechanical de-
different tactical behavior suggests that practitioners should mands during different small-sided games in elite soccer players. Hum
use SSGwith and SSGwithout for different technical/tactical Mov Sci 36: 123–133, 2014.
training aims during practice. In addition, using SSGwith, 17. Halouani J, Chtourou H, Gabbett T, Chaouachi A, Chamari K. Small-
sided games in team sports training: A brief review. J Strength Cond Res
between-position differences highlighted that supplementary 28: 3594–3618, 2014.
high-speed to sprint activities should be considered for 18. Hill-Haas SV, Coutts AJ, Dawson BT, Rowsell GJ. Time-motion char-
overloading such players. This would not be the case for acteristics and physiological responses of small-sided games in elite youth
SSGwithout, during which no between-position difference was players: The influence of player number and rule changes. J Strength Cond
Res 24: 2149–2156, 2010.
found. Therefore, practitioners could manage the ApP using 19. Hill-Haas SV, Dawson B, Impellizzeri FM, Coutts AJ. Physiology of small-
both SSGwith, SSGwithout, and/or supplementary solutions to sided games training in football: A systematic review. Sports Med 41:
plan the whole-session training load during practice, in ac- 199–220, 2011.
cordance with the match intensity distribution for each 20. Hill-Haas SV, Rowsell GJ, Dawson BT, Coutts AJ. Acute physiolog-
metric (36). ical responses and time-motion characteristics of two small-sided
training regimes in youth soccer players. J Strength Cond Res 23:
111–115, 2009.
21. Impellizzeri FM, Marcora SM, Castagna C, et al. Physiological and per-
Acknowledgments formance effects of generic versus specific aerobic training in soccer
players. Int J Sports Med 27: 483–492, 2006.
The authors wish to thank all the subjects of the study for their 22. Impellizzeri FM, Marcora SM, Coutts AJ. Internal and external
committed effort and Mattia Cefis, PhD student at University of training load: 15 years on. Int J Sports Physiol Perform 14: 270–273,
Brescia, for his committed help in the statistical analysis. The 2019.
authors have no conflicts of interest to disclosure. No specific 23. Kelly DM, Drust B. The effect of pitch dimensions on heart rate responses
and technical demands of small-sided soccer games in elite players. J Sci
grant supports the present study and the results do not constitute Med Sport 12: 475–479, 2009.
endorsement of the product by the authors or the NSCA. 24. Köklü Y, Sert Ö, Alemdaroğlu U, Arslan Y. Comparison of the physio-
logical responses and time-motion characteristics of young soccer players
in small-sided games: The effect of goalkeeper. J Strength Cond Res 29:
References
964–971, 2015.
1. Aguiar M, Botelho G, Lago C, Maças V, Sampaio J. A review on the effects 25. Lacome M, Simpson BM, Cholley Y, Lambert P, Buchheit M. Small-sided
of soccer small-sided games. J Hum Kinet 33: 103–113, 2012. games in elite soccer: Does one size fit all? Int J Sports Physiol Perform 13:
2. Bradley PS, Ade JD. Are current physical match performance metrics in 568–576, 2018.
elite soccer fit for purpose or is the adoption of an integrated approach 26. Mallo J, Navarro E. Physical load imposed on soccer players during
needed? Int J Sports Physiol Perform 13: 656–664, 2018. small-sided training games. J Sports Med Phys Fitness 48: 166–171,
3. Casamichana D, Castellano J, Diaz AG, Gabbett TJ, Martin-Garcia A. 2008.
The most demanding passages of play in football competition: A com- 27. Martin-Garcia A, Casamichana D, Diaz AG, Cos F, Gabbett TJ. Posi-
parison between halves. Biol Sport 36: 233–240, 2019. tional differences in the most demanding passages of play in football
4. Casanova N, Palmeira-DE-Oliveira A, Pereira A, et al. Cortisol, testos- competition. J Sports Sci Med 17: 563–570, 2018.
terone and mood state variation during an official female football com- 28. Martin-Garcia A, Gomez Diaz A, Bradley PS, Morera F, Casa-
petition. J Sports Med Phys Fitness 56: 775–781, 2016. michana D. Quantification of a professional football team’s external
5. Castagna C, D’Ottavio S, Cappelli S, Araújo Póvoas SC. The effects of load using a microcycle structure. J Strength Cond Res 32:
long sprint ability-oriented small-sided games using different ratios of 3511–3518, 2018.

Copyright © 2022 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Area Per Player to Replicate 4-minPeak (2022) 00:00 | www.nsca.com

29. Oliva-Lozano JM, Gomez-Carmona CD, Rojas-Valverde D, Fortes V, 34. Riboli A, Coratella G, Rampichini S, Cé E, Esposito F. Area per player in
Pino-Ortega J. Effect of training day, match, and length of the microcycle small-sided games to replicate the external load and estimated physio-
on the worst-case scenarios in professional soccer players. Res Sports Med: logical match demands in elite soccer players. PLoS One 15: 1–15, 2020.
1–14, 2021. Epub ahead of print. 35. Riboli A, Coratella G, Rampichini S, Limonta E, Esposito F. Testing
30. Olthof SBH, Frencken WGP, Lemmink KAPM. Match-derived relative protocol affects the velocity at VO2max in semi-professional soccer
pitch area changes the physical and team tactical performance of elite players. Res Sports Med 30: 1–11, 2021.
soccer players in small-sided soccer games. J Sports Sci 36: 1557–1563, 36. Riboli A, Esposito F, Coratella G. The distribution of match activities
2018. relative to the maximal intensities in elite soccer players: Implications for
31. Olthof SBH, Frencken WGP, Lemmink KAPM. Match-derived rela- practice. Res Sports Med: 1–12, 2021. Epub ahead of print.
tive pitch area facilitates the tactical representativeness of small-sided 37. Riboli A, Olthof SBH, Esposito F, Coratella G. Training elite youth soccer
games for the official soccer match. J Strength Cond Res 33: 523–530, players: Area per playe in small-sided games to replicate the match de-
2019. mands. Biol Sport 39: 579–598, 2022.
32. Owen AL, Wong DP, Paul D, Dellal A. Effects of a periodized small-sided 38. Riboli A, Rampichini S, Cè E, et al. Effect of ramp slope on different
game training intervention on physical performance in elite professional methods to determine lactate threshold in semi-professional soccer play-
soccer. J Strength Cond Res 26: 2748–2754, 2012. ers. Res Sports Med 27: 326–338, 2019.
33. Riboli A, Ce E, Rampichini S, et al. Comparison between continuous and 39. Riboli A, Semeria M, Coratella G, Esposito F. Effect of formation, ball in
discontinuous incremental treadmill test to assess velocity at VO2max. play and ball possession on peak demands in elite soccer. Biol Sport 38:
J Sports Med Phys Fitness 57: 1119–1125, 2017. 195–205, 2021.

Copyright © 2022 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
View publication stats