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

net/publication/357351648

Contextualised peak periods of play in English Premier League matches

Article in Biology of Sport · December 2021

DOI: 10.5114/biolsport.2022.112083

CITATIONS READS

0 931

9 authors, including:

Wonwoo Ju Dominic Doran

Liverpool John Moores University Liverpool John Moores University
6 PUBLICATIONS 1 CITATION 85 PUBLICATIONS 2,011 CITATIONS

SEE PROFILE SEE PROFILE

Antonio José Gómez-Díaz Andrés Martín García

FC Barcelona FC BARCELONA
41 PUBLICATIONS 1,134 CITATIONS 17 PUBLICATIONS 229 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Monitoring Metabolic Power and Training Load in Elite Soccer. View project

CONTEXT & SCIENCE IN ELITE FOOTBALL View project

All content following this page was uploaded by Wonwoo Ju on 27 December 2021.

The user has requested enhancement of the downloaded file.

 Contextualised
Original Paper peak periods of play DOI: https://doi.org/10.5114/biolsport.2022.112083

Contextualised peak periods of play in English Premier League

matches
AUTHORS: Wonwoo Ju1, Dominic Doran1, Richard Hawkins2, Antonio Gómez-Díaz3, Andres Corresponding author:
Martin-Garcia4, Jack D Ade5, Andy Laws6, Mark Evans6, and Paul S Bradley7 Wonwoo Ju
The Research Institution for
Sport and Exercise Sciences at
1
The Research Institution for Sport and Exercise Sciences at Liverpool John Moores University, Liverpool, England, UK Liverpool John Moores University,
2 Liverpool, England, UK
Football Medicine and Science Department at Manchester United Football Club, Manchester, UK
3 E-mail: [email protected]
The Poland National Football Team, Polish Football Association, Warsaw, Poland
4
FC Barcelona Sports Performance Department, Barcelona, Spain
5 ORCID:
Liverpool FC Academy, Liverpool Football Club, Liverpool, UK
6 Wonwoo Ju
Department of Computer Science, Liverpool John Moores University, Liverpool, UK 0000-0003-3860-0178
7
Football Science Consultant, UK
Dominic Doran
0000-0002-4753-4195
ABSTRACT: The present study aimed to determine the physical-tactical trends of elite players/teams during
Richard Hawkins
peak 1-, 3- and 5-min periods of match-play. A total of 50 English Premier League matches (n = 583 player
0000-0002-5339-0394
observations) were analysed by coding the players’ physical-tactical activities through the synchronisation of
tracking data and video. The contextualised data showed that during the peak periods (i.e., the most demanding Andrés Martín-García
0000-0002-7341-4664
passage of play), players/teams covered the largest distances for ‘Recovery Run’ (28–37%) out of possession
and ‘Support Play’ (9–13%) in possession. In the following periods, players covered less high-intensity distance Antonio Gómez-Díaz
0000-0001-7151-6133
versus the average with a more pronounced decline in the next 1-min period than longer duration 3- and 5-min
periods (48% vs ~25–30%, ES: 0.4–0.5, P < 0.01); team data showed similar trends with different relative Jack Ade
0000-0002-3627-9256
patterns (31% vs 17–30%, ES: 0.5–0.8, P < 0.01). After peak periods, players/teams performed 20–53% less
high-intensity distances for ‘Covering’ and ‘Recovery Run’ (ES: 0.2–0.7, P < 0.01) out of possession. However, Andy Laws
0000-0002-5469-4322
players covered 28–91% less distance for ‘Run with Ball’ (ES: 0.1–0.5, P < 0.05) when in possession. Some
Paul S Bradley
physical-tactical actions exhibited inconsistency in different time durations of the next periods; however, these
0000-0003-4399-1759
physical-tactical data were position-specific. This may signify that each position has certain physical-tactical
actions to execute even after the peak periods, especially when they are tactically required to do so. As the
data demonstrates unique physical-tactical trends of players/teams during the peak and next periods of play,
this could help practitioners prescribe position- and player-specific drills, and better understand transient
decrements in high-intensity running after intense passages of play.

CITATION: Ju W, Doran D, Hawkins R et al. Contextualised peak periods of play in English Premier League Key words:
matches. Biol Sport. 2022;39(4):973–983. Match performance
Peak periods
Transient decrements
Physical-tactical data
Received: 2021-09-29; Reviewed: 2021-11-20; Re-submitted: 2021-11-20; Accepted: 2021-12-05; Published: 2021-12-30. Football

INTRODUCTION
Time-motion analysis has been widely used for profiling the match coaches prescribe specific drills that mimic these intensified periods
performance of elite players [1–4]. Practically, the activity profiles of matches more effectively.
derived from match-play are used for designing training drills [5, 6]. Football (soccer) is a team sport where player’s physical and
However, previous studies have mainly analysed the average physi- tactical actions are influenced by both opponent and teammate’s
cal demands [3, 7, 8], which underestimates locomotive demands actions [15]. Nevertheless, previous studies included only player
of players [9]. Hence, greater attention has been paid to the physical performances to understand individual patterns rather than actual
demands during the peak period (i.e., the most intense period of team trends [11, 13, 17]. Quantifying individual players limits our
a match) [9–13]. Although peak performance data have been prac- understanding of a team’s collective performance during match-play.
tically used as a benchmark to devise football-specific drills [6, 14], Thus, analysing team trends during intensified periods of play could
issues exist when attempting to directly translate these into specific provide new insights and help with team-based drill prescrip-
drills as the context of play is completely omitted from any of the tions [5, 14]. No research, to the best of our knowledge, has at-
studies that have quantified match-play peak periods [15, 16]. Thus, tempted to observe the peak physical demands for team perfor-
tactical context should be fused with physical metrics to help mances. Therefore, analysing team’s collective physical-tactical

Biology of Sport, Vol. 39 No4, 2022 973

 Wonwoo Ju et al.

performances could add insights into how teams collectively perform perfect (κ = 0.94), respectively. The novel filter isolated high-inten-
physical-tactical actions during intensified periods of play together sity activities reaching speeds > 19.8 km · h−1 for a minimal dwell
with individual player data. time of 1 s [24].
Several studies have investigated not only peak periods of play The researcher completed 350 hours of coding to analyse 50 com-
over different time durations (e.g., 1-, 3-, 5-min) but the 5-min petitive matches and 1,265 player observations within 20 different
periods after intense periods during match-play to examine transient teams. For individual player’s analysis, only outfield players who had
decrements in high-intensity running compared to the match aver- completed the entire match in the same position were included
age [1, 3, 18]. The immediate declines in physical performance (583 player observations). This consisted of 179 Central Defensive
during the next 5-min periods have been ascribed to fatigue induced players (CDP), 147 Wide Defensive players (WDP), 167 Central
by the activities during peak periods. Although it is highly complex, Midfield players (CMP), 54 Wide Offensive players (WOP), and
there seems to be several contributing factors that cause fatigue (e.g., 36 Central Offensive players (COP). However, all of the player’s
muscle acidosis and reduced muscle creatine phosphate) after intense contextualised performances for each match were summarised to
periods [19]. However, temporary declines in high-intensity running analyse team performances (players who were subbed in or out were
are not necessarily linked to fatigue but could be due to pacing included; 100 match observations). All data were analysed for the
strategies/tactical alterations [20] and/or less playing opportuni- duration of each half, including stoppage time. Prior to analysis, all
ties [21]. To potentially understand ‘HOW’ players/teams alter their original data were anonymised to ensure confidentiality. Ethical ap-
tactical behaviour during the phases that follow intense periods of proval was granted by Liverpool John Moores University (LJMU)
match-play, amalgamating physical and tactical performance data research ethics committee.
could be a solution [22].
Previous research examining transient decrements in high-inten- Match Control and Data Balance
sity running in the next period after the most demanding passage of To improve the scientific rigor of the research design, matches were
a match had several limitations [1–3, 18]. Most studies used a pre- arbitrarily selected while simultaneously controlling situational factors
defined period (e.g., 0–5, 5–10 min etc.), which can under or over- (e.g., team/opponent standards, locations, and seasonal phases) [25].
estimate the physical demands during the peak and the following Thus, the number of matches for each parameter was initially bal-
periods, respectively [23]. Thus, it is more advisable to use a rolling anced. Matches were excluded if goal differential was > 3 and
average technique (distance covered from every time point) to provide a player dismissal occurred since these influence match running
a more precise estimation of physical demands during such peri- performances [20, 26].
ods [11, 13, 17]. Moreover, studies investigating transient decrements
used only a 5-min interval for the next period after the most intense The Integrated Approach of Match Performance
period of play, which could omit brief changes immediately after High-intensity actions isolated by the novel filter were synchronised
intense actions [2, 18]. Hence, using shorter durations of the next with video footage of all players throughout matches to code the
period after the most intense passage of play may be more advanta- tactical purpose of each action (Table 1). All coding occurred using
geous to understand short-term fluctuations. Therefore, the present QuickTime Player (Apple Inc, Cupertino, California) to watch video
study aimed to determine the physical-tactical profiles of elite play- and then categorise tactical actions.
ers/teams during peak 1-, 3- and 5-min periods of high-intensity The coding process was as follows: high-intensity actions with
running and the subsequent periods of each time duration during one tactical action were classified as a single action with dual tacti-
match-play. cal actions being coded as a hybrid action. High-intensity actions
with more than three tactical actions were classified as ‘Other’. If the
MATERIALS AND METHODS high-intensity action consisted of 70–90% of the primary and
Match Analysis and Player/Team Data 10–30% of the secondary action, it was classified as a hybrid action.
Match physical-tactical data were collated from the 2018–19 Eng- But if it was made up of 50–60% of the primary and 40–50% of
lish Premier League season using an integrated approach and a new the secondary action, then it was classified as ‘Other’. As hybrid
filter established for this research. Players’ behaviours were captured actions are a combination of the primary and secondary actions [27],
by cameras situated at roof level during matches and their physical- single action events and the primary tactical movements of the hybrid
tactical actions were manually coded using the integrated approach. actions were combined to simplify data outputs.
The validity and reliability of the integrated approach and the novel
filter used were previously verified by Ju et al. [22]. The validity of Physical-Tactical Performance for The Peak Period
the integrated approach demonstrated a strong agreement between Using a rolling average method, the peak periods of high-intensity
the responses of both UEFA qualified coaches and performance running during matches for three different time durations (1-, 3-, and
analysts versus the gold standard responses (~92%), and its inter- 5-min) were determined [11]. These durations were selected firstly,
and intra-observer reliability was a strong (κ = 0.81) to almost to facilitate a more detailed examination of temporal changes than

974
 Contextualised peak periods of play

TABLE 1. The descriptions of the variables within the integrated approach.

Variables Description
In Possession

Push up Pitch Player moves up the pitch to play offside and/or to squeeze to a higher line.
Break into Box Player enters the opposition’s penalty box to receive the ball (typically receive ball from a cross – ball in
front and wide).
Run in Behind/Penetrate Player attacks space behind, overtakes and/or unbalances the opposition defence (typically ball is behind).
Over/Underlap Player runs from behind to in front of the player on the ball or receiving the ball.
Run with Ball Player moves with the ball either dribbling with small touches or running at speed with fewer ball touches.
Move to Receive/Exploit Player moves to receive a pass from a teammate or to create/exploit space (typically come short or move
Space wide to receive ball).
Support Play Player supports from behind/level by trying to engage in offensive/transition play (typically during fast
transitions).
Out of Possession

Interception Player cuts out pass.

Recovery Run Player runs back towards their own goal to be goal side when out of position.
Covering Player moves to cover space or an opposition player while remaining goal side.
Close Down/Press Player runs directly towards opposition player on or receiving the ball,
or towards space or players not on/receiving the ball.
Unclassifiable
Other All other variables that could not be categorised by the above.

using only a 5-min time interval [18, 28], and secondly to correspond RESULTS
with the typical duration of training drills [17]. The next period after Contextualised Peak Periods – Individual Trends
the peak of each time duration was used to evaluate physical perfor- During the peak 1-, 3- and 5-min periods, the players covered
mance decrements by comparing them with the average of the match 28–34% and 22–25% of the high-intensity distance (67 ± 19 m,
that the period occurred in [3]. In addition, this allowed exploration 92 ± 28 m, and 113 ± 36 m, respectively) performing ‘Recovery
of how players/teams changed their tactical behaviour after intense Run’ and ‘Covering’ actions, respectively. In possession, the largest
periods of play. The mean distances of matches were calculated by proportion of the high intensity distance (11%) was covered for
averaging distances covered in all of the 1-, 3-, and 5-min periods ‘Support Play’. In the next 1-, 3- and 5-min periods, the players
excluding stoppage time [3, 18]. Nevertheless, when the amount of experienced a deficit of 48%, 30%, and 25%, respectively, in
the remaining time during the following intense period was not equiv- high-intensity distance compared to the match average (ES:
alent to the peak period, the related data were removed from analysis. 0.4–0.5, P < 0.01). Out of possession the players covered
22–44%, 34–43%, and 27–45% less high intensity distance for
Statistical Analyses ‘Covering’ (ES: 0.2–0.3, P < 0.01), ‘Recovery Run’ (ES: 0.2–0.3,
Data are expressed as the mean ± standard deviation. All statistical P < 0.01), and ‘Close Down/Press’ (ES: 0.1–0.2, P < 0.05),
analyses were conducted using IBM SPSS Statistics for Mac OS X, respectively, compared to the match average whilst also perform-
version 26 (IBM Corp., Armonk, N.Y., USA). Data normality was ing 28–91% less ‘Run with Ball’ distance when in possession (ES:
verified by Shapiro-Wilk and Kolmogorov-Smirnov tests. Differences 0.1–0.5, P < 0.05). Table 2 illustrates the average distance per
between 1-, 3-, or 5-min periods within a game were determined action with the number of actions across various positions during
using one-way analysis of variance (ANOVA) with repeated measures. the peak periods.
Differences between playing positions were determined using one-way
ANOVA. In the event of a significant difference, Bonferroni post hoc Contextualised Peak Periods – Team Trends
tests were used to identify any localised effects. Statistical significance During the peak 1-, 3- and 5-min periods, the teams covered
was set at P < 0.05. Effect size (ES) for the meaningfulness of the 28–37% and 22–23% of the high-intensity distance (420 ± 82 m,
difference was determined as follows: trivial (≤ 0.2), small 646 ± 125 m, and 842 ± 154 m, respectively), performing
(> 0.2–0.6), moderate (> 0.6–1.2), large (> 1.2–2.0) and very ‘Recovery Run’ and ‘Covering’ actions, respectively. However, they
large (> 2.0–4.0) [29]. covered the largest proportion of their high intensity distance for

Biology of Sport, Vol. 39 No4, 2022 975

976
TABLE 2. Average distance (m) per physical-tactical action and average number of actions during the peak periods per game across positions.
Average In Possession Out of Possession
HIR Distance
Position Time No. Actions Overall
(m/min) SP MTR/ES OVL/UDL RWB RIB/PEN BIB CD/PRE COV RR
[min-max]
1 min 55 ± 17* 2 ± 1 [1–4] 73 (1) 29 ± 18 (6) 0 (0) 25 ± 13 (19) 0 (0) 18 ± 11 (5) 15 ± 7 (7) 30 ± 15 (119) 48 ± 17 (95) 38 ± 17 (273)
*
CDP 3 min 24 ± 7 3 ± 1 [1–5] 73 (1) 26 ± 13 (9) 0 (0) 22 ± 11 (29) 0 (0) 25 ± 16 (3) 17 ± 7 (15) 24 ± 10 (209) 40 ± 17 (119) 30 ± 12 (417)
5 min 17 ± 5* 3 ± 1 [1–7] 0 (0) 23 ± 13 (12) 46 (1) 20 ± 10 (25) 0 (0) 18 ± 11 (5) 17 ± 7 (15) 22 ± 8 (272) 37 ± 17 (141) 27 ± 10 (518)
#
1 min 76 ± 18 2 ± 1 [1–4] 38 ± 18 (43) 30 ± 17 (22) 47 ± 18 (18) 28 ± 13 (17) 21 ± 9 (10) 23 ± 7 (4) 16 ± 6 (20) 30 ± 14 (72) 41 ± 17 (92) 36 ± 12 (327)
#
WDP 3 min 35 ± 8 4 ± 1 [2–7] 31 ± 17 (65) 29 ± 14 (25) 34 ± 16 (29) 23 ± 14 (25) 23 ± 12 (18) 23 ± 8 (5) 15 ± 5 (34) 26 ± 12 (125) 36 ± 16 (123) 30 ± 9 (490)
#
5 min 26 ± 7 5 ± 1 [3–8] 30 ± 15 (80) 22 ± 10 (43) 32 ± 15 (33) 25 ± 12 (36) 21 ± 10 (15) 18 ± 7 (8) 15 ± 4 (49) 24 ± 10 (161) 34 ± 14 (142) 28 ± 8 (624)
1 min 68 ± 17 2 ± 1 [1–5] 41 ± 20 (41) 22 ± 13 (21) 54 ± 33 (3) 31 ± 13 (27) 30 ± 13 (9) 22 ± 11 (5) 21 ± 10 (33) 31 ± 17 (80) 39 ± 19 (94) 37 ± 18 (337)
CMP 3 min 32 ± 10 3 ± 1 [1–7] 35 ± 18 (57) 24 ± 10 (25) 41 ± 44 (3) 27 ± 13 (29) 24 ± 12 (18) 27 ± 11 (13) 21 ± 12 (57) 26 ± 13 (120) 32 ± 16 (140) 30 ± 11 (514)
5 min 23 ± 7 4 ± 1 [1–7] 34 ± 17 (56) 24 ± 10 (38) 33 ± 39 (5) 25 ± 11 (38) 22 ± 10 (22) 22 ± 12 (14) 19 ± 10 (72) 25 ± 11 (149) 31 ± 15 (159) 28 ± 9 (627)
#
1 min 76 ± 16 2 ± 1 [1–4] 34 ± 16 (15) 35 ± 14 (22) 0 (0) 30 ± 17 (16) 30 ± 11 (12) 27 ± 20 (2) 19 ± 7 (12) 26 ± 10 (10) 34 ± 14 (28) 33 ± 10 (129)
WOP 3 min 36 ± 7 4 ± 1 [2–6] 30 ± 17 (19) 29 ± 15 (28) 0 (0) 29 ± 17 (30) 28 ± 10 (20) 27 ± 11 (7) 22 ± 10 (28) 28 ± 14 (16) 32 ± 17 (37) 29 ± 9 (201)
#
5 min 27 ± 5 5 ± 2 [2–9] 29 ± 16 (29) 29 ± 15 (34) 19 ± 9 (2) 22 ± 10 (30) 24 ± 9 (22) 24 ± 12 (7) 18 ± 8 (23) 24 ± 10 (30) 31 ± 17 (39) 27 ± 9 (245)
1 min 71 ± 14 2 ± 1 [1–4] 38 ± 21 (8) 25 ± 15 (13) 0 (0) 32 ± 19 (4) 33 ± 15 (16) 25 ± 13 (6) 27 ± 11 (22) 31 ± 15 (7) 41 ± 15 (4) 32 ± 12 (84)
COP 3 min 32 ± 7 4 ± 1 [2–6] 44 ± 17 (14) 23 ± 12 (18) 0 (0) 26 ± 15 (9) 24 ± 12 (19) 25 ± 12 (7) 27 ± 10 (29) 36 (1) 32 ± 17 (2) 29 ± 11 (117)
5 min 25 ± 6 5 ± 1 [2–8] 33 ± 18 (21) 23 ± 14 (11) 0 (0) 27 ± 15 (10) 26 ± 14 (25) 21 ± 13 (8) 22 ± 9 (42) 26 ± 9 (4) 37 ± 7 (3) 26 ± 7 (141)
1 min 67 ± 19 2 ± 1 [1–5] 39 ± 19 (108) 29 ± 15 (84) 48 ± 20 (21) 29 ± 14 (83) 29 ± 13 (47) 23 ± 11 (22) 21 ± 10 (94) 30 ± 15 (288) 42 ± 18 (313) 36 ± 15 (1150)
Overall 3 min 31 ± 9 3 ± 1 [1–7] 34 ± 18 (156) 26 ± 13 (105) 35 ± 19 (32) 25 ± 14 (122) 25 ± 11 (75) 26 ± 10 (35) 20 ± 10 (163) 25 ± 12 (471) 35 ± 17 (421) 31 ± 9 (1739)
5 min 23 ± 7 3 ± 1 [1–9] 32 ± 16 (186) 24 ± 12 (138) 34 ± 18 (41) 24 ± 11 (139) 24 ± 11 (84) 21 ± 11 (42) 18 ± 8 (201) 23 ± 10 (616) 34 ± 16 (484) 23 ± 7 (2155)

CDP: Central Defensive Player; WDP: Wide Defensive Player; CMP: Central Midfield Player; WOP: Wide Offensive Player; COP: Central Offensive Player. HIR: High-intensity running.
SP: ‘Support Play’, MTR/ES: ‘Move to Receive/Exploit Space’, OVL/UDL: ‘Overlap/Underlap’, RWB: ‘Run with Ball’, RIB/PEN: ‘Run in Behind/Penetrate’, BIB: ‘Break into Box’, CD/PRE: ‘Close
Down/Press’, COV: ‘Covering’, RR: ‘Recovery Run’. ‘Push up Pitch’ and ‘Interception’ were excluded due to the small number of actions. Physical-tactical average distances are reported as
mean ± SD (m) and numeral in parenthesis indicates the total number of physical-tactical actions performed. *Less high-intensity running distance than other positions (P < 0.01). #Greater
high-intensity running distance than CMP (P < 0.05).
Wonwoo Ju et al.
 Contextualised peak periods of play

FIG. 1. Team performance; number of physical-tactical actions and players involved during the peak and next 1-, 3-, 5-min periods.
Numbers above the bars indicate mean values. Dotted lines indicate before-after values. *Difference from peak period (P<0.01).

FIG. 2. Central Defensive Player; contextualised distances at high-intensity in the peak 1-, 3-, and 5-min periods during the match,
the subsequent period (next), and the match average (mean). *Difference from match average for ‘Covering’ (P<0.05). #Difference
from match average for and ‘Recovery Run’ (P<0.01). ◇Difference from match average for ‘Run in Behind/Penetrate’ (P<0.01).
◆
Difference from match average for ‘Break into Box’ (P<0.01).

Biology of Sport, Vol. 39 No4, 2022 977

 Wonwoo Ju et al.

FIG. 3. Wide Defensive Player; contextualised distances at high-intensity in the peak 1-, 3-, and 5-min periods during the match, the
subsequent period (next), and the match average (mean). *Difference from match average for ‘Recovery Run’ (P<0.05). #Difference
from match average for and ‘Support Play’ and ‘Run with Ball’ (P<0.05). ◇Difference from match average for ‘Break into Box’ and
‘Run in Behind’ (P<0.01). ◆Difference from match average for ‘Move to Receive/Exploit Space’ (P<0.05).

FIG. 4. Central Midfield Player; contextualised distances at high-intensity in the peak 1-, 3-, and 5-min periods during the match,
the subsequent period (next), and the match average (mean). *Difference from match average for ‘Covering’ (P<0.01). #Difference
◇
from match average for and ‘Close Down/Press’ (P<0.01). Difference from match average for ‘Break into Box’ (P<0.01). ◆Difference
◻
from match average for ‘Run with Ball’ (P<0.05). Difference from match average for ‘Move to Receive/Exploit Space’ (P<0.01).
■
Difference from match average for ‘Support Play’ (P<0.05).

‘Support Play’ (12–13%) in possession. In the next 1-, 3- and Contextualised Peak Periods – Position-Specific Trends
5-min periods, the teams had a deficit of 31%, 30%, and 17%, Central Defensive Player
respectively, in high-intensity distance compared to the match During the peak 1-, 3-, and 5-min periods, CDP performed ~80% of
average (ES: 0.5–0.8, P < 0.01). The teams covered 20–41% their high-intensity distance (55 ± 17 m, 72 ± 22 m, and 86 ± 26 m,
and 32–53% less high-intensity distance for ‘Covering’ and ‘Re- respectively) out of possession whilst covering 39–49% of the distance
covery Run’, respectively, compared to the match average (ES: for ‘Recovery Run’ and 36–45% for ‘Covering’ (Figure 2). CDP covered
0.4–0.7, P < 0.01). Figure 1 shows the frequency of high-inten- greater high-intensity ‘Covering’ distance than WOP and COP during
sity actions and the numbers of players involved during the peak all of the peak periods (ES: 0.7–1.4, P < 0.01) whilst also perform-
and next periods. ing more ‘Recovery Run’ distance than COP (ES: 0.9–1.0, P < 0.01).

978
 Contextualised peak periods of play

FIG. 5. Wide Offensive Player; contextualised distances at high-intensity in the peak 1-, 3-, and 5-min periods during the match, the
subsequent period (next), and the match average (mean). *Difference from match average for ‘Interception’ (P<0.01). #Difference from
match average for and ‘Move to Receive/Exploit Space’ (P<0.01). ◇Difference from match average for ‘Support Play’ (P<0.05) and
‘Run with Ball’ (P<0.01). ◆Difference from match average for ‘Run in Behind/Penetrate’ (P<0.05).

FIG. 6. Central Offensive Player; contextualised distances at high-intensity in the peak 1-, 3-, and 5-min periods during the match, the
subsequent period (next), and the match average (mean). *Difference from match average for ‘Recovery Run’ (P<0.01). #Difference
from match average for and ‘Covering’ (P<0.01). ◇Difference from match average for ‘Close Down/Press’ (P<0.05). ◆Difference from
match average for ‘Break into Box’ (P<0.01). ◻︎Difference from match average for ‘Run in Behind/Penetrate’, ‘Move to Receive/Exploit
Space’, and ‘Run with Ball’ (P<0.01). ■Difference from match average for ‘Support Play’ (P<0.05).

Wide Defensive Player Central Midfield Player

During the peak 1-, 3-, and 5-min periods, WDP covered 28–34% During the intensified 1-, 3-, and 5-min periods, CMP covered
and 20–23% of their high-intensity distance (76 ± 18 m, 29–33% and 21–22% of their high-intensity distance (68 ± 17 m,
104 ± 24 m, and 132 ± 35 m, respectively) for ‘Recovery Run’ 96 ± 29 m, and 116 ± 34 m, respectively) for ‘Recovery Run’ and
and ‘Covering’, respectively whilst they covered 14% of the distance ‘Covering’, respectively, while they performed 12–15% of the distance
for ‘Support Play’, and 6–7% for ‘Over/Underlap’ (Figure 3). In pos- for ‘Support Play’ (Figure 4). During all of the peak periods, CMP
session WDP performed greater high-intensity ‘Support Play’ distance performed greater high-intensity ‘Recovery Run’ distance compared
than CDP (ES: 0.7–1.0, P < 0.01) and ‘Over/Underlap’ than other to COP (ES: 0.8–1.0, P < 0.01) whilst also covering greater ‘Support
positions (ES: 0.4–0.7, P < 0.01). Play’ distance than CDP (ES: 0.7–0.8, P < 0.01).

Biology of Sport, Vol. 39 No4, 2022 979

 Wonwoo Ju et al.

Wide Offensive Player that all outfield players have to collectively perform some bouts of
During the peak 1-, 3-, and 5-min periods WOP covered 19–23% high-intensity actions during intensified periods of competition as
of their high-intensity distance (76 ± 16 m, 107 ± 22 m, and a team. Nevertheless, such data provide only a rudimentary insight
134 ± 26 m, respectively) for ‘Recovery Run’ whilst they performed on physical performance.
14–19% of their high-intensity distance for ‘Move to Receive/Exploit The contextualised data revealed that out of possession ~20–35%
Space’, 10–14% for each ‘Support Play’ and ‘Run with Ball’, and of the high-intensity distance was covered by players for each ‘Recov-
8–10% for ‘Run in Behind/Penetrate’ when in possession (Figure 5). ery Run’ and ‘Covering’ whilst in possession ~10% was covered for
During all of the peak periods WOP performed more high-intensity ‘Support Play’. This may indicate that peak periods occur during a fast
distances for ‘Run in Behind/Penetrate’ and ‘Move to Receive/Exploit transition phase since such actions as ‘Recovery Run’ and ‘Support
Space’ than CDP, WDP, and CMP (ES: 0.5–1.6, P < 0.01) whilst Play’ are commonly performed when the ball is quickly moved defen-
also covering greater ‘Run with Ball’ distance than CDP and WDP sively or offensively during a quick transition [15]. This could also be
(ES: 0.4–0.8, P < 0.05). supported with the team performance data where teams produced
high-intensity ‘Recovery Run’ and ‘Support Play’ actions the most out
Central Offensive Player of possession and in possession, respectively. This could be due to
During the peak 1-, 3-, and 5-min periods, COP covered 23–25% of players/teams executing more high-intensity actions during decisive
their high-intensity distance (71 ± 14 m, 96 ± 21 m, and phases of play than normal situations [30]. Nevertheless, since the
126 ± 28 m, respectively) for ‘Close Down/Press’ whilst they ran present study did not analyse phases of play (e.g., attack-to-defence
14–20% and 12–20% of the distance for ‘Run in Behind/Penetrate’ transition phases), it is difficult to fully conclude whether intensified
and ‘Support Play’, respectively (Figure 6). COP covered more high- periods take place during fast transition phases or not. Thus, future
intensity ‘Close Down/Press’ distance than other positions during all studies should attempt to condense contextualised actions into the
of the peak periods (ES: 0.4–2.5, P < 0.05). In possession, COP phases of play to provide additional granularity.
performed greater high-intensity ‘Run in Behind/Penetrate’ distance However, the contextualised data during the peak periods were
during all of the peak periods compared to other positions (ES: 0.4–2.2, position-specific. For instance, the key high-intensity tactical actions
P < 0.05) whilst also covering more distance for ‘Break into Box’ during the peak periods for CDP were ‘Covering’ and ‘Recovery Run’.
than CDP (ES: 0.7–0.9, P < 0.01) and WDP (ES: 0.6, P < 0.05). This is possibly due to one of their main defensive duties, which is to
defend the space left behind particularly when a turnover in possession
DISCUSSION occurs [31]. In addition to these, ‘Support Play’ was another main
The present study is the first to consider the contextualised high- physical-tactical action for WDP and CMP, but there was a bespoke
intensity distance covered during peak 1-, 3-, and 5-min periods of action for WDP (‘Over/Underlap’). This clearly demonstrates their
match-play and the following periods of each duration using a rolling attacking responsibilities during the peak periods. For instance, WDP
average technique for individual and team performances. The con- and CMP should perform ‘Support Play’ to become involved in the
textualised data now provide important insights into how players/ attacking/defence-to-attack transition phase to produce a promising
teams tactically perform in relation to high-intensity efforts during attacking threat [7, 15]. Furthermore, the key high-intensity tactical
peak periods and how they altered their physical-tactical behaviour activities for WOP were ‘Move to Receive/Exploit Space’, ‘Run in Be-
during the following periods. Nonetheless, some physical-tactical hind/Penetrate’, and ‘Support Play’, and ‘Run with Ball’ when in pos-
actions demonstrated inconsistency in different time durations of the session and ‘Recovery Run’ when out of possession. By contrast, ‘Close
next periods, and these physical-tactical data were position-specific. Down/Press’, ‘Run in Behind/Penetrate’, and ‘Support Play’ were the
This may indicate that each position has certain physical-tactical key high-intensity tactical activities for COP. The data clearly exhibit
actions to perform even after intensified periods of play, especially their specific tactical roles during intensified periods. For example,
when tactically required to do so. COP should aggressively close down/press the opponent to make it
Numerous studies have examined match running performances in hard for them to advance their attacking play or regain possession
peak periods of play to provide an insight into intensified discrete when out of possession [32] whilst they should also perform attacking
periods (e.g., 1-min) [1, 3, 4, 11, 13, 18]. Supported by previous actions (e.g., ‘Run in Behind/Penetrate’) to create promising chances
studies [2, 4, 11, 18], data demonstrates that CDP demonstrated when in possession [7]. Such position-specific tasks could be used to
the lowest locomotive demands whilst WDP and WOP exhibited the replicate intensified periods during training matches as the 11v11 train-
largest physical demands during intense periods. Additionally, the ing matches could offer the players resources to train the most de-
present study for the first time analysed the peak periods for team manding episodes of match-play [6]; however, it should be acknowl-
performances. Data indicate that during the peak 1-, 3-, and 5-min edged that it is unlikely to provide the necessary ‘overload’ desired at
periods, almost all of the outfield players in different playing positions times. Additionally, these position-specific trends could be easily trans-
(9–10 players) were involved collectively covering high-intensity dis- lated into training sessions using the average distance per physical-
tances of ~400 m, ~650 m, and ~850 m, respectively. This indicates tactical action and average number of actions during the peak periods

980
 Contextualised peak periods of play

(Table 2). For example, whilst CMP is driving through the middle covering ~20–55% less high-intensity distances for ‘Covering’ and
running with the ball at high-intensity (~20 m), WDP could perform ‘Recovery Run’ during all of the subsequent periods compared to the
a high-intensity over/underlapping action (~35 m) in a wide area from match average. Since both players and teams consistently ran less
the middle to the final third. Once CMP pass the ball to WDP, they high-intensity distance for such variables, this may denote that they
could produce a ‘Break into Box’ action (~20 m) and WDP could tend to be defensively in a good tactical position/formation when out
cross the ball at the end of the action. Both then produce ‘Recovery of possession after intense periods whilst covering less high-intensi-
Run’ actions (~30–40 m) to get goal side of the ball to replicate fast ty distance. This could be due to the defensive phases of play being
transition phases of the peak periods. Although this type of work has more physically taxing [34]. In possession, players performed
been previously attempted to replicate player’s physical demands ~30–90% less ‘Run with Ball’ distances in the next periods compared
whilst simultaneously reflecting position-specific game situations [5, to the match average, which may be due to such actions (e.g., drib-
14], they used average physical demands, which may underestimate bling) causing an increased energy cost when compared to running
the true match demands [9]. This could also be supported by the fact without the ball [35] or due to player time running with the ball
that the average distance per action performed during intensified being small [36]. However, even with such context it is still challeng-
periods is greater (e.g., ~35 m during the peak 1-min period, Ta- ing to fully explain why transient decrements occur in subsequent
ble 2) than that performed during the entire match (~20 m) [7]. periods. Thus, the systematic use of video to check each player and
Thus, these contextualised peak distance data could help practitioners to add more layers of information together with effective playing time
better prescribe not only position- but player-specific drills with the (e.g., only in-play time) may provide clearer insights into how players/
true peak demands. teams alter their physical-tactical actions after intense periods [34].
Decrements that follow after the peak periods have been previ- Data demonstrates that physical-tactical trends during the subse-
ously examined alongside analysing intensified periods of play [1–3, 18]. quent periods were also position-specific. For instance, WDP covered
However, most studies used a predefined technique (e.g., 0–5, ~30–50% less high-intensity distance for ‘Recovery Run’ in all of the
5–10 min etc.), which can under or overestimate the true physical next periods compared to the match average. ‘Recovery Run’ is when
demands of the peak period and the subsequent period, respective- players run back toward own goal to get goal side of the ball when
ly [28]. Moreover, different speed thresholds for high-intensity running out of position [22], therefore this might specify that WDP modulates
−1
during the peak periods have been used (> 14.0–19.8 km · h ), their physical-tactical performances by being less involved in the at-
which makes it difficult to compare across studies. Despite these short- tacking/transition phase during the next periods. That said, it would
comings, the general consensus from previous research was that the be of greater interest if measuring the ability of the player to be involved
transient decrements in high-intensity running occur after intense pe- in the subsequent attack after the tactical modulation to evaluate the
riods, which agrees with the findings of the present study. Addition- effectiveness of the player. However, certain physical-tactical actions
ally, the present study analysed shorter durations of the next periods exhibited inconsistency in different time durations of the next periods.
to evaluate more detailed temporal changes after the peak periods of For instance, COP covered ~80–100% less high-intensity distances
match-play. Data demonstrated that players experienced more pro- for ‘Break into Box’ in the next 1- and 5-min periods compared to the
nounced reductions in high-intensity running in the next 1-min period match average; however, they covered ~20% more distances for this
(~50%) compared to the next 3- and 5-min period (~25–30%) after action during the next 3-min period. This may indicate that players
the peak periods of play. This prominent short-term fluctuation during selectively produce high-intensity running particularly when tactically
the next 1-min period seems likely due to less energy available from required to do so. Yet, it is still difficult to draw conclusions since
creatine phosphate hydrolysis since creatine phosphate concentrations performances are influenced by context (e.g., no need to perform or
could be significantly diminished after some bouts of high-intensity choosing not to perform physical-tactical actions). Thus, more context
actions [33]. That said, similar trends for team performances were should be provided to better understand ‘WHY’ players perform less
observed during the subsequent periods to individual performances. physical-tactical actions during the subsequent period after the
This might indicate that teams briefly modify their collective tactical peak passage.
behaviour after intensified periods during matches, which could also
be supported by the number of high-intensity actions and players in- Limitation
volved in the following periods (Figure 1). Yet, it could be due to reduced Firstly, the present study did not quantify acceleration/deceleration
playing time (e.g., ball out of play); thus, it is difficult to fully determine efforts during matches. Although most of these efforts do not reach
without context whether transient decrements are down to fatigue or high-intensity speed thresholds, they are very frequent during match-
tactical alterations/pacing strategies or reduced playing opportunities. es and are extremely taxing mechanically [37]. Thus, these actions
The present study for the first time provides important insights should be incorporated and contextualised to have a comprehensive
on how players/teams alter their tactical behaviour during the next understanding of the true physical demands with tactical purposes.
1-, 3-, and 5-min periods through integrating the physical-tactical Moreover, although contextualised data include high-intensity running
metrics. Players/teams changed their tactical performances by activities with a single (~75%) and hybrid tactical (~15%) actions

Biology of Sport, Vol. 39 No4, 2022 981

 Wonwoo Ju et al.

with unclassified movements as ‘Other’ (~10%), the present study CONCLUSIONS

combined the singular actions and the primary action of the hybrid The contextualised distance data could help coaches and/or practi-
actions to simplify data output. Thus, future research should evaluate tioners better prescribe position- and player-specific training drills,
hybrid actions to provide more transparency and insight to practitio- and may help improve the understanding of transient decrements in
ners. Furthermore, since the present study analysed general posi- high-intensity running during the subsequent periods. However, since
tional data (e.g., WDP and CMP) and these were derived from differ- it is still complex due to numerous influencing factors, additional
ent formations (e.g., 4-3-3, 3-5-2, etc.), this could have impacted context should be provided to have a better understanding of the
data due to formations/player playing style influencing match perfor- transient decrements in high-intensity running.
mance [38, 39]. Therefore, it is warranted to evaluate the effects of
formation or different playing styles of players (e.g., central defensive Conflict of interest
or attacking midfielders) on match physical-tactical performances. The authors have no conflict of interest to declare.

REFERENCES
1. Bradley P, Sheldon W, Wooster B, A Comparison of Match Demands Using 19. Mohr M, Krustrup P, Bangsbo J. Fatigue
Olsen P, Boanas P, Krustrup P. Ball-in-Play versus Whole Match Data in in soccer: a brief review. J Sports Sci.
High-intensity running in English FA Professional Soccer Players of the English 2005; 23(6):593–599.
Premier League soccer matches. J Sports Championship. Sports. 2021; 9(6):76. 20. Bradley P, Noakes TD. Match running
Sci. 2009; 27(2):159–168. 10. Castellano J, Martín-García A, performance fluctuations in elite soccer:
2. Di Mascio M, Bradley PS. Evaluation of Casamichana D. Most running demand indicative of fatigue, pacing or situational
the most intense high-intensity running passages of match play in youth soccer influences? J Sports Sci. 2013;
period in English FA premier league congestion period. Biol Sport. 2020; 31(15):1627–1638.
soccer matches. J Strength Cond Res. 37(4):367–373. 21. Carling C, Dupont G. Are declines in
2013; 27(4):909–915. 11. Martín-García A, Casamichana D, physical performance associated with
3. Mohr M, Krustrup P, Bangsbo J. Match Gómez-Díaz A, Cos F, Gabbett TJ. a reduction in skill-related performance
performance of high-standard soccer Positional differences in the most during professional soccer match-play?
players with special reference to demanding passages of play in football J Sports Sci. 2011; 29(1):63–71.
development of fatigue. J Sports Sci. competition. J Sports Sci Med. 2018; 22. Ju W, J Lewis C, Evans M, Laws A,
2003; 21(7):519–528. 17(4):563. S Bradley P. The validity and reliability of
4. Oliva-Lozano JM, Rojas-Valverde D, 12. Novak AR, Impellizzeri FM, Trivedi A, an integrated approach for quantifying
Gomez-Carmona CD, Fortes V, Coutts AJ, McCall A. Analysis of the match physical-tactical performance.
Pino-Ortega J. Worst case scenario match worst-case scenarios in an elite football Biol Sport. 2022; 39(2):253–261.
analysis and contextual variables in team: Towards a better understanding 23. Oliva-Lozano JM, Martín-Fuentes I,
professional soccer players: a longitudinal and application. J Sports Sci. 2021; Fortes V, Muyor JM. Differences in
study. Biol Sport. 2020; 39(16):1850–1859. worst-case scenarios calculated by fixed
37(4):429–436. 13. Riboli A, Semeria M, Coratella G, length and rolling average methods in
5. Ade JD, Drust B, Morgan OJ, Bradley PS. Esposito F. Effect of formation, ball in professional soccer match-play.
Physiological characteristics and acute play and ball possession on peak Biol Sport. 2021; 38(3):325–331.
fatigue associated with position-specific demands in elite soccer. Biol Sport. 24. Carling C, Le Gall F, Dupont G. Analysis
speed endurance soccer drills: production 2021; 38(2):195–205. of repeated high-intensity running
vs maintenance training. Sci Med Footb. 14. Bradley P, Martín-García A, Ade J, performance in professional soccer.
2021; 5(1):6–17. Gómez-Díaz A. Position Specific & J Sports Sci. 2012; 30(4):325–336.
6. Martín-García A, Castellano J, Positional Play Training in Elite Football: 25. Barnes C, Archer DT, Hogg B, Bush M,
Gómez-Díaz A, Cos F, Casamichana D. Context Matters. Football Medicine & Bradley PS. The evolution of physical and
Positional demands for various-sided Performance. 2019:31–35. technical performance parameters in the
games with goalkeepers according to the 15. Bradley P. FOOTBALL DECODED: Using English Premier League. Int J Sports
most demanding passages of match play Match Analysis & Context to Interpret the Med. 2014; 35(13):1095–1100.
in football. Biol Sport. 2019; Demands: Self-Published; 2020. 26. Carling C, Bloomfield J. The effect of an
36(2):171–180. 16. Carling C, McCall A, Harper D, Bradley PS. early dismissal on player work-rate in
7. Ade J, Fitzpatrick J, Bradley PS. Comment on: “The Use of Microtechnology a professional soccer match. J Sci Med
High-intensity efforts in elite soccer to Quantify the Peak Match Demands of Sport. 2010; 13(1):126–128.
matches and associated movement the Football Codes: A Systematic Review”. 27. Bradley P, Ade JD. Are Current Physical
patterns, technical skills and tactical Sports Med. 2019; 49(2):343–345. Match Performance Metrics in Elite Soccer
actions. Information for position-specific 17. Casamichana D, Castellano J, Fit for Purpose or Is the Adoption of an
training drills. J Sports Sci. 2016; Gómez-Díaz A, Gabbett TJ, Integrated Approach Needed? Int J Sports
34(24):2205–2214. Martín-García A. The most demanding Physiol Perform. 2018; 13(5):656–664.
8. Bradley P, Carling C, Gómez-Díaz A, passages of play in football competition: 28. Varley MC, Elias GP, Aughey RJ. Current
Hood P, Barnes C, Ade J, et al. Match a comparison between halves. Biol Sport. match-analysis techniques’
performance and physical capacity of 2019; 36(3):233–240. underestimation of intense periods of
players in the top three competitive 18. Fransson D, Krustrup P, Mohr M. Running high-velocity running. Int J Sports
standards of English professional soccer. intensity fluctuations indicate temporary Physiol Perform. 2012; 7(2):183–185.
Hum Mov Sci. 2013; 32(4):808–821. performance decrement in top-class 29. Batterham AM, Hopkins WG. Making
9. Mernagh D, Weldon A, Wass J, Phillips J, football. Sci Med Footb. 2017; meaningful inferences about magnitudes.
Parmar N, Waldron M, et al. 1(1):10–17. Int J Sports Physiol Perform. 2006;

982
 Contextualised peak periods of play

1(1):50–57. 34. Castellano J, Errekagorri I, Los Arcos A, J Sports Med. 2013; 34(1):34–39.
30. Faude O, Koch T, Meyer T. Straight Casamichana D, Martín-Garcia A, 38. Bradley PS, Carling C, Archer D,
sprinting is the most frequent action in Clemente F, et al. Tell me how and where Roberts J, Dodds A, Di Mascio M, et al.
goal situations in professional football. you play football and I’ll tell you how The effect of playing formation on
J Sports Sci. 2012; 30(7):625–631. much you have to run. Biol Sport. 2022; high-intensity running and technical
31. Modric T, Versic S, Sekulic D. Position 39(3):607–614. profiles in English FA Premier League
Specific Running Performances in 35. Reilly T. Motion analysis and soccer matches. J Sports Sci. 2011;
Professional Football (Soccer): Influence physiological demands. Science and 29(8):821–830.
of Different Tactical Formations. Sports. soccer: Routledge; 2003. p. 67–80. 39. Dellal A, Chamari K, Wong DP,
2020; 8(12):161. 36. Carling C. Analysis of physical activity Ahmaidi S, Keller D, Barros R, et al.
32. Lucchesi M. Pressing. Spring City, PA: profiles when running with the ball in Comparison of physical and technical
Reedswain; 2004. a professional soccer team. J Sports Sci. performance in European soccer
33. Haff GG, Triplett NT. Essentials of 2010; 28(3):319–326. match-play: FA Premier League and La
strength training and conditioning 4th 37. Varley MC, Aughey RJ. Acceleration Liga. Eur J Sport Sci. 2011;
edition: Human kinetics; 2015. profiles in elite Australian soccer. Int 11(1):51–59.

Biology of Sport, Vol. 39 No4, 2022 983

View publication stats