Content uploaded by Ricardo Duarte
Author content
All content in this area was uploaded by Ricardo Duarte on Jun 10, 2014
Content may be subject to copyright.
Collective intelligence: An incursion into the tactical performance of football teams
Ricardo Duarte & Telmo Frias
Faculty of Human Kinetics, Portugal
Keywords: performance analysis, collective motion variables, collective behaviour of
teams.
Abstract
The purpose of this paper is to present an alternative approach to teams’ tactical
performance based on principles of complexity sciences. In this view, the collective
behaviour of a team is a process arising from the coordinated actions of individual players
that work together to achieve the successive goals of the game (e.g., pressing collectively
to close spaces and recover the ball and then, opening space to organise the attack). This
kind of collective intelligence includes some critical features such as: i) players behaving
as a unit, ii) teams showing a personal identity and structural organisation, and iii),
players coordinating differentiated tasks between them. These principles are introduced
using examples extracted from competitive matches and qualitative reports from players
and coaches. Based on these fundaments, some innovative measures of teams’ tactical
performance are suggested, such as: i) the surface area occupied by the teams, ii)
distance between teams’ centre, iii) stretch index of the teams, iv) length and width of the
“block”, v) ratio length/width of a team, and so on. Using a specifically conceived software
application (TeamSense 1.0), video animations and graphical data from competitive
match-play of professional football teams are presented in order to demonstrate the utility
of these tools to enhance tactical performance of teams. Finally, it is argued that football
coaching and professional performance analysis departments can benefit from the
adoption of this type of performance analysis.
Introduction
This paper presents an alternative approach to the analysis of teams’ tactical
performance. This perspective proposes that the collective behaviour of a team can be
regarded as a kind of a highly distributed collective intelligence among the players. This
proposal contrasts with the common view that the coach acts as a higher processor in a
centralised control hierarchy. The idea of collective intelligence is analysed using
qualitative data extracted from players and coaches’ reports, and also discussing the
notion of teams regarded as super-organisms. Technological developments and
innovative measures are introduced as key issues to improve the analysis of teams’
tactical performance.
Collective Intelligence: Teams as Superorganisms
How players decide during the game? Some studies developed in team sports have been
demonstrated that individual movements of players during the game are highly
coordinated, forming a coherent collective behaviour (Bourbousson, Sève, & McGarry,
2010a, 2010b; Passos et al., 2011). This relation between individual and team behaviours
revealed that players’ decisions and actions are very dependent of the whole behaviour of
the team, as well as team’s behaviour depends also on the individual behaviours of each
player. This property of complex systems (as a football team, for instance) is called as
interdependency (Bar-Yam, 2004). In this sense, the performance of a football team
during the game can be viewed as a collective intelligent behaviour not centralised in any
player (or coach) awareness, but distributed among all the players. This kind of
intelligence is expressed by the collective behaviours of the teams, which emerge from
the coordinated actions of the players.
But, what constrain players to coordinate their actions? It seems that the
coherence in players’ behaviour resulted from the fact that they share, not only the same
general aims (e.g., score more goals than the opposite team), but particularly some
specific momentary sub-goals (e.g., cooperating with a teammate to close a penetration
space). Thus, the coherent behaviours of players create a relative stable structural
organisation and a personal identity for the team. In socio-biology, animal groups with
these features are called as superorganisms. The collective wisdom of these
superorganisms (e.g., colonies of ants, swarms of honeybees, flocks of birds) seems also
to arise from the simple complementary behaviours of each individual that typically
perform different tasks (Seeley, 2002; Hölldobler, & Wilson, 2009). Due the cooperative
nature of players’ relations within a team, the concept of superorganism seems very
suitable to explain the collective intelligence and decision making processes of the
football teams as a whole (Duarte et al, 2011; Marsh et al, 2006). For example, André
Villas-Boas (FC Porto’s head coach) advocates that players have a crucial role in the
collective decisions of the team. He said: “we respect the unpredictability of the game. We
believe that for the teams express all their quality, players need to have freedom to decide
on the pitch. We don’t act as dictator of the choices. The game is unpredictable and we
need to respect this” (in Alvarenga, 2011). These statements revealed that this coach
seems to view the collective behaviour of its team as dependent of the changing
conditions of the game.
Specifically, superorganismic entities as football teams also have some critical
features such as: i) players behaving as a unit, ii) teams showing a personal identity and
structural organisation, iii), players coordinating differentiated tasks between them, and iv),
teams undergoing a clearly adaptive cycle of growth (Duarte et al., 2011). In this sense,
the interdependence between team and player behaviours was at the core of Xavi
Hernández’s (central midfield of Barcelona) thoughts about its own performance:
“honestly, the team is the one who gives and takes me. My game needs teammates. I’m
not anyone if nobody creates space ahead, and other opens a short passing line...without
teammates my football doesn’t make sense” (in Cano, 2009, pp.91). Also, when Pep
Guardiola (Barcelona’s head coach) was asked about if Lionel Messi was the best player
in the world, he answers “I don’t discuss things in this way. I don’t analyse the player
alone, I try to understand him within the general context of the team” (in Cano, 2009,
pp.79). These thoughts seem to demonstrate how players functionally behave as a unit
and how they need to coordinate their individual actions to improve simultaneously team
and individual performance effectiveness.
Considering team’s behaviour as influenced by the features of the players and the
interactions they tend to develop during the game, it is possible to argue that teams have
a kind of ‘fingerprint’ (i.e. a personal identity of the team). This identity is based on a
structural organisation defined by the positional distribution of the players in the field and
the trends of relationship between players as a function of the game context. In this sense,
José Mourinho (Real Madrid’s head coach) seems to agree with the idea that a relative
ordered structural organisation is the basis for players ‘decide with freedom’ during the
game: “it’s important for a player to know that in certain position he has a teammate, and
from a geometrical point of view they have something built on the field that allows them to
decide and regulate their actions” (in Cano, 2009, pp.137). He emphasises also the need
to focus team’s preparation in the improvement of structural organisation: “for me, the
most important is to have some behavioral principles that give organisation to the team.
My first weeks of preparation systematically focus on tactical organisation, always to
structure and improve the collective performance” (Oliveira et al, 2006).
Another interesting point to briefly discuss is the idea that the collective behaviour
of a team undergoes a clearly adaptive cycle of growth. That is, time seems to improve
the effectiveness of teams’ tactical performance. Xabi Alonso (Real Madrid’s central
midfield) referred this idea when he talks about its coming to the team: “when I arrived, it
was difficult for everyone to have an idea how to play together. When to press ahead,
when retain the ball...we were very irregular. With time, we know each other better and
now, we go on the pitch with an identity. The sensation is the team was growth” (in Cano,
2009, pp.126). However, is it the team growth only dependent of time? It is arguably that
an accurate diagnostic and monitoring of teams’ behaviours seems to contribute to the
enhancement of their tactical performance. In the next section, some concerns about the
tactical performance of teams will be discussed.
Tactical Performance of Teams
Recent technological developments in tracking systems allow capturing the individual
coordinated movements of all the players that together constrain the emergence of the
collective behaviour of a team. These sophisticated commercial systems (e.g., ProZone,
Amisco, GPSports) obtain time-series of positional data of the players from the pitch.
Thus, it is possible to reconstruct the two-dimensional players’ movements and the
collective motion of the teams.
Based on this type of data, some innovative measures of tactical performance
have been proposed in the last years. For example, abrupt changes in surface area
occupied by the teams can reveal adjustments of their structural organisation (e.g.,
decreasing the covered area to close spaces and increase density of players when a
team loses the ball, Figure 1).
Other measures of teams’ collective performance suggested in literature are the
displacement and distance between teams’ centre, stretch index of the teams, length and
width of the “block”, ratio length/width of a team, predominant regions of players’ action
(Schölhorn, 2003; Frencken, & Lemmink, 2008; Folgado, 2010; Bourbousson et al.,
2010b; Lames, Ertmer, & Walter, 2010; Duarte et al., 2011).
In order to improve the time consuming of the computation procedures, a
specifically conceived software application (TeamSense 1.0) was developed in Matlab
environment (The MathWorks Inc., Natick, MA, USA).
Figure 1. The surface area of the defensive team is lower than the offensive team. Changes in ball
possession seem to influence the areas covered by teams (Source: Ribot, 2006).
This application uses positional data as input files to perform the analyses of
collective/tactical motion variables, as well as graphical displays (time plots) and 2D video
animations. Figure 2 presents a single photogram from the 2D video animation of a
match.
Figure 2. Photogram of the teams’ initial formation in the start of the match.
Performance analysis studies using this type of collective variables during the
match are still starting. Figure 3 shows exemplar data from a single match of two
professional football teams (Duarte et al., 2011).
Figure 3. Variations in the area covered by each team along the 15-min periods of the match.
After some initial stability, the home team displayed higher surface area values
than the visiting team in the third, fourth and fifth periods. However, in the last 15-mins
period, the visiting team overpasses the home team, maybe due to the instability created
by a goal suffered at minute 75, which created a disadvantage in match-score (Duarte et
al., 2011).
Concluding remarks and applications
The proposed alternative approach identified some features that allowed regarding teams
as complex collective entities or superorganisms, in which the collective functioning
expresses the emergent collective intelligence of each team (a kind of ‘team fingerprint’).
These ideas were strengthened using qualitative reports from top players and coaches.
Based on these fundaments, some innovative measures of tactical performance were
proposed, as well as presented some exemplar data from professional football teams.
The use of this emerging type of performance analysis seems to be promising in
the improvement of accurate diagnostics and monitoring of teams’ tactical performance in
competition and training settings. It seems also to be clear that this type of performance
analysis provide information to which players and coaches are very sensitive. Thus,
professional performance analysis departments, and football coaching in general, may
use this approach to enhance teams’ performance.
References
Alvarenga, V. H. (2011, January 21). Villas-Boas grato a Robson e Mourinho: «Eu ia para jornalista
desportivo» [Villas-Boas grateful to Robson and Mourinho]. Retrieved January 21, 2011, from
http://www.maisfutebol.iol.pt/fcporto/fc-porto-villas-boas-porto-liga-mourinho/1227565-1304.html.
Bar-Yam, Y. (2004). Making things work. Cambridge, MA: NECSI Knowledge Press.
Bourbousson, J., Sève, C., McGarry, T. (2010). Space-time coordination patterns in basketball: Part 1. Intra-
and inter-couplings among player dyads. Journal of Sports Sciences, 28(3), 339-347.
Bourbousson, J., Sève, C., McGarry, T. (2010). Space-time coordination patterns in basketball: Part 2. The
interaction between the two teams. Journal of Sports Sciences, 28(3), 349-358.
Cano, O. (2009). El modelo de juego del FC Barcelona. Una red de significado interpretada desde el
paradigma de la complejidad [The game model of FC Barcelona. A network of meaning understood
from the paradigm of complexity]. España: MCSports.
Duarte, R., Araújo, D., Davids, K., Folgado, H., & Marques, P. (2011). Do professional football teams
behave like superorganisms? Paper presented at the 13th European Congress of Sport Psychology,
12-17 July, Funchal, Portugal.
Duarte, R., Araújo, D., Davids, K., Folgado, H., Marques, P., & Ferreira, A. (2011). In search for dynamical
patterns of teams’ tactical behaviours during the match. Paper presented at the 7th World Congress on
Science & Football, 26-30 May, Nagoya, Japan.
Folgado, H. (2010). Towards an understanding of youth football teams tactical performance by analysis of
collective positional variables during small-sided games [Unpublished master thesis]. Vila Real,
Portugal: UTAD.
Frencken, W.G.P. & Lemmink, K.A.P.M. (2008) Team kinematics of small-sided soccer games; a systematic
approach. In T. Reilly, & F. Korkusuz (Eds.), Science and Football VI (pp 161-170). New York:
Routledge
Hölldobler, B., Wilson, E. O. (2009). The Superorganism: the beauty, elegance, and strangeness of insect
societies. New York: W.W. Norton & Company, Inc.
Lames, M., Ertmer, J., & Walter, F. (2010). Oscillations in football - order and disorder in spatial interactions
between the two teams. International Journal of Sport Psychology, 41, 85-86.
Marsh, K. L., Richardson, M. J., Baron, R. M., & Schmidt, R. C. (2006). Contrasting approaches to
perceiving and acting with others. Ecological Psychology, 18, 1-37.
Oliveira, B., Amieiro, N., Resende, N., Barreto, R. (2006). Mourinho. Porquê tantas vitórias? [Mourinho.
Why so many triumphs?]. Lisboa: Gradiva.
Passos, P., Milho, J., Fonseca, S., Borges, J., Araújo, D., Davids, K. (2011). Interpersonal Distance
Regulates Functional Grouping Tendencies of Agents in Team Sports. Journal of Motor Behavior, 43(2),
155-163.
Ribot, J. C. (Producer). (2006). Football, l’intelligence collectif [Football, collective intelligence]. [TV
documentary]. France: Mosaïque Films.
Schöllhorn, W. I. (2003). Coordination dynamics and its consequences on sport. International Journal of
Computer Science in Sport, 2, 40 – 46.
Seeley, T. D. (2001). Decision making in superorganisms: how collective wisdom arises from the poorly
informed masses. In G. Gigerenzer & R. Selten (Eds), Bounded rationality: the adaptive toolbox
(pp249-261), Cambridge, MA: MIT Press.