Elsevier

Human Movement Science

Volume 36, August 2014, Pages 199-216
Human Movement Science

Predicting the lateral direction of deceptive and non-deceptive penalty kicks in football from the kinematics of the kicker

https://doi.org/10.1016/j.humov.2014.04.004Get rights and content

Highlights

  • The capacity of kickers to disguise their intentions was analyzed in the penalty kick.

  • Local and compound kinematic variables correlated highly with shot direction.

  • Kinematic variables were highly correlated at timings that are useful to goalkeepers.

  • Some variables were affected by deception before ball contact but not at ball contact.

  • Results are interpreted with the hypothesis of non-substitutability of genuine action.

Abstract

This study addresses the utility of the kinematics of penalty takers for goalkeepers in association football. Twelve professional and semi-professional players shot to one side of the goal with (deceptive condition) or without (non-deceptive condition) simulating a shot to the opposite side. The body kinematics of the penalty takers were registered with motion-capture apparatus. Correlation and regression techniques were used to determine the relation between the shot direction and aspects of the penalty taker’s kinematics at different moments. Several kinematic variables were strongly correlated with shot direction, especially those related to the lower part of the body. Some of these variables, including the angle of the non-kicking foot, acquired high correlations at time intervals that are useful to goalkeepers. Compound variables, here defined as linear combinations of variables, were found to be more useful than locally defined variables. Whereas some kinematic variables showed substantial differences in their relation to ball direction depending on deception, other kinematic variables were less affected by deception. Results are interpreted with the hypothesis of non-substitutability of genuine action. The study can also be interpreted as extending the correlation and regression methodology, often used to analyze variables defined at single moments, to the analysis of variables in a time continuous fashion.

Introduction

In sports such as handball, basketball, or football, there are severe time constraints for athletes while perceiving and acting. In addition, players are pressured to achieve high levels of precision. A match event in which these spatial and temporal constraints are particularly evident is the penalty kick in association football (Lopes, Araújo, Peres, Davids, & Barreiros, 2008). One of the aspects of the penalty kick that has received attention concerns the information that goalkeepers use to anticipate the direction of the ball (Diaz et al., 2012, Dicks et al., 2010a). With the current experiment we aim to contribute to the knowledge about this aspect of the penalty kick, focusing on the information available in the kinematics of the penalty taker and on the role of deception.

Before we describe the purpose of the experiment in more detail, we briefly review previous results. A crucial issue is the time at which goalkeepers commit themselves to a side. Dicks, Davids, and Button (2010) reported average ball flight times between 590 and 648 ms and average goalkeeper movement times between 750 and 1085 ms. Hence, goalkeepers who base the direction of their dives on the first part of the ball trajectory are likely to start moving too late, especially if one takes into account that a small perceptual-motor delay must exist. The findings reported by Dicks et al. (2010) therefore support the common claim that goalkeepers should initiate their movements before ball contact, which means that they should not rely exclusively on information from the ball trajectory (Dicks et al., 2011, Franks and Harvey, 1997).

As an alternative source of information, goalkeepers may use the kinematics of their opponent before ball contact to anticipate the direction of the shot. This gives rise to two questions. First, which kinematic variables are good predictors of ball direction? And second, which kinematic variables are actually used by goalkeepers? A substantial body of work has addressed the second question, using self-reports (Kuhn, 1988), occlusion paradigms (Dicks et al., 2010b, Smeeton and Williams, 2012), and, most particularly, gaze-registration methods (Button et al., 2011, Dicks et al., 2010a, Piras and Vickers, 2011, Savelsbergh et al., 2005). Areas that goalkeepers have been claimed to focus on include the penalty takers’ hips, the non-kicking foot, and the region between the ball and the kicking leg (i.e., ‘visual pivot’; Piras & Vickers, 2011).

Less research has addressed the first question, about how useful the candidate kinematic variables actually are. Franks and Harvey (1997) analyzed videos of penalties in FIFA World Cup competitions. They concluded that several kinematic variables have a high reliability at or immediately before ball contact. These variables include the inward or outward knee rotation of the kicking leg and the point of contact on the ball. However, given the time constraints for goalkeepers, Franks and Harvey considered that variables should be detectable and have a high reliability a certain time interval before ball contact. This led them to consider the final placement (i.e., pointing direction) of the non-kicking foot as the most useful variable. They reported that this variable has a reliability of about 80% and that it can be detected from about 150–200 ms before ball contact.

Studies by Lees and Owens (2011) and Diaz et al. (2012) also concerned the information value of candidate kinematic variables. These studies were more sophisticated than the one reported by Franks and Harvey (1997) in the sense that motion-capture equipment was used to register the kinematics of the penalty takers, allowing more advanced methods to analyze the reliability of the candidate variables. The three studies (i.e., the ones by Franks & Harvey, Lees & Owens, and Diaz et al.) agree in pointing toward the orientation of the non-kicking foot as a relatively reliable source of information around 200–250 ms before ball contact. Lees and Owens also reported hip rotation (as projected on the horizontal plane) and hip and ankle flexion as significant indicators of shot type and shot direction. Diaz et al. presented results for several locally defined kinematic variables. In addition, they concluded that global or distributed information might be useful. At 200 ms before ball contact, for instance, one of the sources of distributed information considered by Diaz et al. had a reliability of 77%. An emphasis on distributed information is consistent with research in other sports (e.g., Abernethy et al., 2001, Huys et al., 2008, Ward et al., 2002).

The experiments reported by Lees and Owens (2011) and Diaz et al. (2012), however, are not without shortcomings. First, penalty takers were asked to shoot the ball into a smaller-than-standard size goal (Lees & Owens) or into a small (2.43-m wide) canvas substituting a goal (Diaz et al.). Second, penalties were shot from a distance shorter than the regular 11 m. Third, no goalkeepers were present during the penalty kicks. The results of these studies can hence be generalized to penalty kicks in match situations only if one assumes that these aspects do not affect the kinematics of the kicks, or, more precisely, if one assumes that they do not affect the reliability of the considered variables as information sources. We think that the relevance of this assumption warrants further research. The importance of representative designs has also been emphasized in previous penalty kick studies (Button et al., 2011, Dicks et al., 2010a, Lopes et al., 2008; cf. Araújo, Davids, & Hristovski, 2006).

In addition, the experimental designs of Lees and Owens (2011) and Diaz et al. (2012) did not consider the issue of deception, even though deception was mentioned in the respective discussions as being an important issue (cf. Dicks et al., 2010b, Smeeton and Williams, 2012). Suppose that a goalkeeper relies on the orientation of the non-kicking foot. If penalty takers know this, they may try to deceive the goalkeeper by kicking the ball in the opposite direction of that to which the non-kicking foot is oriented. Some aspects of the kicking action, however, need to be established in order to kick the ball in a particular direction, meaning that it is likely that some higher-order or distributed kinematic variables remain specific to the kick direction intended by penalty takers. This is captured by the hypothesis of the non-substitutability of genuine action: In trying to produce an unnatural movement pattern, one may successfully produce some of the kinematic details of the genuine action, but typically not all the complexity of the action that is needed to convince the perceiver that the action is genuine (Richardson and Johnston, 2005, Runeson and Frykholm, 1981, Runeson and Frykholm, 1983; for sports applications see Jackson, Warren, & Abernethy, 2006).

To summarize, the combined literature states that goalkeepers’ actions are at least partly based on the kinematics of the penalty takers before ball contact. To analyze the information contained in the available kinematic variables, and to analyze which variables are actually used, it seems indispensable to register the movements of penalty takers during penalty kicks. Useful work in this regard has been done by Lees and Owens (2011) and Diaz et al. (2012), but further advances may be achieved by performing experiments in more representative situations and by considering the issue of deception. We asked professional and semi-professional players to take penalties in a situation with a goalkeeper and with a standard-size goal. The penalty takers used deceptive and non-deceptive strategies and their movements were registered with an infrared movement-registration system.

A further aspect of our study that we consider a contribution with regard to previous work in the field of penalty kicks is our data analysis, which is inspired by work concerning variable use in other tasks (Jacobs et al., 2001, Michaels and de Vries, 1998). We assess the information available in single kinematic variables with the correlations between the kinematic variable and ball direction, reasoning that more informative variables have higher correlations. Likewise, the information contained in combinations of variables (which may also be referred to as compound variables, higher-order variables, or distributed variables) is assessed with multiple regressions with ball direction as dependent variable. Our application of correlation and regression methodology has the particularity that the values of the kinematic variables change continuously during the approach of the penalty takers to the ball. Our analyses therefore extend the use of correlation and regressions analyses from applications with single-moment variables to temporarily extended variables (cf. Jacobs, Vaz, & Michaels, 2012). In addition to using ball direction as dependent variable, we performed correlation and regressions with dive direction as dependent variable, in order to estimate the variables used by goalkeepers to predict kick direction.

Section snippets

Participants

The penalty takers were twelve male professional and semi-professional players (Mage = 21.2 years; SD = 4.6 years). Eight penalty takers were senior players of a team in the Portuguese National Second Division. Four penalty takers were junior players that played in the same team as the senior players or alternated between that team and the junior team. The junior team played in the Portuguese National Junior Second Division. In association football, the penalty kicks taken by a team are mostly taken

Results

In this section we consider (1) measures concerning penalty kick outcome, (2) correlations between single kinematic measures and penalty kick direction, (3) multiple regressions with several kinematic measures predicting penalty kick direction, (4) predictions of dive direction, and (5) percentages of correctly predicted penalty kick directions.

Discussion

The aims of the present study were (1) to determine the information value of single kinematic variables, related to the movement of penalty takers, as predictors of the direction of the penalty kick, (2) to determine the information value of compound kinematic variables, and (3) to determine the effect of deception on the information value of these kinematic variables. In contrast to previous studies (Diaz et al., 2012, Lees and Owens, 2011), our experiment was performed with a regular-size

Acknowledgments

The authors wish to thank all volunteers for their help during data collection and Marta Mendes for her collaboration on many aspects of the study and for the supervision of the volunteers. We thank Chris Button and Raoul Huys for their valuable comments on earlier versions of the manuscript. The authors also wish to thank to the Portuguese Foundation for Science and Technology for the financial support to this research manuscript.

References (36)

  • D. Araújo et al.

    The ecological dynamics of decision making in sport

    Psychology of Sport & Exercise

    (2006)
  • M. Dicks et al.

    Individual differences in the visual control of intercepting a penalty kick in association football

    Human Movement Science

    (2010)
  • R.C. Jackson et al.

    Anticipation skill and susceptibility to deceptive movement

    Acta Psychologica

    (2006)
  • B. Abernethy et al.

    Expertise and the perception of kinematic and situational probability information

    Perception

    (2001)
  • C. Button et al.

    Statistical modelling of gaze behaviour as categorical time series: What you should watch to save soccer penalties

    Cognitive Processing

    (2011)
  • P.A. Cabe et al.

    Characterizing perceptual learning using regression statistics: Development of a perceptual calibration index

    American Journal of Psychology

    (2010)
  • J. Cohen et al.

    Applied multiple regression/correlation analysis for the behavioral sciences

    (2003)
  • G.J. Diaz et al.

    Anticipation from biological motion: The goalkeeper problem

    Journal of Experimental Psychology: Human Perception and Performance

    (2012)
  • M. Dicks et al.

    Examination of gaze behaviours under in situ and video simulation task constraints reveals differences in information pickup for perception and action

    Attention, Perception, & Psychophysics

    (2010)
  • M. Dicks et al.

    Availability of advance visual information constrains association-football goalkeeping performance during penalty kicks

    Perception

    (2010)
  • M. Dicks et al.

    Deception, individual differences and penalty kicks: Implications for goalkeeping in association football

    International Journal of Sports Science & Coaching

    (2011)
  • B.R. Fajen et al.

    Learning to control collisions: The role of perceptual attunement and action boundaries

    Journal of Experimental Psychology: Human Perception and Performance

    (2006)
  • I.M. Franks et al.

    Cues for goalkeepers: High-tech methods used to measure penalty shot response

    Soccer Journal

    (1997)
  • J.J. Gibson

    The ecological approach to visual perception

    (1979)
  • R. Huys et al.

    Global information pickup underpins anticipation of tennis shot direction

    Journal of Motor Behavior

    (2009)
  • R. Huys et al.

    On the dynamic information underlying visual anticipation skill

    Attention, Perception, & Psychophysics

    (2008)
  • D.M. Jacobs et al.

    Individual differences and the use of nonspecifying variables in learning to perceive distance and size: Comments on McConnell, Muchisky, and Bingham (1998)

    Attention, Perception, & Psychophysics

    (2001)
  • D.M. Jacobs et al.

    Lateral interception I: Operative optical variables, attunement, and calibration

    Journal of Experimental Psychology: Human Perception and Performance

    (2006)
  • Cited by (41)

    • Affordance-based control in deceptive and non-deceptive penalties in soccer goalkeeping: Gender matters!

      2022, Psychology of Sport and Exercise
      Citation Excerpt :

      This conception is similar as in more cognitive approaches to deception (e.g., Jackson & Cañal-Bruland, 2019), but distinguishes itself because the effectiveness of deception is also assumed to depend on the goalkeeper’s action capabilities. Thus, by virtue of the stringent time constraints in the penalty goalkeeping, goalkeepers have to dive before the ball is being kicked (Dicks et al., 2010a; Navia et al., 2017) and would presumably use early information from the runup of the taker to predict future ball direction (Dicks et al, 2010c; Dicks et al, 2011), which is less reliable and more likely to encompass misleading information if the kicker tries to deliberately deceive the kicker (Lopes et al., 2014). Consequently, it has been proposed that initiating the lateral dive early is likely to make the goalkeeper more susceptible to deception (Dicks et al., 2011).

    • Kinematic patterns underlying disguised movements: Spatial and temporal dissimilarity compared to genuine movement patterns

      2017, Human Movement Science
      Citation Excerpt :

      Our data reveal that the highest spatial dissimilarities between disguised and non-disguised throws occurred in the final movement phase (close to ball release). This effect has often been reported in the literature (Brault et al., 2010; Lopes et al., 2014), thus, kinematic information available in the final movement phase is plausibly difficult to deploy when adapting one’s own motor behavior to the observed action. Consequently, it might still be argued that expert athletes are highly capable of disguising their action intentions in their movements when trying to manipulate an opponent’s anticipatory process.

    View all citing articles on Scopus
    View full text