This article was downloaded by: [Vrije Universiteit, Library]
On: 10 June 2011
Access details: Access Details: [subscription number 907218092]
Publisher Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 3741 Mortimer Street, London W1T 3JH, UK
Ergonomics
Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713701117
Anticipation and visual search behaviour in expert soccer goalkeepers
Geert J. P. Savelsberghab; John Van der Kampa; A. Mark Williamsc; Paul Wardd
a
Research Institute for Fundamental and Clinical Human Movement Sciences, Vrije Universiteit
Amsterdam, Amsterdam, The Netherlands b Institute for Biophysical and Clinical Research into
Human Movement, Manchester Metropolitan University, Manchester, UK c Research Institute for Sport
and Exercise Sciences, Liverpool John Moores University, Liverpool, UK d Learning Systems Institute,
Florida State University, USA
To cite this Article Savelsbergh, Geert J. P. , Van der Kamp, John , Williams, A. Mark and Ward, Paul(2005) 'Anticipation
and visual search behaviour in expert soccer goalkeepers', Ergonomics, 48: 11, 1686 — 1697
To link to this Article: DOI: 10.1080/00140130500101346
URL: http://dx.doi.org/10.1080/00140130500101346
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf
This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.
Ergonomics,
Vol. 48, Nos. 11 – 14, 15 September – 15 November 2005, 1686 – 1697
Anticipation and visual search behaviour
in expert soccer goalkeepers
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
GEERT J. P. SAVELSBERGH*yz, JOHN VAN DER KAMPy,
A. MARK WILLIAMSx and PAUL WARD}
yResearch Institute for Fundamental and Clinical Human Movement Sciences,
Vrije Universiteit Amsterdam, Amsterdam, The Netherlands
zInstitute for Biophysical and Clinical Research into Human Movement,
Manchester Metropolitan University, Manchester, UK
xResearch Institute for Sport and Exercise Sciences, Liverpool John Moores University,
Liverpool, UK
}Learning Systems Institute, Florida State University, USA
A novel methodological approach is presented to examine the visual search
behaviours employed by expert goalkeepers during simulated penalty kick
situations in soccer. Expert soccer goalkeepers were classified as successful or
unsuccessful based on their performance on a film-based test of anticipation
skill, thereby allowing an intra-group comparison of visual search behaviour
on the task. The anticipation test required participants to move a joystick in
response to penalty kick situations presented on a large screen. The
proportion of penalties saved was assessed as well as the frequency and time
of initiation of joystick corrections. Visual search behaviour was examined
using a portable eye movement registration system. The successful experts
were more accurate in predicting the height and direction of the penalty kick,
waited longer before initiating a response and appeared to spend longer
periods of time fixating on the non-kicking leg compared with the nonsuccessful experts.
Keywords: Ball flight; Eye movement; Fixation; Penalty kicks
1. Introduction
Successful performance in sport requires skill in perception as well as the efficient and
accurate execution of movement patterns (see Williams et al. 1999, Savelsbergh et al.
2002). The awareness that skilled perception precedes appropriate action has led
researchers to examine its role in sport performance. For example, researchers using the
temporal occlusion paradigm have shown that experts are superior to novices in using
predictive information (‘advance cues’) from an opponent’s body movements to guide
their anticipatory responses (see Abernethy 1987, Williams and Burwitz 1993, Abernethy
et al. 2001). In this approach, participants are presented with filmed sequences that are
*Corresponding author. Email: G_J_P_savelsbergh@fbw.vu.nl
Ergonomics
ISSN 0014-0139 print/ISSN 1366-5847 online ª 2005 Taylor & Francis
http://www.tandf.co.uk/journals
DOI: 10.1080/00140130500101346
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
Visual search in goalkeepers
1687
representative of their customary view of the action. These film clips are selectively edited
to provide a varying extent of predictive information from the opponent’s body
movements and the ball’s flight path, with participants being required to predict the
end result of the sequence observed. The expert performer’s superiority over the
novice has been demonstrated in a range of sports, including soccer (Keller et al. 1979,
Neumaier et al. 1987, McMorris et al. 1993, Williams and Burwitz 1993, McMorris and
Colenso 1996).
Although researchers have attempted to determine the important sources of information
utilized by expert goalkeepers during the soccer penalty kick, findings are somewhat
contradictory (for detailed reviews, see Williams 2000, Van der Kamp 2001). On the one
hand, experimental evidence suggests that movement of the hips, kicking leg and trunk
just before and during contact is important (Tyldesley et al. 1982, Williams and Burwitz
1993). On the other hand, others have argued, based on data from notational analysis,
that the orientation of the non-kicking foot just before ball contact is more predictive
(Franks and Hanvey 1997). However, researchers have typically relied on potentially less
direct measures of information pick-up, such as verbal or written reports (Williams and
Burwitz 1993) or event occlusion techniques, and there have been few attempts to record
goalkeepers’ visual behaviour using eye movement registration techniques.
Savelsbergh et al. (2002) used a novel methodological approach to examine skill-based
differences in anticipation and visual search behaviour during the penalty kick in soccer.
Expert goalkeepers who played semi-professional soccer (second division of the National
League and highest amateur league) in the Netherlands and novice goalkeepers were
required to move a joystick in response to penalty kick situations presented on film. The
expert goalkeepers stopped more penalties and were generally more accurate in predicting
the direction of the penalty kick, waited longer before initiating a response and made
fewer corrective movements with the joystick. In addition, visual search behaviour was
examined using an eye-movement registration system. The expert goalkeepers employed a
more efficient search strategy involving fewer fixations of longer duration to less disparate
areas of the display. The novices spent longer periods of time fixating on the trunk, arms
and hips, whereas the experts spent more time fixating the head and found the kicking leg,
non-kicking leg and ball areas to be more informative, particularly as the moment of
foot – ball contact approached.
The majority of researchers interested in visual search behaviour in sport have
attempted to identify differences in point-of-gaze as a function of skill, experience or
sometimes age. The customary approach has been to compare different levels of expertise
(i.e. experts vs. novices) in order to identify consistent skill-based differences in visual
search strategies. Typically, researchers have neglected to examine whether successful
performers employ different visual search patterns than unsuccessful performers within a
group where the participants are presumed to have a similar level of expertise. In order to
be able to make such a comparison, one has to create groups using a within-task criterion
such as the number of penalties saved (Whiting 1986). The need to identify intra-group
differences is particularly important if one acknowledges that not all expert goalkeepers
are successful in stopping penalty kicks (Franks and Hanvey 1997, Van der Kamp 2001).
For instance, an analysis of penalty kicks within a sample of expert goalkeepers (German
Bundesliga) shows that the difference between successful and unsuccessful goalkeepers in
stopping a penalty kick amounts to 30% (see figure 1; Van der Kamp 2001). A withingroup comparison, therefore, may disclose subtle differences in visual search behaviour
that may help to reveal the determinants of successful performance. These determinants
may have remained obscure in studies that relied exclusively on expert – novice
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
1688
G. J. P. Savelsbergh et al.
Figure 1. Frequency distribution of the percentage of penalty kicks stopped by
goalkeepers in the German Bundesliga between 1963 and 1997. Only goalkeepers who
faced 10 or more penalty kicks were selected. As a result the distribution represents 2615
out of the total of 3102 penalty kicks faced by 91 goalkeepers. Both skewness (i.e. 0.488,
SE = 0.253) and kurtosis (i.e. 0.216, SE = 0.500) are smaller than twice the standard
error, the frequency data therefore are normally distributed. Hence, an unsuccessful
penalty saver can be defined as goalkeeper who stops 5% or fewer penalty kicks
( p0.05 = 4.8%), whereas a successful penalty saver stops 34% or more ( p0.95 = 34.1)
(adapted from Van der Kamp, 2001; raw data from Kropp and Trapp, 1999).
comparisons. The aim of this study was to examine whether there are differences in visual
search behaviour within a group of expert-level goalkeepers. Goalkeepers were classified
as successful or unsuccessful based on their performance on a film-based test of
anticipation skill involving the soccer penalty kick. It was anticipated that this withingroup comparison would help resolve the present debate regarding the key predictive
source(s) of information used by goalkeepers when attempting to anticipate the direction
of the penalty kick.
2. Method
2.1. Participants
Sixteen goalkeepers (mean age = 25.7, SD = 7.1 years) playing in one of the three highest
leagues in The Netherlands (Premier, First division and semi-professional division)
Visual search in goalkeepers
1689
provided informed consent prior to participating and were free to withdraw from testing
at any stage. Data from seven of these players were reported previously by Savelsbergh
et al. (2002). Participants were subsequently divided into two groups, successful (SE) and
non-successful (NE) experts, based on their ability to predict direction correctly in a test
of anticipation skill as detailed in the Results section.
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
2.2. Test film
The test film was produced in conjunction with PSV Eindhoven Football (soccer) Club.
Ten youth players (mean age = 18.9, SD = 1.5 years) taking penalty kicks were filmed
from the goalkeeper’s perspective. The film clips were recorded using a digital video
camera (Canon XM 1) positioned in the middle of the goal at a height of 1.77 m.
A 2.42 m 6 1.50 m sailcloth was hung from a regulation cross-bar to indicate the area to
which the players had to shoot. Six different target areas (0.81 m61.50 m) were painted on
the sailcloth. The players were asked to try to disguise the intended target of the penalty
kick, as they would in a normal competitive situation. Each film clip included the penalty
taker’s approach to the ball, their actions prior to and during ball contact and the first
portion of ball flight. Two penalties were recorded in each target location for every player,
providing a total of 120 trials. A microphone was attached to the sailcloth to indicate the
moment at which the ball crossed the goal line, whilst a second microphone was positioned
near the penalty spot to record the moment of ball – foot contact. These two temporal
measures were employed to calculate the flight time and velocity for each penalty kick. The
average ball flight time was 648 ms and the mean ball velocity was 16.84 m s71.
2.3. Apparatus
The film clips were back-projected (EIK CC-7000), using a reflective surface to increase
image size, onto a large screen (2.29 m 6 2.27 m) positioned 3.45 m from the participant.
The experimental layout was the same as that reported by Savelsbergh et al. (2002). The
image of the penalty taker subtended a visual angle of approximately 88 at foot – ball
impact, thereby closely simulating the real image size and distance between the
goalkeeper and the penalty spot.
The response movements performed by the participants were recorded using a handheld joystick. The joystick (Dual Axis Farnell M11Q61P) was positioned at waist height
just in front of the participant. The joystick signal in millivolts was amplified and stored
on computer by means of LABVIEW (version 5.1) and could be moved through 3608.
The film clip and the joystick were synchronized in milliseconds by means of a 5 V signal
that marked the start and end of the film clip.
Visual search behaviours were recorded using an eye – head integration (EHI) system
that included an Applied Science Laboratories (ASL) 4000SU eye-tracker and an
Ascension Technologies magnetic head tracker (model: 6DFOB). The EHI is a videobased monocular system that measures eye line-of-gaze using head-mounted optics. The
system works by collecting three pieces of information: displacement between the left
pupil and corneal reflex (reflection of the light source from the surface of the cornea),
position of the eye in the head and position and orientation of the head in space. The
relative position of these features is used to compute visual point-of-gaze with respect to a
pre-calibrated nine-point grid projected onto the scene plane. A simple eye calibration
was performed to verify point-of-gaze before each participant was tested. The calibration
was checked following each block of ten trials. The data were superimposed onto the
1690
G. J. P. Savelsbergh et al.
scene in the form of a positional cursor to highlight point-of-gaze. This image was then
stored using a video recorder for further analysis. The data were subjected to a frameby-frame analysis using a PAL standard video recorder (Panasonic AG7330) at 50 Hz.
The accuracy of the system was + 18 visual angle. System precision (i.e. amount of
instrument noise in the eye position measure when the eye is perfectly stationary) was
better than 0.58 in both the horizontal and the vertical direction.
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
2.4. Procedure
After obtaining informed consent, the participants were positioned behind the joystick.
They had to anticipate the direction of each penalty kick quickly and accurately by
moving the joystick as if to intercept the ball. If the joystick was positioned in the correct
location at the moment the ball crossed the goal line, the penalty was judged a successful
save. Participants were allowed to use the joystick to make corrections to their initial
decision as the penalty kick evolved. No feedback was given as to their performance on
each trial. Before the penalties were presented, a test was undertaken to determine
whether there were baseline differences between the two groups in simple reaction time.
Instead of a penalty clip, a star was presented at one of six possible locations and the
participant had to move the joystick as quickly as possible to the correct position. A star
was randomly presented at each location four times, providing a total of 24 test trials.
These trials helped the participants to familiarize themselves with the movements of
the joystick.
After the reaction time test, five practice trials were carried out using the penalty clips
to familiarize the participants with the experimental protocol. Thirty film clips were then
presented to the participants, five penalties in each of six locations. These film clips were
chosen from the original sample of 120 trials by a panel of three experienced soccer
coaches as being representative of typical penalty kick scenarios. An equal mix of rightand left-footed penalty takers was included. The end location of the penalties was
completely randomized, but kept in the same order for each participant.
2.5. Dependent variables and analysis
2.5.1. Reaction time. The reaction time was defined as the time period between the onset
of the star stimulus and the initiation of the joystick movement (in ms). The start of the
movement was defined as the moment at which the velocity of the joystick exceeded 5%
of its peak velocity. This period was intended as a baseline measure of reaction time.
2.5.2. Anticipation test. The following measures were recorded from the anticipation test:
. Penalties saved – the percentage of trials for which the joystick was positioned in the
correct location at the moment the ball crossed the goal line.
. Correct side – the percentage of trials for which the joystick was positioned on the
correct side (i.e. right or left judgement) at the moment the ball crossed the goal line.
. Correct height – the percentage of trials for which the joystick was positioned at the
correct height (i.e. high or low judgement) at the moment the ball crossed the goal line.
. Proportion of corrections – the percentage of trials where corrective movements (i.e. a
joystick movement that was initially directed toward an incorrect location) of the
joystick occurred before the ball passed the goal line. This dependent variable was
interpreted as an indication of a change in the participant’s decision.
1691
Visual search in goalkeepers
. Time of initiation of joystick movement – the time when the participant began to move
the joystick relative to foot – ball contact by the penalty taker (in ms). A minus sign
indicates that the joystick was moved before foot – ball contact.
2.5.3. Visual search data
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
. Percentage viewing time – the amount of time participants spent fixating various areas
of the display when attempting to anticipate ball direction. Eight fixation location
categories were used, including at the head, hips, kicking leg and non-kicking leg. If
the fixation was not in one of these areas, it was unclassified.
. Search rate – a fixation was defined as the period of time when the eye remained
stationary within 1.58 of movement tolerance for a period equal to, or greater than,
120 ms (see Williams et al. 1999). From this measure, the following was calculated: the
number of visual fixations, the number of areas fixated and the mean fixation duration
per trial.
Each dependent measure was analysed separately using a one-way analysis of variance
in which ‘group’ was the between-participants factor. With regard to percentage viewing
time, we were interested only in a specific and limited number of contrasts based on
earlier reported research (Savelsbergh et al. 2002), for instance, identifying whether group
differences existed for time spent fixating each body segment only rather than contrasting,
for instance, time spent fixating on the kicking leg for the SE group compared with time
spent fixating on the trunk for the NE group. Partial Z2 was also calculated for each main
effect as a measure of meaningfulness.
3. Results
In table 1 the percentage of penalties stopped is presented separately for each goalkeeper.
On the basis of these performance scores, two groups of participants were created for
further analyses. A cumulative method was used to determine the probability that a
penalty could be stopped by randomly selecting from one of six possible locations. The
method showed that only if a participant stopped nine or more penalties was the
probability that a gambling strategy was invoked 5% or less (see table 2). Table 1 shows
that participants 11 – 16 stopped less than nine penalty kicks (i.e. 5 30%). It can be
reliably concluded that the strategy adopted by these participants was no better than one
Table 1. The expert goalkeepers ranked according to the percentage of penalties stopped.
P refers to participants playing in the Premier League, F to goalkeepers of the First division
and S to semi-professional goalkeepers [the latter group of participants was also included in
the Savelsbergh et al. (2002) study].
Rank
1
2
3
4
5
6
7
8
Division
Score
Rank
Division
Score
F
S
F
S
F
S
S
F
63
57
50
43
37
37
33
33
9
10
11
12
13
14
15
16
S
F
S
P
F
S
F
P
33
30
26
23
23
20
17
17
1692
G. J. P. Savelsbergh et al.
Table 2. Probability table for randomly selecting one out of six locations. The number of
locations is the number of locations correctly selected at random with 30 repetitions. The
probability of occurrence is the probability that a particular number of locations is correctly
selected at random (e.g. the probability that five locations are correctly selected at random is
19%). The cumulative probability is the probability that more than a particular number of
locations are correctly selected at random [e.g. the probability that five or more locations are
correctly selected at random (i.e. 0.575) is the total sum of the probabilities of occurrence of
five and more locations that are correctly selected].
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
Number of correct locations
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Probability of occurrence
Cumulative probability
0.004
0.025
0.073
0.136
0.184
0.192
0.160
0.109
0.063
0.030
0.012
0.004
0.001
0.000
0.000
2.161075
4.061076
6.661077
9.561078
1.261078
1.361079
1.2610710
1.0610711
7.1610713
4.2610714
2.0610715
7.7610717
2.3610718
4.9610720
6.7610722
4.5610724
1
0.995
0.970
0.897
0.760
0.575
0.383
0.223
0.113
0.050
0.019
0.006
0.002
0.000
0.000
2.661075
4.761076
7.761077
1.061077
1.361078
1.461079
1.3610710
1.1610711
7.6610713
4.4610714
2.0610715
7.9610717
2.3610718
4.9610720
6.8610722
4.5610724
where locations were randomly selected, and therefore, these expert goalkeepers
comprised the NE group. The second group of experts comprised participants 1 – 6,
who were much more successful in stopping penalties (37 – 63%). This group was labelled
the SE group. In the remainder of this section, these two groups are compared on
anticipation and visual search.
3.1. Anticipation test
The mean group performance variables are presented in table 3. Not surprisingly,
given the within-task criterion, the SE group saved significantly more penalties
[F(1,10) = 31.64, p = 0.000, Zp2 = 0.760] and were more accurate in predicting the height
1693
Visual search in goalkeepers
Table 3. The dependent measures recorded on the anticipation test across groups
(mean + SD). The novice data are from Savelsbergh et al. (2002).
Experts
Penalties stopped (%)
Correct height (%)
Correct side (%)
Proportion of corrections (%)
Time of initiation of joystick
Movement (ms)a
Reaction time (ms)
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
a
Successful
Non-successful
Novices
47.8 + 10.9
49.4 + 9.3
93.3 + 6.9
22.5 + 9.4
7230 + 69.0
21.1 + 4.0
30.6 + 8.2
68.9 + 16.3
21.1 + 8.9
7359 + 110
25.9 + 10.8
32.6 + 8.2
71.4 + 8.2
38.5 + 15.3
7479 + 292
238 + 33.2
262 + 21.2
257 + 46.4
A minus sign indicates that the joystick is moved before foot–ball contact.
Table 4. Fixation duration, number of fixation locations and number of fixations
(mean + SD). The novice data are from Savelsbergh et al. (2002).
Experts
Fixation duration (ms)
Number of fixation locations
Number of fixations
Successful
Non-successful
Novices
501 + 129
3.0 + 0.4
3.5 + 0.6
529 + 129
2.9 + 0.6
3.3 + 0.8
430 + 76
3.1 + 0.5
4.0 + 0.5
[F(1,10) = 13.827, p = 0.004, Zp2 = 0.580] and side of the penalty kicks [F(1,10) = 11.41,
p = 0.007, Zp2 = 0.533] than the NE group.
The SE group initiated the joystick movement nearer (7230 ms) to the foot – ball
contact (i.e. later in the run-up) than the NE group (7359 ms) [F(1,10) = 5.90, p = 0.035,
Zp2 = 0.371]. These results were not attributable to differences in general reaction time
between groups assessed on the star reaction test [F(1,10) = 2.25, NS]. Finally, no
differences were revealed between the two groups for the percentage of corrective
movements of the joystick [F(1,10) = 0.75, NS].
3.2. Visual search data
The visual search variables are presented in table 4. The measures of search rate were
analysed separately using a one-way analysis of variance in which ‘group’ (UE, SE) was
the between-participants factor. There were no significant differences between the SE
group and the UE group for fixation duration [F(1,10) = 0.14], number of fixations
[F(1,10) = 0.38] and the number of areas fixated per trial [F(1,10) = 0.26].
Percentage viewing time is presented graphically in figure 2. Pre-planned comparisons
performed on the percentage viewing time data using separate one-way analyses of
variance for each fixation location (head, hips, kicking leg and non-kicking leg) with
‘group’ (NE, SE) as between-participants factor showed that the SE group spent more
time fixating the non-kicking leg [F(1,10) = 5.50, p = 0.041, Zp2 = 0.355] than the NE
group, whereas a trend for the NE group to look longer at the head was found [F(1,10) =
3.64, p = 0.08, Zp2 = 0.267]. Moreover, in the case of the SE group the proportion of time
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
1694
G. J. P. Savelsbergh et al.
Figure 2. The percentage time spent viewing each fixation location for the successful
and non-successful expert as a function of locations.
designated as ‘unclassified’ was larger than for the NE group [F(1,10) = 5.16, p = 0.047,
Zp2 = 0.340]. There were no significant differences between the groups with respect to the
regions hips [F(1,10) = 0.17] and kicking leg [F(1,10) = 0.01].
4. Discussion
The aim of this study was to examine differences in anticipation and visual search
behaviour between successful and non-successful expert goalkeepers in saving soccer
penalty kicks. The successful experts were significantly more accurate in predicting the
height and direction of the penalty kick, waited longer before initiating a response and
spent longer periods of time fixating on the non-kicking leg in comparison to the
unsuccessful experts. It appears that the difference between successful and less successful
expert performers in stopping a penalty is determined by a combination of when to
initiate a response and attention to the non-kicking leg.
In order to obtain a complete picture with respect to the role of visual search and
anticipatory performance in saving penalty kicks, the data from the successful and less
successful expert goalkeepers were compared with those collected from novice
goalkeepers in a previous study by Savelsbergh et al. (2002). To facilitate the comparison
process we have presented data from the novice group in tables 3 and 4. No significant
differences between the NE group in the current study and the novice group used by
Savelsbergh et al. (2002) were found, with the exception of percentage of joystick
corrections [a one-way ANOVA comparing the NE-group of the current study with the
novices of the Savelsbergh et al. (2002) study revealed a significant effect only for
proportion of corrections; F(1,11) = 5.98, p = 0.032, Zp2 = 0.352].
Although there were no differences in general visual search characteristics such as mean
duration, number of fixations or number of fixation locations across the two expert
groups in the present study, the experiment did reveal a difference between the groups
with respect to the particular regions fixated. This difference is not in the amount of
information, but in the nature of the information picked up by the goalkeepers. The nonsuccessful experts showed a trend to fixate more on the head, whereas the successful
experts fixated more on information from the non-kicking leg. The present findings
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
Visual search in goalkeepers
1695
corroborate the ideas of Franks and Hanvey (1997), who argued that the orientation of
the non-kicking leg is the optimal source of information available prior to foot – ball
contact. According to Franks and Hanvey (1997), the non-kicking foot is oriented such
that it points towards the ball’s likely destination, and appears to be reliable in 80% of
penalty kicks. The positioning of the non-kicking leg occurs at 200 – 250 ms prior to ball
contact (Franks and Hanvey 1997), which makes it the most suitable source of
information. Alternative sources of information (e.g. orientation of the kicking leg, foot –
ball contact) may be more predictive but may not leave the goalkeeper sufficient time to
make the initial anticipatory decision required to save the penalty. When experts fixate
other areas, it is likely to act as a confirmatory source of information as well as to allow
subtle on-line adaptations and modifications to take place within the scope of the
available constraints (e.g. to make changes in hand position).
It is important to stress that the assumed importance of the non-kicking leg as a
determinant of successful performance in stopping a penalty kick does not conflict with
the observations from previous studies that have identified the head, kicking leg and/or
foot – ball contact as the most important sources of information (Tyldesley et al. 1982,
Williams and Burwitz 1993, Savelsbergh et al. 2002). The expert groups in these studies
probably included expert goalkeepers using less successful strategies. For example, two of
the seven expert goalkeepers in the Savelsbergh et al. (2002) study were designated to the
NE group (see table 1). The NE group performed at a similar level and showed a similar
search rate to the novice group (table 3). It is not unlikely therefore that the expert groups
in previous studies may have been ‘contaminated’ with goalkeepers using a gambling
strategy, which would have led to the identification of sources of information that were
either less reliable or occurred very late during the run-up, leaving the goalkeeper
insufficient time to stop the penalty. In contrast, the current study shows that the
differences in anticipation skill between successful and less successful goalkeepers is
probably due to subtle differences in visual search behaviour based around the extraction
of information from the non-kicking leg. Manipulation of information sources is
required, perhaps by spatial occlusion, to further assess whether it is the non-kicking leg
per se, or whether it is an even more subtle source of information, for instance the relative
motion between the kicking and non-kicking leg, that makes the difference between
successful and less successful penalty stoppers.
The successful expert goalkeepers initiated their joystick movements later or nearer to
foot – ball contact than the NE group and novices. The clear difference between successful
and non-successful expert goalkeepers in the initiation of the movement would be
expected, given that the potential sources of information become more specific (and hence
more reliable) to the ball’s destination with the unfolding of the run-up. In the film clips
we used, the non-kicking leg is positioned around 350 ms before foot – ball contact.
Considering the initiation of the joystick (230 ms before contact) and visual search data
of the SE group, one can infer that the successful expert goalkeepers picked up this
specific information prior to the initiation of the response. As a result of using such a
highly specific and reliable source of information, the need to correct the response was
minimized, as shown by the smaller proportion of corrected movements (at least when
compared with novices). Similar strategies have been reported when attempting to
intercept a ball in flight (see Oudejans et al. 1997, Rodrigues et al. 1999). Oudejans et al.
(1997) showed that, when attempting to catch fly balls, expert catchers initiated their
movements towards the ball later and made fewer corrective actions prior to interception.
In the same vein, in a one-handed catching experiment, a straighter trajectory of the hand
and an interception point later in time in order to ‘buy extra time’ was found for a
1696
G. J. P. Savelsbergh et al.
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
successful catch (Laurent et al. 1994). It can be concluded that the expert performer tries
to extract information as long as possible – but not too long – before initiating the hit,
catch or joystick movement. Because the detected information is more specific, the need
to correct the movement is diminished.
In conclusion, the present study discerned two strategies among expert goalkeepers in
their attempts to save a penalty kick. Like novices, the non-successful expert goalkeepers
appeared to use a gambling strategy, in which they initiated their action relatively
early. In contrast, the successful expert goalkeepers employed a distinct anticipation
strategy and initiated their actions relatively late in the run-up of the kicker. The
critical difference for success was the use of information related to the non-kicking leg
to decide the ball’s destination. Finally, we would like to stress that penalty saving is
a multifaceted skill. While predicting the direction of the shot is a decisive factor
for a successful save, it is only one out of several important factors in stopping a
penalty kick.
Acknowledgement
We thank Theo de Haan for pointing out the cumulative method as a way to define the
unsuccessful experts group. We also acknowledge the assistance of Jora Wolfgram and
Marije Muthert, who helped with data collection and analysis.
References
ABERNETHY, B., 1987, Anticipation in sport: a review. Physical Education Review, 10, 5 – 16.
ABERNETHY, B., GILL, D.P., PARKS, S.L. and PACKER, S.T. 2001, Expertise and the perception of kinematic and
situational probability information. Perception, 30, 233 – 252.
FRANKS, I.M. and HANVEY, T., 1997, Cues for goalkeepers. High-tech methods used to measure penalty shot
response. Soccer Journal, 42, 30 – 33.
KELLER, D., HENNEMANN, M.C. and ALGRIA, J., 1979, Fussball; der Elfmeter. Analyse des spezifischen
vorbereitenden Anpassungsverhaltens zur Durchfuhrung einer sportlichen Bewegung. Leistungssport, 9,
394 – 398.
KROPP, M. and TRAPP, A., 1999, 35 Jahre Bundesliga-Elfmeter (Kassel: Agon Sportverlag).
LAURENT, M., MONTAGNE, G. and SAVELSBERGH, G.J.P., 1994, The control and coordination of one-handed
catching: the effect of temporal constraints. Experimental Brain Research, 101, 314 – 322.
MCMORRIS, T. and COLENSO, S., 1996, Anticipation of professional soccer goalkeepers when facing right- and
left-footed penalty kicks. Perceptual and Motor Skills, 82, 931 – 934.
MCMORRIS, T., COPEMAN, R., CORCORAN, D., SAUNDERS, G. and POTTERS, S., 1993, Anticipation of
soccer goalkeepers facing penalty kicks. In Science and Football II, T. Reilly, J. Clarys and A. Stibbe
(Eds), 250–253 (London: E & FN Spon).
NEUMAIER, A., POEL, H. TE and STANDTKE, V., 1987, Zur Antizipation des Elfmeterschusses aus der Sicht des
Torwarts. Leistungsport, 17, 23 – 32.
OUDEJANS, R.R.D., MICHAELS, C.F. and BAKKER, F.C., 1997, The effects of baseball experience on movement
initiation in catching fly balls. Journal of Sports Sciences, 15, 587 – 595.
RODRIGUES, S.T., VICKERS, J.N. and WILLIAMS, A.M., 1999, Two visual systems and temporal pressure in table
tennis. Journal of Sport and Exercise Psychology, 21, p. S91.
SAVELSBERGH, G.J.P., WILLIAMS, A.M., VAN DER KAMP, J. and WARD, P., 2002, Visual search, anticipation and
expertise in soccer goalkeepers. Journal of Sports Sciences, 20, 279 – 287.
TYLDESLEY, D.A., BOOTSMA, R.J. and BOMHOFF, G., 1982, Skill level and eye movements patterns in a sportoriented reaction time task. In Motorik- und Bewegungsforschung, H. Rieder, K. Bos, H. Mechling and
K. Reische (Eds), 290 – 29 (Schorndorf: Hofmann).
VAN DER KAMP, J., 2001, De kunst van het stoppen van de strafschop. [The art of stopping a penalty kick.]
Sportpsychologie Bulletin, 12, 53 – 61.
WHITING, H.T.A., 1986, Isn’t there a catch in it somewhere? Journal of Motor Behavior, 18, 486 – 491.
Visual search in goalkeepers
1697
Downloaded By: [Vrije Universiteit, Library] At: 13:15 10 June 2011
WILLIAMS, A.M., 2000, Perceptual skill in soccer: implications for talent identification and development. Journal
of Sports Sciences, 18, 1 – 14.
WILLIAMS, A.M. and BURWITZ, L., 1993, Advance cue utilization in soccer. In Science and Football II, T. Reilly,
J. Clarys and A. Stibbe (Eds), 239 – 243 (London: E & FN Spon).
WILLIAMS, A.M., DAVIDS, K. and WILLIAMS, J.G., 1999, Visual Perception and Action in Sport (London and New
York: E & FN Spon).