Physical Characteristics, Physiological Attributes, and On-Field Performances of Soccer Goalkeepers

International Journal of Sports Physiology and Performance, 2011, 6, 509-524 2011 Human Kinetics, Inc. Physical Characteristics, Physiological Attributes, and On-Field Performances of Soccer Goalkeepers Gal Ziv and Ronnie Lidor The soccer goalkeeper (GK) is required to perform strenuous actions during practice sessions and actual games. One of the objectives of those professionals who work with GKs is to obtain relevant information on physical characteristics and physiological attributes of GKs, and to use it effectively when planning training programs for them. This article has three purposes: (a) to review a series of studies (n = 23) on physical characteristics, physiological attributes, and on-field performances of soccer GKs; (b) to outline a number of methodological limitations and research concerns associated with these studies; and (c) to suggest several practical recommendations for soccer coaches who work with GKs. Four main findings emerged from our review: (a) professional adult GKs usually are over 180 cm tall and have a body mass of over 77 kg; (b) studies on agility and speed produced mixed results, with some showing similar values between GKs and field players and others showing reduced performance in GKs; (c) GKs usually have higher vertical jump values when compared with players playing the various field positions; (d) GKs cover approximately 5.5 km during a game, mostly by walking and jogging. Four methodological limitations and research concerns associated with the reviewed studies were discussed, among them the lack of a longitudinal approach and the lack of on-field performance studies. Three practical recommendations are made for coaches, one of which is that coaches should adopt a careful approach when selecting testing protocols and devices for the assessment of GKs physiological attributes. Keywords: physical fitness, exercise test, strength and conditioning programs The goalkeeper (GK) in soccer is required to demonstrate a high level of proficiency in various actions related to both defensive and offensive aspects of the game. 1 Defensive acts, such as catching a shot, dealing with the high cross, deflecting, positioning near the goal area, punching, and saving in one-on-one situations, as well as offensive acts, such as distribution of the ball and kicking the ball, are performed repeatedly in games and practices. In order to execute the abovementioned defensive and offensive tasks accurately and rapidly, the GK should participate in appropriate preparation training programs. Gal Ziv and Ronnie Lidor are each with the Zinman College of Physical Education and Sport Sciences, Wingate Institute, Israel, and with the Faculty of Education, University of Haifa, Israel. 509

510 Ziv and Lidor Those professionals who work with GKs should obtain relevant information on the physical characteristics and physiological attributes of GKs. This knowledge can be used effectively in planning better training programs for GKs for different phases of the season. Additional information on on-field performances by the GK in actual games can also help coaches plan their training programs, by creating better simulations of how the GK should perform during games. The purpose of this article was threefold: (a) to review a series of studies (n = 23) on physical characteristics, physiological attributes, and on-field performances of soccer GKs national team players, professional players playing for major leagues in Europe, and elite youth players; (b) to outline a number of methodological limitations and research concerns associated with the reviewed studies; and (c) to suggest several practical recommendations for soccer coaches and strength and conditioning coaches who work with GKs. A search was conducted using computerized databases (PubMed and SPORT- Discus) for peer-reviewed articles in the English language. All articles that had data on youth and adult soccer GKs were collected. The criteria for exclusion were as follows: pooled data for players playing different positions, articles dealing with injuries in GKs, and articles focusing on psychological preparation of GKs. An initial search in PubMed for the term goalkeeper* provided 106 results, out of which 31 were relevant to our investigation. A subsequent search in SPORTDiscus provided 113 results, out of which 20 had not been presented in the PubMed search. A follow-up search in PubMed with the addition of the term soccer to the term goalkeeper* provided 62 results, out of which three were new. Searches were conducted during the month of December 2009. A follow-up search was conducted in July 2010 and one new relevant article was found. Altogether, 55 articles on GKs were found. Out of those, 23 articles matched our criteria and were included in our review. Physical Characteristics A summary of the physical characteristics of soccer goalkeepers is presented in Table 1. In general, adult GKs are taller than 180 cm. Only three studies found height values of less than 180 cm. In one of these studies, 2 the participants were under the age of 14; in another study, 3 the participants were adolescents with an average age of 16.8, and in the third study, 4 data were combined for both adult players and adolescent players. Concerning body mass, when these three articles are not included, the average body mass of GKs is over 77 kg. Therefore, it can be observed that professional adult goalkeepers are usually over 180 cm tall and have a body mass of over 77 kg. As can be seen in Table 1, only 8 out of the 16 studies reported percent body fat (%BF). Most values were between 12 and 14% fat, while one study reported 20% fat. 5 This was the only study that used bioelectric impedance rather than skinfold measurement or hydrostatic weighting. Both skinfold measurements and bioelectric impedance can accrue error. Percent body fat estimated from skinfold measurements can have a standard error of the estimate (SEE) of 3.4 3.9% fat. 6 The accuracy of bioelectric impedance can be compromised by a number of factors, such as instrumentation, subject factors (eg, hydration level, exercise), technician s skill (eg, placement of sensors, body position), environmental factors (eg, tempera-

Characteristics of Soccer Goalkeepers 511 ture), and the prediction equation used. 7 For example, dehydration can reduce 5 kg from a person s fat-free mass during measurement. 7 Hence, %BF values should be compared with caution among studies. Only one study 3 used hydrostatic weighting as a method for estimating %BF, and a value of 13.7% fat was found. This value coincides with the values found when skinfold measurement was performed. When compared with other players who play different positions, GKs appear to be taller and heavier. In Davis et al s study of the English first and second division leagues, 8 GKs were found to be heavier (86.1 ± 5.5 kg) than fullbacks (75.4 ± 4.6 kg), midfielders (73.2 ± 4.8 kg), and forward (76.4 ± 7.2 kg). In another study, of players from the first league in Croatia, 5 GKs shared similar height and mass with defensive players but were taller by approximately 3 cm and heavier by approximately 4 kg than midfielders and forward. These findings were supported by another study of Croatian players, reporting that GKs were taller, heavier, and had higher %BF than attackers, defenders, and midfielders. 9 Several other studies partially or fully support these data. 4,10 13 In contrast to these studies, one study of a premier league football club in England found no differences in physical characteristics of players playing different positions. 14 However, it is possible that the small number of players in each group that participated in this study precluded a significant finding: only two GKs, seven defenders, 10 midfielders, and five strikers were tested. Despite the lack of statistical significance, the differences between means in height and body mass were important from a practical perspective. Height and body mass values of the GKs were 185 ± 12 cm and 86.3 ± 12.1 kg, respectively, compared with those of defenders (180 ± 7 cm and 82.5 ± 6.6 kg, respectively), midfielders (178 cm and 78.3 ± 6.7 kg, respectively), and strikers (180 ± 6 cm and 80.9 ± 9.3 kg, respectively). A careful calculation of simple effect sizes using the larger standard deviation values shows small to moderate effect sizes. For example, an effect size of.42 was found in stature between GKs and defenders, and an effect size of.58 between GKs and midfielders. For body mass, an effect size of.66 was found between GKs and midfielders, and.31 between GKs and defenders. Lastly, only one study that examined physical characteristics over a full season was found. In this study, Carling and Orhand 15 found that body mass, fat-free mass, and %BF remained stable in GKs from the French League 1 professional soccer club. However, while changes in body mass were far from being statistically significant (P =.93), changes in %BF failed significance testing only after the use of a Bonferroni multiple comparisons correction. Before that correction, changes in %BF were significant (P =.04, effect size =.53). Aerobic Profile Physiological Attributes A summary of the methodologies used and the VO 2 max values reported for GKs is presented in Table 2. Goalkeeper VO 2 max was reported in three studies of youth players 2 4 and in five studies of adult players. 8 10,16,17 In one of the studies on youth players, participants had a wide range of ages (from 14.7 ±.2 to 21.5 ± 1.9 y). 4 The VO 2 max values were estimated using Astrand s nomogram from a submaximal aerobic test on a cycle ergometer, and were reported for the group

Table 1 A summary of the physical characteristics of soccer GKs (means ± SD) Study Participants Level of competition Height (cm) Mass (kg) Fat (%) FFM (kg) Adults Acar et al 32 n = 9 Turkey and Cyprus premier league 186.0 ± 5.0 79.4 ± 7.6 N/A N/A n = 6 2nd and 3rd leagues Arnason et n = 24 17 teams in the Icelandic elite and first n = 68 GKs from 76 clubs in the English Premier 185.2 ± 4.7 81.4 ± 7.7 12.3 ± 5.3 S 71.4* al 10 division Bloomfield Premier League: Premier League: N/A N/A et al 11 League, German Bundesliga, Ital- 188 ± 4 83.3 ± 6.9 n = 50 ian Serie A, Spanish Premiera Division Bundesliga: Bundesliga: 189 ± 4 82.2 ± 6.2 n = 60 Serie A: 186 ± 4 n = 56 Premiera Div.: 185 ± 4 Serie A: 79.1 ± 5.5 Premiera Div.: 81.1 ± 4.3 Carling & N/A GKs from a French League 1 club N/A Start preseason: Orhant 15 81.75 ± 3.81 Davis et al 8 n = 13 English GKs from 8 clubs of the 1st and 2nd divisions End preseason: 83.80 ± 3.08 Midseason: 83.33 ± 4.73 End midseason: 81.47 ± 4.01 Start preseason: 11.63 ± 2.96 End preseason: 12.67 ± 2.51 Midseason: 11.68 ± 3.28 End midseason: 10.27 ± 2.25 Start preseason: 72.20 ± 2.69 End preseason: 73.15 ± 1.97 Midseason: 73.10 ± 2.47 End midseason: 73.04 ± 1.76 End of season: End of season: End of season: 83.28 ± 3.46 12.28 ± 2.79 72.99 ± 1.69 N/A 86.1 ± 5.5 13.3 ± 2.1 S 74.6* (continued) 512

Table 1 (continued) championship in 182.1 80.1 20.2 B 63.9* Study Participants Level of competition Height (cm) Mass (kg) Fat (%) FFM (kg) Duraskovic n = 88 GKs from the 2002 world 186.4 ± 5.3 81.6 ± 6.3 N/A N/A et al 33 in Japan and Korea Guner et al 12 n = 20 GKs from Turkish premier league 187.2 ± 3.7 84.32 ± 6.97 N/A N/A Hencken & n = 2 GKs from premiership football club 185 ± 12 86.3 ± 12.1 N/A N/A White 14 England Matkovic n = 7 GKs from the first league in Croatia et al 5 Ramadan & n = 2 1982 Kuwaiti world cup soccer team N/A N/A 13.0 ±.2 S N/A Byrd 17 Sporis et al 9 n = 30 GKs from 12 teams in the first national league in Croatia 185 ± 3.1 81 ± 2.3 14.2 ± 1.9 S 69.5* Sutton et al 13 n = 8 GKs from English premier league 190 ± 3 91.2 ± 4.6 N/A N/A Taskin 20 n = 42 GKs from different leagues of Turkey 186 ± 3 77.62 ± 3.26 N/A N/A Adolescents Gil et al 4 n = 29 GKs from one club in Spain including Adults and the senior team and cadet teams adolescents, ages 14.5 21.5 y 179.5 ± 5.9 73.95 ± 7.9 12.2 ± 1.7 S 64.9 # Tahara et al 3 n = 6 GKs from one high school in Japan 177.8 ± 3.9 71.4 ± 3.5 13.7 ± 4.1 H 61.6 ± 2.9 Adolescents, ages 16.8 ± 1.1 y Wong et al 2 n = 10 GKs from an under-14 y league in Children Hong Kong 169 ± 6 54.6 ± 7.3 N/A N/A Note. FFM = fat-free mass; *Not available in original paper calculated by authors; #Data provided separately for muscle mass, bone mass, and residual mass; B Bioelectric impedance; S Skinfold measurements; H Hydrostatic weighting. 513

Table 2 A summary of the methodologies used and the VO2max values (ml O2 kg 1 min 1 ) reported for GKs (means ± SD) Study Participants Level of Competition Method of Measurement VO2max Adults Arnason et al 10 n = 24 17 teams in the Icelandic elite and first maximal treadmill test 57.3 ± 4.7 division Aziz et al 16 n = 16 Local top professional soccer league in field test repeated sprint test 50.2 ± 5.3 Singapore Davis et al 8 n = 13 English GKs from 8 clubs of the 1st and field test multistage fitness test 56.4 ± 3.9 2nd divisions Ramadan & Byrd 17 n = 2 1982 Kuwaiti world cup soccer team maximal cycle ergometer test 48.0 ± 4.4 Sporis et al 9 n = 30 GKs from 12 teams in the first national maximal treadmill test 50.5 ± 2.7 league in Croatia Adolescents Gil et al 4 n = 29 GKs from one club in Spain including the Ages, 14.5 21.5 y senior team and cadet teams estimation using Astrand s nomogram from a submaximal aerobic test on a cycle ergometer 48.4 ± 11.1 Tahara et al 3 n = 6 GKs from one high school in Japan maximal treadmill test 54.2 ± 4.5 Adolescents, ages, 16.8 ± 1.1 y Wong et al 2 n = 10 GKs from an under-14 y old league in Age under 14 y Hong Kong maximal treadmill test 55.1 ± 8.5 514

Characteristics of Soccer Goalkeepers 515 as a whole without differentiating between ages. The GKs VO 2 max values were 48.41 ± 11.10 ml O 2 kg 1 min 1. These values were significantly lower than those of field players, which ranged from 57.7 to 62.4 ml O 2 kg 1 min 1. An analysis of percentiles of VO 2 max revealed that 25% of the goalkeepers had a VO 2 max of 39.21 ml O 2 kg 1 min 1, which is a low value for a soccer player of any position, and 10% of GKs had a VO 2 max of 30.4 ml O 2 kg 1 min 1, which is a value lower than the recommended aerobic capacity for sedentary men. 4 Tahara et al 3 reported greater VO 2 max values (54.2 ± 4.5 ml O 2 kg 1 min 1 ) in a group of six GKs between the ages of 16 and 18 from a high school in Japan. 3 The VO 2 max was measured using a maximal treadmill test. Similar to the study conducted by Gil et al, 4 the VO 2 max values of the Japanese GKs were lower than those measured in field players (61.4 ± 5.3 ml O 2 kg 1 min 1 ). Lastly, VO 2 max values of 10 youth GKs (age = 13.4 ±.7 y) were 55.1 ± 8.5 ml O 2 kg 1 min 1. 2 The VO 2 max measurement was performed using a maximal incremental treadmill test. In this study, no differences were indicated between the GKs and field players. The low VO 2 max values found in the study by Gil et al 4 can be partially explained by the fact that these values were estimated from a submaximal cycle ergometer test rather than a treadmill test. The VO 2 max values on a cycle ergometer tend to be lower than those achieved on a treadmill, especially for those accustomed to running and unaccustomed to cycling. 18,19 In adults, VO 2 max values of 48.0 ± 4.4 ml O 2 kg 1 min 1 were found in two GKs of the 1982 Kuwaiti World Cup Soccer Team. 17 Values for outfielders were higher by approximately 3 8 ml O 2 kg 1 min 1. These values were obtained from a maximal cycle ergometer test. Sporis et al observed similar VO 2 max values in a group of GKs from the professional First National League in Croatia. 9 In this study a maximal, incremental treadmill test was used, and the GKs had a mean VO 2 max of 50.5 ± 2.7 ml O 2 kg 1 min 1. Values for outfielders were higher by approximately 9 12 ml O 2 kg 1 min 1. Differences between the GKs and outfielders were also found in players from teams in the Icelandic elite and first division leagues. Arnason et al 10 used a maximal, incremental treadmill test, and VO 2 max values of the GKs were 57.3 ± 4.7 ml O 2 kg 1 min 1. Values for outfielders were higher by approximately 5 6 ml O 2 kg 1 min 1. 10 In two studies, VO 2 max values were estimated from field tests. In the first study, 8 a multistage fitness test was performed for estimating VO 2 max values. Mean VO 2 max for the GKs was 56.4 ± 3.9 ml O 2 kg 1 min 1, which was lower by 4 ml O 2 kg 1 min 1 than that of outfielders (60.4 ± 3.0 ml O 2 kg 1 min 1 ). In the second study, 16 VO 2 max was estimated from a repeated sprint test in players from a local professional soccer league in Singapore. The VO 2 max values were 50.2 ± 5.3 ml O 2 kg 1 min 1 in the GKs and 54.3 ± 3.4 ml O 2 kg 1 min 1 in outfielders. While variations exist in the values reported in these studies, the GKs had lower VO 2 max values when compared with outfielders. These values range from approximately 48 to 56 ml O 2 kg 1 min 1. Looking at aerobic capacity from a different perspective, Guner et al 12 conducted a modified maximal incremental treadmill test in 197 players from the Turkish Premiere League. The protocol included 30 s pauses between intervals for blood sampling. Running velocities and heart rates were reported for four blood lactate concentrations: 2, 2.5, 3, and 4 mmol L 1. The GKs ran at significantly lower velocities at each blood lactate concentration compared with the outfielders.

516 Ziv and Lidor For example, at 2 mmol L 1 the GKs ran at 11.33 km h 1, whereas the rest of the players ran at velocities ranging from 12.93 to 13.62 km h 1, and at 4 mmol L 1 the GKs ran at 14.15 km h 1, whereas the rest of the players ran at velocities ranging from 15.56 to 16.05 km h 1. No differences in heart rates for each blood lactate concentration were found between players playing different positions. The VO 2 max of GKs was indeed found to be lower than that of field players; however, improving VO 2 max in GKs would be beneficial. For example, a high aerobic capacity can delay fatigue by indirectly improving recovery between bouts of intense movements. It is possible that slow long-distance endurance training is over-emphasized in GKs. Such training is very limited in improving aerobic capacity. Instead, placing an emphasis on high-intensity interval training can help improve GKs VO 2 max. Agility and Speed A summary of the agility and speed characteristics of GKs is presented in Table 3. Agility and speed are integral to the game of soccer. Not only field players require these physiological attributes, but also GKs need to be agile, as their role requires quick movements and directional changes. 1 Goalkeepers appear to share some of the agility and speed attributes of field players. For example, in a study of 243 soccer players from different leagues in Turkey, Taskin 20 reported performances of a 30-m sprint, a four-line sprint, and a speed-dribbling test. The GKs in this study performed as well as field players in the 30-m sprint (4.26 ±.13 s) and the fourline sprint (14.19 ±.26), but were slower than field players in the speed dribbling test (21.14 ±.58). The effect size as calculated from the data in this study was.79. Similarly, a study that assessed repeated sprint ability found that the fastest sprint times were similar in GKs and field players of a local professional soccer league in Singapore. 16 However, the total sprint time was slower in the GKs than that of the field players. The effect size of the difference, as calculated based on the data in this study, was.67. The repeated sprint test in this study included eight 20-m sprints with 20 s of active recovery between the sprints. These results suggest that even though sprint ability is similar in GKs and field players, the GKs fatigue faster. One possible reason for this difference is that the GKs had lower aerobic capacity. Other possible reasons could include local muscular fatigue, environmental factors (eg, dehydration), and psychological factors. Similar results were found in youth soccer players. Wong et al s 2 study of 70 under-14 players competing at the highest level in Hong Kong found that GKs 10-m (2.06 ±.12 s) and 30-m (4.92 ±.32) sprint performance was similar to that of field players. Unlike the previous studies, in a study of 241 soccer players in Spain, 4 30-m sprint times were slower in GKs (3.82 s) when compared with forward (3.61), but similar to those of midfielders (3.68) and defenders (3.70). This was also true for a modified 30-m sprint test using 10 cones that were spread every 3 m. Standard deviation values were not reported, and therefore the effect size could not be calculated. Similarly, results of a 20-m sprint in Croatian elite players were slower in GKs (3.51 ±.9 s) than in defenders (3.36 ±.6), midfielders (3.43 ±.8), and attackers (3.28±.7). 9 However, the effect sizes calculated from the data available were low (.09.25). Lastly, 60-m shuttle run sprint times were slower in GKs (12.71 ±.42 s) when compared with forward (12.19 ±.30) in a sample of 135 players from the

Table 3 A summary of the agility and speed characteristics of GKs (means ± SD) Study Participants Level of Competition Method of Measurement Results Adults Aziz et al 16 n = 16 Local top professional soccer league in Singapore Davis et al 8 n = 13 English GKs from 8 clubs of the 1st and 2nd divisions Taskin H 20 n = 42 GKs from different leagues in Turkey 20 m repeated sprint test Fastest sprint 3.12 ±.09 s Total sprints time 26.00 ±.91 s (slower than FP) 60 m shuttle sprint test 12.71 ±.42 s (slower than forward but similar to rest of FP) 30 m sprint 4.26 ±.13 s Four-line sprint 14.19 ±.26 s Speed dribbling 21.14 ±.58 s (slower than FP) Adolescents Gil et al 4 n = 29 GKs from one club in Spain 30 m flat sprint 3.82 ± N/A Ages 14.5 21.5 y including the senior team and 30 m sprint with 10 cones spread 4.97 ± N/A cadet teams every 3 m (both values slower than forward but similar to rest of FP) Wong et al 2 n = 10 GKs from an under-14-year-old league in Hong Kong 10 m sprint 2.06 ±.12 s Age under 14 y 30 m sprint 4.92 ±.32 s Note. FP = field players. 517

518 Ziv and Lidor first and second divisions in England. 8 This difference represented a high effect size of over 1.0. We believe that the differences in effect sizes between a straight sprint test (.09.25) and a shuttle-run test (over 1.0) when comparing GKs to field players represent a selection process of GKs, and are not due to deficiencies in the GKs training. While GKs are required to have excellent agility skills, these skills are different from those represented in a 60-m shuttle run. In a 60-m shuttle run test, good speed and quick changes of direction are needed in order to perform effectively. Goalkeepers are required to have excellent skills of changing directions, mainly because they must perform well in a relatively small space around the goal. It should be noted that none of the abovementioned studies reported values of effect sizes and observed power; we calculated simple effect sizes based on the means and SDs where available. However, these calculations are only an estimation, as pooled SDs were not calculated. Moreover, in some instance, effect sizes should have been computed from the original ANOVAs, and this was not possible without access to the raw data. Power values are also important in understanding whether the results of the statistical procedure represent a real difference or a lack of difference, or are due only to chance. Power and Strength A summary of the power and strength characteristics of GKs is presented in Table 4. Good vertical jump (VJ) skills are important for GKs, as they are often required to leap vertically to catch or deflect the ball. Three studies reported VJ values of soccer players. 2,9,10 In Arnason et al s study of players from the Icelandic elite and first divisions, a squat jump (SJ) and a countermovement jump (CMJ) were performed. 10 In both jumps, the players hands were held at the waist. The GKs SJ mean value was 35.8 ± 5.3 cm and the CMJ mean value was 38.0 ± 5.6 cm. These values were similar to those of the field players in the study. However, a leg extensor power test was also conducted in this study. The GKs had higher power values (1451 ± 233 W) than all the values of the field players pooled together (1349 ± 196 W). The fact that higher leg extensor power did not result in higher VJ values can be explained by the higher body mass of GKs (81.4 ± 7.7 kg) compared with that of the field players (76.5 ± 6.6 kg). A CMJ test was also performed in Wong et al s study of a group of under-14 players competing in the highest level of competition in Hong Kong. 2 The CMJ values of the GKs were 52.5 ± 5.7 cm; these values were similar to those of the field players. Although it was not mentioned whether the hands were kept at the waist or were allowed to move freely throughout the CMJ, the results suggest that the hands were allowed to move freely. This can explain, at least in part, the better performance of the under-14 players in this study as compared with the jumping scores of adult players in the study by Arnason et al. 10 Indeed, in a study of 11 adult volleyball players, a difference of 14 cm was recorded between VJ without arm swing and VJ with arm swing. 21 In another study a difference of approximately 6 was observed between the two types of VJ. 22 In contrast to these two studies, Sporis et al 9 reported better VJ performance in GKs compared with field players from the first national league in Croatia. The SJ values of the GKs were 46.8 ± 1.5 cm compared with 42.3 ± 2.1 cm in the defenders, 41.49 ± 4.0 cm in the midfielders, and 44.2 ± 3.2 cm in the attackers. Similarly, the CMJ values of the GKs were 48.5 ± 1.5 cm compared with

Table 4 A summary of the power and strength characteristics of GKs (means ± SD) Study Participants Level of competition Method of measurement Results Adults Arnason et al 10 n = 24 17 teams in the Icelandic elite SJ-NS 35.8 ± 5.3 cm and first division CMJ-NS 38.0 ± 5.6 cm Leg extensor squat power test 1451 ± 233 W (higher than FP) Davis et al 8 n = 13 English GKs from 8 clubs of the 1st and 2nd divisions Oberg et al 23 n = 16 GKs from 12 teams of an adult soccer division Sporis et al 9 n = 30 GKs from 12 teams in the first national league in Croatia Adolescents Wong et al 2 n = 10 GKs from an under-14-year-old Age under 14 y league in Hong Kong Isokinetic measurements of knee extension and flexion torque at 120 s -1 2.77±.53 N m kg -1 WAnT Peak power 14.79 ± 1.93 W kg -1 Isokinetic measurements of knee 292 ± 41.8 N m extension and flexion torque at: Isometric 271 ± 24.3 N m 30 s -1 152 ± 18.5 N m 180 s -1 similar to defenders but higher than rest of FP SJ 46.8 ± 1.5 (higher than FP) 48.5 ± 1.5 (higher than FP) CMJ 52.5 ± 5.7 FP = field players; SJ-NS = squat jump with no arm swing; CMJ-NS = countermovement jump with no arm swing; CMJ = countermovement jump no mention of arm movement; SJ = squat jump no mention of arm movement; WAnT = Wingate Anaerobic Test. 519

520 Ziv and Lidor 44.2 ± 1.9 cm in the defenders, 44.26 ± 2.1 cm in the midfielders, and 45.3 ± 3.2 cm in the attackers. No information was provided regarding whether or not hands were held at the waist throughout the jumps. It is unclear from these three studies whether GKs indeed perform better at VJ than field players. Additional studies are needed in order to clarify this issue. In two studies, knee flexion and extension torques were measured using an isokinetic apparatus. 8,23 In one of these studies, 23 knee extension torque was higher in GKs and defenders than in midfielders and forward. These differences disappeared at a velocity of 180 s 1 when absolute values were divided by body surface area. It is possible that dividing the absolute values by body mass would eliminate all differences, especially since GKs tend to have a higher body mass than field players, as indicated earlier. However, body mass values were not reported. In the other study, 8 knee extension torque values were reported relative to body mass, and no differences between GKs and field players were observed. Whether absolute or relative values are of importance depends on the task performed. In the case of GKs, it is suggested that relative values are more important, as the work (eg, jumping to catch the ball) is done against the body mass. On-Field Performance Information on patterns of movements performed by GKs in actual games, as well as information on technical aspects of these movements, can help coaches optimize the planning of their training. A time-motion analysis (data related to the patterns of movements demonstrated by the GK) and a notational analysis (data related to the technical aspects of the performed acts) can provide detailed information on the walking or running velocities and the distances covered by GKs during games, or on the technical level of the acts they performed (eg, jumping, kicking). While studies using time-motion analyses and notational analyses are in abundance for soccer field players (e.g., 24 29 ), only two studies assessing the actions of GKs were found. 30,31 In one study, Di Salvo et al 31 monitored the activity profile of 62 GKs playing in the English Premier League during 109 games. They found that GKs covered 5,611 ± 613 m during a game. Most of this distance was covered during walking (4,025 ± 440 m) and jogging (1,223 ± 256 m), with only a small portion covered during running (221 ± 90 m), high-speed running (56 ± 34 m), and sprinting (11 ± 12 m). Most sprints covered a distance of less than 10 m. It appears that for the GK, high-intensity movements are separated by longer walking and jogging periods that allow for recovery. However, this study did not consider high intensity actions such as clearing, controlling, and saving. An attempt was made to record the actions of GKs when they performed defensive maneuvers during the World Cup Tournament that was held in Japan and Korea in 2002. 30 In this study, Di Baranda et al conducted an analysis of 34 GKs in 54 games. The GKs mostly intervened in the penalty area (44.4%), followed by the goal area (17.7%) and outside the penalty area (6.6%). They performed 23.4 defensive technical actions per game. Among the most common actions were saving (ie, catching or blocking a kick), foot control (controlling the ball with the feet and trying to pass to a teammate), and clearing out (deflecting a ball that cannot be caught by fisting or kicking it). The physical actions of the GKs were comprised

Characteristics of Soccer Goalkeepers 521 of 6.2 ± 2.7 dives per game, 3.8 ± 2.3 jumps per game, and 18.7 ± 6 displacements (forward, sideways, and backward) per game. In addition, displacements were the most common physical actions demonstrated by the GKs before performing a technical action. It should be noted that in both of these studies, no physiological data (eg, heart rate) were collected. It is difficult to collect data on GKs during actual games; coaches are not enthusiastic about placing devices/instruments on the GK s body that may distract their attention from performing their main tasks. However, such data could help increase our understanding of the physiological load imposed on GKs throughout the game, as well as clarify instructional issues related to strength and conditioning programs developed specifically for GKs. Methodological Limitations and Research Concerns Four methodological concerns and research limitations associated with the reviewed studies are discussed, as follows. (a) The lack of a longitudinal approach. In most of the studies described in our review, the physical and physiological tests were performed on the GKs only once. No replicated measurements across different phases of the training program (ie, the preparation, competition, and transition phases) were collected. In order to systematically examine the physical characteristics and physiological attributes of GKs, a longitudinal approach should be adopted by researchers. In this approach, a number of variables are measured in the same group of athletes, not only during different phases of a given season, but also across a number of seasons. This database will help the researcher/coach to better plan short- and long-term training programs for the GKs. For example, a longitudinal study on one group of beginning GKs playing in a soccer academy should be conducted over a number years, with data collected on physical characteristics such as height, body mass, and %BF, and physiological attributes, such as agility and speed, and power and strength. (b) The lack of on-field performance data. Most of the studies discussed in our review were conducted with information collected on the physical characteristics and physiological attributes of GKs under sterile settings (eg, laboratory conditions). On-field performance data were collected in only two studies. However, no data on the physiological responses of the GKs during actual games were made available in these studies. A systematic analysis of the main actions demonstrated by GKs during the game, as well as of their physiological responses during actual games, may provide soccer coaches and strength and conditioning coaches with more relevant information on aspects related to their training programs. (c) The lack of experimental studies. Most of the reviewed studies were of a descriptive nature. In order to improve training programs for elite GKs, training methods used by coaches should be based on scientific evidence. Comparisons among different training programs can be made effectively only in experimental settings, where the intervention is compared with other interventions and to a control condition. For example, a study that examines the effectiveness of

522 Ziv and Lidor different training methods for improving agility and speed in adult GKs can be conducted to help the coach select the most appropriate method/s for his or her players. Another study can be performed to examine the influence of one training regime on strength in GKs of different age groups, in order to help coaches select the most appropriate regime for a given age group. (d) The lack of reporting on effect sizes and observed power. In most of the reviewed studies, statistical values such as effect sizes and observed power were not made available. It is true that effect sizes can be estimated based on the reported means and standard deviations. However, researchers are expected to calculate effect sizes associated with their collected data, and not leave this to the reader. In order to conceptually understand the practical significance of the findings emerging from the studies on physiological attributes of GKs, information on the effect sizes and observed power is essential. Practical Applications Based on the studies reviewed in the current article, three practical implications are suggested for coaches who work with GKs. (a) Data on the physical characteristics, and particularly the physiological attributes, of GKs should be collected at several points during the season. More specifically, data should be gathered in the preparation phase, the competition phase, and the transition phase of the season. This information will enable the coach to effectively assess the physical and physiological state of the GKs during different phases of the season. (b) A careful approach should be adopted when selecting testing protocols and devices for the assessment of GKs physiological attributes. The selected protocol(s) should be used at each point of testing, so that an appropriate comparison between the achievements of the GK at each point can be made. In addition, if a comparison with the achievements of other GKs is made, it would be advantageous to use data that were collected using the same apparatus and protocols. For example, if it is the aim of coaches to assess the VJ ability of their GKs as demonstrated in actual practice sessions and games, it is proposed that they select a CMJ where the hands are allowed to move freely. The same protocol should be used each time the test is given to the GKs (eg, in the preparation phase and the competition phase of the season). In addition, the same version of the test should be used by coaches when comparing jumping achievements of their GKs and those of other GKs. (c) Relevant knowledge on physical characteristics and physiological attributes of GKs should be used in the processes of talent identification and team selection. When coaches attempt to identify young soccer players who have the potential to become GKs, or to select a GK for their teams, they should take into account what is already known about physical and physiological characteristics of the GKs. It is true that the eyes of the coach can help him or her make a good decision in the processes of talent identification and team selection, however relevant information from the studies reviewed in our article can improve their decision making.

Characteristics of Soccer Goalkeepers 523 (d) Specificity is of great importance when planning training programs for GKs. Although we did not focus on training programs for GKs in our review, it is recommended that coaches plan specific training programs for GKs. In some of the reviewed studies the physiological attributes of GKs (eg, power and strength) were found to be different than those of field players. Therefore, training programs that reflect the specific professional needs of GKs should be planned by coaches. For example, high-intensity interval training can replace slow long distance endurance training in order to improve maximal aerobic capacity. In addition, shuttle runs should be avoided and more specific agility drills emphasizing quick changes of directions around the goal line should be preferred. Acknowledgments The authors would like to thank Dinah Olswang for her editorial assistance during the preparation of this manuscript. No sources of funding were used to assist in the preparation of this review. The authors have no conflicts of interest that are directly relevant to the content of this review. References 1. Welsh A. The soccer goalkeeping handbook The authoritative guide for players and coaches. Chicago, IL: Master Press; 1999. 2. Wong PL, Chamari K, Dellal A, Wisloff U. Relationship between anthropometric and physiological characteristics in youth soccer players. J Strength Cond Res. 2009;23:1204 1210. 3. Tahara Y, Moji K, Tsunawake N, et al. Physique, body composition and maximum oxygen consumption of selected soccer players of Kunimi High School, Nagasaki, Japan. J Physiol Anthropol. 2006;25:291 297. 4. Gil SM, Gil J, Ruiz F, Irazusta A, Irazusta J. Physiological and anthropometric characteristics of young soccer players according to their playing position: relevance for the selection process. J Strength Cond Res. 2007;21:438 445. 5. Matkovic BR, Misigoj-Durakovic M, Matkovic B, et al. Morphological differences of elite Croatian soccer players according to the team position. Coll Antropol. 2003;27(Suppl 1):167 174. 6. Whaley MH, Brubaker PH, Otto RM, eds. ACSM s Guidelines for Exercise Testing and Prescription. 6th ed. Philadelphia, PA: American College of Sports Medicine; 2006. 7. Heyward VH, Wagner DR. Applied body compostion assessment. Champaign, IL: Human Kinetics; 2004. 8. Davis JA, Brewer J, Atkin D. Pre-season physiological characteristics of English first and second division soccer players. J Sports Sci. 1992;10:541 547. 9. Sporis G, Jukic I, Ostojic SM, Milanovic D. Fitness profiling in soccer: physical and physiologic characteristics of elite players. J Strength Cond Res. 2009;23:1947 1953. 10. Arnason A, Sigurdsson SB, Gudmundsson A, Holme I, Engebretsen L, Bahr R. Physical fitness, injuries, and team performance in soccer. Med Sci Sports Exerc. 2004;36:278 285. 11. Bloomfield J, Polman R, Butterly R, O Donoghue P. Analysis of age, stature, body mass, BMI and quality of elite soccer players from 4 European Leagues. J Sports Med Phys Fitness. 2005;45:58 67. 12. Guner R, Kunduracioglu B, Ulkar B, Ergen E. Running velocities and heart rates at fixed blood lactate concentrations in elite soccer players. Adv Ther. 2005;22:613 620.

524 Ziv and Lidor 13. Sutton L, Scott M, Wallace J, Reilly T. Body composition of English Premier League soccer players: influence of playing position, international status, and ethnicity. J Sports Sci. 2009;27:1019 1026. 14. Hencken C, White C. Anthropometric assessment of Premiership soccer players in relation to playing position. Eur J Sport Sci. 2006;6:205 211. 15. Carling C, Orhant E. Variation in body composition in professional soccer players: interseasonal and intraseasonal changes and the effects of exposure time and player position. J Strength Cond Res. 2010;24:1332 1339. 16. Aziz AR, Mukherjee S, Chia MY, Teh KC. Validity of the running repeated sprint ability test among playing positions and level of competitiveness in trained soccer players. Int J Sports Med. 2008;29:833 838. 17. Ramadan J, Byrd R. Physical characteristics of elite soccer players. J Sports Med Phys Fitness. 1987;27:424 428. 18. Basset FA, Boulay MR. Specificity of treadmill and cycle ergometer tests in triathletes, runners and cyclists. Eur J Appl Physiol. 2000;81:214 221. 19. Basset FA, Boulay MR. Treadmill and cycle ergometer tests are interchangeable to monitor triathletes annual training. J Strength Cond Res. 2003;2:110 116. 20. Taskin H. Evaluating sprinting ability, density of acceleration, and speed dribbling ability of professional soccer players with respect to their positions. J Strength Cond Res. 2008;22:1481 1486. 21. de Ruiter CJ, Vermeulen G, Toussaint HM, de Haan A. Isometric knee-extensor torque development and jump height in volleyball players. Med Sci Sports Exerc. 2007;39:1336 1346. 22. Maffiuletti NA, Dugnani S, Folz M, Di Pierno E, Mauro F. Effect of combined electrostimulation and plyometric training on vertical jump height. Med Sci Sports Exerc. 2002;34:1638 1644. 23. Oberg B, Ekstrand J, Moller M, Gillquist J. Muscle strength and flexibility in different positions of soccer players. Int J Sports Med. 1984;5:213 216. 24. Anastasiadis S, Anogeianaki A, Anogianakis G, Koutsonikolas D, Koutsonikola P. Real time estimation of physical activity and physiological performance reserves of players during a game of soccer. Stud Health Technol Inform. 2004;98:13 15. 25. Bangsbo J, Mohr M, Krustrup P. Physical and metabolic demands of training and match-play in the elite football player. J Sports Sci. 2006;24:665 674. 26. Burgess DJ, Naughton G, Norton KI. Profile of movement demands of national football players in Australia. J Sci Med Sport. 2006;9:334 341. 27. Di Salvo V, Baron R, Tschan H, Calderon Montero FJ, Bachl N, Pigozzi F. Performance characteristics according to playing position in elite soccer. Int J Sports Med. 2007;28:222 227. 28. Krustrup P, Mohr M, Ellingsgaard H, Bangsbo J. Physical demands during an elite female soccer game: importance of training status. Med Sci Sports Exerc. 2005;37:1242 1248. 29. Stroyer J, Hansen L, Klausen K. Physiological profile and activity pattern of young soccer players during match play. Med Sci Sports Exerc. 2004;36:168 174. 30. De Baranda PS, Ortega E, Palao JM. Analysis of goalkeepers defence in the World Cup in Korea and Japan in 2002. Eur J Sport Sci. 2008;8:127 134. 31. Di Salvo V, Benito PJ, Calderon FJ, Di Salvo M, Pigozzi F. Activity profile of elite goalkeepers during football match-play. J Sports Med Phys Fitness. 2008;48:443 446. 32. Acar M, Keles C, Balyan M, Ozkol Z, Karamizrak O. Comparison of jumping and anaerobic performance in Turkish and Cypriot professional soccer goalkeepers. (abstract) J Sports Sci. 2004;22:521. 33. Duraskovic R, Joksimovic A, Joksimovic S. Weight-height parameters of the 2002 world football championship participants. Phys Educ Sport. 2004;2:13 24.