Soccer is unequivocally one of the most popular sports in the world (López-Segovia et al., 2011). A survey conducted in 2006 showed that there were a total of 265 million male and female soccer players and 5 million referees and officials or 4% of the world's population actively involved in the sport (FIFA, 2006). According to the International Federation of Association Football (FIFA), 90% of all registered players were male, with younger soccer players comprising the greatest proportion (54.7%) of all registered male players (FIFA, 2006).

Soccer players require a moderate to high levels of aerobic and anaerobic power, good agility, and a variety of technical and tactical skills to bolster the likelihood of their success in the sport (Carling et al., 2009; Reilly et al., 2000). In order to respond to the physical demands of play, training components such as anaerobic power, speed, and agility should be incorporated into training (Amiri-Khorasani et al., 2011; Steffen et al., 2008). Prior to participation, warm-ups are generally required to “ready” oneself for training or matches (Sander et al., 2013). Warm-up programs essentially constitute mild or moderate exercise types that are geared toward enhancing performance (Sander et al., 2013; Sotiropoulos et al., 2010; Subasi et al., 2008) by increasing blood flow and the elasticity and neuron activity of the muscles (Skof and Strojnik, 2007).

Several studies have considered the effects of warm-up programs on soccer performance; such effects include but are not limited to, improvements in the time to complete a 20m sprint (Kilding et al., 2008; Sander et al., 2013; Steffen et al., 2008), jump height (Kilding et al., 2008; Steffen et al., 2008) and time to complete the Illinois agility test (Amiri-Khorasani et al., 2010; Kilding et al., 2008). Two widely used soccer specific warm-up programs are the 11+ and HarmoKnee. Kiani et al., 2010 designed the HarmoKnee warm-up program for soccer players. It essentially consists of a warm-up and muscle activation, along with balance, strength and core stability exercises. The FIFA Medical and Research Centre (F-MARC) developed the 11+ warm-up program for soccer players. The 11+ program, an advanced version of the 11 program, includes running, strength, plyometric and balance components (Soligard et al., 2008). Only a small number of studies have focused on the 11+ and HarmoKnee. The main focus of these studies, was to test their effect on strength (Brito et al., 2010; Daneshjoo et al., 2012a) balance (Daneshjoo et al., 2012b), and prevalence of injuries (Soligard et al., 2008). However, there are earlier studies on the physical performance of soccer players using the 11 program (Kilding et al., 2008; Steffen et al., 2008). Kilding and coworkers (2008) observed an improvement in leg power (3-step jump and countermovement jump) and 20m sprint, although they reported no changes in body mass, agility (Illinois test) and core stability of young soccer players. In contrast, Steffen et al., 2008 showed no significant differences between the 11 and control groups in terms of vertical jump, 40m sprint running, speed tests (10m with and without a ball) and soccer skill tests among adolescent female soccer players. Discovery of these discrepancies has led FIFA to develop the 11+ program that incorporates an increase in intensity and exercise components to potentially improve the physical performance of soccer players. Hence, further research into the effects of this advanced version of the 11+ program on the physical performances of soccer players is merited.

Both the 11+ and HarmoKnee are multifaceted soccer specific warm-up programs which aim to improve the general fitness of soccer players as well as considering intrinsic risk factors such as strength and balance. Generally, the exercises included in multifaceted warm-up programs have been shown to improve athletic performance among male soccer players (Fradkin et al., 2010). Specifically, improvements have been reported in counter movement jump height (Sotiropoulos et al., 2010), isokinetic strength (Brito et al., 2010; Daneshjoo et al., 2012a) and sprint ability (Fletcher and Jones, 2004). Core stability exercises are another important component included in multifaceted warm-up programs. These core exercises have been shown to improve technical skills and total awareness of the game (Holm et al., 2004; Leetun et al., 2004; Paterno et al., 2004). Thus, it would be expected that a multifaceted warm-up program should improve physical performance. Notably, a systematic review of warm-up, running speed, jumping and skill performances studies (1966-2008) reported that 79% (of 32 articles) showed improvements in jump height, time- to- complete sprint tests and technical skill performances while only 3% suggested no changes (Fradkin et al., 2010).

To date, no research has investigated the effects of the 11+ and HarmoKnee warm-up programs on the physical performance of young male professional soccer players. On the other hand, it is unknown whether the 11+ and HarmoKnee warm-up programs would enhance the physical capacity of professional soccer players. Therefore, we aimed to test the effects of the warm-up programs on the performance of young professional male soccer players.

Thirty-six healthy, young (n = 36) male professional soccer players (aged between 17 and 20) who regularly took part in soccer training sessions and whose medical histories were devoid of major lower limb injury or illness participated. Three top Iranian professional teams were selected. All participants were outfield players and members of one of the three professional teams. During the 2011 competitive season, all of these players participated in the Iranian premier official league championships (First Division). All subjects participated in common training consisting of technical and tactical drills such as passing, shooting, dribbling and heading drills. In addition, they played in small-sided games using 5×5 m and 10×10 m square grids. The players were discouraged from engaging in other physical activities. The participants were divided equally into 3 groups, namely the 11+, HarmoKnee and control group (n = 12 per group) (Table 1). Each group had an almost equal distribution of playing positions, including defenders, mid-fielders and strikers. Goalkeepers were excluded from this study due to the different physical and physiological demands placed upon them, and their training is very different to those of other players (Arnason et al., 2004; Gil et al., 2007; Ziv and Lidor, 2011). The study was approved by the Institute of Research Management and Monitoring, University of Malaya and the Sports Centre Research Committee.

Towards the middle of the 2011 season, the coaches and team managers from the three professional teams were invited to a four- hour instruction course aimed at introducing the intervention programs. Three under-21 (U21) teams participated in this study, each one comprised approximately 30 professional players (matched at pre-test), from which 12 players were randomly picked to participate in the study. A one-way analysis of variance (ANOVA) did not show any significant differences between the groups during the pre-test for all variables measured (p ? 0.05). Each subject performed only one of the selected warm-up programs.

Prior to the commencement of the intervention programs, all players attended a workshop to learn the correct methods to perform the exercises. This workshop was conducted for each team separately. None of the teams knew about the exercises the other teams were performing. They received video instructions and illustrations on the intervention programs. The subjects were briefed on the procedures they would undergo before being asked to read and sign a consent form. All of the procedures were carried out in accordance with the Declaration of Helsinki with respect to human research. All the training sessions were supervised by one of the researchers to ensure they conformed to the standards of compliance and to make certain that correct technique was used. Verbal encouragements were given throughout the training period to help the subjects to concentrate on the quality of their movements. The exercise programs commenced on 15^th April 2011 and were completed on 15^th June 2011 (24 sessions).

The 11+ consisted of three parts, the first of which involved running exercises (part 1). The second part covered six exercises, all of which comprised three levels of difficulty and were aimed at improving strength, balance, muscle control and core stability (part 2). The third and final part consisted of advanced running exercises (part 3). The varying levels of difficulty enhanced the program's efficiency and enabled coaches and players to individually adapt to the program. The 11+, which replaced the usual warm-up training, took approximately 20-25 minutes to complete. All of the exercises (27 exercises) focused on core stability, neuromuscular control, eccentric hamstring strength and agility (Table 2). This group performed the 11+ exercises three times per week.

HarmoKneewarm-up program included five parts: warm up, muscle activation, balance, strength and core stability, all of which can be combined and performed in a regular soccer training session (Table 3). Total program duration was 20 to 25 minutes (Kiani et al., 2010). Similar to the 11+, HarmoKnee was also performed three times per week.

The control group was asked to continue with their regular training and warm-up sessions throughout the study period. In addition, prior to the commencement of the study, the control group was assured that they would receive the intervention program in the subsequent season.

Five standard soccer specific tests (in an indoor hall with synthetic flooring) were performed at the same time of the day under standardized environmental conditions (19-22°C). All tests were carried out between 8 am and 11 am. The time spent on each test was recorded (with a resolution of 0.01 s) using a handheld stopwatch (Sportline 240, Sportline,Inc, Hazleton, Pa) which was operated by one of the researchers. Performance tests were conducted by a different member of the research team. The tester was blinded to the type of intervention the players participated in. A standardized 15-min warm-up, including running, low-intensity cycling, calisthenics and dynamic stretching was performed before the tests.

The pre-testing was conducted one week prior to the first day of training, while the post-test was recorded three days after the final training session. All tests were conducted in the same order for each player during the pre- and post-tests.

After warming up with ball-based exercises, the players performed a 20m shuttle run both with and without a ball to assess their coordinated dribbling under time pressure and speed. The tests were based on a straight dribbling test developed by Reilly and Holmes, 1983, (intraclass correlation coefficients (ICC) = 0.95) where five cones were placed in a straight line 2.8, 4.8, 6.0, 8.0 and 10.0m from the start line (perpendicular to the line). The players were instructed to dribble around alternate obstacles until the fifth cone was circled, and then return through the course in a similar fashion as fast as they could. The starting position was from an upright position. The test was completed when the player in control of the ball, crossed the finish line (Figure 1). Two successful complete trials, one with and one without the ball, were recorded, with the best results for each test being chosen for analysis (Steffen et al., 2008).

The 20m single sprint is a standard test for assessing soccer players' running speed (Rahnama et al., 2009). The participants performed two 20m trials, with a 3-min recovery period in between. The players started from a standing position, and the timing system was triggered as soon as they left the starting mat. The best attempt was used for the analysis.

The vertical jump test has been shown to be a reliable test in measuring the jump performance of soccer players (ICC = 0.97), (Christou et al., 2006). Each subject performed two practice jumps prior to testing to ensure proper jumping techniques. They were instructed to jump vertically (initiated from a knee flexion of 90°), and execute a maximum vertical jump while swinging their arms actively. Jump height was determined using a measuring tape fastened to a dark paper on which each subject's pre-test and post-test jump chalk prints would be clearly recorded. Each subject performed 2 practice jumps and the best score was used for analysis (Vanderford et al., 2004; Christou et al., 2006).

This test is a standard test with high reliability (ICC=0.97) in terms of assessing soccer players' skill and accuracy in kicking a ball (Reilly and Holmes, 1983). Players were required to kick a ball from a wall and then trap or kick the ball on the rebound as many times as possible within a 30-sec period (Figure 2). The subjects were allowed to kick the ball from the air or ground while avoiding the use of their arms or hands. Each subject performed 3 sets of this test, with the best attempt used for the analysis (Vanderford et al., 2004; Reilly et al., 2007).

The Illinois agility test is commonly used in measuring agility in soccer (Kilding et al., 2008; Katis and Kellis, 2009; Amiri-Khorasani et al., 2010). The reliability of this test has been reported to be high (ICC = 0.85) (Katis and Kellis, 2009). The length of the zone is 10m, while the width (distance between the start and finish points) is 5m. Four cones were placed in the center of the testing area at a distance of 3.3m from one another. Four cones were used to mark the start, finish and two turning points. The subjects started the test lying face down, with their hands at shoulder level. The trial started on the “go” command, and the subjects began to run as fast as possible. The trial was completed when the players crossed the finish line without having knocked any cones over. Three trials were performed by every subject with the best score (time) used for analysis (Kilding et al., 2008; Amiri-Khorasani et al., 2010).

A two-way (3 × 2) ANOVA design with repeated measures was applied to examine the possible interaction between groups (the 11+ vs HarmoKnee vs control groups) and the time (pre- and post-tests) intervention phases. As for statistical significance, the post- hoc Bonferroni test was conducted to determine pair wise differences. The data met all the assumptions for linear statistics and the Levene's test was used to assess the homogeneity of variance between the groups. In order to determine the relationship between the 20m single sprint and agility, the Pearson correlation was used. The reliability of each test was assessed by the intraclass correlation coefficient (ICC). The effect sizes of each variable were tested using Cohen's d [d = (M1 - M2 / √ ((SD1²×SD2²) / 2)] within each group (0.1, 0.3, and 0.5 as small, medium, and large effects, respectively) and partial eta (η) squared between groups (0.01 = small effect, 0.06 = medium effect, and 0.14 = large effect), (Cohen, 1988; Pallant, 2007). Statistical significance was set at p < 0.05.

The means of pre- and post-tests of the groups are presented in Table 4. High reliability was observed in the 10m speed test with a ball (ICC = 0.98). In the 10m speed test with the ball, the results did not show any significant differences between the pre-and post-tests (F_2,33 = 0.642, p = 0.429). The results did, however, indicate significant differences between groups (F_2,33 = 4.9, p = 0.013). The partial eta squared statistic indicated a large effect size (0.23) between the groups. The Bonferroni post-hoc test showed significant differences between the control and the 11+ (p = 0.048), as well as HarmoKnee (p = 0.021).

The reliability of the 10m speed test without a ball was high (ICC = 0.96). There was no significant main effect difference between the pre- and post-tests (F_2,33 = 0.700, p = 0.409). However, significant differences between the groups (F_2,33 = 5.66, p = 0.008) were observed. The partial eta squared statistic indicated a large effect size (0.26) between the groups. The results from the Bonferroni post hoc test indicated a significant difference between the control and the 11+ (p = 0.031), as well as HarmoKnee (p = 0. 012).

The 20m single sprint test demonstrated high reliability (ICC = 0.91). No significant main effect difference was found between the pre- and post-tests (F_2,33 = 4.06, p = 0.052), but the results showed significant differences between the groups (F_2,33 = 7.89, p = 0.008). The partial eta squared statistic indicated a large effect size (0.25) between the groups. The Bonferroni post-hoc test indicated significant differences between the control and the 11+ (p = 0.023), as well as HarmoKnee (p = 0.019).

High reliability was found in the vertical jump test (ICC = 0.98). There was a significant main effect between the pre- and post-tests in the vertical jump (F_2,33 = 15.2, p = 0.000). The partial eta squared statistic indicated a large effect size (0.31) between the groups. The results showed a significant increase only in the 11+ group from pre- to post-tests by 3.7% in jump height. The Cohen's d indicated a small effect (0.25) between times in the 11+. No significant differences were found between the groups (p ? 0.05).

The Wall-Volley test showed a high degree of reliability (ICC=0.97). There was a significant main effect between the pre- and post-tests (F_2,33 = 21.173, p = 0.000). The partial eta squared statistic indicated a large effect size (0.39) between times. Both the 11+ (p = 0.000) and HarmoKnee (p = 0.002) showed significant increases in scores within the groups, with 5.4% and 5.2%, respectively. The Cohen's d indicated a large effect in the 11+ (0.59) and HarmoKnee groups (0.53). However, the results did not show any significant differences between the groups (p ? 0.05).

The reliability of Illinois agility test was found to be high (ICC = 0.96). A significant main effect was observed between the pre- and post-tests (F_2,33 = 7.12, p = 0.012). The partial eta squared statistic indicated a large effect size (0.18) between times. The results indicated a significant decrease in times within the 11+ (p = 0.001) by 1.7%. The Cohen's d indicated a large effect in the 11+ (0.64). Significant differences were found between the groups (F_2,33 = 11.94, p = 0.000). The partial eta squared statistic indicated a large effect size (0.42) between the groups. The post-hoc Bonferroni test indicated significant differences between the 11+ (p = 0.000) and HarmoKnee (p = 0.007) compared to the control groups. Moreover, the finding showed a moderate correlation in the 11+ group between the results of Illinois test and the 20m single sprint (r = 0.539).

The aim of this study was to investigate the effect of the 11+ and HarmoKnee warm-up programs on the physical performance of young professional male soccer players. Significant differences between the control group and the 11+ as well as HarmoKnee in the speed tests (10m with and without a ball and 20m sprint) were found. The 11+ and HarmoKnee both showed trends for improvement; however, the changes were not significantly different (p ? 0.05) between pre-to-post time points in the 10m speed tests (with and without a ball) and 20m single sprint. From the present results it is difficult to conclude that these warm-up programs can improve speed performance. Of interest, however, is that Kilding et al., 2008, using the 11 program, reported changes among non-professional male soccer players (mean age 10.4 ± 1.4 y) in a 20m sprint following 30 sessions (6 weeks). This might have been due to the number of sessions (30) employed, mean age of the sample used and their level of skills. It can be argued that since professional soccer players already possess highly developed physical characteristics (Cometti et al., 2001), thus little improvement will be demonstrated. Moreover, speed performance is largely dependent on genetic influences such as muscle fiber type, with only relatively small improvements observed with training programs (Ross et al., 2001; Tønnessen et al., 2011). The present findings suggest that both of the warm-up programs are in need of some modifications to enhance players' running speed. Sander et al., 2013 showed that five minutes of nonspecific running, coordination exercises, stretching and acceleration runs (within a warm-up) improved 10m and 30m sprint performance in young male soccer players. Repeated shuttle sprint training of 15-20m over a 7-week period improved the 30m speed test in adolescent professional male soccer players (Buchheit et al., 2010). Moreover, studies showed that plyometric training of forward and lateral hopping, shuffle, ladder drills, box and depth jumps (Michailidis, et al. 2013), 40-cm hurdle jumps and 40-cm drop jumps (Chelly et al., 2010) enhanced speed performance in soccer players. Hence, we suggest adding additional training elements such as plyometric training, ladder drills and repeated shuttle sprints in both programs to fully realize the enhancement of speed performance. Further studies examining the effects of these two modified warm-up programs on the sprint performance of professional soccer players are warranted.

The comparison in the vertical jump within the groups showed a significant increase in the post-test only in the 11+ group (3.7 %); however, the effect size was small. Kilding et al., 2008 examined the suitability of the 11 warm-up program for young soccer players and reported that leg power (3-step jump and countermovement jump) increased significantly (3.4%). Conflicting results were reported after a ten-week training program using the 11 warm-up program on female adolescent soccer players, where no difference in jump height was observed (Steffen et al., 2008). These differences could be attributed to such factors as the gender and age of the subjects. Although the findings of Kilding and his colleagues (2008) are quite similar to those of the present study, the skill levels of the subjects were different. They employed non-professional male soccer players, whereas professional players were employed in this study. In the present study, we used the 11+, which is an advanced version of the 11 (Soligard et al., 2008). Based on the magnitude of change in jump performance and comparing the level of the players from the two studies, the 11+ program would seem superior to the 11.

Our results indicated an improvement in Wall-Volley skill test within the groups. Both the 11+ and HarmoKnee showed significant increases in scores (5.4% and 5.2% respectively). The effect size analysis also indicated large effects in both groups. Therefore, it can be concluded that there is a difference between the two warm-up programs compared to the control group. Previous studies have reported positive effects of different training programs on soccer skill tests (Christou et al., 2006; Billot et al., 2010). Billot et al., 2010 found that the kicking skill of 20 male amateur soccer players improved after a 3-week electro stimulation training program on the quadriceps muscle. The results of this study suggested that the warm-up training programs positively influenced the players' performance in the Wall-Volley skill test. Many factors influence performance on sport-specific skill tests, including physiological factors (sprint, strength, power), neural control of movement and perceptual-cognitive skills such as anticipation and visual search strategies (McMillan et al., 2005; Russell et al., 2011). It is likely that the multifaceted warm-up programs have the potential to increase the stimulus of these factors.

The results indicated a decrease in time taken for Illinois test in the 11+ by 1.7%, whereas no difference was observed in HarmoKnee group. The Cohen's d indicated a large effect size in the 11+ program. These results seem to suggest that the 11+ may be more beneficial in improving agility in young male soccer players than the HarmoKnee program. The above findings are in agreement with those of an earlier study that investigated the effect of 10-week intervention training in young male soccer players (Simek Salaj et al., 2007). The experimental group reported significant changes in the Y20 agility test. Agility is a complex ability which depends on coordination, the mobility of the joint system, dynamic balance, and strength, as well as stabilizing and suppressing strength and speed (Simek Salaj et al., 2007). One potential explanation for the present results could be the fact that our intervention programs included the essential components to improve agility. The Illinois test reflects the performance of combined speed and agility (Katis and Kellis, 2009; Vescovi and VanHeest, 2010). Motor skills common to soccer such as sprinting and jumping actions, have similarities in muscular, biomechanical and kinematic characteristics (Vescovi and McGuigan, 2008; Zajac, 2002) and are closely associated (Vescovi and McGuigan, 2008). During the acceleration period, instantaneous leg power is required to propel the whole body mass to maximum speed in the shortest possible time (Farlinger et al., 2007). Pauole et al., 2000 reported a moderate correlation (r = 0.53) between the agility and sprint time (36.58 m; 40 yd) in 105 college-aged male athletes. In this study, a moderate correlation (r = 0.54) was also reported in the 11+ group between Illinois and 20m single sprints.

There are a few limitations to the present study that should be addressed. This study was performed at mid-season (competition season), which means that some players might have played more games than others and potentially had a higher baseline fitness level, which could have affected any potential changes in performance. In order to account for this limitation, we chose soccer players that had played in more competitions, as evidenced by their playing history and recommendations by the coaches. Another limitation was the time measuring device used in the present study. To account for this, time was measured by an experienced researcher, which generated the high reliability of all tests in the present study.

It can be concluded that neither the 11+ nor HarmoKnee improved players' performance in the speed test (10m with and without a ball) and 20m single sprint between pre-to- post time points. The 11+ did improve vertical jump, Wall-Volley and Illinois agility test performance, while the HarmoKnee group showed a significant increase in performance on Wall-Volley test.