Validation of the Loughborough Soccer Passing Test in Young Soccer Players

Abstract

Le Moal, E, Rué, O, Ajmol, A, Abderrahman, AB, Hammami, MA, Ounis, OB, Kebsi, W, and Zouhal, H. Validation of the Lloughborough soccer passing test in young soccer players. J Strength Cond Res 28(5): 1418–1426, 2014—The aim of this study was to examine the validity and reliability of the Loughborough Soccer Passing Test (LSPT) in adolescent soccer players. Eighty-seven players, aged 14–17 years, were recruited according to their playing level: elite (n = 44), sub-elite (n = 22), and non-elite (n = 21). Two attempts of the LSPT were performed at baseline. Players then completed 10 attempts over 3 weeks to familiarize themselves with the test. Subsequently, 2 main trials, separated by 1 week, were performed; the mean of the 2 attempts was recorded as the performance score. After familiarization, the performance scores showed significant differences (p < 0.01) between elite (40.3 ± 8.3 seconds), sub-elite (58.1 ± 10.2 seconds), and non-elite players (66.6 ± 11.7 seconds). There was low-to-moderate reliability between trials with sub-elite (r = 0.35, p < 0.05) and non-elite players (r = 0.47, p < 0.05), but very good for elite players (r = 0.96, p < 0.05). Scores at baseline were better (p < 0.05) for elite players (51.0 ± 9.3 seconds) compared with sub-elite (60.8 ± 8.2 seconds) and non-elite players (69.0 ± 11.1 seconds). The LSPT seems to be a valid and reliable protocol to assess differences in soccer skill performance in adolescent players and can distinguish players according to their playing level. The LSPT was able to distinguish different abilities without players undergoing any familiarization with the test, thus enabling it to be used for talent identification purposes.

Introduction

The process of talent identification is an important aspect in sports research, especially considering the high financial stakes of modern sport (^23,27). The aim of this process was to identify gifted players early and to permit them to be part of high-level practice through an organized program.

In soccer, an activity that requires physical (⁵) and psychological (¹⁶) attributes and where technical and tactical skills are invaluable (²²), a multivariate approach is needed for early detection of talented players (²⁰). A typical talent identification scheme for soccer was proposed by Reilly et al. (²⁰) comprising 15 anthropometric measurements, 8 physiological measurements, 3 psychological tests, and 2 skill tests (shooting and dribbling slalom) developed by Reilly and Holmes (¹⁹). Several validated tests are currently being used to evaluate the physical prowess of players. For example, the Yo-Yo intermittent recovery test is a valid measure of fitness performance in soccer (¹⁴), and the field test of Kemi et al. (¹²) measures and so represents a more soccer-specific alternative for the traditional treadmill tests (^4,15). The assessment of skill is rarely included when the fitness of players is monitored (¹). Indeed, there is a dearth of studies on skill performance within the research literature, which seems remarkable when it is readily acknowledged that the successful execution of skill is the most important aspect of soccer play.

The classical definition of skill is “the learned ability to bring about predetermined results with maximum certainty often with the minimum outlay of time or energy or both” (¹³). Soccer requires the execution of many aspects of skill in a dynamic context. Therefore, although there are some “closed skills” (e.g., taking a free kick), soccer is predominantly an “open skill” game (¹³). In other words, a player might have good patterns of movement (technique), but if he does not perform the right action at the right time (skill), then he becomes an almost “useless player” (¹³). Therefore, isolating 1 aspect of the game, for example, passing or shooting from a static situation (and typically in a rested stated), may make it an execution of “technique” rather than “skill” per se (¹).

It is not surprising that most tests that purport to measure skill performance in soccer have examined closed skills (^9,21). Hoare and Warr (⁹) used slalom dribbling to observe players' skill performance, but the static position of cones did not allow an interactive environment to demonstrate skill per se. Rösch et al. (²¹) asked the participants to shoot at targets and examined the accuracy of shooting. Different tasks allowed observation of long-passing ability, accuracy, and ball speed. Nevertheless, the majority of soccer skills are open skills, performed in an unpredictable situation, with outside factors dictating how and when the skill is performed (¹³). The difficulty is to provide a test with enough uncertainty to replicate game conditions and with enough precision in the protocol for the test to be reliable.

The Loughborough Soccer Passing Test (LSPT) is a reliable and valid test, which assesses the multifaceted aspects of soccer skill including passing, dribbling, control, and decision making (^1–3). These authors observed that elite male and female players performed significantly better than their non-elite counterparts, confirming the criterion validity of the test (^2,3). Because talent identification process takes place between 12 and 15 years of age (²⁷), it is important to know if the LSPT is appropriated to the skill of adolescent players. O'Regan et al. (¹⁷) used a modified version of the LSPT because the original test conditions were not suitable for their players (aged 12 years). Nevertheless, for better comparison and performance tracking between young and adult players, it may be better to keep the same LSPT conditions. Nevertheless, to the best of our knowledge, the validity and reliability of LSPT in adolescent soccer players have not yet been assessed.

To obtain reliable results, one of the most important aspects to validate a test, subjects are required to have multiple attempts to familiarize themselves with the procedures before habituation occurs (⁷). The authors who initially validated the LSPT for male and female players reported a small order effect between trials (^2,3). They recommended that future studies wishing to examine soccer skill using the LSPT should engage the highest level of soccer players at their disposal and require them to attend at least 1 (and preferably 2) familiarization sessions (²). In talent identification conditions, the familiarization period may not always be possible, especially in situations where players from diverse backgrounds come together. Therefore, the results without familiarization have to be able to discriminate players according to the playing level.

To date and to the best of our knowledge, the LSPT has been validated for use in male and female adults but not in young soccer players. This use may help to improve objectivity of the talent identification process and decision making regarding the recruitment of young soccer players. Consequently, the purpose of this study was to validate the LSPT in young soccer players. The second aim was to assess the impact of familiarization on LSPT performance to examine if this test could be used for talent identification processes when players are unable to practice the test on multiple occasions.

Methods

Experimental Approach to the Problem

Talent identification is an important process allowing young soccer players to reach high-level performance as adults. Identification and development schemes are multifactorial, and although many studies exhibit an in-depth test battery to assess anthropometry, physiology, and psychology attributes, there is a lack of research regarding skills evaluation (interactive patterns of movement) corresponding with modern soccer aspects. The LSPT is a multifaceted test and is appropriate to assess interactive and unpredictable soccer skills. However, the test has only been validated for adult players, and, therefore, the aim of this study was to validate the LSPT in young players. Eighty-seven young male soccer players, of differing abilities, were familiarized with the test and then undertook 2 main trials to assess the validity and reliability of the LSPT. The secondary aim was to examine the effect of familiarization on LSPT performance across the varying levels of players. Therefore, the test was conducted twice at baseline (before familiarization), and reliability and validity were examined.

Subjects

Eighty-seven young male players (mean ± SD; age = 15.2 ± 0.9 years; stature = 171.0 ± 7.0 cm; body mass = 58.5 ± 7.5 kg; soccer playing experience = 6.9 ± 0.3 years) volunteered for this study after obtaining parental consent (Table 1). The local Ethics Committee of the University of Rennes II (CCPPRB) approved the experimental protocol.

Table 1: Anthropometric characteristics of the players.^*

The participants were grouped according to their playing level as given below:

• Elite (National Training Squad, n = 44; age = 15.3 ± 0.7 years; stature = 175.2 ± 5.1 cm; body mass = 64.4 ± 5.7 kg). This group of elite players was selected from 6 regional centers of excellence throughout the country. Typically, 800 young male soccer players per year apply for these centers with approximately 240 (30 players per center) being selected after talent identification procedures. They practice soccer 11 months a year, for at least 5 years, and undergo 5 training sessions and 1 competitive game per week, in addition to their school physical education classes.
• Sub-elite (French Second and Third Division players, n = 22; age = 15.3 ± 1.1 years; stature = 172.0 ± 7.4 cm; body mass = 60.0 ± 8.9 kg).
• Non-elite (French Seventh Division players, n = 21; age = 14.8 ± 1.1 years; stature = 165.7 ± 8.4 cm; body mass = 51.0 ± 8.0 kg).

The participants were from a range of outfield playing positions and were involved in regular training and match play. After completion of a screening questionnaire, written informed consent and parental consent was obtained from all participants. The data collection took place during March/April (toward the end of the French season), so teams were well established.

Procedures

Layout of the Loughborough Soccer Passing Test

Figure 1 illustrates the layout of the LSPT. Four soccer-specific plastic rebound boards were placed as shown on each of the 4 lines marking the 12 × 9.5-m grid (to the inside of the boards). Before placement, 4 colored target areas (red, blue, green, and yellow; 0.6 × 0.3 m) were attached in the middle of each board. In addition, a 0.1 × 0.15-m black piece was painted in the middle of the target areas. Colored cones were used to distinguish the different zones. Participants started with the soccer ball by the central cone, and the examiner started timing the test (using data acquisition software especially designed for this study for more efficient data collection) from the moment the ball was played out of the inner rectangle. The specific color was called out just before the participant completed the current pass.

Figure 1: Schematic representation of the Loughborough Soccer Passing Test.

The same examiner was used so as to eliminate interexperimenter variability. The sequence of passes was determined by 1 of 8 trial orders that were randomly generated by the software so that each trial consisted of 8 long (green and blue) and 8 short (yellow and red) passes.

Participants were informed that passes could only be executed from within the passing area. Furthermore, the players were informed that for the best performance on the LSPT, they would have to perform the test as quickly as possible while making the fewest mistakes. The time was recorded when the last pass was completed. A further role of the examiner was to record penalty time points accrued during the trials. Thus, the examiner stood in such a position that all 4 target areas could be viewed.

Penalty time was awarded for the following errors:

• Five seconds for missing the bench completely or passing the wrong target.
• Three seconds for missing the target area.
• Three seconds for handling the ball.
• Two seconds for passing the ball from outside of the designated area.
• Two seconds for the ball touching any cone.
• One second for every second taken over the allocated 43 seconds to complete the test.

Additionally, 1 second was deducted from the total time if the ball hit the 10-cm strip in the middle of the target.

All trials took place within an indoor sports hall. Table 2 shows the schedule of experimental procedures for this study. On each occasion, players came to the sports hall as part of their normal training squad at the same time of day. The sports hall was maintained at an environmental temperature of between 18 and 20°, and the participants did not receive any fluids during the trials. A 10-minute standardized warm-up consisting of jogging, passing, and sprinting exercises preceded the trials. All players then took part in regular training (on the adjacent indoor court) and were systematically taken out of the training session to perform the LSPT in groups of 3 players at a time. Each player took turns in completing LSPT while the 2 other were required to undertake short passes on an adjacent area.

Table 2: Planning of the experimentation period (4 weeks).^*

During the prefamiliarization (Pre-Fam) session, participants were given a brief explanation of how to perform the LSPT and then allowed 4–5 passes (per player) under test conditions to briefly practice the test. Then they performed 2 recorded trials (Pre-Fam Trial 1 and Pre-Fam Trial 2) separated by approximately 10 minutes within the Pre-Fam session. After this initial session, participants undertook 5 training sessions, during a period of 3 weeks (Table 2), to become fully accustomed with the LSPT. For each training session, players performed 2 LSPT trials; the mean of both attempts were communicated to players so that they could evaluate their own performances.

After this period of habituation, participants completed 2 main trials (Main Trial 1 and Main Trial 2), each separated by 1 week. Within each trial, participants undertook 2 attempts of the LSPT and the mean of both was recorded as the performance score. Between these 2 main trials, no further opportunity for practice was given to participants. The main trials were conducted at the same time of the day to avoid possible circadian variations, but nutrition and hydration status of subjects were not monitored.

Statistical Analyses

One-way analysis of variance (ANOVA) was used to examine anthropometric characteristics between groups. To assess the validity of the LSPT, 2-way ANOVA was used to determine whether there were any differences in skill scores between the 2 trials of the Pre-Fam session, and between Main Trial 1 and Main Trial 2. Post hoc analyses for both types of ANOVA were conducted using Holm-Bonferroni adjusted paired t-tests.

Reliability was assessed using Pearson's correlation (r) and coefficient of variation (CV). The strength of correlations was based on the values derived from the study by Vincent (²⁴), that is, 0.5 (low), 0.7 (moderate), and 0.9 (strong). The “95% absolute levels of agreement” (limits of agreement; Bland and Altman (⁶)) was the final method of assessing reliability. To compare the measurements by taking into account heteroscedastic errors through dimensionless ratios, the “ratio limits of agreement” (RLOA), using log-transformed measurements, were used as well. The results are presented as mean ± SD values. Statistical significance was accepted at p ≤ 0.05.

Results

After Familiarization

Validity was assessed by comparing performances between elite, sub-elite, and non-elite players (Table 3). There was a main effect of group for LSPT performance scores with elite players (40.3 ± 8.3 seconds) demonstrating 44.2% and 65.2% better performance than sub-elite (58.1 ± 10.2 seconds, p < 0.001) and non-elite (66.6 ± 11.7 seconds, p < 0.001) players, respectively (Figure 2). Sub-elite players also showed a 14.6% better performance score compared with non-elite players (p < 0.001). The aspect of time-only distinguished elite (35.9 ± 4.7 seconds), semi-elite (42.2 ± 3.3 seconds; p < 0.001), and non-elite (46.0 ± 5.8 seconds; p < 0.001) players, and semi-elite from non-elite time players (p < 0.001). Penalty time scores were also significantly lower for elite (4.3 ± 5.8 seconds) compared with sub-elite (15.9 ± 9.0 seconds; p < 0.001) and non-elite (20.6 ± 8.6 seconds; p < 0.001) players. Sub-elite players also showed lower penalty time scores than non-elite (p < 0.001) players.

Table 3: The LSPT scores for the 3 groups.^*

Figure 2: Diagrammatic representation of Loughborough Soccer Passing Test (LSPT) performance scores before familiarization: Data are sorted and plotted for each subject of each playing level.

With regard to reliability, no differences were found between Main Trial 1 and Main Trial 2 for all groups for time (p = 0.29–0.75), penalty (p = 0.14–0.54), and global (p = 0.09–0.85) performance scores. There were moderate-to-strong correlations between Main Trial 1 and Main Trial 2 for elite players for the 3 components (r = 0.73–0.96, p < 0.01). Low-to-moderate but significant correlations were found for non-elite players for the 3 components (r = 0.47–0.50; p < 0.05). Although a moderate correlation was found for sub-elite players for the time aspect (r = 0.77; p < 0.01), there were no significant correlations for penalty and global performance scores (r = 0.22–0.35; p > 0.16). The CV was lower for the time aspect than global performance time for all groups (1.21–1.84% vs. 1.79–3.88%). The limits of agreement (LOA, Table 5) for elite players were low and consistent for the 3 components (±8.2–9.9 seconds).

Table 5: Paired t-test data, correlation (r), coefficient of variation (CV, %), absolute limits of agreement (LOA), and ratio limits of agreement (RLOA) for LSPT performance after familiarization.^*

The LOA values were low for the time aspect in sub-elite players (±5.4 seconds) but higher for penalty and global performance (±22.5–23.9 seconds). The random error within the non-elite group was lower for the time aspect (±11.5 seconds) compared with penalty and global performance components (±16.4 and ±23.8 seconds). The mean (bias) ratio multiplied or divided by the agreement ratio indicated greater reliability for the time component as opposed to global performance for sub-elite and non-elite players (×/÷ 1.16 and ×/÷ 1.33 vs. ×/÷ 1.51 and ×/÷ 1.65) but not for elite players (×/÷ 1.39 for time aspect and ×/÷ 1.28 for global performance).

Before Familiarization

Loughborough Soccer Passing Test performances before familiarization are presented in Table 4. No differences were found between Pre-Fam Trial 1 and Pre-Fam Trial 2 with elite players for time (39.7 ± 3.5 seconds vs. 38.4 ± 3.9 seconds, p = 0.245), penalty time (12.4 ± 6.6 seconds vs. 11.5 ± 8.2 seconds, p = 0.697), and global performance (52.1 ± 8.1 seconds vs. 49.9 ± 10.6 seconds, p = 0.447) components. With sub-elite players, time score was not significantly different (43.5 ± 3.2 seconds vs. 46.0 ± 5.6 seconds, p = 0.07) but penalty score (11.4 ± 7.2 seconds vs. 20.7 ± 7.3 seconds, p < 0.01) and global performance score (54.9 ± 7.6 seconds vs. 66.7 ± 8.8 seconds, p < 0.01) were significantly different (Figure 3). The time, penalty time, and global performance scores were significantly different with non-elite players (46.1 ± 4.4 seconds vs. 49.3 ± 5.6 seconds, p < 0.05; 16.9 ± 7.1 seconds vs. 25.9 ± 9.0 seconds, p < 0.01, and 63.1 ± 10.3 vs. 75.3 ± 11.8 seconds, p < 0.01, respectively). There was a main effect of group for LSPT global performance scores with elite players (51.0 ± 9.3 seconds) demonstrating 19.2 and 35.5% better performance than semi-elite (60.8 ± 8.2 seconds, p < 0.002) and non-elite (69.0 ± 11.1 seconds, p < 0.001) players, respectively; semi-elite players also performed better than non-elite players (p < 0.009).

Table 4: Paired t-test data, correlation (r), coefficient of variation (CV), absolute limits of agreement (LOA), and ratio limits of agreement (RLOA) for LSPT performance without familiarization.^*

Figure 3: Diagrammatic representation of Loughborough Soccer Passing Test (LSPT) performance scores after familiarization: Data are sorted and plotted for each subject of each playing level.

Time only was sufficient to distinguish elite (39.0 ± 3.6 seconds) from semi-elite (44.7 ± 4.4 seconds; p < 0.001) and from non-elite (47.6 ± 5.0 seconds; p < 0.001) players; semi-elite and non-elite scores were also different (p < 0.05). Penalty time scores were not significantly lower for elite (11.9 ± 7.4 seconds) compared with semi-elite (16.0±7.2 seconds; p = 0.169) but were lower for elite and semi-elite compared with non-elite (21.4 ± 8.0 seconds; p < 0.001 and p < 0.05, respectively) players.

There were moderate correlations between trials for global performance scores (r = 0.69–0.82, p < 0.01) for elite and non-elite players but a lower correlation for semielite players (r = 0.47, p < 0.05). Penalty time showed moderate correlations for all groups (r = 0.55–0.73, p < 0.01) and correlations were good for time only with elite players (r = 0.65, p < 0.01) but not for semi-elite and non-elite players (r = 0.33–0.34, p = 0.13). The CV was lower for the time aspect than global performance time for all groups (3.9–5.6% vs. 7.2–15.9%). The LOA for the time-only aspect was the smallest (±6.1–11.8 seconds; Table 4), indicating that this has the least amount of random error, with penalty and global performance time showing slightly greater error (±12.1–13.6 seconds and ±13.2–16.7 seconds, respectively). The random error within the elite group for the time only (±6.1 seconds) was lower than other groups but not for penalty and global performance time. Mean (bias) ratio multiplied or divided by the agreement ratio indicated greater reliability for the time component as opposed to global performance for elite and sub-elite players (×/÷ 1.20 and ×/÷ 1.3 vs. ×/÷ 1.46 and ×/÷ 1.34) but not for non-elite players (×/÷ 1.33 for time aspect and ×/÷ 1.25 for global performance).

Discussion

The main aim of this study was to examine the validity and reliability of the LSPT for use with adolescent soccer players. We found good validity and reliability, especially for elite players. The second aim was to examine whether the LSPT could discriminate players according to their playing level in talent identification conditions, that is, without any previous familiarization. The results showed better scores for the elite group of players.

After familiarization with the test, elite players demonstrated better performance than their semi-elite and non-elite counterparts, thus confirming the criterion validity of the LSPT. They exhibited better performance scores than sub-elite (45%) and non-elite players (64%), and sub-elite players had better performance than non-elite players (15%). As soccer-playing ability was perceived to be higher with the highest playing level, these scores highlight the construct validity of the LSPT. The LSPT is also able to distinguish elite players with their counterparts by analyzing various aspects of soccer skill performance (gross motor performance with the time-only score and accuracy using penalty time). Indeed, elite players took 17 and 28% less time to complete the test compared with sub-elite and non-elite players and accrued 272 and 380% less penalty time than sub-elite and non-elite players, respectively. Sub-elite players also demonstrated better scores than non-elite players (12% for time and 29% for penalty time). The reliability measures for elite players were superior to sub-elite and non-elite players (Table 5). Therefore, the LSPT can be recommended for use where the examination of soccer skill performance of adolescent players is required.

Because of the few studies that have used the LSPT in young players, it is difficult to compare our findings. Impellizzeri et al. (¹⁰) used the LSPT with 26 junior soccer players (mean age = 17.8 years) to examine the effects of aerobic interval training on the decline in short-passing ability caused by a short bout of high-intensity intermittent activities. Time scores ranged 44–49 seconds and global performance was around 59–60 seconds in the unfatigued condition and around 62–68 seconds in the fatigued condition. These LSPT performance scores corresponded to our sub-elite players, but no information was provided concerning the playing level of the subjects in the study by Impellizzeri et al. (¹⁰).

Compared with adult men (³), the elite young players in our study demonstrated a better global performance after habituation (40.3 ± 8.0 seconds vs. 43.6 ± 4.5 seconds). This result is surprising because these adolescent players were 15 years old and still developing their physical and motor skill learning abilities (^25,26). It is possible that the university-level subjects of Ali et al. (³) were inferior to our international age–level players or that the differences were because of disparities in scoring between investigators. Nevertheless, results of young elite players without familiarization (50.8 ± 7.4 seconds, Table 3) were inferior to these adults, implying that the adult players of Ali et al. (³) may not have received adequate practice attempts to become fully familiarized with the LSPT. Indeed, the fact that Ali et al. (³) reported a trial order effect, where players improved performance from the first trial to the second trial, helps support this assertion. They concluded that a few more attempts to habituate participants with the test would reduce this learning effect. In our study, 10 training attempts seemed to suppress the trial order effect. Our results suggest that with more practice, adult players would show better scores, and also better reliability.

Nevertheless, the LSPT is able to detect differences between playing level without a familiarization of 12 trials in young players (Table 3). The reliability coefficients are certainly weaker, with trial order effects apparent, indicating that players were still undergoing learning effects. More specifically, elite players reduced their global performance score (−2.2 seconds), but sub-elite and non-elite players had higher scores (+11.7 seconds and +12.2 seconds, respectively) from Pre-Fam Trial 1 to Pre-Fam Trial 2. It is possible that sub-elite and non-elite players tried to improve their time taken to the detriment of accuracy (+10 seconds and +9 seconds for penalty time, respectively). This difference of skill comportment is probably because of a different stage of learning within our players. Indeed, Fitts and Posner (⁸) described 3 stages of motor learning: cognitive, associative, and autonomous. During the cognitive stage, an individual may have a general idea of the movement required for a task but might not be sure how to execute that movement. Performance during this stage is likely to be highly variable with a large number of errors and with slow movements. Once the learner has acquired the basic movement pattern, the second, or associative, phase of learning begins. This phase is characterized by more subtle movement adjustments. The movement outcome is more reliable, and the movements are more consistent from trial to trial. Automaticity of motor learning occurs in the third stage, the autonomous stage. At this stage, the motor skill has been learned, and little cognitive effort is required to execute it. Automaticity is evident when a motor skill can be performed while engaging in another task, such as walking and talking or playing the piano and singing (²⁸). Evidence from neuroscience indicates that less brain activation is required when automaticity of movement has been achieved (¹⁸), suggesting that fewer attentional demands are required. The fact that elite players are closer to the autonomous stage than other players could explain the lower SDs of results, by having a more reliable movement pattern, and a lower time score by having a more economical movement and a greater attention reserved for other aspects of performance, as looking at the next target.

Figure 2 shows a diagrammatic representation of LSPT performance scores without familiarization. When looking at elite data, we observed a linear progression for 36 players and 8 players with higher results (>60 seconds). We can assume that these players possess high physical capabilities that counterbalance their lack of technical abilities. If we did not have any information on their training in the past, a ranking was sufficient to “detect” most of the elite players. Indeed, except the 8 elite players with higher scores than other elite players, only 5 or 6 sub-elite or non-elite players demonstrated similar performance than elite players.

These findings confirm that soccer performance is multifaceted, and a single test is not sufficient to evaluate a player's ability. With our players, a threshold of 53 seconds would permit discrimination of most of the elite players but also to detect sub-elite and non-elite players with good technical abilities but whose physical abilities are not (yet?) satisfactory to play at elite level.

Figure 3 shows the data representation after familiarization. With a threshold at 49 seconds, we can discriminate 43 of 44 elite subjects and only 6 sub-elite players. The 8 players who had the highest scores seem to have good learning abilities because only 1 subject demonstrated a “bad” performance compared with other elite subjects.

As the LSPT was able to distinguish players according to their playing level, this opens up the possibility of using the test to examine passing skill performance in young players. From a practical perspective, the LSPT would be relatively easy to set up as part of a battery of talent identification tests: it requires minimal preparation, few materials, and relatively efficient with regard to time. Moreover, our results indicate that minimal familiarization is necessary to be able to distinguish skilled players of different abilities. The results of Impellizzeri et al. (¹⁰) suggest the LSPT should be performed in an unfatigued state for optimal performance scores. Finally, the LSPT permits the assessment of some of the multifaceted aspects of soccer skill including passing, dribbling, control, and decision making (³). Coaches can therefore evaluate several skills that are not evident with other tests (^9,20). The LSPT is particularly adapted with the context of talent identification where differences of maturation are high with adolescents, and so traditional tests (²⁰) disadvantage players with a lateness of maturation. For example, most of the selected players (French players) were born in the first part of year, so they were more mature than their counterparts (¹¹).

Although talent identification processes are used within elite centers/programs to discover future elite players, the process can also be undertaken at lower levels such as regional school centers and local clubs who may need to evaluate or grade players. Our results suggest that the LSPT could be used for this purpose because of its ability to distinguish between sub-elite and non-elite players.

Future investigations should examine learning effects on LSPT performance after 1, 2, 3, and more trials, until stable scores are obtained. Moreover, in addition to characterizing players according to playing level, the LSPT can be used to examine how soccer skill performance evolves with age.

Practical Applications

The LSPT was shown to be a valid and reliable method of examining soccer skill performance in young players. This test exhibits better reliability within the elite group. Moreover, the LSPT is also able to distinguish players of different playing levels with minimal familiarization. Collectively, these data support the use of the LSPT for talent identification purposes in young soccer players.

Acknowledgments

The authors would like to thank the soccer players for their participation in this study. The results of the present study do not constitute endorsement of the product by the authors or the National Strength and Conditioning Association. The authors declare no conflict of interest. E. Le Moal and O. Rué contributed equally to this work.