INTRODUCTION

Football (soccer) is a physically demanding, multidirectional sport characterised by repeated high-intensity actions such as sprinting, accelerating, decelerating, and rapid changes of direction, often performed under physical contact.²¹ These demands contribute to distinct physical and biomechanical profiles across playing positions, with midfielders typically exhibiting greater endurance due to higher total distance covered, forwards emphasising explosive speed and agility, and defenders relying on strength, positioning, and rapid reactive movements.^{1, 4} Consequently, positional roles represent an important determinant of physical performance and training needs in football.

The FMS is widely used to assess fundamental movement competency, offering insight into movement quality, potential asymmetries and limitations that may influence both performance and injury risk.^{24, 36, 37, 45} Analysing associations between FMS scores, athletic performance measures and positional demands may enhance practitioner’s understanding of how movement competency aligns with the sport’s specific physical requirements, if at all.⁹ Extending this analysis across sex and age groups may also be relevant, as developmental, biomechanical and physiological differences may substantially affect both movement competency and athletic performance outcomes.¹⁹

Despite the growing complexity of modern football and the increasing emphasis on individualized training, evidence linking the FMS with key performance indicators - such as sprint speed, agility, and vertical jump - across playing positions remains somewhat limited.^{27, 51} For example, Kraus et al.²⁷ synthesised 34 FMS-related studies and concluded that, although the FMS can be reliable when administered by experienced raters (> 100 trials), the total composite score has limited capacity to predict athletic performance. Notably, only a subset of the literature addressed performance prediction (12 studies in their review) and much of the available evidence was not designed to test moderation by contextual variables such as positional roles in team sports (e.g., defender vs. midfielder). Consequently, the current knowledge base provides insufficient resolution regarding whether movement-quality-performance relationships are consistent across positions or whether they vary according to the distinct sprint-jump demands of each role. In parallel, Verheul et al.⁵¹ highlighted that field-based quantification of biomechanical loads remains challenging and that commonly available monitoring approaches may not capture tissue- and structure-specific loading pathways with sufficient precision. This limitation is highly relevant to positional profiling in football, where between-position differences are expressed through both external performance outputs (e.g., sprinting and jumping) and the underlying mechanical demands that produce them. Collectively, these papers indicate that existing evidence is constrained by i) limited performance-prediction strength for the FMS composite score and ii) methodological barriers to capturing football-relevant mechanical demands in the field, thereby leaving an important gap regarding position-stratified FMS, performance relationships, particularly when also considering sex and age group. Moreover, existing research has not sufficiently examined whether these relationships vary according to sex and age, nor whether positional responsibilities mediate the association between movement quality and performance.^{2, 14} Previous studies suggest that FMS scores differ across demographic groups ^{34, 35, 37, 41} yet their direct connection to football-specific physical performance measures is still unclear.

This lack of integrated evidence constrains coaches, sports scientists and rehabilitation professionals in developing evidence-based, position-specific training and injury-prevention strategies. Given that agility, speed and power are critical determinants of football performance¹² and that positional roles impose distinct biomechanical and physical demands,^{11, 43} a comprehensive examination of FMS in relation to performance, position, sex and age is warranted. Accordingly, this study examined the relationships between FMS scores, physical performance outcomes (sprint speed, vertical jump and agility) and playing positions in football, while accounting for sex and age differences. It further explores positional variations in FMS profiles and examines sex- and age-related trends in movement quality, aiming to provide preliminary evidence that may assist practitioners in interpreting movement and performance profiles across competitive groups. In addition, recognising that the FMS comprises distinct movement domains, this study explored whether theory-driven modified composite scores emphasising lower-limb mobility and trunk control demonstrated stronger associations with sprint and jump performance than the full FMS composite score. This secondary objective was designed to examine whether movement tasks more closely related to locomotor and lumbo-pelvic function provide a more performance-relevant representation of movement quality in football.

Methods

Experimental Design

This study employed a cross-sectional, non-interventional comparative design to examine differences in functional movement quality and physical performance across competitive categories and playing positions in football. Players were stratified by sex/age group (girls U17, women’s first team, men’s first team) and by self-reported primary playing positions (i.e., goalkeeper, defender, midfielder or attacker). Primary outcome variables included FMS scores, modified lower-body (mFMS-LB) and lower-body plus trunk (mFMS-LB+T) composite score, vertical jump performance - squat jump (SJ), countermovement jump with arms locked (CMJ-AL) and countermovement jump with free arms (CMJ-AF) - and linear sprint performance (10m and 30m, where available). Between-group comparisons were conducted across competitive categories and positional subgroups, while correlational analyses examined associations between movement quality and performance measures across the pooled sample. The design was observational in nature and limited to a single testing time point; therefore, analyses were restricted to group comparisons and cross-sectional relationships without inference of causality or training effects.

Participants

The sample included 53 competitive football players: Girls U17 (n = 14), Women First Team (n = 26), and Men First Team (n = 13). Participants had a mean age of 19.66 ± 5.20 years (range: 14-33), height of 172.36 ± 8.75 cm, body mass of 64.14 ± 11.29 kg, and 6.62 ± 4.29 years of structured football training (≥ 4 sessions/week). Playing positions comprised of defenders (37.7%), midfielders (28.3%), attackers (26.4%) and goalkeepers (7.5%), with 75.5% right-foot dominant. No participant reported a musculoskeletal injury within the previous six months. All athletes were classified as Tier 3 (Highly Trained/National Level) according to the Participant Classification Framework,³⁸ based on competitive league participation. Inclusion criteria were active competitive participation in the preceding season, regular training (≥ 4 sessions/week), a minimum of five years of club-organised football training, at least one year of structured strength and conditioning training and testing experience. Exclusion criteria included recent lower-limb injury or surgery, current pain or functional limitation, neurological or vestibular disorders and medication affecting neuromuscular function. Participants avoided high-intensity activity for 48 hours prior to testing. The study was non-interventional, approved by the institutional review board (RES2024/016), and conducted with written informed consent (parental consent for minors), with all data anonymised for subsequent analyses.

Procedures

The study adopted a comparative–analytical design using standardised FMS and physical performance assessments to ensure methodological consistency. Participants were stratified by sex, age, and playing position to examine demographic- and position-specific patterns. All tests were administered by experienced sport scientists with prior certification and practical experience in FMS scoring and performance testing, using standardised instructions, demonstrations and a uniform warm-up protocol to minimise intra- and inter-rater variability. Testing sessions were video recorded for scoring verification and quality control. Players were grouped by self-reported primary playing position (goalkeeper, defender, midfielder, or attacker) and sex/age category for analysis.

Functional movement screen (FMS)

Functional movement quality was assessed using the FMS, administered according to standardised, previously published protocols.^{31, 32, 40} The FMS consists of seven movement tasks (deep squat, hurdle step, in-line lunge, shoulder mobility, active straight leg raise, trunk stability push-up, rotary stability), each scored on a 0–3 ordinal scale (3 = performs movement correctly without compensation; 2 = completes movement with compensation; 1 = unable to complete movement; 0 = pain during movement). For bilateral tests, both sides were assessed and the lower score was retained for subsequent analysis, in accordance with established scoring procedures. Individual task scores were then summed to derive a composite score ranging from 0 to 21. All test set-ups (e.g., hurdle height, tibial-length–based heel-to-toe distance in the in-line lunge) followed standard FMS kit specifications and published administration guidelines.^{31, 32} Each participant performed three standardised attempts per task (and per side for bilateral tests), with the highest score from the attempts retained for each side. For tests with separate left and right-side scores, the lower score between the left and right sides was then recorded as the final score for that task. The shoulder mobility, trunk stability push-up, and rotary stability tests included their standard clearing screens, whereby any pain reported during the test itself or clearing screen resulted in a score of 0 for that assessment. Subsequent analyses were conducted using the composite FMS score.^{7, 8, 36}

Modified functional movement screen composites

In addition to analyses using the full composite FMS score (0-21), two secondary, theory-driven modified composites were computed to examine whether movement tasks primarily assessing lower-limb mobility and lumbopelvic control demonstrated stronger associations with sprint and jump performance. These modified composites excluded the shoulder mobility task, not because upper-extremity function is irrelevant to sprinting, but because the FMS shoulder mobility screen predominantly evaluates glenohumeral range of motion rather than dynamic arm-swing coordination or force contribution during locomotion. The modified scores were therefore constructed to isolate movement constructs more directly aligned with lower limb force production and propulsion.

The first modified composite, termed the modified lower-body FMS (mFMS-LB; 0-12), included the deep squat, hurdle step, in-line lunge and active straight leg raise. These tasks were selected as they predominantly assess lower-limb mobility, unilateral stability, inter-limb coordination and pelvic control during functional movement patterns. Each task was scored on the standard 0-3 scale and individual scores were summed to produce a composite ranging from 0 to 12, with higher values indicating superior lower-extremity movement quality.

The second modified composite, termed the modified lower-body plus trunk FMS (mFMS-LB+T; 0-18), comprised the four lower-body tasks described above in addition to the trunk stability push-up and rotary stability tests. These additional tasks were included to capture sagittal- and transverse-plane trunk stability and lumbopelvic control, which are theoretically implicated in force transmission and sprint-jump performance. Scores from the six tasks were summed to yield a composite ranging from 0 to 18.

Athletic performance tests

Lower-limb power and sprint ability were assessed using vertical jump and linear sprint tests. Vertical jump height was measured with the My Jump 2 application during the squat jump, countermovement jump with hands on hips (CMJ-AL) and countermovement jump with free arms (CMJ-AF) a method shown to be valid and reliable in various populations.^{42, 44, 47} The CMJ-AL was included to assess lower-limb power while minimizing the contribution of arm swing, whereas the CMJ-AF reflects a more sport-specific jumping action in which arm swing contributes to impulse generation and superior jump height. Sprint performance was evaluated using a 30 m linear sprint with split times at 10 m and 30 m. Participants performed three maximal sprint trials, each separated by three minutes of passive recovery, with the fastest trial retained for subsequent data analysis. Timing was obtained from 4K (60 fps) video recordings analysed with Kinovea software²⁵, which has been shown to be a valid and reliable method for sprint assessment.¹⁰

All trials were video recorded for verification and quality control. Each FMS task was performed for three standardised attempts (and bilaterally where applicable), following detailed verbal instructions and a visual demonstration. The highest score achieved, in accordance with FMS scoring criteria, was retained for analysis. Prior to testing, all participants followed a standardised 10-minute dynamic warm-up consisting of: i) three minutes of low-intensity jogging; ii) dynamic mobility exercises for the hip, knee, and ankle (leg swings, walking lunges, inchworms; 2 × 10 repetitions each); iii) progressive sprint drills (2 × 20m at ~60% and ~80% perceived maximal effort); and iv) two submaximal practice jumps. The warm-up sequence and order were identical for all participants and supervised by the research team to ensure consistency. All participants completed FMS and vertical jump testing; whilst sprint testing was not performed by the men’s team due to logistical constraints.

Statistical Analysis

For performance tests involving multiple trials (sprint and jump assessments), relative reliability was evaluated using two-way random-effects intraclass correlation coefficients (ICC [2,1]) with 95% confidence intervals. Absolute reliability was quantified using the coefficient of variation (CV, %) and standard error of measurement (SEM). ICC values were interpreted as: < 0.50 = poor, 0.50-0.74 = moderate, 0.75-0.89 = good, ≥ 0.90 = excellent reliability.²⁶ Descriptive statistics are reported as mean ± standard deviation (SD), together with 95% confidence intervals (95% CI). Statistical significance was set at p < 0.05. Prior to inferential analyses, assumptions of normality and homogeneity of variances were examined using the Shapiro-Wilk and Levene’s tests, respectively. Sprint performance differences between girls U17 and women’s first team players were analysed using independent-samples t-tests for both 10m and 30m sprint times. Because of unequal group sizes, standardized mean differences were calculated using Hedges g; a bias-corrected estimator of effect size, with corresponding 95% confidence intervals.^{18, 28} Effect sizes (Hedges g and Pearson’s r) were interpreted using the following magnitude thresholds: < 0.20 = trivial; 0.20–0.49 = small; 0.50.79 = moderate; ≥ 0.80 = large. Comparisons across competitive categories (girls U17, women’s first team, men’s first team) and by playing position, were performed using one-way analyses of variance (ANOVA). ^{6, 29} When significant main effects were observed, Bonferroni-adjusted post hoc tests were applied. For consistency across analyses, effect sizes for pairwise comparisons were expressed as Hedges g with 95% confidence intervals. Bivariate associations among training experience, FMS composite scores (full and modified), jump performance (countermovement jump with arms free, countermovement jump with arms locked, squat jump), and sprint performance (10m and 30m) were examined using Pearson’s product–moment correlation coefficients. For each association, r values, 95% confidence intervals and p-values were reported. The magnitude of correlations was interpreted using the same thresholds described above (applied to absolute r values). Correlation magnitudes were interpreted using the same thresholds described above (applied to absolute r values) to facilitate consistent interpretation across effect-size families.^{46, 48}

Results

This section reports descriptive and inferential statistics for functional movement quality and performance outcomes across cohorts and where applicable, by playing position. Figures and tables provide full descriptive distributions and group means. A consolidated overview of cohort differences is presented in Table 1. Women’s first team players outperformed the girls U17 in sprint performance (both distances), while the men’s first team players demonstrated the highest jump performance across all jump tests. FMS scores were higher in both senior groups compared to the U17 girls group, with comparable values between women’s and men’s first teams. Sprint comparisons involving men were not possible due to missing sprint data, as outlined in the methods.

UKSCA-IQ

Applications of technology in S&C coaching

This content is available free to registered users

Additional resources