This study employed a stratified sampling method to recruit 85 adolescent competitive swimmers based on the World Aquatics Point Scoring System 2025 for the 50 m freestyle event. This system is internationally recognized for standardizing performance across age and gender categories and was used here to classify swimmers into performance-based groups. Group I: 23 athletes with ≥621 points. Group II: 42 athletes scoring 440–620 points, indicative of intermediate-level performance typically seen in regional or national developmental athletes. Group III: 20 athletes with <440 points, reflecting early-stage or lower-performing competitive athletes. The full anthropometric and performance characteristics are detailed in Table 1. Although all participants were national-level swimmers, the groups differed in average age and training history (e.g., Group I: 16.2 years, 9.0 years of training vs. Group III: 14.0 years, 4.5 years). This classification approach is consistent with prior literature using FINA-based stratification to assess swimmer development and performance (Crowley et al., 2017; Keiner et al., 2021). Sex distribution and pubertal maturation status were not recorded, which may contribute to unmeasured heterogeneity in physical development and training responsiveness. Grouping was based solely on performance level (FINA points), without adjustment for maturational stage or sex. All participants underwent comprehensive protocol briefings before providing written informed consent. The study protocol was approved by the Science Research Ethics Sub-Committee of the School of Sports Science, South China Normal University (Approval No. SCNU-SPT-2025-012).

Strength capacities can be classified through biomechanical, physiological, and anatomical perspectives. For comprehensive evaluation of adolescent short-course front crawl swimmers, this study adopts an anatomical framework dividing strength capacities into Upper Limb Strength, Core Stability, and Lower Limb Strength. The investigation employed a biphasic testing framework comprising seven land-based physiological assessments administered consecutively with standardized 5-min active recovery intervals, followed by aquatic performance evaluation 24 h post-terrestrial testing.

Swimming performance was evaluated through a maximal-effort 50 m front crawl sprint. Each swimmer performed a continuous two-length swim (2 × 25 m) in a 25 m pool. Prior to testing, swimmers completed a standardized warm-up protocol: 10 min of dynamic land-based activation, 15 min of in-water preparation, and a 10-min rest period. The trials were conducted individually to ensure maximal effort and eliminate drafting or pacing effects, in line with previous protocols (Jones et al., 2018; Born et al., 2020). Swimmers performed a standard dive start from the platform, and time was recorded using handheld stopwatches by experienced personnel positioned at the finish wall. Although manual timing introduces a small margin of error, this method has been commonly used in similar studies with acceptable reliability (Keiner et al., 2021; Zebura et al., 2023). To control for effort and consistency, tests were repeated if the performance fell below 90% of the swimmer’s personal best or if a false start occurred. The best of two valid trials was retained for analysis, following protocols similar to those used by Tan et al. (Tan et al., 2024).

To examine the combinatorial effects of strength variables on swimming performance, we applied fsQCA, a method rooted in Boolean algebra and set theory. FsQCA identifies configurations of conditions that are sufficient or necessary for a specific outcome, making it well-suited for investigating complex, multi-causal phenomena such as sport performance (Wang et al., 2022). Unlike traditional regression-based approaches that assume linear, additive, and symmetric effects, fsQCA accommodates causal asymmetry and equifinality, acknowledging that athletes may reach similar performance levels through distinct combinations of physiological attributes. This approach has previously been used in athletics research, for example, in javelin throwing, where athletes with contrasting physical profiles achieved equivalent performance through different factor combinations (Wang et al., 2022). FsQCA does not require large samples for statistical inference, as it relies on set-theoretic logic rather than probabilistic assumptions. According to methodological guidelines (Ragin, 2009), small-to-medium-N studies, typically ranging from 10 to 100 cases, are considered appropriate and even ideal for generating meaningful configurational results. Our sample of 85 swimmers falls within this range, offering sufficient diversity to identify relevant condition combinations. This sample size is consistent with other recent fsQCA applications in sport performance contexts, such as Wang et al. (2022), who analyzed elite javelin throwers using a comparable number of cases. We applied a minimum raw consistency threshold of 0.80 and a PRI (Proportional Reduction in Inconsistency) threshold of 0.75, as recommended by Ragin (Ragin, 2009) and Schneider & Wagemann (Schneider and Wagemann, 2012). These thresholds are commonly used in medium-N studies and offer a balance between explanatory power and solution diversity. Although some variables (e.g., medicine ball throw) appeared less frequently in the final configurations, we retained them due to their established role in swimming literature as indicators of upper-body power. This helped ensure a comprehensive assessment of relevant strength domains.

The dataset comprising all performance metrics was systematically processed for configuration analysis using fsQCA 3.0 software to examine the complex causal relationships between land-based strength parameters and 50 m front crawl swimming performance, with the latter designated as the outcome variable and the former as conditional variables. Calibration procedures, which function to operationalize abstract concepts into quantifiable fuzzy-set membership scores, transformed both outcome and conditional variables into membership degrees ranging from 0 (full non-membership) to 1 (full membership). Given the absence of standardized empirical cut-offs in the strength–swimming performance domain, we adopted a percentile-based calibration approach following guidelines from Ragin (2009) and Schneider and Wagemann (2012). Specifically, we used the 75th percentile for full membership, 50th percentile (median) for the crossover point, and 25th percentile for full non-membership. This approach ensures sample-specific calibration without imposing arbitrary or externally mismatched thresholds. While no universal benchmarks exist for these strength measures in adolescent swimmers, this strategy is widely applied in small-to-medium-N fsQCA studies when theoretical thresholds are unavailable. Additionally, to assess the robustness of our findings, we performed sensitivity checks by adjusting calibration thresholds ±5% for selected variables. These adjustments yielded similar configurations and consistency scores, confirming the stability of the original results.

For calibration purposes only, additional swimming-specific strength metrics were collected using two instruments: (Amaro et al., 2017) a tethered-swimming ergometer to assess peak and mean power output during a 15-s all-out front crawl effort, and (Lopes et al., 2021) an underwater force mounted on the pool wall to measure wall push-off force during a standard start. These data were not included in the fsQCA models but were used to validate the dry-land strength measures through correlational comparisons (Table 5).

Prior to conducting configuration effect analysis, we performed a necessity examination of individual conditional variables through fsQCA 3.0 by computing consistency and coverage metrics. Consistency, measured on a 0–1 scale, quantifies the directional association strength between conditional variables and the outcome variable, with values approaching 1.0 indicating that the conditional variable’s presence systematically corresponds to outcome occurrence across cases. Coverage estimates the explanatory scope by evaluating the proportion of outcome-positive cases where the conditional variable is present. Diagnostic thresholds followed convention from comparative sports analytics: conditional variables exceeding 0.9 consistency threshold were considered potentially necessary, whereas values below this cutoff indicated requirement for synergistic multi-factor configurations. A value of less than 0.9 indicates that the variable explains the outcome variable together with other conditional variables. The calculation shows that each condition variable showed a certain degree of correlation in the cases of excellent sports performance, but the performance in the non-high-sports performance cases is not enough to establish the necessary condition. Therefore, it is necessary to explore the way to achieve excellent front crawl swimming results and to comprehensively investigate the multiple synergies of multiple factors (Table 6).

The analytical protocol established parameter thresholds with raw consistency at 0.80, PRI consistency at 0.75, and minimum case frequency of 1, incorporating the foundational assumption in counterfactual analysis that both presence and absence of individual land-based strength training modalities could contribute to elite freestyle performance outcomes. Core conditions were identified through the comparative nesting of intermediate and parsimonious solutions, defined as variables appearing in both solution types, while peripheral conditions existed solely in intermediate solutions. This methodological framework revealed six distinct causal configurations (Table 7), with overall solution consistency reaching 0.929, surpassing theoretical sufficiency thresholds. The aggregate coverage of 0.727 indicated that these multi-factor pathways explained 72.7% of high-performance cases.

The study conducted sensitivity checks on optimal freestyle performance configurations by recalibrating anchor points to 80%, 50%, and 20% percentiles. Core configurations demonstrated sustained robustness, with statistically insignificant variations in solution consistency and coverage metrics across threshold adjustments. A formal subset consistency analysis revealed hierarchical relationships among configuration patterns, satisfying necessity conditions for causal robustness, thereby validating the methodological dependability of conclusions (Table 8).

The S1a configuration is characterized by back squat, pull-ups, and grip strength as core condition variables, with bench press and eight-level core stabilization as peripheral factors, while medicine ball chest throws and vertical jump are non-essential. This combination reflects a strength profile focused on posterior chain engagement and trunk control. Similarly, the S1b configuration identifies back squat, pull-ups, and grip strength as foundational requirements, with bench press and vertical jump as peripheral contributors. Although the peripheral variables differ slightly, both S1a and S1b prioritize similar physical qualities. These configurations may support freestyle performance through complementary strength contributions at different race phases. For example, the back squat develops lower-extremity triple-extension power, which is relevant for block starts and turn thrusts, as shown by Zebura et al. (Zebura et al., 2023), who reported associations between lower-body maximal strength and start velocity in elite swimmers (Zebura et al., 2023). Pull-up strength reflects latissimus dorsi function and contributes to sprint performance and swimming speed (Pérez-Olea et al., 2018; Pardo-Atarés et al., 2024). Although grip strength remains less studied, it may contribute to stable hand positioning during propulsion. While our findings suggest a possible functional synergy among these strength components, we did not directly assess neuromuscular coordination or in-water force timing. Future research using EMG or kinetic analyses could clarify how dry-land strength qualities influence race-segment-specific force application. Differential adaptations emerge through configuration-specific emphasis: S1a prioritizes trunk rigidity via Progressive Plank to maintain streamlined postures during start entries and glide phases, whereas S1b leverages vertical jump training’s stretch-shortening cycle properties to amplify explosive start capabilities. These composite adaptations typify elite short-course front crawl swimming athletes combining robust foundational strength with refined hydrodynamic technique.

The S2a configuration identifies bench press, pull-ups, and vertical jump as core condition variables, with progressive plank and back squat classified as peripheral factors, while medicine ball chest throws and grip strength are considered non-determinant; this configuration accounts for approximately 48.7% of elite freestyle performance cases (3.4% uniquely explained). The S2b configuration similarly establishes bench press, pull-ups, and vertical jump as core elements, both configurations share an emphasis on upper-body pushing (bench press), vertical explosiveness (jump), and pulling strength (pull-ups), suggesting an integrated contribution to both start and stroke phases. Strength adaptations drive performance optimization: Kwok et al. (Yu Kwok et al., 2021) suggested that bench press enhances pectoralis major activation critical for stroke efficiency (Yu Kwok et al., 2021), while Born et al. (Born et al., 2020) suggested that vertical jump training improves start and turn acceleration via explosive triple extension (Born et al., 2020). Pull-ups ensure kinetic chain integrity through latissimus dorsi-centric contraction patterns, optimizing early-phase high-elbow catch mechanics complemented by jump-derived torso stiffness improving mid-pull stabilization. Start phases exhibit superior block velocity due to vertical jump engagement, while turns exploit ankle plantarflexion recoil for wall push-off enhancement. These configurations collectively represent athletes with refined stroke mechanics and start-turn proficiency, demonstrating that integrating vertical explosion (jump training) with horizontal thrust generation (bench press/pull-ups) creates a multidimensional power matrix surpassing traditional squat-dominated training paradigms, which is particularly advantageous for sprint-centric performance optimization.

The S3 configuration emphasizes a broader base of full-body power, combining bench press, back squat, grip strength, and vertical jump. This configuration is unique in excluding pull-ups, indicating that some athletes may rely more on raw power than on pulling-specific strength. Progressive plank appears peripherally, while medicine ball throw is excluded. This configuration explains approximately 15% of elite freestyle performance cases, with 1.5% uniquely attributable to it. Athletes aligned with this configuration emphasize whole-body coordination, typically exhibiting higher kicking frequencies and superior starting capabilities. Mechanistically, bench press enhances horizontal propulsion by optimizing upper-limb thrust during the stroke phase, while back squat strengthens lower-body explosive power critical for start and turn performance. Vertical jump training improves vertical takeoff force, enhancing block start efficiency, and grip stability minimizes hydrokinetic hand slippage during propulsion. The exclusion of pull-ups (neglecting latissimus dorsi engagement) and medicine ball throws (suppressing trunk rotation dynamics) prioritizes direct power output over coordinated multi-planar movements. Progressive plank peripherally supports basic torso stability but does not dominate adaptability. This configuration may benefit short sprints where immediate force output is key, but its limited coverage suggests that omitting trunk and back synergy reduces its broader effectiveness. Athletes benefiting exclusively from S3 often possess inherent morphological or technical peculiarities, such as disproportionate limb leverage ratios or a neuromuscular preference for linear force transmission, that align with this simplified kinetic chain strategy. Consequently, S3 represents a specialized, supplementary training archetype, contrasting with the comprehensive S1 and S2 configurations favored by most elite performers due to their integrated power-coordination frameworks.

In the S4 configuration, pull-ups, progressive core stabilization, and vertical jumps are identified as core condition variables, while bench press, back squat, and medicine ball chest throws are classified as marginally non-existent conditions. This configuration, though less frequent, may represent swimmers with highly developed posterior chains and core control, compensating for less traditional push-based strength. This indicates the S4’s highly limited applicability but efficacy in specialized cases. Its intrinsic mechanism prioritizes back pulling strength and core stability combined with vertical explosive power while disregarding traditional thrust-oriented training. It may cater to athletes with specific technical profiles—those relying predominantly on back musculature rather than pectoralis major for propulsion and emphasizing vertical jump-driven starts over horizontal thrust. The uniqueness of the 2.6% cases likely stems from athletes possessing distinct physiological traits: superior latissimus dorsi development with underdeveloped pectoral muscles, restricted shoulder mobility, exceptional core stability, or exceptional vertical takeoff velocity requirements during starts despite deficiencies in other physical attributes. These athletes may achieve performance through high-frequency technical execution, leveraging back-dominant mechanics at the cost of conventional force transmission pathways. The S4 configuration offers an alternative training paradigm for athletes with atypical physical characteristics, utilizing their anatomical peculiarities to circumvent traditional training limitations.

Across configurations S1 to S4, several recurring patterns emerge. Pull-ups appear in nearly all solutions, highlighting the importance of upper-body pulling strength, particularly for underwater propulsion. Vertical jump also features prominently, underlining the role of lower-limb explosiveness for starts and wall push-offs. Progressive plank is a frequent core or peripheral condition, reflecting the value of core stability for maintaining hydrodynamic position. In contrast, medicine ball throw appears only peripherally or is excluded entirely, suggesting its limited contribution in this cohort. These findings suggest that multiple high-performance profiles exist: some athletes excel via upper-body dominance, others through lower-body explosiveness, or through balanced core–limb coordination. Recognizing these configurations allows coaches to tailor training toward the individual athlete’s strengths while avoiding overgeneralized programs. Among the identified configurations, S1a/S1b and S2a/S2b should be prioritized in training practice, as they accounted for most high-performance cases and included core strength variables such as pull-ups, bench press, and vertical jump. These combinations reflect a balance between upper-body pulling, pushing, and lower-body explosive power, offering a robust foundation for improving sprint performance in adolescent swimmers. Coaches should particularly focus on these elements when designing generalized training protocols before individual profiles are fully assessed.

The study identified critical strength configurations that optimally enhance short-distance front crawl performance: Configurations S1a/S1b emphasize back squats, pull-up capacity, and handgrip strength as core factors, with peripheral variables influencing outcomes. Configurations S2a/S2b prioritize bench press performance and vertical jump metrics to optimize stroke efficiency and start/turn acceleration. Configuration S3 focuses on whole-body coordination dynamics, while Configuration S4 combines back muscle pulling strength, core stability, and vertical explosive power. Elite performance requires meeting strength baselines: simultaneous attainment of threshold levels in upper limb propulsion strength, core stability, and lower limb explosiveness.

Previous dry-land strength studies have typically applied linear correlation or regression methods to examine the relationship between isolated strength metrics and swimming performance. For instance, Amaro et al. (2017) and Crowley et al. (2017) reported significant associations between upper-body strength (e.g., bench press, pull-up) and sprint times. More recently, Zebura et al. (2023) found strong correlations between lower-body strength and 5–15 m starts velocities in elite swimmers. While these findings highlight important predictors, such approaches assume linear and additive effects. In contrast, our fsQCA results show that multiple, distinct configurations of strength variables can lead to excellent performance, revealing the presence of causal complexity and functional equivalence not captured by traditional models.

While current analyses primarily focus on strength training, unexamined technical factors may interact with strength adaptations. Additionally, the cross-sectional design of the study limits causal inference, as it does not allow us to determine whether strength improvements directly cause performance gains. Individual variability in training responsiveness underscores the necessity for personalized programming. Existing correlation-based findings should be supplemented with longitudinal studies to establish causal relationships between strength training and swimming performance. In addition, all dry-land strength tests in this study were performed within a single session with fixed 5-min rest intervals. Although this approach ensured procedural consistency, it may have introduced cumulative neuromuscular fatigue, particularly following maximal-effort tasks such as the bench press and back squat, which could have influenced later performance scores and affected the resulting configurations. Furthermore, swim trials were hand-timed, which may introduce minor measurement errors. While care was taken to standardize timing procedures, human reaction variability can affect the precision of short-distance swim trials. Another limitation is the lack of maturation control. Given the adolescent age range of the participants, differences in biological development (e.g., pubertal stage) could influence both strength levels and swimming efficiency, potentially confounding the relationship between strength and performance. Additionally, sex distribution was not recorded, which limits the ability to interpret strength and performance differences across groups. Given known sex-related differences in strength development during adolescence, this omission represents a further source of uncontrolled heterogeneity. Future research should expand to diverse age groups and swimming disciplines to validate configuration universality. The integration of strength development with technical skill refinement warrants exploration to design training protocols that concurrently enhance power output and movement economy. Moreover, the present study did not incorporate technical or psychological factors such as stroke mechanics, motor coordination, or mental readiness, all of which are known to influence swimming performance and interact with neuromuscular traits. The exclusion of relevant physical characteristics, such as flexibility and joint mobility, further limits the ecological validity of the findings. Including these variables would enhance the comprehensiveness of the strength–performance relationship. Future models should adopt a multidimensional approach that includes such variables to provide more holistic and actionable insights (Tedeschi, 2012).

These recommendations should, however, be interpreted considering the study’s methodological constraints. The sample, although sufficient for fsQCA, was regionally limited and may not fully represent broader adolescent swimmer populations. Future research should include longitudinal tracking of strength training adaptations and their impact on swimming performance to strengthen causal interpretation. Validation of the current findings across other swimming events (e.g., backstroke, butterfly), age groups, and sex categories is also warranted to assess generalizability. Furthermore, combining fsQCA with supervised machine learning models (e.g., decision trees, random forests) could support the development of practical diagnostic tools to classify athletes into strength-performance configurations and personalize training interventions. Finally, although this study identified multiple configurations of strength traits associated with high-level performance, their overlapping nature may limit immediate practical application. Without diagnostic tools to classify individual athletes into specific configurations, the translation into personalized training plans remains a conceptual proposition. In addition, the configurational model was not cross-validated on an independent sample, which limits its generalizability. Future research should aim to replicate these findings in separate cohorts and explore decision trees or supervised models to facilitate configuration-based training guidance.