Assessing the biomechanical characteristics of defensive movements in elite boxers of different competitive levels

Abstract

Background

The purpose of this study was to identify biomechanical patterns that determine the efficiency and stability of performing defensive movements in highly qualified boxers, and to establish relationships between the level of athletic training and the nature of motor strategies.

Methods

The study was conducted at the National Boxing Training Centre in Xianyang, Shaanxi Province, China, from January to December 2023. 28 athletes took part in the experiment, divided into two groups according to the level of competitive experience. The methodology included the recording of spatiotemporal, kinematic, kinetic, and electromyographic parameters when slipping, ducking, and performing sidesteps under conditions that simulated real combat.

Results

The results showed that international boxers had a high amplitude and speed of movement of the centre of mass (CM) (by 12.3% and 18.6% higher), increased angular velocity in the hip joints (by 23.2°/s), and increased muscle activation of key segments of the torso and limbs (up to + 22.8%). There were more compact and reproducible patterns of muscle activity, a lower delay in peaks of effort (-22 ms), a decrease in peak moments of force in the joints (up to -16.5%), high load symmetry (89.2%) and reduced coefficients of variability, indicating the stability of the motor programme. National-level boxers have pronounced asymmetries, high inter-repetition variability, and slower reactive responses, which theoretically may increase biomechanical strain.

Conclusion

The findings emphasise that biomechanical maturity, expressed in stable intermuscular coordination, symmetry and variability, is closely related to the competitive level. The practical significance lies in the use of these parameters for biomechanical monitoring, personalisation of the training process, and early detection of functional limitations that can affect the reliability and mechanical efficiency of the technique in combat conditions.

Introduction

Modern boxing, one of the most dynamic and injury-prone sports, requires great motor action accuracy, fast decision-making, and motor pattern stability under sensory load. Slips, ducks, and sidesteps in boxing are crucial to battle safety and efficiency, especially for professional athletes. Even though boxers receive extensive technical and tactical training, biomechanical mechanisms behind successful defence methods are still poorly understood. Understanding which motor qualities improve defensive manoeuvres and minimise theoretical biomechanical overloads, asymmetries, and instability in the musculoskeletal system is crucial.

Technical proficiency and the athlete’s ability to recreate consistent and economical biomechanical patterns determine defensive movement effectiveness. Zhong and Bu [1] showed that weak control over the center of mass (CM) increases the probability of losing balance during slips, which violates posture and opens vulnerable areas for assault. Neuromuscular coordination matters too. Liu et al. [2] stressed the need of synchronising thoracic and lower extremity muscle activation to prevent abrupt dislocations. Beránek et al. [3] discovered that optimal deep torso muscular activation minimises lower limb joint strain during protective actions.

Kinematic and kinetic characteristics that indicate technique efficiency and safety have been prioritised in biomechanical research. Limballe et al. [4] reviewed defensive movement speed and amplitude and found that counterattacking techniques preferred shorter, reproducible movements. Lambert et al. [5] found that experienced boxers have decreased spatial characteristic fluctuations due to motor program consistency, represented in low inter-repetition variability. Wu et al. [6] identified a link between asymmetry and lumbar pain in athletes, indicating that asymmetries enhance the risk of microtraumatisation.

Despite these foundational studies, previous methods often modelled single movements in the lab, failed to compare data with elite athletes, or omitted critical parameters like angular velocities in large joints, muscle effort peak delays, and the complex relationship between variability and muscle symmetry [7, 8]. To address these gaps, we hypothesized that international-level boxers, compared to national-level athletes, would demonstrate: (1) greater kinematic efficiency (higher amplitude and speed of center of mass movement); (2) more stable and symmetric neuromuscular activation with shorter peak effort delays; and (3) optimized biomechanical profiles, indicated by lower peak joint forces and greater motor symmetry.

Thus, the need for a comprehensive study of the biomechanical characteristics of defensive movements in highly qualified boxers becomes obvious both from the standpoint of improving competitive efficiency and understanding theoretical mechanical stress. The purpose of the study was to identify biomechanical patterns that determine the efficiency and mechanical stability of performing defensive movements in highly qualified boxers. The objectives of the study included the analysis of spatiotemporal, kinematic, kinetic, and electromyographic parameters, the establishment of differences in motor strategies between boxers of different levels, and the assessment of the relationship between symmetry, variability, and stability of the motor programme.

Materials and methods

The study was conducted between January and December 2023 at the Chinese National Boxing Training Centre in Xianyang, Shaanxi Province, China [9]. The study involved 28 elite boxers (18 men and 10 women) between the ages of 19 and 31, who were included based on their current national-level athletic qualifications, absence of injuries during the last six months, and consent to participate. Athletes with chronic musculoskeletal disorders and neurological disorders were excluded. All participants were divided into two groups depending on the level of competitive experience: national (n = 15) and international (n = 13). This distribution allowed assessing the impact of the level of training on the biomechanics of defensive movements. The study protocol was reviewed and approved by the Sports Medicine Expert Commission of the Chinese National Boxing Training Centre (Protocol No. CNBC-2022-41) and complied with the institution’s internal requirements for conducting research involving athletes. Limb dominance was determined based on the preferred kicking and stabilising leg; 22 athletes (78.6%) were right-leg dominant and 6 (21.4%) left-leg dominant, with no between-group differences (p = 0.77). Fighting stance (orthodox/southpaw), verified by coaching staff and competition records, was distributed as 21 orthodox (75.0%) and 7 southpaw (25.0%), without significant group differences (p = 0.86). Training experience averaged 7.8 ± 2.1 years in the national group and 11.6 ± 2.7 years in the international group (p < 0.001).

The biomechanical assessment was performed using the Vicon Nexus 2.10 three-dimensional motion analysis system (Vicon Motion Systems, UK), which includes 12 infrared cameras and two platforms for recording the reaction force of the AMTI OR6-7 support (Advanced Mechanical Technology (AMTI), USA). The system has been calibrated according to the manufacturer’s standard protocol. Additionally, Noraxon MyoMotion inertial sensors (Noraxon, USA) were used to evaluate muscle activity recorded in real time. Surface electromyography was performed using the Delsys Trigno system (Delsys, USA), which allowed assessing the activation level of four key muscles: m. rectus femoris, m. biceps femoris, m. obliquus externus abdominis, and m. deltoideus. Electrodes were placed in accordance with the SENIAM (Surface ElectroMyoGraphy for the Non-Invasive Assessment of Muscles) guidelines following standard skin preparation (shaving, abrading, and cleaning with alcohol). Electromyography (EMG) data were recorded at a sampling frequency of 1,500 Hz. Signal processing was performed offline. Raw EMG signals were band-pass filtered (20–450 Hz, 4th order Butterworth filter) to remove movement artifacts and high-frequency noise, full-wave rectified, and then smoothed using a root mean square (RMS) moving window of 50 ms to create a linear envelope. To allow for between-subject comparisons, EMG amplitudes were normalized to a percentage of Maximum Voluntary Contraction (%MVC). MVCs were obtained for each muscle prior to the dynamic trials through three 5-second maximal isometric contractions against manual resistance, separated by 2-minute rest intervals.

These muscles were selected to capture the principal neuromuscular contributions to boxing defensive actions across the lower limb, trunk, and upper limb segments. The m. rectus femoris and m. biceps femoris represent key knee/hip extensor-flexor activity responsible for rapid lowering/raising, braking, and re-acceleration during slips, ducks, and lateral steps, thereby reflecting lower-limb control of centre-of-mass displacement. The m. obliquus externus abdominis was monitored as a primary trunk rotator/lateral stabiliser supporting torso control and efficient inter-segmental coordination, while the m. deltoideus was included to represent shoulder-girdle activation involved in maintaining guard position and coordinating upper-limb counterbalance during evasive manoeuvres. While pelvic stabilizing muscles play a well-documented role in lower-body stability, they were not included in the present recording protocol. This decision was driven by the methodological focus on primary multi-joint movers driving explosive vertical and lateral displacements, as well as practical limitations regarding equipment channel availability and the risk of electrode displacement under the athletes’ standard competition boxing trunks during high-amplitude maneuvers. Future investigations should incorporate gluteal electromyography to provide a more comprehensive assessment of lumbo-pelvic stabilization during boxing defenses.

Each participant performed three defensive movement conditions: slip, duck, and sidestep. Testing was conducted in response to standardised attacks delivered by a sparring partner of a comparable weight category according to a predefined protocol. For each defensive condition, the sparring partner executed straight attacking actions directed to the head at a controlled and coach-monitored tempo, and the boxer responded with the instructed defensive movement immediately after attack initiation. Each movement type was performed in 3 valid repetitions, resulting in 9 analysed trials per athlete (3 slips, 3 ducks, and 3 sidesteps). A rest interval of 20–30 s was provided between repetitions and 2 min between movement conditions to minimise fatigue-related effects. The order of the three defensive conditions was standardised across participants. The tests were carried out in a room with controlled lighting and a temperature of 21–23 °C at a relative humidity of 40–50%. The floor was covered with a non-slip rubber coating. Before testing, all athletes completed a standardised 15-minute warm-up. Reflective markers were placed on anatomical landmarks of the lower limbs, torso, and upper limbs according to the Plug-in Gait protocol [10]. For the principal analysis, parameters from the three defensive conditions were averaged to characterise the athlete’s overall defensive biomechanical profile.

Kinematic data was recorded at a frequency of 200 Hz, platform data – at a frequency of 1,000 Hz, and electromyography (EMG) – at a frequency of 1,500 Hz. The parameters to be analysed included the amplitude and velocity of the CM, angular velocity in the hip and knee joints, symmetry of movements, coefficient of variation, and muscle coordination. Biomechanical indicators of joint overload, including peak moments of force and the time of onset of maximum effort, were treated as candidate indicators of joint overload and potential injury-related mechanical stress. In addition, kinetic asymmetry was determined using the symmetry index based on the moment of force in both legs.

Data processing was carried out using Visual3D Professional v6 software (C-Motion, USA) [11], EMGworks Analysis v4 (Delsys, USA) [12] and MATLAB R2022a (MathWorks, USA) [13]. Descriptive statistics are reported as mean ± standard deviation. Prior to inferential testing, the distributional assumptions for each outcome were evaluated using the Shapiro-Wilk test and visual inspection of Q-Q plots, while homogeneity of variance was assessed using Levene’s test. Between-group differences were examined using independent-samples t-tests (or Welch’s t-test when variances were unequal) and one-way analysis of variance with Bonferroni-adjusted post hoc comparisons where appropriate; when normality assumptions were not satisfied, the Mann-Whitney U test was applied as a sensitivity analysis. In addition to p-values, effect sizes were reported for all primary contrasts: Cohen’s d (positive values indicate higher values in the international group), alongside 95% confidence intervals. Given the study focus on potential injury-related biomechanical surrogates (e.g., joint moments, asymmetry indices, and variability coefficients), the interpretive emphasis was placed on the magnitude of effects rather than statistical significance alone. Because prospective injury incidence was not recorded in the present design, claims regarding ‘injury risk markers’ are framed as indirect inferences based on established biomechanical theory; future work should link these indicators to injury outcomes using correlational and predictive modelling.

Results

Kinematic analysis of the pooled defensive-movement dataset (combined slips, ducks, and sidesteps) revealed between-group differences in centre-of-mass (CM) displacement amplitude, CM movement speed, hip angular velocity, and variability of angular velocity (Table 1). International-level boxers demonstrated higher CM displacement amplitude (25.6 ± 2.8 cm vs. 22.8 ± 3.2 cm; p = 0.018; Cohen’s d = 0.93; 95% CI for mean difference: 0.5 to 5.1 cm), higher CM movement speed (133.2 ± 9.4 cm/s vs. 112.4 ± 10.9 cm/s; p < 0.001; d = 2.03; 95% CI: 12.9 to 28.7 cm/s), and higher hip angular velocity (205.7 ± 18.9°/s vs. 182.5 ± 21.6°/s; p = 0.006; d = 1.14; 95% CI: 7.5 to 38.9°/s). The difference in knee angular velocity did not reach statistical significance (179.8 ± 19.6°/s vs. 168.3 ± 24.1°/s; p = 0.092; d = 0.52; 95% CI: −5.5 to 28.5°/s). The coefficient of variation of angular velocity was lower in the international group (11.1 ± 3.3% vs. 15.2 ± 4.7%; p = 0.013; d = − 1.00; 95% CI: −7.2 to − 1.0%).

Table 1.

Kinematic parameters of defensive movements in athletes of various training levels

Indicator	National level (n = 15)	International level (n = 13)	p-value	Cohen’s d	95% CI for mean difference
CM displacement amplitude (cm)	22.8 ± 3.2	25.6 ± 2.8	0.018	0.93	0.5 to 5.1
CM movement speed (cm/s)	112.4 ± 10.9	133.2 ± 9.4	< 0.001	2.03	12.9 to 28.7
Angular velocity of the hip joint (°/s)	182.5 ± 21.6	205.7 ± 18.9	0.006	1.14	7.5 to 38.9
Angular velocity of the knee joint (°/s)	168.3 ± 24.1	179.8 ± 19.6	0.092	0.52	−5.5 to 28.5
Coefficient of variation of angular velocity (%)	15.2 ± 4.7	11.1 ± 3.3	0.013	-1.00	−7.2 to − 1.0

Cohen’s d is calculated as (international − national)/SDpooled; positive values indicate higher values in the international group. The 95% confidence interval refers to the mean difference between groups (international − national)

Source: compiled by the authors

The observed between-group effect sizes for the kinematic variables ranged from d = 0.52 to d = 2.03. Relative to the national group, the international group showed an 18.6% higher CM movement speed, a 12.3% higher CM displacement amplitude, a 23.2°/s higher hip angular velocity, and a 4.1%-point lower coefficient of variation of angular velocity. Knee angular velocity was numerically higher in the international group, although this difference was not statistically significant.

Electromyographic analysis based on the pooled data from slips, ducks, and sidesteps showed higher activation amplitudes in the international group across all recorded muscles and a shorter activation peak delay (Table 2). Mean m. rectus femoris activation was 72.6 ± 7.5%MVC in the international group versus 63.7 ± 8.2%MVC in the national group (p = 0.021; d = 1.13; 95% CI: 2.8 to 15.0). For m. biceps femoris, the corresponding values were 67.3 ± 8.7%MVC and 58.9 ± 9.1%MVC (p = 0.037; d = 0.94; 95% CI: 1.5 to 15.3). For m. obliquus externus abdominis, the values were 75.1 ± 6.2%MVC and 61.2 ± 6.5%MVC (p < 0.001; d = 2.18; 95% CI: 9.0 to 18.8). For m. deltoideus, the values were 68.3 ± 7.1%MVC and 55.6 ± 7.8%MVC (p < 0.001; d = 1.70; 95% CI: 6.9 to 18.5). Activation peak delay was lower in the international group (62 ± 14 ms vs. 84 ± 17 ms; p = 0.008; d = − 1.40; 95% CI: −34.0 to − 10.0 ms).

Table 2.

Electromyographic parameters during defensive movements

Muscle	National level (%MVC)	International level (%MVC)	p-value	Cohen’s d	95% CI for mean difference
m. rectus femoris	63.7 ± 8.2	72.6 ± 7.5	0.021	1.13	2.8 to 15.0
m. biceps femoris	58.9 ± 9.1	67.3 ± 8.7	0.037	0.94	1.5 to 15.3
m. obliquus externus abdominis	61.2 ± 6.5	75.1 ± 6.2	< 0.001	2.18	9.0 to 18.8
m. deltoideus	55.6 ± 7.8	68.3 ± 7.1	< 0.001	1.70	6.9 to 18.5
Activation peak delay (ms)	84 ± 17	62 ± 14	0.008	-1.40	−34.0 to − 10.0

Muscle activation is reported as a percentage of maximal voluntary contraction (%MVC); activation peak delay is reported in milliseconds (ms). Cohen’s d is calculated as (international − national)/SDpooled; positive values indicate higher values in the international group. The 95% confidence interval refers to the mean difference between groups (international − national)

Source: compiled by the authors

For the electromyographic variables, effect sizes ranged from d = 0.94 to d = 2.18, while activation peak delay showed d = − 1.40. Compared with the national group, the international group demonstrated higher activation amplitudes in m. rectus femoris (+ 8.9%MVC), m. biceps femoris (+ 8.4%MVC), m. obliquus externus abdominis (+ 13.9%MVC), and m. deltoideus (+ 12.7%MVC), together with a 22 ms lower activation peak delay.

Kinetic analysis of the combined defensive trials showed lower peak joint moments and shorter time to peak force in the international group, together with a higher symmetry index (Table 3). Peak knee joint moment was 124.6 ± 15.8 N·m in the international group and 142.3 ± 19.6 N·m in the national group (p = 0.017; d = − 0.99; 95% CI: −31.5 to − 3.9 N·m). Peak hip joint moment was 148.2 ± 18.7 N·m and 168.7 ± 22.4 N·m, respectively (p = 0.009; d = − 0.99; 95% CI: −36.5 to − 4.5 N·m). Time to peak force was 85 ± 18 ms in the international group versus 112 ± 21 ms in the national group (p < 0.001; d = − 1.37; 95% CI: −42.2 to − 11.8 ms). The symmetry index was higher in the international group (91.6 ± 5.2% vs. 82.1 ± 7.4%; p < 0.001; d = 1.47; 95% CI: 4.6 to 14.4%).

Table 3.

Kinetic load and symmetry indicators when performing defensive movements

Indicator	National level	International level	p-value	Cohen’s d	95% CI for mean difference
Peak moment of force in the knee joint (N·m)	142.3 ± 19.6	124.6 ± 15.8	0.017	-0.99	−31.5 to − 3.9
Peak moment of force in the hip joint (N·m)	168.7 ± 22.4	148.2 ± 18.7	0.009	-0.99	−36.5 to − 4.5
Time to reach peak force (ms)	112 ± 21	85 ± 18	< 0.001	-1.37	−42.2 to − 11.8
Symmetry index (%)	82.1 ± 7.4	91.6 ± 5.2	< 0.001	1.47	4.6 to 14.4

Cohen’s d is calculated as (international − national)/SDpooled; positive values indicate higher values in the international group. The 95% confidence interval refers to the mean difference between groups (international − national)

Source: compiled by the authors

The between-group effects were consistently large (Cohen’s d ≈ − 1.37 to 1.47). The analysis of kinetic load parameters and motion symmetry revealed critical differences between national and international athletes, which are of direct importance for assessing biomechanical stability and theoretical mechanical strain. Boxers of the international category demonstrated lower peak moments of force in the large joints of the lower limb, which may reflect a different distribution of mechanical load during defensive manoeuvres. A particularly pronounced decrease in peak hip moment (by 12.1%) may indicate more coordinated whole-body movement and more effective use of the kinetic chain. However, lower peak joint moments should not be interpreted as direct evidence of reduced injury risk, since joint loading is task-specific and injury mechanisms depend on multiple biomechanical and contextual factors. Therefore, these findings are better understood as indicators of movement strategy and load organisation. The relationship between these kinetic timing and symmetry indicators is illustrated in Fig. 1.

Fig. 1.

Indicators of biomechanical stability: Kinematic symmetry index vs. Time to reach peak force

International athletes reached peak effort 27 ms earlier, which may reflect more time-efficient temporal coordination and faster neuromuscular adjustment during defensive actions. In contrast, delayed activation in national athletes may indicate less efficient timing of stabilisation. However, these temporal differences should be interpreted as performance-related correlates rather than direct indicators of joint protection or injury risk. The symmetry index was significantly higher in the international group (91.6% vs. 82.1%), indicating balanced load distribution and stable motor patterns. Lower symmetry in national athletes suggests functional imbalances and greater biomechanical strain, potentially predisposing to chronic overload. Overall, kinetic parameters in international athletes reflect advanced adaptation to competitive stress and more efficient, reliable defensive mechanics.

Thus, the results of the analysis confirm that kinetic indicators – peak moments of force, time of onset of effort and the symmetry index – are not only characteristics of technical performance, but also candidate biomechanical surrogates that have been associated with injury risk in the broader sports biomechanics literature. However, their predictive validity in boxing requires prospective injury surveillance and multivariable modelling.

The analysis of variability and symmetry parameters in the pooled defensive dataset demonstrated between-group differences in the coefficient of variation of CM movement, coefficient of variation of angular velocity, kinematic symmetry index, and symmetry fluctuation between repetitions (Table 4). Accordingly, the results presented in Tables 1, 2, 3 and 4 should be interpreted as reflecting the athletes’ general defensive biomechanical profile rather than movement-specific differences between slips, ducks, and sidesteps. International-level boxers showed lower coefficient of variation of CM movement (8.9 ± 2.4% vs. 12.7 ± 3.1%; p = 0.002; d = − 1.36; 95% CI: −5.9 to − 1.7%), lower coefficient of variation of angular velocity (11.1 ± 3.3% vs. 15.2 ± 4.7%; p = 0.013; d = − 1.00; 95% CI: −7.2 to − 1.0%), higher kinematic symmetry index (89.2 ± 5.6% vs. 78.5 ± 6.8%; p < 0.001; d = 1.71; 95% CI: 5.9 to 15.5%), and lower symmetry fluctuation between repetitions (6.7 ± 2.5% vs. 11.4 ± 3.9%; p = 0.004; d = − 1.41; 95% CI: −7.2 to − 2.2%).

Table 4.

Indicators of variability and symmetry of defensive movements in athletes of various levels

Indicator	National level	International level	p-value	Cohen’s d	95% CI for mean difference
Coefficient of variation of CM movement (%)	12.7 ± 3.1	8.9 ± 2.4	0.002	-1.36	−5.9 to − 1.7
Coefficient of variation of angular velocity (%)	15.2 ± 4.7	11.1 ± 3.3	0.013	-1.00	−7.2 to − 1.0
Kinematic symmetry index (%)	78.5 ± 6.8	89.2 ± 5.6	< 0.001	1.71	5.9 to 15.5
Symmetry fluctuation between repetitions (%)	11.4 ± 3.9	6.7 ± 2.5	0.004	-1.41	−7.2 to − 2.2

Cohen’s d is calculated as (international − national)/SDpooled; positive values indicate higher values in the international group. The 95% confidence interval refers to the mean difference between groups (international − national)

Source: compiled by the authors

A study of elite boxers reveals that international-level athletes possess significantly greater motor stability and functional symmetry (89.2%) than national competitors. Their low movement variability reflects well-developed, repeatable defensive motor programs. This biomechanical consistency is crucial for maintaining balance, efficiently transferring momentum, and avoiding disorganization during sudden attacks.

Conversely, national boxers exhibit higher variability and lower symmetry (78.5%), indicating unstable motor habits. This asymmetry, likely due to insufficient training of the non-dominant side or intermuscular imbalances, increases their susceptibility to unilateral overload and potential micro-injuries, ultimately impairing their mechanical consistency and defensive readiness over time. In sports biomechanics, such parameters are often discussed as theoretical contributors to cumulative stress applied to lower extremity and lumbo-pelvic joints.

Thus, the results show that high symmetry and low movement variability may be considered indicators of biomechanical maturity and functional consistency in defensive actions. However, describing them as prerequisites for safety should be avoided, as the present study did not directly assess injury outcomes. Monitoring these measures throughout training helps track the athlete’s progress and detect underlying dysfunctions that limit protective action effectiveness and may theoretically contribute to mechanical overload.

Discussion

Internationally competitive boxers have a more economical and repeatable defensive movement profile than national-level athletes, with differences in centre-of-mass control, joint angular mechanics, neuromuscular timing, and kinetic load distribution [14, 15]. Kinematically, increased centre-of-mass displacement at faster speed may indicate the ability to quickly reposition while maintaining postural control, which is essential to time-pressure evasion. In other settings, high-level athletes use proximal joint control to make sharp or evasive movements, as shown by Mausehund and Krosshaug [16] and Donnelly et al. [17]. In this data, these kinematic traits indicate movement strategy and technical organization, not injury reduction. Higher hip (and knee) angular velocities may indicate faster defensive manoeuvre start and better inter-segmental coordination. In injury-focused study examining elite boxing joints, Matharu et al. [18] expressed the worry that task specificity and measurement variability can limit the interpretative value of angular velocity data. Thus, angular velocities should be considered part of a multi-parameter biomechanical profile rather than a diagnostic criterion.

Neuromuscular findings support the idea that elite defensive performance requires anticipatory trunk and shoulder-girdle stability and coordinated lower-limb drive. Research by De Donato et al. [19] and Çemç et al. [20] suggests that experienced athletes possess more efficient and timely activation patterns during high-speed interactions and repeated hits. However, Kolodziej et al. [21] describe how increased EMG amplitudes may indicate compensatory co-contraction in suboptimal mechanics, making them not necessarily helpful. The combination of low variability and strong symmetry suggests optimised coordination rather than compensation in this dataset. Shorter delays to peak activation and faster peak effort may aid repositioning and joint stability during evasive actions. According to biomechanical descriptions from Huo [22] and Bathe et al. [23], greater timing and motor control reduce exposure to potentially harmful loading patterns during dynamic tasks. These temporal differences should be viewed as performance-relevant correlates rather than validated injury predictors because injury events were not observed and the design was not prospective. Boxing-specific surveillance work by Dande et al. [24] highlights the need to connect laboratory-derived markers to real-world injury incidence.

Lower peak joint moments and quicker force development may imply better mechanical load distribution across segments and kinetic chain use. McCarthy et al. [25] and Sethi [26] have pushed similar interpretations of repeated impact exposure and high-risk profiles in boxing, as have Attar et al. [27] in discussions of structured preventative programs for the sport. lower peak moments may indicate a plan to limit excessive joint loads while retaining functional speed and accuracy, not lower capability. Stable motor programs and balanced load sharing are indicated by motor symmetry and reduced inter-repetition variability. Hamoongard et al. [28] stress that cognition and motor training improving coordination and control can change biomechanical risk variables in dynamic sports actions. In top sports, prediction and classification methods increasingly use symmetry- and variability-derived characteristics along with fitness measures, however Martins et al. [29] stress that their usefulness depends on outcome definition and validation context. Broad biomechanical syntheses by Vancini et al. [30] also suggest that some asymmetry and variability may be useful, supporting flexible responses under uncertainty.

Within this interpretive framework, the lower variability observed in higher-level athletes may reflect automation that supports efficient decision-to-action coupling under rapid opponent-driven perturbations. Economic analyses of risk and return in professional boxing by Davids et al. [31] implicitly recognise that recurrent exposure and reinjury considerations can shape training and competition behaviours. At the mechanistic level, evidence from other upper-limb loading tasks presented by Motlagh and Lipps [32] suggests that fatigue and altered coordination can modify joint mechanics and potentially increase mechanical stress, underscoring why repeatability and timing may matter for safe execution. Nevertheless, epidemiological syntheses of amateur boxing injuries by Alevras et al. [33] caution against inferring individual risk from isolated biomechanical correlates without prospective follow-up and multivariable modelling.

Consequently, statements about ‘injury risk reduction’ should be limited to theory-consistent interpretation: the present associations are compatible with mechanisms that could plausibly reduce excessive joint loading or unilateral overload, but they do not establish causality or predictive validity. Conceptual and ethical discussions on boxing-related harm by Bartlett et al. [34] underscore that injury and neurological outcomes are influenced by a complex set of exposure, behavioural and regulatory factors that extend beyond movement mechanics alone. Moreover, work in Olympic combat sports by Lambert et al. [35] indicates that injury patterns vary by discipline and competition context, reinforcing the need for sport- and setting-specific validation of any proposed markers.

It should also be noted that the inclusion of both male (n = 18) and female (n = 10) athletes may have contributed to additional inter-individual variability in certain biomechanical and electromyographic parameters due to known sex-related differences in anthropometric structure, muscle mass distribution, and neuromuscular activation strategies. However, as the primary comparisons in the present study were conducted according to competitive level rather than sex, and athletes were matched by qualification and weight category, the observed between-group differences are more plausibly explained by training status and motor maturity than by sex-specific characteristics. Nevertheless, future investigations with larger sex-stratified samples are necessary to determine whether defensive motor patterns demonstrate sex-dependent adaptations at the elite level.

Thus, the results of this study emphasise the critical importance of the integral assessment of kinematic, kinetic, and electromyographic parameters in the analysis of defensive movements in boxing. The differences found between the groups confirm that indicators of motor symmetry, variability, temporal activation structure, and mechanical load distribution can serve as reliable markers of athletic fitness and the basis for individualising training programmes. An integrated approach to assessing the biomechanics of defensive actions allows optimising efficiency and minimising traumatic risks, which is of particular importance in conditions of high competitive load.

Conclusions

The study revealed significant differences in the kinematic, kinetic, and electromyographic characteristics of defensive movements among boxers of various levels of athletic skill, thereby expanding the understanding of the mechanisms of motor organisation in combat interaction. The data obtained indicate that international athletes demonstrate more pronounced motor stability, high symmetry of movements, and coordinated and predictable activation of key muscle groups, which indicates the presence of stable motor programmes and developed strategies for sensorimotor coordination. These features are directly related to the overall quality of defensive manoeuvres, including slips, ducks, and sidesteps. However, because the present analysis pooled the three movement types, the findings should be interpreted as evidence of general defensive biomechanical stability rather than as movement-specific effects.

Qualitative differences in the movement of the CM, the angular velocity of joints, and the moment of muscle strength suggest that the international level of training is associated with greater efficiency, accuracy, and consistency of movements. An increase in the speed and amplitude of the CM movement (by 18.6% and 12.3%, respectively), together with a decrease in the variability of these parameters, may indicate a more automated and repeatable motor response during defensive actions. Increased muscle activation values in key segments (by 14.1–22.8%), combined with a shorter delay in peak effort and lower peak joint moments, may reflect a distinct neuromuscular control strategy and more coordinated pre-movement stabilisation. At the same time, elevated EMG amplitudes should be interpreted cautiously, since they may also arise from compensatory activation or increased co-contraction.

Significant differences in kinetic parameters, in particular, a decrease in peak moments of force in the hip and knee joints while achieving effort earlier, indicate a rational distribution of load among international athletes. High symmetry indicators in both kinetic and kinematic characteristics indicate a balanced motor pattern and the absence of functional asymmetries, which is theoretically advantageous for mitigating excessive joint stress. On the contrary, athletes at the national level demonstrate instability of motor patterns, asymmetry, and high variability, which may be conditioned by the limited amount of training and competitive experience.

The practical significance of the results lies in the possibility of using the obtained biomechanical and electromyographic indicators as objective criteria for assessing the level of training of athletes and monitoring their functional state. These data can serve as a basis for the development of individualised training programmes aimed at improving motor stability, movement symmetry, and defence efficiency.

A limitation of the study is that it did not consider possible cognitive and perceptual aspects of responding to attacking stimuli; consequently, reduced variability in centre-of-mass movements may partly reflect attentional focus rather than coordination-related motor stability alone. Additionally, the sample size is relatively small in relation to the extensive number of analyzed biomechanical variables, and the present analysis does not report reliability measures for these biomechanical variables, which should be acknowledged when interpreting the results. In future studies, it is advisable to supplement the analysis with neurocognitive parameters and to incorporate visual deprivation/occlusion or dual-task conditions to disentangle attentional-sensory contributions from coordinative control, establish test-retest reliability metrics for the biomechanical variables, as well as to expand the sample to include athletes of various age and gender categories.

Authors’ contributions

Conception and design: X.W. Acquisition, analysis, or interpretation of data: G.K. Drafting the work or revising it critically for important intellectual content: X.W., G.K. Final approval of the manuscript: both authors.

Funding

None.

Data availability

The authors confirm that the data supporting the findings of this study are available in the article.

Declarations

Ethics approval and consent to participate

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. A study was approved by Ethics Commission of the, National University of Ukraine on Physical Education and Sport December 21, 2022, No 8593-A.

Consent for publication

Informed consent was requested from all individual participants included in the study.

Competing interests

The authors declare no competing interests.