Kickboxing is a combat sport that demands a combination of technical, tactical, and psychophysical abilities (1). K1 is a specific discipline within this sport, characterized by high-intensity combat that allows boxing punches, kicks, and knee strikes. Therefore, the outcome of matches depends on specific technical-tactical (TE-TA) structures and other anthropometric characteristics (2–4). Two domains, what the fighter does (TE-TA structures) and who the fighter is physically (anthropometry), are closely interconnected. Body morphology can influence both the selection and effectiveness of TE-TA structures (5). Moreover, in combat sports, researchers have demonstrated strong associations between motor and functional abilities and both TE-TA structures' effectiveness and time-motion analysis results up to 90% in some cases (6). Together, these findings suggest that performance results emerge from the interaction of TE-TA structures and an athlets physical attributes. Furthermore, the focus of sports science is often to define the characteristics that explain and form a model of performance (7). A higher frequency of certain TE-TA structures has been observed among more successful fighters (8, 9). The present study addresses this gap by developing a logistic regression model that incorporates both offensive and defensive TE-TA structures together with anthropometric variables, with the aim of estimating the likelihood of victory in high-level K1 bouts. This combined approach is important because it moves from describing successful fighters to predicting match outcomes based on modifiable TE-TA structures and relatively stable morphological characteristics, thereby providing coaches and analysts with a more powerful tool for performance profiling and decision-making in elite K1 kickboxing.

Beyond identifying the frequency and impact of these structures on wins and losses, it is necessary to develop a model that reliably characterizes a successful K1 kickboxer competing at a high level. Predictive and explanatory models have incorporated various components of biophysical characteristics in relation to athletes' success and quality (1, 10, 11). Our study proposes a prognostic-explanatory model that estimates a fighter's likelihood of success in high-level K1 kickboxing matches. This study includes a wide range of TE-TA structures and develops a logistic regression model specific to K1 kickboxing, an approach that has not been reported in earlier research at this level of detail. This research aims to contribute to a deeper understanding of the key determinants of success in elite-level kickboxing bouts and provide guidelines for optimizing training processes and managing fight tactics and prognostics.

Based on previous findings, we hypothesized that a higher frequency of certain offensive TE-TA structures would have a positive impact, and also, overreliance on defensive structures might lower the odds of winning. Anthropometric characteristics, especially height, were considered to be a quality for greater odds of success in fighting sports.

As reported in the Methods, inter-rater reliability was confirmed [ICC(2,1) = 0.79; Pearson's r = 0.83–0.86], indicating good reliability (12).

Table 2 presents the basic descriptive statistics for all analyzed structures, mean ± SD, and 95% Confidence Interval (CI).

Among the analyzed technical-tactical structures, hand blocks against punches (V_16) stood out as the most prevalent defensive technique. The anthropometric data revealed considerable variability in fighter height and weight, which may contribute to diverse fighting styles and strategies within the super heavyweight division.

The model identified several significant predictors of victory. Positive effects per unit increase were found for successful strikes: rear-hand uppercuts to the head (V_6_1; 23.0% increase in the odds of winning), lead-hand hooks to the head (V_9_1; 17.3% increase in the odds of winning), and low rear-leg roundhouse kicks to the lead leg (V_30_1; 12.6% increase in the odds of winning). Negative predictors included defensive structures: each hand block against leg strikes (V_15; 10.4% decrease in the odds of winning), and each hand block against punches (V_16; 4.2% decrease in the odds of winning). Furthermore, fighter height exhibited a positive linear relationship with each additional centimetre, associated with a 9.8% increase in the odds of victory.

The ROC analysis yielded an Area Under the Curve (AUC) of 0.84, indicating the model's high discriminative power in distinguishing between wins and losses (Figure 1). An AUC greater than 0.80 is considered strong (13). Furthermore, the Hosmer-Lemeshow test is pointing to a good model fit (χ² = 9.93; p > 0.05). Table 3 displays the results of the binary logistic regression model, including coefficients, odds ratios, and significance levels for all predictors.

This study identifies key technical, tactical, and anthropometric factors that are associated with winning or losing in K1 kickboxing matches. Descriptive statistics show a high degree of variability in both striking structures and anthropometric characteristics among super heavyweight fighters. Similar observations have been reported in prior studies focused on the morphological characteristics of super heavyweight kickboxers (10). Logistic regression identified several key structures significantly influencing the odds of victory or defeat. The model demonstrated strong discriminative power (AUC = 0.84), confirming its practical applicability. The rear-leg low roundhouse kick to the opponent's lead leg showed a strong positive association with winning. Each successful execution of this technique increased the odds of victory by 12.6%, which highlights the importance of low kicks in modern K1 kickboxing, not only as scoring techniques but also for impairing the opponent's mobility throughout a fight (2, 8). Similarly, precise punches, particularly the lead-hand hook to the head and the rear-hand uppercut to the head, proved to be strong success predictors. Each additional execution of these techniques increased the odds of winning by 17.3% and 23%, respectively. These findings suggest the need to develop advanced hand techniques and apply them effectively during critical moments of a fight (14, 15). Conversely, an increased number of hand blocks against punches and hand blocks against leg strikes was associated with decreased odds of victory by 4.2% and 10.4%, respectively. This suggests that excessive reliance on defensive strategies, particularly passive blocking, may be counterproductive. Such patterns may reflect a lack of initiative or tactical dominance, allowing the opponent to control the rhythm and flow of the fight. However, the negative association between frequent blocking and winning should be interpreted as associative rather than causal. High frequencies of hand blocks might reflect reactive behavior under sustained pressure, indicating that a fighter is in a disadvantageous position and forced into a defensive role. Thus, frequent blocking may be a consequence of being dominated rather than a direct cause of losing, highlighting the importance of maintaining offensive initiative to control the rhythm of the bout. These results are consistent with research emphasizing the importance of offensive engagement and tempo control (16). Among anthropometric characteristics, height emerged as a significant predictor of success. Each additional centimeter increased the odds of winning by 9.8%. Increase in winning odds per additional centimetre of height should, however, be interpreted with caution. Greater height is typically accompanied by longer reach, different body mass distribution, and potentially distinct tactical preferences, and those advantages are demonstrated in striking martial arts (1, 17, 18). Taller fighters may more easily maintain distance and exploit longer reach in both offense and defense. Nevertheless, it is important to note that technical proficiency can compensate for the lack of height, underscoring the importance of developing key TE-TA structures.

These findings have several practical implications for training. Modeling a successful kickboxer must include structures that reflect technical-tactical abilities, as well as training load that affects biomechanical and motor performance (19). Coaches and fighters should prioritize the development of effective low roundhouse kicks, precise head-level punches, and combinations involving these key techniques. Training should also emphasize proactive defense and minimizing passive blocking, with a strong focus on movement and positioning. While height may be an advantage in fighter selection, it should not overshadow the critical importance of TE-TA preparation. However, the retrospective design and the restriction of the sample to K1 rules and the absolute weight class can be defined as limitations that can shorten the generalizability of these findings to other combat sports and weight divisions. Despite the historical nature of the dataset, the analyzed bouts capture competitive patterns that are fundamental in contemporary K1 kickboxing, especially regarding fundamental TE-TA structures.

Although measurement reliability was confirmed through high ICC and Pearson correlations, the subjective nature of TE-TA annotation cannot be eliminated. Future research should incorporate larger and more diverse samples, automated match analysis tools, and explore the influence of psychosomatic and broader anthropological factors on match outcomes.

In summary, this study supports practical guidelines for training and tactical planning in elite K1 kickboxing, laying the groundwork for more advanced predictive modeling in combat sports.

Although the present findings are derived from elite-level competition, the applied modeling approach has relevance for understanding performance indicators across different stages of athlete development. Future research should extend such frameworks to longitudinal and comparative designs in order to examine how technical–tactical efficiency and anthropometric characteristics interact over time within developmental and talent development pathways.

The proposed model provides valuable insight into the technical-tactical structures and anthropometric characteristics of top-level K1 fighters. This study is distinguished by its analysis of a wide range of variables and the development of a new predictive performance model that influences match outcomes in K1 kickboxing. The model's high discriminative capacity (AUC = 0.84) confirms its reliability for predicting match outcomes and guiding the development of optimal fight strategies.

Notably, the model identifies a specific combination of techniques as low roundhouse kicks and precise head-level punches, as critical to success, while simultaneously highlighting the potential disadvantages of an excessively defensive approach.

The accumulation of successful techniques, such as a low rear-leg roundhouse kick to the opponent's lead leg, a lead-hand hook to the head, and a rear-hand uppercut to the head, along with greater height, significantly increased the odds of victory. Conversely, techniques like hand block against punches and hand block against leg strikes were associated with a greater risk of defeat. These findings can be used to adjust training and competitive strategies, allowing coaches and athletes to base their decisions on clearly defined predictors of success. This mathematical model can serve as a practical tool for this purpose.

This study possesses several notable strengths, including the analysis of a relatively large sample of elite K1 bouts, the development of a novel predictive model with high discriminative power (AUC = 0.84), and the methodological innovation of dichotomous coding for strike effectiveness. The manual notational analysis, conducted with good inter-rater reliability, allowed for a granular examination of TE-TA structures. However, certain limitations must be acknowledged. The retrospective design and restriction to the K1 super heavyweight division limit the generalizability of findings to other rule sets, weight classes, or female athletes. The model's predictive accuracy (AUC = 0.84) was evaluated on the same dataset used for model development, without external or internal validation. Future studies should employ cross-validation techniques or independent test datasets to confirm generalizability. Although calculated reliability measures were strong, the inherent subjectivity of manual video analysis remains a constraint. Failure to account for within-fighter correlation may have resulted in underestimated standard errors and overstated statistical significance. While the direction and magnitude of effects are theoretically consistent with prior literature, the precision of estimates should be interpreted with caution. Among anthropometric characteristics, height emerged as a significant he analyzed TE-TA structures remain the fundamental core of contemporary K1 kickboxing. While tactical nuances and pacing may have evolved, the underlying biomechanical and strategic principles of effective striking and anthropometric advantages likely retain their relevance. Furthermore, while the model identifies associative predictors, the causal mechanisms linking frequent blocking to losing require further investigation (e.g., is blocking a cause or a consequence of being dominated?). Future research should employ prospective designs, incorporate automated tracking technologies, and integrate psychological and physiological variables to build more comprehensive performance models.