Judo is a Japanese martial art and an Olympic and Paralympic combat sport, introduced at the Tokyo Olympic Games in 1964 and the Seoul Olympic Games in 1988. Competitors obtain victory by throwing their opponents on their back from the standing position or by causing submission by neck choke, elbow armlock, or immobilization in ground combat (1). Contest involves open and complex skills, presents irregular intervals of effort and pause, and is characterized as an intermittent activity (2). The physiologic demands and employed metabolic mechanisms differ substantially between the dynamic and calmer phases of the contest (3), leading to specific physiologic responses with specific implications for training and conditioning and for tactical preparation (4). Several previous studies [e.g., reviewed in Barreto et al. (5)] have analyzed the number of high-activity events or the duration of high-activity phases vs. calmer periods of contest. Such studies have generally been based on assessments of video materials with the subjective classifications of phases into high or low-activity classes. From these types of studies of any sport, several interesting concepts for training and competition have emerged. One of them is the “activity profile” (6–9). Overall, it refers to the individualized performance profiles of athletes through the analysis of the movement pattern that they express during competition.

Accelerometers can detect and record the body accelerations (BA) for motion along three orthogonal axes (vertical, mediolateral, and anteroposterior) in arbitrary acceleration units (a.u. or counts) (10). In this way, these devices provide a proxy measure of the mechanical work done by a moving body and wearable versions have been widely used to measure physical activity in clinical/laboratory and free-living environments (11). In combat sports, Kirk et al. (12) and Del Vecchio et al. (13) used wearable accelerometers to evaluate BA in a simulated mixed martial arts (MMA) contest and in several physical activities in a taekwondo training session (e.g., combat simulation, etc.), respectively. In addition, accelerometry has also been used in judo in the topics of injury prevention [e.g., (14–17),] and technique training [e.g. (18,19),]. However, as yet, there have been no studies quantifying in-contest BAs for judo athletes. This topic represents a potential area of exploration that would provide a valuable resource to objectively determine numerous dynamic aspects of this sport and provide a wealth of unknown information concerning the mechanical parameters of judo contestants that, alongside physiological data, might be incorporated into training programs. Of course, judo competitions are organized based on sex and weight categories, and competitors across distinct weight classes exhibit variations in technical and tactical aspects, as well as in physiology, performance, and body composition (20–22). Thus, to ensure that the fullest, most detailed picture is obtained, and for the BA data to be applied accurately, the quantification in-contest BAs for judo athletes must also be specified according to these categories.

Thus, the goals of this study were to determine the BÁs profile of the male and female weight divisions in a judo contest and to ascertain the involvement of the vertical, mediolateral, and anteroposterior axes in it, in a sample of national and international level judo athletes. It was hypothesized that two different BA profiles of the judo contest would be identified: the male and female athletes of the light and middle weight divisions and the male and female ones of the heavy weight divisions. It was also expected that athletes of all weight divisions would mainly express their BA during the contest in the anteroposterior and the vertical axes, with the BA of the anteroposterior axis being the significantly predominant one.

With respect to the BA, a significant sex-by-weight interaction was found (Table 2), with the highest BA score found for the ML category. The highest BA value for women was recorded for the FL category, this being slightly higher than the lowest male value. Large and statistically significant differences were found between BAs for ML and MH (p < 0.001, r = 0.73, moderate effect), MM and MH (p < 0.001, r = 0.73, moderate effect), FL and FH (p < 0.001, r = 0.68, moderate effect), and FM and FH (p = 0.003, r = 0.61, moderate effect) (Table 2, Figure 1).

Regarding the involvement of the three axes, a significant sex-by-weight interaction was observed for the anteroposterior axis (Table 2). Athletes of all weight divisions mainly performed their BA in the anteroposterior and the mediolateral axes, with athletes of both sexes in the light and middle weight divisions having higher BA values than those in the heavy division (Table 2, Figure 2). Differences among weight divisions in the three axes were also observed. However, the one-sample Pearson's chi-squared (χ²) test did not show any of the axes to be significantly predominant in any weight division during the contest (p = 0.400, p = 0.631, p = 0.844, p = 0.749, p = 0.644, and p = 0.895, for ML, MM, MH, FL, FM, and FH, respectively). The Monte Carlo method suggested the scenarios with BA involving all axes as the most likely, with a slight preference of the anteroposterior and mediolateral ones, across all weight divisions for both sexes (Table 3). Thus, scenarios in which BA shows a clearly dominant axis appear to have only a marginal probability of occurrence (Table 3).

Concerning physiological variables, the HR at the end of the contest did not manifest a significant difference among the weight divisions, the % HRmax at the end of the contest presented one significant difference between the ML and FH (p < 0.001, r = 0.59, moderate effect), the BLa 1 min after the end of the contest also showed several significant differences (MM vs. FM, p = 0.043, r = 0.46, moderate effect; MM vs. FH, p = 0.009, r = 0.55, moderate effect; MH vs. FL, p < 0.001, r = 0.73, moderate effect; MH vs. FH, p < 0.008, r = 0.56, moderate effect), and the RPE at the end of the contest did not display significant differences (Table 2).

The goals of this study were to determine the BA profile of the judo contest of the male and female weight divisions and to ascertain the involvement of the vertical, mediolateral and anteroposterior axes in it. Results allow the identification of three BA profiles in judo; one for ML and MM, another one for FL and FM and the third one for MH and FH. In relation to the axes' involvement, findings revealed that athletes of all weight divisions tend not to favor any particular axis during the contest, since all of them are almost equally involved. The Monte Carlo method simulations suggested as the most likely scenarios those with BA involving all axes, with a slight preference of the anteroposterior and mediolateral ones.

The hypothesis that two different BA profiles of the judo contest would be identified was not confirmed; three different BA profiles were detected. To date, no study has determined the BA of the judo male and female contests. Kirk et al. (12) determined the external load experienced by the MMA athletes during a simulated MMA contest comprising three rounds of 5 min with 1 min rest between rounds, from the magnitudes of accelerations in the mediolateral, vertical and anteroposterior axes (224.32 ± 26.59 a.u./counts). External load or load are the concepts used in Sports Sciences to refer to the BA (reported as a.u./counts) of a specific physical activity. In the study of Kirk et al. (8), athletes wore Minimax X3, triaxial accelerometers secured to the torso using in a neoprene harness over the T3–4 vertebrae (the epoch configuration was not specified). Del Vecchio et al. (13) determined the BA of a sample of non-adult taekwondo athletes (expressed as VM) in several physical activities, in a taekwondo training session (e.g., combat simulation, etc.). They used an ActiGraph wGT3X+ device that was located on at the waist of the athletes, with a 5-epoch configuration and with the Freedson's algorithm for children (33) to process data and their result in the contest simulation was a BA of 530.44 ± 240.42 a.u./cpm. Given the underlying mechanisms driving the differences of BA between light, middleweight, and heavy divisions and between sexes found in the present study, they remain unknown. Thus, further research is needed to ascertain the factors contributing to these phenomena. However, these results suggest firstly, that weight affects the dynamics of the contest and secondly, that the way it influences the dynamics of the contest varies between sexes. A study by Sterkowicz-Przybycień et al. (21) with elite judo athletes, found significant differences in combat-phase times dependent on weight category and sex divisions. Additionally, in a study of male and female judo cadets, Miarka et al. (34) identified sex differences in the frequencies and timings of judo combat actions, types of techniques used, and numbers of penalties received. Thus, it seems likely that the strategic priorities in judo differ across weight divisions and between the sexes and this goes some way to explaining the results of the present study.

About the contribution of three axes in the judo contests, the hypothesis that athletes of all weight divisions would mainly express their BA during the contest in the anteroposterior and the vertical axes, being the BA of the anteroposterior axis the significantly predominant one, was not confirmed; athletes of all sexes and weight divisions produced BA on all axes equally. This finding should be considered in the design of training programs. To the authors' knowledge, the current literature contains no studies examining the issue of BA directional demands in judo contests. Santos et al. (36) evaluated the effects of judo contest on the athletes' postural control and physiological loading before, during and after a simulated contest. They observed the greatest effects manifested on the anteroposterior axis, which is not in accordance with this study. Here again, the core processes behind these outcomes remain undiscovered; therefore, additional investigation is necessary to identity their determinants. Regarding the results of the study of Santos et al. (35) and those of the Monte Carlo method of this study, the dominant trend in the contest could be that the anteroposterior axis could have a slightly higher participation than the rest axes. Nevertheless, this tendency might vary during specific contest moments (e.g., when an athlete is guarding an advantage, emphasizing the mediolateral axis).

All physiological variables examined here are in line with results from previous studies. This suggests that the athletes who participated in this study did indeed perform to their fullest abilities during the study's contests and thus, the analysis of these contests provides a useful contribution to the scientific literature. For instance, the maximum HR measured during the contests in this work were between 170 ± 9 and 182 ± 5 b/min (88.4 ± 5.2 and 94.7 ± 2.9% of the HRmax), which compares favorably with values measured by Sbriccoli et al. (36) in senior male and female judo athletes, respectively, in simulated contests [181 ± 11 b/min (94.8% of HRmax) and 176 ± 6 (92.87% of HRmax)]. Similarly, the BLa values seen in this study (from 8.9 ± 1.8 to 14.3 ± 4.1 mmol/L) are also in agreement with previous research which, after simulated judo contests are reported as being from 7.5 ± 5.1 to 18.1 ± 4.4 mmol/L (37, 38). Lastly, end-of-contest RPE, which in this study is recorded as from 15.9 ± 2.1 to 17.8 ± 1.5 a.u. is also in the same range as the value for this variable seen in other research. For example, Chtourou et al. (39) obtained values between 14 and 14.5 a.u. for regional and national level athletes after a 5-min simulated judo contest; Stavrinou et al. (40) reported 19.1 ± 0.78 a.u. in cadets after the last of a set of four 4-min contests; and Santos et al. (35) recorded 16.1 ± 0.4 a.u. for regional and national athletes after a 7-min simulated judo contest.

This study suggests that demands on judo athletes in a contest, measured in terms of body accelerations, differ by sex and weight classes with three different profiles; light and middle weight males, light and middle weight females, and heavy weight for both. Hence, professionals should tailor the training regimens to accommodate these distinct profiles within the weight and gender classes. In addition, this study also suggests that judo contest motion data spans equally along the vertical, mediolateral and anteroposterior axes, with the most probably scenario of a slight preference of the anteroposterior and mediolateral axes. Thus, professionals should design training tasks which could enclose all these situations to better prepare athletes for the dynamics of contests.