Exploratory machine learning in high-level Jiu-Jitsu athletes suggests a review of categories and their rules based on anthropometric and handgrip strength data

Abstract

Jiu-jitsu is the basis of mixed martial arts. In competitions, athletes are separated by age, gender, weight, and rank. Athletes promote successive gripping movements, demonstrating the importance of handgrip strength (HGS) in this modality. The objective of the present study was to evaluate whether, by considering HGS, the competitive categories established in jiu-jitsu are well divided. This is a cross-sectional, descriptive, and observational study. The sample consisted of 206 competing jiu-jitsu athletes. Anthropometric and HGS assessments were performed, along with descriptive statistics of the sample characteristics. Dissimilarity measures between observations of the study variable HGS were calculated, along with the effect size and Z score. Observing the grouping by weight categories, featherweight, and lightweight categories were dissimilar to the others, suggesting that, taking HGS as a basis, it makes sense to separate the athletes into two categories for absolute, one for featherweight and lightweight and another for the other categories. By age, there was only a similarity between Master 2 and 3 categories, suggesting a division of the absolute into three age categories. When investigated according to body mass index, there is a similarity between all categories. The HGS of jiu-jitsu athletes represents a significant difference and potential for athletes' performance. The present study suggests a review of weight and age categories in jiu-jitsu competitions, favoring competitions with more homogeneous categories and greater competitiveness and broad competition without discrepancies in the strength of its practitioners.