Objectivity and reliability of the Judo Attack System Software

Abstract

In judo matches, athletes perform different technical-tactical actions according to age, sex, and competitive level. To identify these differences, the Judo Attack System Software (JASS) was developed specifically to assess the combat attack system, consisting of the tori (athlete executing the action) and uke (athlete receiving the action) approach, tori’s grip, final displacement, stance (laterality) and direction of attack. To assess the reliability of the software, the Robotic Process Automation (RPA) method was used and resulted in 100% of correct records. To verify the objectivity of the analyses, 40 international matches were used to determine inter-rater and intra-rater agreement. The Intraclass Correlation Coefficient (ICC) demonstrated inter-rater and intra-rater agreement between 0.89 (good) and 1 (excellent) for total attacks, scoring attacks and no-scoring attacks. For the 39 variables analyzed, Cohen’s Kappa index (κ) showed that 74.3% obtained classification almost perfect, 20.5% substantial, 2.6% fair and 2.6% poor for the intra-rater analysis. In the inter-rater analysis 15.4% had an almost perfect classification, 41% substantial, 25.6% moderate, 12.8% fair, 2.6% slight, and 2.6% poor. The Judo Attack System Software showed good objectivity inter-rater and intra-rater in most of the proposed measures, becoming an auxiliary tool to register the attack systems in judo combat.