IDO MOVEMENT FOR CULTURE

Background. Since 2016, at the University of Montpellier, 3rd-year students (judo specialty) have been analyzing judo combats in a multicriteria way to optimize their coaching, refereeing, teaching, and journalism.
Problem. One of their works consisted in analyzing on a computer a delayed real-time combat (without pause, nor visual or auditory signal) to observe one judoka. The task was to estimate the force ratio between the two judokas to determine if the judoka observed was, by clicking (annotating) next to the video in an orthonormal graph composed of 4 sectors: (dominant-programmed [S1]; programmed-dominated [S2]; dominated-automatic [S3]; automatic-dominant [S4]). The more important the criteria were considered; the more clicks (annotations) were placed on the periphery of the axes.
Aim. To confirm and improve our previous results on this subject: (i) to analyze a recorded video sequence excluding video backtracking (as if they were live, as a coach or a referee), (ii) to identify the reversal of force ratio, (iii) to make repeatable and homogeneous observations.
Method. After 40 min of training, 12 students analyzed 4 times the same combat presenting a reversal of force ratio: the judoka to be observed was first dominant-programmed (many clicks were expected in S1), then was dominated-automatic and lost the fight (many clicks were expected in S3). These students were aged 20.4 ± 1.9 years, brown belt to black belt 2nd dan, practicing 4h to 12h of judo weekly, had followed courses of judo specialty (90h), ICT (18h) and video analysis (10h). Shapiro-Wilk, Kruskal-Wallis and Kolmogorov-Smirnov tests were used to analyze the data collected. A significance level was set at 5%.
Results. The 4 observations were identical and homogeneous. No statistically significant difference in: (i) duration times for the 4 observations (Kruskal-Wallis’s test: p=0.919); (ii) number of clicks per observation (Kruskal-Wallis’s test: p=0.715); (iii) the distributions of the coordinates X and Y of the clicks per bin in the four observations (Kruskal-Wallis’s test: p=0.942 and p=0.968). The percentages of clicks per sector according to time: [S1]=40.2%; [S2]=16.5%; [S3]=32.3%; [S4]=11%) and the evolution of the number of clicks in S1 show the reversal of force ratio. There is a significant difference in the comparison of the distributions of collected and random coordinates (Kruskal-Wallis’s test: p=0.699 and p=0.173 respectively for X and Y coordinates).
Conclusions. These results do not appear to be due to chance. The identification of the reversal of force ratio in combat, the homogeneity and repeatability of the analyses confirm our hypotheses and the previous study in 2019. The transferability of these skills to real-life situations must be verified.