HOW TO PREDICT GYMNASTICS’ RESULTS - A SIMPLE METHOD BASED ON THE 2022 EUROPEAN CHAMPIONSHIPS

Abstract

It is very important for coaches in charge of selection to know the competition level based on the athletes registered, as well as to assess the gymnasts’ capacity to achieve the expected performance. In international artistic gymnastics competitions, access to the all-around (C2), apparatus (C3), and team (C4) finals is determined based on the results of the qualification round (C1). A specialist who competes on only one or two apparatuses may have a chance to win medals on those apparatuses but may also be a disadvantage to the team score. Based on the nominative entries submitted by each competing nation one month prior to the 2022 European Championships held in Munich, we extracted previous individual scores from national and international competitions over a period of six months before the 2022 EC. Individual standard deviations were computed for each competing athlete, and their mean and maximal scores were used as two predictive outcomes. The individual standard deviation on still rings showed a low Q1, median, Q3, and IQR for both junior and senior MAG, while pommel horse exhibited greater variability. Among WAG athletes, those competing in junior and senior vault and floor had lower Q1, median, Q3, and IQR compared to uneven bars and balance beam. When comparing both predictions (mean and maximal values) to the scores obtained in Munich, small differences were observed in terms of random errors. However, a large systematic error indicated a significant overestimation in the maximal prediction, explained by the fact that athletes are not always able to perform at their best. The percentage of gymnasts with a score less than 0.30 points away from the mean prediction ranged from 28.1% (senior floor) to 58.9% (junior still rings) for MAG and from 23.9% (junior balance beam) to 70.7% (senior vault) for WAG athletes. The simple method outlined in this article provides information about the method’s reliability on different apparatuses, as well as insight into the accuracy of score predictions across the apparatuses.