Abstract - Predicting taekwondo winners in high-level competition using ranking scores and country performance scores: an analysis of the 2019 World Taekwondo Championship
Background. The ranking system can be considered a measure of athletes’ performance. In addition, countries with high performance are expected to have better athletes.
Problem and Aim. The present study aimed to verify whether ranking scores and country performance by competitors differ between winners and losers; and whether ranking scores and country performance by competitors could predict winners and losers in a high-level taekwondo competition.
Methods. A total of 759 combats at the 2019 World Taekwondo (WT) Championship were analyzed in the present study. We used as variables the ranking scores and country performance score (CPS) calculated based on the number of Olympic medals won by a given country. The Bayesian t-test was used to estimate the difference in the rank score and CPS between winners and losers, and Support Vector Machine was used to predict the winners.
Results. In summary, the results showed that both the ranking scores and CPS of losing athletes were lower than winners. Moreover the ranking score and CPS of the athlete and their opponent predicted the winners with a moderate degree of accuracy (higher than 0.70) in both the grouped sample and females separated out, but had low accuracy (0.69) in male athletes.
Conclusions. Considering the complexity of a Taekwondo competition, it is possible to interpret that the ranking scores and the CPS of the competitors are important variables in predicting winners in high-level taekwondo competition.