Probability of winning a badminton game: the relationship between unforced errors and winner points

The aim of this study was to establish the probability of winning a game of badminton based on the number of unforced errors and winner points. The variables were quantified in 56 male matches of the 2012 Brazilian Championship. The binomial logistic regression was used to calculate the probability of victory based on the number of unforced error and winner point commited. The number of unforced errors in defeat was significantly higher compared to victory, and the winner points in victory was significantly higher than in defeat (p<0.01). We found that when the number of unforced errors becomes greater than the winner points, the winning probability decreases significantly. The model was able to classify correctly 75.8% of the results and explained 27.4% of the variance of the game outcome. The results of this study provide important information for coaches and athletes to readjust the game strategies based on the athlete's performance during the match.