Abstract

Purpose: This study investigated the predictive capacity of pre-competition emotions, post-competition emotional states, and emotional changes (delta) in determining competitive outcomes.Methods: Thirty-seven athletes (27 male, 10 female; age =24.4±7.3 years) completed the Portuguese version of Sport Emotion Questionnaire (SEQ) before and after a competition. Five emotions (anxiety, sadness, anger, excitement, joy) were analyzed through descriptive statistics, parametricnon-parametric comparisons (α=0.05), effect sizes (Cohen’s d’r’), and machine learning (logistic regressionRandom Forest with 5-fold cross-validation).Results: Pre-competition emotions showed no significant differences between winners (n=18) and losers (n=19), though winners exhibited higher excitement (5.4±4.2 vs. 2.5±6.4; d’=0.55) and joy (5.9±4.9 vs. 3.4±7.0; d’=0.41), with anxiety identified as the top machine learning predictor (importance: 0.278). Post-competition, winners displayed significantly elevated levels across all emotions (p<0.001), particularly excitement (7.4±4.8 vs. −1.2±7.2; d’=1.4) and joy (8.8±4.5 vs. −1.1±8.0; r=0.64). Analysis of emotional deltas revealed winners’ significant increases in excitement (+2.00; p=0.018) and joy (+2.94; p=0.009), versus losers’ pronounced reductions in sadness (−6.89; p=0.002) and anger (−5.53; p=0.003). A predictive model using deltas achieved 78.2% accuracy, with delta joy as the strongest predictor (OR=3.62).Conclusions: Emotional states after competition and their dynamic changes, characterized by amplification in winners and suppression in losers are stronger outcome determinants than pre-competition emotions. Interventions targeting adaptive emotion regulation may enhance athletic performance.