The unpredictable talent selection in youth beach handball players.

This study identified key variables to differentiate selected and non-selected Brazilian young beach handball players using non-linear approaches. A total of 64 players (16.5 ± 0.8 years-old; 33 male) participated in the Brazilian National Team preparatory stage for the Beach Handball World Championships (2017), and were grouped in selected and non-selected players. Participants were assessed for anthropometrics and body composition, countermovement jump, and ball speed throwing. Comparisons between groups were tested, and the outcomes were analyzed using logistic regression (SPSS 25.0; p < 0.05). Network analysis was used to establish systemic non-linear interrelationships between variables, and the Expected Influence was calculated (Rstudio). For male players, increments of 1 kg in fat-free mass and 1 m/s in ball speed, increased the odds of being selected by 44% and 17 times, respectively. For females, 1 cm increase in jump height and in palm diameter, increased the odds of being selected by 29.5%, and 6.7 times, respectively. The non-linear approach showed body mass as the expected influence variable for both selected and non-selected male and female players, differently of those variables observed using the logistic regression. The results reflect the multidimensional predictors of talent in beach handball, and represent important finding for coaches and stakeholders.