Kinematic variables influencing somersault turn time in medley swimmers using artificial neural networks (ANN)

Introduction: Efficient performance in swimming turns is a critical component of competitive success, especially in sprint and middle distance events where turn execution can significantly affect race outcomes. Objective: Turn Time in Medley Swimmers Using Artificial Neural Networks (ANN), Examine selected kinematic and kinetic variables related to joint angles, propulsion characteristics, and centre of mass dynamics during the somersault turn in competitive youth medley swimmers. Analyze the relationships between these biomechanical variables and somersault turn time using Pearson’s correlation. Determine the relative and normalized importance of the investigated kinematic variables in predicting turn time through the application of an Artificial Neural Network (ANN) model. Identify the key biomechanical determinants of turn performance, thereby providing evidence-based insights for technique optimization and training design. Methodology: A descriptive analytical design with a biomechanical approach was employed to investigate the kinematic variables influencing somersault turn time in competitive youth male swimmers competitive medley swimmers aged 14–15 years from Smouha Sporting Club (Egypt) participated in the study. All swimmers were officially registered with the Egyptian Swimming Federation, trained regularly without prolonged interruptions, demonstrated technical proficiency in the investigated turn technique, and showed a high competitive performance level relative to their peers. Discussions: Findings indicate that turn performance is governed by a complex nonlinear interaction among lower-limb joint angles, trunk positioning, and propulsion-related variables, rather than a single kinematic factor. Conclusion: This supports the established understanding that swimming turns are highly coordinated, multi-phase movements in which performance emerges from the interaction of interdependent components.