A pilot study on the validity and fidelity of two virtual reality systems for tennis training across different skill levels

By enabling the manipulation of environmental constraints, including ball speed, trajectories, and court conditions, current virtual reality (VR) tennis training systems offer new opportunities to practice specific skills under controlled conditions. However, these systems remain poorly validated and their effectiveness in improving performance is unclear. Therefore, we examined the consistency between biomechanical variables, success rate, mental workload, presence, and acceptance across real-world and virtual environments implemented using head-mounted display systems with two devices: dual game controllers (Controllers VR) and a single haptic racket (Racket VR). Sixteen tennis players were divided into beginner and expert groups and performed forehand and backhand strokes in each environment. In the real-world condition, a principal component analysis of forehand biomechanical variables revealed a first factorial axis characterized by higher linear and angular velocities. This kinematic pattern is associated with higher success rates ( r = 0.46) and lower overall mental workload ( r = − 0.49). In VR, scores on this axis were systematically lower, suggesting altered biomechanical execution and increased cognitive demands. These effects were partially mitigated in Racket VR, but were more pronounced in expert players. Presence scores in VR were moderate-to-high, suggesting that players may experience a strong sense of presence even when motor behavior is altered toward pantomime-like movements. Overall, this pilot study suggests that the effectiveness of commercial VR solutions for tennis training may be limited, particularly in experts, and highlights the need for longitudinal studies to assess skill transfer from VR training to on-court tennis performance.