Proximal Power Gains in Football: Position-Specific Adaptations Revealed Through Compositional Data Analysis

Proximal Power Gains in Football: Position-Specific Adaptations Revealed Through Compositional Data Analysis

Alvin Sherman Library

Training-induced adaptations in joint power distribution during speed running may enhance performance, but how these changes manifest across different American football position groups remains poorly understood. We investigated the influence of training and player position on joint power composition among players training for the National Football League draft. Forty-eight draft-eligible players performed pre- and post-training camp running trials at 23.4 km/h. We used motion capture software and MATLAB to calculate bilateral peak hip, knee, and ankle joint power. Compositional data analysis (CoDA) was used, in which positive/generated and negative/absorbed power compositions were expressed as isometric log-rations: hip versus knee + ankle (ILR1) and knee versus ankle (ILR2). Linear mixed model analyses assessed the influence of training and position on positive and negative ILR1 and ILR2. Training induced a significant shift toward proximal power generation (increased ILR1) in both limbs (left: p = .004; right: p = .009), regardless of position. Skill consistently exhibited more distal strategies than Big (p = .031). For negative power, a significant group-by-time interaction was observed for left limb ILR1 (p = .037), with Big showing a proximal shift post-training (Δ = 0.303, p = .011). ILR2 values revealed stable knee-to-ankle power rations over time, though Skill demonstrated more distal strategies than Big (p < .001). Training resulted in a proximal shift in joint power distribution, especially in Big players, suggesting enhanced propulsive capacity and eccentric control. CoDA enabled statistically interpretable modeling of joint power redistribution, showing position-specific adaptations that traditional percentage-based methods may obscure.