The fastest bowlers and pitchers, who are able to perform at levels that exceed those of their team-mates, should have a combination of optimal physical characteristics and techniques. The relationship between ball speed and anthropometrical characteristics of 48 professional baseball pitchers was investigated using a hybrid evolutionary algorithm (HEA). The average height and weight of the subjects were 191.4±4.7 cm and 96.2+8.2 kg, respectively. Their mean brachial index, BI, was 74.9±4.1. Performing our HEA analysis led us to discover the best rule sets between different inputs (height, weight, segment lengths) resulting in models relating ball speed to a pseudo-ectomorphic parameter. The robustness of the results was achieved by implementing a parsimonious strategy based on the Akaike criterion as the fitness function. Furthermore, the rule set resulted to be almost always weight greater than constant and after selection of the best pitchers (the lighter ones) the related distribution of speed as a function of BI was shown to be bimodal. These results could be used for talent identification and coaching technique improvement. Our finding could also have some relevance for bio-inspired robotics.
© Copyright 2007 International Journal of Computer Science in Sport. Sciendo. Alle Rechte vorbehalten.