sports biomechanics
...MLP) neural network. The MLP neural network is an artificial neural network and is useful for simulating flight distances of javelins, for specific release parameters. The MLP model allows for the simulated flight distances to be obtained from the input of given release parameters. Release parameters were estimated using three- dimensional analysis of video data The neural network model was composed of the following input parameters: The angle of release, angle of attack, the side angle of attack and the velocity of release. From the input of combinations of release parameters, maps can be generated to predict the distances reached from a throw. The networks proved to be successful in predicting actual flights of javelins to within 5%, showing there to be a mean difference of 2.5% between the model and real throws. The study found the most significant parameter to be the release velocity As release angles Were thought to have had a limited impact on the distances reached. It was also acknowledged that a moderate side angle of attack is favoured to generate Maximum flight distances,as this acts to compensate for the javelins rotation on release. It was concluded that ideal release velocities lied between 27-28ms` and should have an angle of attack between 1 and 3 °C larger than the angle of release. Thus for higher velocities the angle of attack should ...