To elect the winner in amateur boxing, five expert judges observe the boxers to assign "correct hits". These hits are then processed by the automated Scoring Machine System (SMS) to assign scores to each boxer. Because it is based on quick human observation under pressure, this system is prone to error. The purpose of this investigation was to identify the nature of this error defined either as false positive (point assigned to an incorrect hit) or false negative (point not assigned to a correct hit). We analysed videos (25 fps) of 10 boxing bouts performed during the 2008 Beijing Olympic Games using a specialized software for boxing match analysis (www.dfanalysis.com). Slow motion and replay modes were used by an expert in boxing match analysis to establish true correct hits based on the Amateur International Boxing Association`s guidelines. True correct hits were then compared to "correct hits" assigned by the SMS to identify false positives and false negatives. On average, 11.1 ± 3.3 "correct hits" (mean ± SD) were assigned by the SMS. This score was significantly lower than the number of true correct hits established with boxing match analysis (22.9 ± 9.3, P<0.01). Comparison of SMS and boxing match analysis revealed 5.1 ± 1.4 false positives and 16.9 ± 8.4 false negatives. In conclusion, boxing match analysis revealed that SMS significantly underestimates the number of true correct hits because of a high number of false negatives. New training methods to improve judges` ability to assign true correct hits and/or changes in the SMS threshold to assign a score should be tested. Video vom Vortrag unter http://www.youtube.com/watch?v=gS8yXXGMQK4 und http://www.youtube.com/watch?v=EvkK3E0p5Fg
© Copyright 2012 World Congress of Performance Analysis of Sport IX. Veröffentlicht von University of Worcester. Alle Rechte vorbehalten.