The purpose of the current investigation was to validate an automatic path change feature developed for the ProZone3® system using a case study where a single player was filmed during 97 minutes actual match play. The player`s path changes were analysed in detail by a human observation process and compared to the path changes determined by ProZone3®. A secondary aim was to validate the times when the player entered different areas of the pitch that were recorded by ProZone3®. There was a good strength of inter-observer agreement for the human observation process for identifying path changes (ê = 0.71) and for the pitch areas entered by the player (97.0% of areas entered were agreed within 1s). There was a moderate strength of agreement between the human observers and ProZone3® for path changes (ê = 0.41) with many of the disagreements being expected given the limitations of human observation of movement. There was a good strength of agreement between the human observers and ProZone3® for the pitch areas entered (between 95.7% and 96.7% of areas entered being agreed) with almost all of the disagreements being where a player was within 1m of the boarder between 2 pitch areas.
© Copyright 2009 International Journal of Computer Science in Sport. Sciendo. All rights reserved.