Nowadays, many tracking systems in football provide positional data of players but only a few systems provide reliable data of the ball itself. The tracking quality of many available systems suffers from high ball velocities of up to 34 ms-1 and from the occlusion of both the players and the ball. Knowledge about the position and velocity of the football can yield valuable information for players, coaches and the media. To assess the accuracy of the football's velocity provided by the radio-based tracking system RedFIR, we used a ball shooting machine to repeatedly simulate realistic situations at different velocities ranging from 7.9 ms-1 to 22.3 ms-1 in an indoor environment. We then compared velocity estimates for 50 shots at five speed levels with ground truth values derived from light gates by way of mean percentage error (MPE) and Bland-Altman analysis. The speed values had an MPE of 2.6% (-0.49 ms-1)

No abstract