Impellizzeri FM¹, Sassi A¹, Rampinini E¹

1, Human Performance Lab, S.S. MAPEI, Castellanza, Italy

Introduction Commercial semi-automated, video-computerized, image recognition match analysis systems can now supply professional soccer teams with match analysis data on all players just a few hours after the end of the game. Despite the popularity of these systems among professional soccer teams and researchers, no data are available about their validity. Therefore, the aim of this study was to provide preliminary evidence on the accuracy and reliability of one of these commercial match analysis systems (SICS, Bassano del Grappa, Italy). Methods Both radar and video analysis data were collected on a soccer player while performing an intense match phase simulation (350 m) including sprinting, striding and jogging in 4 different zones of the pitch. In a 5th pitch zone, 6 maximal 30-m sprint were also completed. Players velocities measured with the video-analysis system were compared with those measured using a radar system (Stalker, ATS). Data were filtered using a zero lag Butterworth filter.  To avoid the edge effect due to filtering algorithm, the first and last 10% of the data were excluded from the computation of TE. Data from video analysis were tracked twice by two different operators. Accuracy and reliability for low intensity running (<15 km/h, LIR), high intensity running (>15km/h, HIR) and total distance (TD) of the intense match phase simulation were determined using the typical error (TE) expressed as coefficient of variation (Bangsbo, Norregaard et al. 1991; Hopkins 2000). Similarly, TE was calculated for sprinting (>19 km/h) and peak running speed (PS) during the 30-m sprints. The mean TE of the instantaneous velocities measured using the video-match analysis system taken with respect to the radar velocities was also calculated as indicator of overall velocity accuracy. Results During intense match phase simulation accuracy for LIR was 2.1 and 3.1%, for HIR was 3.6 and 5.6% and for TD resulted 1.0 and 1.5%, for the first and second tracking respectively. Reliability was 1.0% for LIR, 3.2% for HIR and 1.0% for TD. During sprinting accuracy was 1.8 and 1.7% for sprinting distance, and 1.3 and 0.9% for PS, for the first and second tracking respectively. Reliability was 3.2% for sprinting and 0.6% for PS. The overall accuracy of displacement data (velocity) was 5.6%. Discussion/conclusion The preliminary results of this study showed high accuracy and reliability of a commercial match analysis system suggesting its validity for the quantification of match-related physical activities in soccer players for practical and research purposes.  Further studies with larger sample size and using 90 min soccer-match simulation are necessary to fully validate video-match analysis systems. References Bangsbo, J., L. Norregaard, et al. (1991). “Activity profile of competition soccer.” Can J Sport Sci 16(2): 110-6. Hopkins, W. G. (2000). “Measures of reliability in sports medicine and science.” Sports Med 30(1): 1-15.

10th Annual Congress of the European College of Sport Science in Lausanne, Switzerland 2006.