Short-term performance prediction in team sports using tracking data

Abstract This research addresses short-term performance prediction in Rugby Union using player tracking data. Rugby’s unique characteristics as a conquest sport with strict offside rules make the ball position a crucial performance indicator. We used event data to identify ball carriers in offensive phase and considered their locations as proxy for the ball. We then aim to predict this performance indicator. We formulated this challenge as a time series prediction problem where player positions (multivariate) predict performance (univariate). Our results suggest using player tracking data improves the prediction ability. Competing Interest Statement The authors have declared no competing interest. Footnotes {arnaud.odet{at}math.univ-toulouse.fr, sebastien.dejean{at}math.univ-toulouse.fr}, {thomas.bechard{at}utoulouse.fr, cristian.pasquaretta{at}utoulouse.fr}