Purpose: Previous research on performance indicators in rugby league (Parmar et al., 2017) has suggested that dimension reduction techniques could be more appropriate when analysing large datasets. Methods: Forty-five rugby league team performance indicators, from all 27 rounds of the 2012. 2013 and 2014 European Super League seasons, collected by Opta. were reduced to 10 orthogonal principal components with standardised team scores produced for each component. This dimension reduction technique resolved the multicollinearity, typically found in sporting data, where performance variables were related to each other, optimising the use of regression analysis. Forced-entry logistic (match outcome) and linear (points difference) regression models were used alongside exhaustive Chi-Square Automatic Interaction Detection (CHAID) decision trees to determine how well each principle component predicted success. Results: The ten principal components were entered into linear and logistic regression models without stepwise methods which explained 81.8% of the variance in point`s difference and classified match outcome correctly ~90% of the time. Results suggested that if a team increased `amount of possession` and 'making quick ground` component scores, they were more likely to win (ß=15.6. OR=10.1) and (ß=7.8. OR=13.3). respectively. Decision trees revealed that making quick ground was an important predictor of match outcome followed by quick play and amount of possession. Discussion: The use of PCA provided a useful guide on how teams can increase their chances of success by improving performances on a collection of variables, instead of analysing variables in isolation. Future studies can create contextual performance profiles to explore variations within and between team performances.
© Copyright 2018 World Congress of Performance Analysis of Sport XII. Veröffentlicht von Faculty of Kinesiology, University of Zagreb, Croatia. Alle Rechte vorbehalten.