Determining, scoring and presenting successful performance in professional rugby league

Performance indicators allow for the objective quantification of performance (Vogelbein, Nopp & Hokelmann, 2014). However, limited PI research for professional rugby league exists, with just one paper published (Woods, Sinclair and Robertson, 2017) although this was conducted on teams from the Australian elite competition, the NRL, with no similar attempts for Europe’s Super League competition. Therefore, this thesis aimed to identify robust indicators of success for professional rugby league teams in super league, which would subsequently allow performances to be scored and assessed graphically through performance profiles. Data from all 27 rounds of the 2012, 2013 and 2014 European Super League seasons were collected by Opta, amounting to 567 matches. Data for 45 action variables was extracted from spreadsheets using Visual Basic for Applications in Microsoft Excel (Excel, v2013, Microsoft Inc., Redmond, USA). To enable clear comparisons between winning and losing teams, draws (n=22) were excluded. Study 1 assessed twenty-four relative variables (home value minus away) using backwards logistic (match outcome) and linear (points difference) regression models alongside exhaustive Chi-Square Automatic Interaction Detection (CHAID) decision trees to identify performance indicators (PIs) and key performance indicators (KPIs). However, some variables which were thought to be important (as identified by previous literature) were removed from the analysis as they did not contribute to the model’s predictive ability as much as others thus calling into question the appropriateness of stepwise methods. Furthermore, unusual results were evident which lead to the conclusion that a suitable dimension reduction technique could be more appropriate to analyse large datasets with multiple variables that could be related to each other. Study 2 utilised principal component analysis to reduce 45 action variables into 10 orthogonal principle components. These components were analysed using backwards and enter methods in logistic and linear regression models alongside CHAID decision trees. This method provided a relevant guide on how teams could improve their performance by improving a collection of variables as opposed to traditional methods which described individual variables. Furthermore, the use of stepwise methods was argued to be less appropriate for sporting performances as some principal components that could relate to success may be removed. Results from both regression models indicated large variations on confidence intervals for beta coefficients and odds ratios, suggesting that the variation of a set of values are more representative of the data analysed, when assessing multiple teams. Therefore, idiographic assessments of performances were suggested to provide relevant information for practitioners, which can be lost through traditional nomothetic approaches, as evidenced in this study. Study 3 utilised the principle component scores to create idiographic performance profiles, according to match venue and match closeness. In addition, a case study was produced assessing two teams’ previous performances, prior to an upcoming game, providing a practical example of how practitioners could utilise this information in their respective environments. Although large variations were evident on profiles, it was suggested that team performances may never stabilise due to the unpredictability of complex sports involving multiple players like rugby league. However it was clear that idiographic profiles provided meaningful and informative assessments of performance which were arguably more relevant for practitioners compared to traditional nomothetic methods. Overall, this thesis facilitated a greater understanding of how rugby league teams perform in Super League, through the use of practical and relevant methodologies that can be utilised by practitioners and coaches who are constantly striving to improve sporting performance. Future research must consider the ‘theory-practice’ gap identified by McKenzie and Cushion (2013) in order to provide simple and relevant answers that practitioners require, which seems to be a principle that has remained elusive thus far