Real time prediction of match outcomes in Australian football.

This study aimed to determine whether machine learning models based on technical performance and not score margin could be used to predict end-of-match outcome of Australian football matches in real-time. If efficacious, these models could be used to generate insights about team performance and support the decision-making of coaches during matches. A database of 168 team technical performance indicators from 829 Australian Football League matches played between 2017 and 2021 was used. Two feature sets (data-driven and data-informed) were used to train and evaluate six models (generalised linear model, random forest and adaboost) on match outcome prediction (Win/Loss) over 120 epochs (a representation of normalised time during each match). All models performed well (mean classification accuracy = 73.5-75.8%) in comparison with a benchmark score-based model (mean classification accuracy = 77.4%). Data-informed feature sets performed better than data-driven in most cases. Classification accuracy was low at the start of a match (45.7-48.8%) but increased to a peak near the end of a match (87.2-92.7%). These findings suggest that any of the employed models can be used to formulate in-match decision support. The model which is best in practice will depend on factors such as time-cost trade-off, feasibility and the perceived value of its suggestions.

A framework to explain the in-match decision-making of elite Australian football coaches.

The ability to make effective decisions is an important function of any football coach, whether during training, team selection, match-day performance or post-match player evaluation. It is not yet known how elite Australian football coaches make decisions during matches, in time-constrained but well-resourced environments. This study is the first to explore the decision-making of elite Australian football coaches during matches, in pursuit of identifying opportunities to improve the translation and implementation of research findings into the competitive match environment. Using semi-structured interviews and thematic analysis, a six-stage framework of the decision-making of elite Australian football coaches during matches was developed. The stages include (1) Opportunity trigger, (2) Understand the opportunity, (3) Determine the need for action, (4) Explore options, (5) Take action and (6) Evaluate the decision. Coaches relied on subjective and objective sources of information and consulted with assistant coaches, performance analysts, and sport scientists. The findings enable researchers to ensure future interventions to improve decision-making during matches are well integrated. They also provide an opportunity for coaches to reflect on their own decision-making process, identifying targeted areas for improvement in their own practice.

The effect of team formation on defensive performance in Australian football.

OBJECTIVES: Understanding the successful characteristics of team formation during different scenarios in Australian Football matches can assist coaches in making important tactical match-day and training decisions. The aims of this study were to explore the outcomes of entries inside 50 m of the goal, in Australian Football and to determine whether there was an association between team formation and team defensive performance after a turnover. DESIGN: Observational. METHODS: Global Positioning System (GPS) data, technical event data and video files from 22 matches in one season were obtained from an elite Australian Football club. Of 1092 forward 50 entries, 392 possession chains that resulted in a turnover were analysed. Variables representing team formation of players at the occurrence of turnover were compared between positive and negative outcomes of the subsequent possession chain. Logistic regression and decision tree modelling were also used to explore associations and variable importance. RESULTS: None of 18 team formation characteristics differed between positive and negative outcomes of turnovers. Multivariate modelling identified that having a team formation with greater width than length made it more likely to result in a positive outcome (Decision tree classification accuracy = 69.5%, AUROC = 0.72). CONCLUSIONS: No single characteristic of team formation affects the outcome of a turnover possession chain, however team formation that was wider than it was long may be associated with a more desirable outcome. The lack of association between most team formation characteristics and defensive outcomes, highlight the risk of over emphasising team formation in tactical planning for some phases of play.