Using cooperative networks to analyse behaviour in professional Australian Football

Understanding passing network characteristics and their link to match outcome in elite Netball

Player interactions in Netball are critical during attacking phases of play to ensure possession is maintained and scoring opportunities are created. This study aims to analyse the characteristics of the passing networks of elite Netball teams and their association with performance outcomes (i.e., win/loss and final margin). Five network metrics used to represent the characteristics of teamwork were calculated for all team performances (n = 112) from one season of professional Netball in Australia. A two-way ANOVA and multiple linear modelling were used to compare characteristics between teams and match outcomes and to predict score margin, respectively. Pass density (F = 65.09, df = 102, p < .001) and pass centrality (F = 7.61, df = 102, p < .01) differed (were higher) in wins/losses. They were also statistically significant contributors (p ≤ .005) to the linear model that predicted a score margin (R2 = .731). Key player centrality and mutual connectedness were different between teams but did not differ by match outcome. The results suggest that, ideally, Netball teams should maximise the number of connections between player pairings, while also relying on a subset of players to be heavily involved in passing sequences. Team cohesion (via passing) therefore appears to be an important measure of team success in elite Netball.

A holistic analysis of collective behaviour and team performance in Australian Football via structural equation modelling

Despite awareness of the importance of quantifying technical, tactical, and physical characteristics of match play, few studies have examined the structural relationship of these aspects in professional sport. Accordingly, this study concurrently examined these components in relation to quarter outcome (n = 272) in Australian Football. The study followed a retrospective longitudinal case study design where one teams' cooperative passing network, skill counts, physical loads, and spatiotemporal behaviours during official Australian Football League games were collected from a period spanning four seasons (2016-2019). A principal components analysis (PCA) and structural equation modelling were used to explore the structural relationships between components and examine the influence on quarter outcome as determined by the point differential (quarter margin). Scoring opportunity and ball movement had direct associations with quarter margin, while unpredictability, uncontested behaviour and physical behaviour did not. Negative associations between uncontested behaviour and scoring opportunity suggest that elevated high-pressure success and a lack of synchrony may positively influence scoring opportunity, a determinant of quarter margin. Further, negative associations between physical behaviour and ball movement suggest that with less physical work, a team's collective ability to transfer possession between teammates is facilitated, offering an interesting dichotomy between skill and physical demands of Australian Football. While hundreds of different metrics are available, the present study was the first to concurrently examine the influence of a variety of match play components on performance outcomes in Australian Football. These results may provide direction for coaches and practitioners when contemplating practice design, tactical strategies, or the development of behaviour through specific training exercises. Game plans and training drills that focus on optimising attacking and low-pressure ball movement coupled with high levels of mutual interaction between teammates may be beneficial for performance.

Assessment of Physical, Technical, and Tactical Analysis in the Australian Football League: A Systematic Review

Background

Elite Australian Football (AF) match-play requires proficiency in physical, technical, and tactical elements. However, when analysing player movement practitioners commonly exclude technical and tactical considerations, failing to recognise the multifactorial nature of AF match-play and providing little context into the movement requirements of the players.

Objectives

This systematic review aimed to identify the physical, technical, and tactical requirements of the Australian Football League (AFL) and to highlight the importance of integrating data from multiple sources when analysing player output.

Methods

A systematic search of electronic databases (CINAHL, PubMed, Scopus, SPORTDiscus, and Web of Science) was conducted from January 2009 to June 2022. Keywords relating to physical, technical, and tactical match requirements were used.

Results

Forty-eight studies met the inclusion criteria. In isolation, physical requirements were the most analysed construct within the AFL (n = 17), followed by technical (n = 9) and then tactical (n = 6). Thirteen studies integrated physical and technical elements, one study integrated technical and tactical elements, one study integrated physical and tactical elements, and one study integrated all three elements. Movement analysis centred around average ‘whole’ match requirements, whereas technical and tactical match analyses focused on key performance indicators of match performance.

Conclusion

While the physical requirements of the AFL have been well documented, there is little understanding of how player technical output and various team tactics influence player movement requirements. Knowledge of how the elements of AF match-play interact with one another could enhance our understanding of match performance and provide a greater resource for training prescription.

Match simulation practice may not represent competitive match play in professional Australian football

Match simulation in team sport should sample representative constraints and behaviours to those observed in competitive matches to enhance near skill transfer. This study compared task constraints (field length, field width, length per width ratio, space per player), time-standardised skill metrics (goals, shots on goal, handballs, kicks, marks, turnovers, tackles, handball proficiency, kick proficiency) and cooperative passing metrics (connectedness, indegree variability and outdegree variability) between match simulation practice and competitive Australian Football League (AFL) games for one professional team. MANOVAs identified activity-related differences for task constraints, skill metrics and cooperative passing networks. During match simulation, goals were scored more frequently, but with less passing actions per minute. Receiving and distributing passing networks were more centralised (reliance on fewer key individuals), with less turnovers and tackles per minute compared to AFL matches. If match simulation is designed to reflect competition, then player and team skill preparation may be compromised. Furthermore, the competing demands in high-performance sport may restrict the degree of representativeness that can be achieved during practice. These findings provide valuable insight and may assist practitioners and/or coaches to understand the value of match simulation practice and to maximise near skill transfer from match simulation to competition.

An assessment of physical and spatiotemporal behaviour during different phases of match play in professional Australian football

Despite advancements in the scale of data available for quantifying the physical and spatiotemporal characteristics of match play, there is an absence of research combining these aspects in professional sport. This study sought to differentiate between phases of play in professional Australian football using novel physical and spatiotemporal metrics. Data was obtained from Australian Football League games to provide new insight into the specific characteristics of each phase of play. A retrospective cross-sectional design was utilised with team's physical and spatiotemporal variables, measured via global navigation satellite system devices. A multinomial logistic regression was conducted to determine which physical and spatiotemporal measures were associated with each phase of play (contested play, defence, offence, set shot, goal reset, umpire stoppage). The addition of the predictors to a model that contained only the intercept significantly improved the fit between the model and data, with the logistic model correctly predicting the phase of play for 63.7% of the cases. This was the first study to concurrently examine differences in physical and spatiotemporal characteristics with respect to phase of play in an Australian football context. Differences in duration, physical and spatiotemporal properties were observed, providing new insight for coaches and subsequently providing direction for conditioning and practice design.

Improving the interpretation of skill indicators in professional Australian Football

OBJECTIVES

This study aimed to provide a simplified, novel method for analysing technical skill involvements in an Australian Football context by reducing the dimensionality of commonly reported skill counts obtained from Australian Football League (AFL) games. This may facilitate their practical use and interpretability.

DESIGN

Retrospective longitudinal design where individual players' technical skill counts were collected over three seasons of official AFL games.

METHODS

Seventy-three skill count values provided publicly by ChampionData® were collected for each match over a three-year analysis period. A principal component analysis was used to reduce the dimensionality of a large number of correlated technical skill indicators into a smaller set of uncorrelated components whilst maintaining most of the variance from the original data set.

RESULTS

The principal component analysis derived four principal components pertaining to high-pressure success, low-pressure success, attacking ball movement ability and scoring ability.

CONCLUSIONS

This study is the first to provide a simplified, novel method for analysing technical skill counts in Australian Football. The derived metrics reveal useful information for coaches and practitioners. This may consequently ease the interpretation of skill count data available to coaches from games, guide opposition analysis, help in the design of representative practice and inform player performance ratings.