Coaching and talent development in esports: a theoretical framework and suggestions for future research

Esports is a growing phenomenon that is capturing the attention of individuals worldwide, and has grown to provide professional and lucrative careers for those who reach the upper echelons. One question that arises, is how esports athletes develop the necessary skills required to improve and compete. This perspective piece opens the door to skill acquisition within esports and how research through an ecological approach can benefit researchers and practitioners as they understand the various perception-action couplings and decision-making challenges faced by esports athletes. We will identify and discuss what constraints look like in esports, the role of affordances, and theorize the implementation of a constraints-led approach in contrasting esports genres. As esports is technology-heavy in nature and generally sedentary, the use of eye-tracking technology is argued to represent an effective method to better understand perceptual attunement between individuals and teams. Future research into skill acquisition in esports is needed to develop a clearer picture of what makes the greatest esports player so great, and how newer players can be developed effectively.

Graph representations for the analysis of multi-agent spatiotemporal sports data

Analyzing tactical patterns in invasion games using multi-agent spatiotemporal data is a challenging task at the intersection of computer and sports science. A fundamental yet understudied problem in this area is finding an optimal data representation for processing athlete trajectories using machine learning algorithms. In the present work, we address this gap by discussing common representations in use and propose Tactical Graphs, an alternative graph-based format capable of producing integrative, contextualized models for machine learning applications. We provide an in-depth, domain-specific motivation of the proposed data representation scheme and show how this approach exploits inherent data traits. We propose Tactical Graph Networks (TGNets), a light-weight, hybrid machine learning architecture sensitive to player interactions. Our method is evaluated with an extensive ablation study and the first comprehensive state of the art comparison between standard feature, state vector, and image-based methods on the same dataset. Experiments were conducted using real-world football data containing short sequences of defensive play labelled according to the outcome of ball winning attempts. The results indicate that TGNets are on par with state-of-the-art deep learning models while exhibiting only a fraction of their complexity. We further demonstrate that selecting the right data representation is crucial as it has a significant influence on model performance. The theoretical findings and the proposed method provide insights and a strong methodological alternative for all classification, prediction or pattern recognition applications in the areas of collective movement analysis, automated match analysis, and performance analysis.

The Dirty League: English Premier League Provides Higher Incentives for Fouling as Compared to Other European Soccer Leagues

Fouling in soccer has been studied from an ethical standpoint as a measure of aggression. However, there is limited research related to fouling for performance. The present study investigated fouling as a factor influencing performance in European soccer leagues. Out of possession fouls (FPGNorm), yellow cards (YCFNorm), and their ratio (YCPFPG) were used as predictors of points (Pts) and goals conceded (GA) at the end of the season using three separate linear regression models. Furthermore, 5-fold cross-validation was used to measure out sample reliability. All the models significantly predicted GA and Pts (p < 0.001). Models predicting GA showed higher reliability than models predicting points. Cross validation (CV) results suggested that FPGNorm and YCPFPG models showed a small standard deviation (SD) in the R² results whereas the results from YCFNorm were not reliable to high SD in the 5-fold CV results. In summary, FPGNorm and YCPFPG seem to predict success (low GA and high Pts) across European soccer leagues, with EPL showing the maximum effect. The findings of the current study and the methodology can be applied to an actual game analysis by coaches in multiple invasion sports. Normalizing for out of possession time is a crucial step for the time spent in particular phases of play, which has not been done in previous research while analyzing 'key performance indices' (KPIs). Normalization can successfully introduce domain-specific knowledge into predictors, which can be used in complex algorithms improving predictions and investigation of underlying mechanisms.

Rugby league ball carrier injuries: The relative importance of tackle characteristics during the European Super League

Rugby league carries a high injury incidence with 61% of injuries occurring at tackles. The ball carrier has a higher injury incidence than the defender, therefore understanding mechanisms occurring during injurious tackles are important. Given the dynamic, open nature of tackling, characteristics influencing tackle outcome likely encompass complex networks of dependencies. This study aims to identify important classifying characteristics of the tackle related to ball carrier injurious and non-injurious events in rugby league and identify the characteristics capability to correctly classify those events. Forty-one ball carrier injuries were identified and 205 matched non-injurious tackles were identified as controls. Each case and control were analysed retrospectively through video analysis. Random forest models were built to (1) filter tackle characteristics possessing relative importance for classifying tackles resulting in injurious/non-injurious outcomes and (2) determine sensitivity and specificity of tackle characteristics to classify injurious and non-injurious events. Six characteristics were identified to possess relative importance to classify injurious tackles. This included 'tackler twisted ball carrier's legs when legs were planted on ground', 'the tackler and ball carrier collide heads', 'the tackler used body weight to tackle ball carrier', 'the tackler has obvious control of the ball carrier' 'the tackler was approaching tackle sub-maximally' and 'tackler's arms were below shoulder level, elbows were flexed'. The study identified tackle characteristics that can be modified in attempt to reduce injury. Additional injury data are needed to establish relationship networks of characteristics and analyse specific injuries. Sensitivity and specificity results of the random forest were 0.995 and 0.525.

Methods of performance analysis in team invasion sports: A systematic review

The objective of this review was to systematically describe the traditional and contemporary data capture and analytic methods employed in performance analysis research in team invasion sports, evaluate the practicality of these methods, and formulate practical recommendations on methods for analysing tactics and strategies in team invasion sports. A systematic search of the databases SPORTDiscus, Web of Science, Scopus, MEDLINE and PubMed was performed. Keywords addressed performance analysis methods and team invasion sports, with all other disciplines of sports science excluded. A total of 537 articles were included in the review and six main themes of research identified. Themes included game actions, dynamic game actions, movement patterns, collective team behaviours, social network analysis and game styles. Performance analysis research has predominantly focused on identifying key performance indicators related to success by analysing differences in game actions between successful and less successful teams. However, these measures are outcome-focused and only provide limited insight into winning team's strategy. Team invasion sports are now viewed as dynamic, complex systems with opposing teams as interacting parts. Strategies and tactics should be analysed using a holistic process-orientated approach by recording dynamic actions, collective team behaviours and passing networks, and viewing them in game styles.

Wearables for Integrative Performance and Tactic Analyses: Opportunities, Challenges, and Future Directions

Micro-electromechanical systems (MEMS) have reduced drastically in size, cost, and power consumption, while improving accuracy. The combination of different sensor technologies is considered a promising step in the monitoring of athletes. Those "wearables" enable the capturing of relevant physiological and tactical information in individual and team sports and thus replacing subjective, time-consuming and qualitative methods with objective, quantitative ones. Prior studies mainly comprised sports categories such as: targeting sports, batting and fielding games as well as net and wall games, focusing on the detection of individual, non-locomotive movements. The increasing capabilities of wearables allow for more complex and integrative analysis expanding research into the last category: invasion sports. Such holistic approaches allow the derivation of metrics, estimation of physical conditions and the analysis of team strategic behavior, accompanied by integrative knowledge gains in technical, tactical, physical, and mental aspects of a sport. However, prior and current researchers find the precise measurement of the actual movement within highly dynamic and non-linear movement difficult. Thus, the present article showcases an overview of the environments in which the wearables are employed. It elaborates their use in individual as well as team-related performance analyses with a special focus on reliability and validity, challenges, and future directions.

What's Next in Complex Networks? Capturing the Concept of Attacking Play in Invasive Team Sports

The evolution of performance analysis within sports sciences is tied to technology development and practitioner demands. However, how individual and collective patterns self-organize and interact in invasive team sports remains elusive. Social network analysis has been recently proposed to resolve some aspects of this problem, and has proven successful in capturing collective features resulting from the interactions between team members as well as a powerful communication tool. Despite these advances, some fundamental team sports concepts such as an attacking play have not been properly captured by the more common applications of social network analysis to team sports performance. In this article, we propose a novel approach to team sports performance centered on sport concepts, namely that of an attacking play. Network theory and tools including temporal and bipartite or multilayered networks were used to capture this concept. We put forward eight questions directly related to team performance to discuss how common pitfalls in the use of network tools for capturing sports concepts can be avoided. Some answers are advanced in an attempt to be more precise in the description of team dynamics and to uncover other metrics directly applied to sport concepts, such as the structure and dynamics of attacking plays. Finally, we propose that, at this stage of knowledge, it may be advantageous to build up from fundamental sport concepts toward complex network theory and tools, and not the other way around.

A comparison of game-play characteristics between elite youth and senior Australian National Rugby League competitions

OBJECTIVES

To compare game-play characteristics between elite youth and senior Australian National Rugby League (NRL) competitions.

DESIGN

Longitudinal observational.

METHODS

The dataset consisted of 12 team performance indicators (e.g., 'all runs', 'offloads' and 'tackles') extracted from all 2016 national under 20 (U20) competition (elite youth; n=372 observations) and National Rugby League (NRL) (elite senior; n=378 observations) matches. Data was classified according to competition (Two levels: U20 and NRL) and modelled using two techniques. Firstly, non-metric multidimensional scaling resolved multivariate competition (dis)similarity, visualised using a two-dimensional ordination. Secondly, a conditional interference (CI) classification tree was grown to reveal the performance indicators most capable of explaining competition level.

RESULTS

Non-metric multidimensional scaling revealed high competition dissimilarity, with U20 and NRL teams orienting distinctive positions on the first dimension of the ordination surface. Five team performance indicators were retained within the CI tree ('all runs', 'tackle breaks', 'tackles', 'missed tackles', and 'kicks'), which correctly classified 79% of the U20 observations and 93% of the NRL observations.

CONCLUSIONS

Multivariate differences between elite youth and senior rugby league competitions were identified. Specifically, NRL game-play was classified by a greater number of 'all runs', and 'tackles' and a lower number of 'missed tackles' relative to the U20 competition. Given the national U20 competition is purported to assist with the development of prospective NRL players, junior coaches may consider training interventions that primarily aid the tackling capacities of players. This may subsequently assist with talent development and player progression in Australian rugby league.

The variable and chaotic nature of professional golf performance

In golf, unlike most other sports, individual performance is not the result of direct interactions between players. Instead decision-making and performance is influenced by numerous constraining factors affecting each shot. This study looked at the performance of PGA TOUR golfers in 2011 in terms of stability and variability on a shot-by-shot basis. Stability and variability were assessed using Recurrence Quantification Analysis (RQA) and standard deviation, respectively. About 10% of all shots comprised short stable phases of performance (3.7 ± 1.1 shots per stable phase). Stable phases tended to consist of shots of typical performance, rather than poor or exceptional shots; this finding was consistent for all shot categories. Overall, stability measures were not correlated with tournament performance. Variability across all shots was not related to tournament performance; however, variability in tee shots and short approach shots was higher than for other shot categories. Furthermore, tee shot variability was related to tournament standing: decreased variability was associated with better tournament ranking. The findings in this study showed that PGA TOUR golf performance is chaotic. Further research on amateur golf performance is required to determine whether the structure of amateur golf performance is universal.

Big data and tactical analysis in elite soccer: future challenges and opportunities for sports science

Until recently tactical analysis in elite soccer were based on observational data using variables which discard most contextual information. Analyses of team tactics require however detailed data from various sources including technical skill, individual physiological performance, and team formations among others to represent the complex processes underlying team tactical behavior. Accordingly, little is known about how these different factors influence team tactical behavior in elite soccer. In parts, this has also been due to the lack of available data. Increasingly however, detailed game logs obtained through next-generation tracking technologies in addition to physiological training data collected through novel miniature sensor technologies have become available for research. This leads however to the opposite problem where the shear amount of data becomes an obstacle in itself as methodological guidelines as well as theoretical modelling of tactical decision making in team sports is lacking. The present paper discusses how big data and modern machine learning technologies may help to address these issues and aid in developing a theoretical model for tactical decision making in team sports. As experience from medical applications show, significant organizational obstacles regarding data governance and access to technologies must be overcome first. The present work discusses these issues with respect to tactical analyses in elite soccer and propose a technological stack which aims to introduce big data technologies into elite soccer research. The proposed approach could also serve as a guideline for other sports science domains as increasing data size is becoming a wide-spread phenomenon.

Sport science integration: An evolutionary synthesis

The aim of the paper is to point out one way of integrating the supposedly incommensurate disciplines investigated in sports science. General, common principles can be found among apparently unrelated disciplines when the focus is put on the dynamics of sports-related phenomena. Dynamical systems approaches that have recently changed research in biological and social sciences among others, offer key concepts to create a common pluricontextual language in sport science. This common language, far from being homogenising, offers key synthesis between diverse fields, respecting and enabling the theoretical and experimental pluralism. It forms a softly integrated sports science characterised by a basic dynamic explanatory backbone as well as context-dependent theoretical flexibility. After defining the dynamic integration in living systems, unable to be captured by structural static approaches, we show the commonalities between the diversity of processes existing on different levels and time scales in biological and social entities. We justify our interpretation by drawing on some recent scientific contributions that use the same general principles and concepts, and diverse methods and techniques of data analysis, to study different types of phenomena in diverse disciplines. We show how the introduction of the dynamic framework in sport science has started to blur the boundaries between physiology, biomechanics, psychology, phenomenology and sociology. The advantages and difficulties of sport science integration and its consequences in research are also discussed.

Attack Coverage in High-Level Men's Volleyball: Organization on the Edge of Chaos?

Change is pervasive, but emerging patterns are occasionally detectable through analysis of systemic behaviors. Match analysis uses these patterns in order to reduce the degree of improvisation and to optimize the training process. However, it is possible that certain game phases elude systematic patterning. In this vein, our aim was to analyze the case of attack coverage in men’s volleyball, as we suspected it would elude systematic patterning and has received negligible attention in scientific research. We analyzed the occurrence of attack coverage in 4544 plays of the 2011 Volleyball World League. A Chi-square test with residual adjusted values was applied to explore significant associations between variables. A Monte Carlo correction was applied, as some cells had n<5. Effect sizes were determined using Cramer’s V. Overall, attack coverage occurred in 3.89% of ball possessions, and 23 distinct structures emerged. These structures lacked significant associations with the game complex, setting zone, and effect of attack coverage. Conversely, attack coverage structures showed significant associations with the attack zone and tempo, with very strong effect sizes (V=0.472 and V=0.521, respectively). As certain attack zones are deeply associated with attack tempo, it is apparent that quicker attack plays affect attack coverage structuring, promoting the formation of less complex structures. Ultimately, attack coverage structures seem to depend on momentary constraints, thereby rendering rigid systematization impracticable. Still, we contended that a principle-based approach might be suitable. This invites researchers to rethink how to interpret game regularities.

Explaining match outcome in elite Australian Rules football using team performance indicators

The relationships between team performance indicators and match outcome have been examined in many team sports, however are limited in Australian Rules football. Using data from the 2013 and 2014 Australian Football League (AFL) regular seasons, this study assessed the ability of commonly reported discrete team performance indicators presented in their relative form (standardised against their opposition for a given match) to explain match outcome (Win/Loss). Logistic regression and decision tree (chi-squared automatic interaction detection (CHAID)) analyses both revealed relative differences between opposing teams for "kicks" and "goal conversion" as the most influential in explaining match outcome, with two models achieving 88.3% and 89.8% classification accuracies, respectively. Models incorporating a smaller performance indicator set displayed a slightly reduced ability to explain match outcome (81.0% and 81.5% for logistic regression and CHAID, respectively). However, both were fit to 2014 data with reduced error in comparison to the full models. Despite performance similarities across the two analysis approaches, the CHAID model revealed multiple winning performance indicator profiles, thereby increasing its comparative feasibility for use in the field. Coaches and analysts may find these results useful in informing strategy and game plan development in Australian Rules football, with the development of team-specific models recommended in future.