The Role of Contextual Game Knowledge in Decision-making within Esports

Abstract

This thesis investigates the mechanisms of expertise and decision-making in esports, focusing on how contextual game knowledge informs expert behaviour. While extensive research has examined expertise across traditional domains including traditional sports, esports, defined broadly as competitive video gaming, empirical research into the mechanisms of expert performance and decision-making in esports remains limited. This is even though esports present a unique and nascent domain in which expertise can be observed and measured. Central to this thesis is the following question: What is the role of contextual game knowledge in decision-making by expert esports competitors? Domain experts leverage structured mental representations and extensive repositories of domain specific knowledge to make rapid and accurate decisions during a task, by anticipating and recognising emergent scenarios through assessment of environmental cues and other contextual factors. One theoretical framework prominent in the assessment of rapid decision-making within time-pressured domains is Naturalistic Decision Making (NDM), which posits that decision-making is facilitated through a recognition-primed process that focuses on rapid assessment of the situation and the selection for the first suitable action as opposed to other frameworks that frames decision-making as a more deliberative process and where action selection is focused on the most optimal solution. However, applications of NDM within the field of games and esports is limited.
Following a pragmatic paradigm, grounded in Design Science Research (DSR), this thesis conducted a convergent mixed methods approach to the methodology that conducts two separate quantitative and qualitative studies, with findings discussed in conjunction after both studies were completed. Both studies examined competitive Team Fortress 2, a class-based first-person shooter. Interviews with expert video game players, following the semi-structured Critical Decision Method (CDM) interview protocol, found that decision-making within the game relies on predefined action templates, segmentation of play, and predefined roles with value and expectancies attributed to them. A hierarchical level of analysis framework is proposed, the Meta/Macro/Microgame model, which breaks actions and decisions into three temporal levels of play. Separately, a hand-based notational analysis captured behavioural patterns linked to distinct player actions, revealing that dynamic game states such as holding contextual advantages and control over objectives influenced player behaviour. Team-coordinated actions are more strongly associated with success, although individual high-risk plays remain impactful. Synthesis of findings from both studies suggest that expert performance is underpinned by cue-driven recognition processes as theorised by the Recognition Primed Decision (RPD) model. Contextual game knowledge enables experts to form structured default plays that adapt dynamically to shifting scenarios. Methodologically, the integration of CDM and notational analysis offers a novel framework for esports research, facilitating a multi-layered analysis of cognition and behaviour. While limitations around generalisability and positionality are acknowledged, the research conducted advances theoretical, methodological, and practical understandings of expertise in esports.

Date of Award	14 Jul 2026
Original language	English
Awarding Institution	University of Portsmouth
Supervisor	Peter Howell (Supervisor), Martina Navarro (Supervisor) & Brett Stevens (Supervisor)