Dynamic difficulty adjustment approaches in video games: a systematic literature review

Mortazavi, Fatemeh; Moradi, Hadi; Vahabie, Abdol-Hossein

doi:10.1007/s11042-024-18768-x

Cited by 6 publications

(2 citation statements)

References 106 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Roohi et al, 2021). Embedded in shipped games and run "online," difficulty-adjustment systems, AI directors, adaptive AI opponents, and matchmaking systems all detect or predict the player's current skill and then select opponents, opponent actions, or features that present an assumed competence-optimal challenge level (Graepel & Herbrich, 2006;Mortazavi, Moradi, & Vahabie, 2024;Paraschos & Koulouriotis, 2023), e.g., finding an equally matched opponent team for an online multiplayer match. Experience-driven procedural content generation systems (Yannakakis & Togelius, 2011) similarly generate whole game levels that are predicted to present optimal challenges to a player.…”

Section: Applications In Computational Design Tools and Interfacesmentioning

confidence: 99%

“…The less specified prior theory, the less constrained the possibility space of computational implementations, the harder it might become to find effective, robust, or globally (near-)optimal solutions. This current lack is evident in the wide variety of documented automated playtesting and difficulty-adjustment systems that in the main claim to implement theory on, e.g., competence satisfaction or flow, but involve algorithms that variously track and optimise for task performance, skill growth, or something else entirely (Albaghajati & Ahmed, 2020;Mortazavi et al, 2024;Paraschos & Koulouriotis, 2023). Here, the need for and value of specifying SDT to the level of computational formalisms found in CB-IMs is immediately apparent.…”

Section: Applications In Computational Design Tools and Interfacesmentioning

confidence: 99%

See 1 more Smart Citation

Advancing Self-Determination Theory With Computational Intrinsic Motivation: The Case of Competence

Lintunen,

Ady,

Guckelsberger

et al. 2024

Preprint

View full text Add to dashboard Cite

Computational models are powerful tools to formalise psychological theories, render them severely testable, and embed them into digital applications. Yet they have seen little uptake in Self-Determination Theory (SDT). Here we demonstrate how SDT's conception of competence and optimal challenge can benefit from computational modelling informed by AI research on computational intrinsic motivation (IM). We surface underspecification in present verbal definitions, challenging the construct validity of common operationalisations and impeding computational implementations. We identify four separate verbal facets of competence and match them to distinct computational IM formalisms. These leverage formal accounts of novelty, diversity, or progress to drive skill acquisition, yielding optimal challenge-seeking behaviour and other competence dynamics postulated by SDT. We argue that these accounts specifically and computational IM more widely highlight inconsistencies within, and can concretise SDT's verbal articulations. We gently introduce motivation researchers to computational modelling and IM, complement model intuition with formal detail, and provide practical pointers for their use in psychological research and application.

show abstract

Section: Applications In Computational Design Tools and Interfacesmentioning

confidence: 99%

Section: Applications In Computational Design Tools and Interfacesmentioning

confidence: 99%

Advancing Self-Determination Theory With Computational Intrinsic Motivation: The Case of Competence

Lintunen,

Ady,

Guckelsberger

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

A framework for neurophysiological experiments on flow states

Durcan,

Holland,

Bhattacharya

2024

Commun Psychol

View full text Add to dashboard Cite

Csikszentmihalyi’s concept of the “flow state” was initially discovered in experts deeply engaged in self-rewarding activities. However, recent neurophysiology research often measures flow in constrained and unfamiliar activities. In this perspective article, we address the challenging yet necessary considerations for studying flow state’s neurophysiology. We aggregate an activity-autonomy framework with several testable hypotheses to induce flow, expanding the traditional “challenge skill balance” paradigm. Further, we review and synthesise the best methodological practices from neurophysiological flow studies into a practical 24-item checklist. This checklist offers detailed guidelines for ensuring consistent reporting, personalising and testing isolated challenge types, factoring in participant skills, motivation, and individual differences, and processing self-report data. We argue for a cohesive approach in neurophysiological studies to capture a consistent representation of flow states.

show abstract