Procedural content generation (PCG) has made substantial progress in shaping static 2D/3D geometry, while most methods treat gameplay mechanics as auxiliary and optimize only over space. We argue that this limits controllability and expressivity, and formally introduce High-Dimensional PCG (HDPCG): a framework that elevates non-geometric gameplay dimensions to first-class coordinates of a joint state space. We instantiate HDPCG along two concrete directions. Direction-Space augments geometry with a discrete layer dimension and validates reachability in 4D (x,y,z,l), enabling unified treatment of 2.5D/3.5D mechanics such as gravity inversion and parallel-world switching. Direction-Time augments geometry with temporal dynamics via time-expanded graphs, capturing action semantics and conflict rules. For each direction, we present three general, practicable algorithms with a shared pipeline of abstract skeleton generation, controlled grounding, high-dimensional validation, and multi-metric evaluation. Large-scale experiments across diverse settings validate the integrity of our problem formulation and the effectiveness of our methods on playability, structure, style, robustness, and efficiency. Beyond quantitative results, Unity-based case studies recreate playable scenarios that accord with our metrics. We hope HDPCG encourages a shift in PCG toward general representations and the generation of gameplay-relevant dimensions beyond geometry, paving the way for controllable, verifiable, and extensible level generation.
In this study, we explore the reasoning capabilities of Large Language Models (LLMs) within the context of the social communication game Werewolf, aiming to evaluate their performance in managing complex system states commonly found in computer games. Our agent architecture gathers data, refines them into detailed information, and plans actions based on this knowledge. To demonstrate the feasibility of using LLM based agents in computer games, we developed a simulation and evaluation tool using the Unity game engine. This software enables users to experiment with various LLMs and agent architectures and to measure model performance within the application. For evaluation, we tested three models: GPT-3.5 Turbo, Mistral-7B-OpenOrca, and Nous-Hermes-Llama2-13B. The results show that even smaller models can perform reasonably well in Werewolf. However, their error rate is significantly higher, highlighting the need for additional software modules or fine-tuning to improve their accuracy.
Core game design concepts such as player experience, balancing, level design, design patterns, and others are important elements when learning game design. One of the most common teaching techniques is to have students play, create, or modify small board or card games. This hands-on approach allows them to practice their design skills in an iterative and interactive way. Analog games are uniquely suited for this purpose as they are much easier to deploy and allow faster tweaking than digital ones. We wanted to overcome these hurdles and developed a suite of four digital games that allow experimentation and modification in accessible and interactive ways: JumpCrafter, FPS Balancing, RaceKart, and Pac-Mon. We have used these games for several years in our classrooms and collected qualitative feedback from 164 students over the course of 12 interventions. In order to identify important qualities of these games, we conducted a reflexive thematic analysis and have constructed themes that highlight four salient elements: The Unexpected Complexity; The Value of Experimentation; The Diversity in Designing; and The Trouble with Perspectives. We explore how these themes affect the continued use and development of our games and provide seven design considerations for future development of what we call "game design games".
Previous research investigated whether and how people from different countries differ in their playing behavior by clustering a large sample of players (n = 12,279 from 26 countries) based on responses to a survey capturing 29 constructs including key demographics (e.g., age, gender), gaming habits (e.g., playtime, spending, appreciation), and cultural attitudes (e.g., motivation, perspectives on community, perceived stressors). The earlier study employed country means and hierarchical clustering to identify eight distinct clusters. Clustering algorithms entail a certain amount of arbitrary decisions. Based on this work, the current study attempts to assess cluster stability by identifying clusters of similarity but this time based on a different set of metrics (cosine similarity) and a different clustering algorithm. Despite the methodological shift, the analysis again identified eight optimal clusters, which closely align with those from the previous study, evidence that the dataset can be segmented in stable clusters. However, some notable differences emerged, providing additional insights into cross-cultural gaming behaviors.
In this short paper we introduce the first version of a 'spell simulator' for 1 Billion Spells, an in development action game, and explore the simulator's utility for conducting large scale exploration and evaluation of that game's spell creation system. Using this simulator we conduct initial experiments to explore specific developer concerns about the game which would be challenging to explore with conventional playtesting. We argue that our initial exploration demonstrates the potential of this approach in evaluating content generation systems for games that have large player explorable possibility spaces.
