Beyond Playtesting: A Generative Multi-Agent Simulation System for Massively Multiplayer Online Games

2025-12-02

AI Summary

Traditional MMO game optimization relies on costly online A/B testing or low-fidelity offline simulation, failing to accurately model player reasoning and causal responses to design interventions. Method: We propose the first LLM-based generative multi-agent simulation system, jointly training player agents and environment dynamics models via supervised fine-tuning and reinforcement learning on real gameplay logs. Contribution/Results: Our approach achieves causally consistent, interpretable, and high-fidelity player behavior modeling—surpassing long-standing fidelity bottlenecks in conventional simulation. Experiments demonstrate strong behavioral alignment with real players (average similarity >0.89) and principled, causally grounded responses to game-mechanic interventions. This work establishes an efficient, low-cost, and trustworthy offline optimization framework for MMO numerical balancing and mechanic design.

Technology Category

• Multiagent Systems: Mechanism Design
• Search and Optimization: Sampling/Simulation-based Search
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions

Abstract

Optimizing numerical systems and mechanism design is crucial for enhancing player experience in Massively Multiplayer Online (MMO) games. Traditional optimization approaches rely on large-scale online experiments or parameter tuning over predefined statistical models, which are costly, time-consuming, and may disrupt player experience. Although simplified offline simulation systems are often adopted as alternatives, their limited fidelity prevents agents from accurately mimicking real player reasoning and reactions to interventions. To address these limitations, we propose a generative agent-based MMO simulation system empowered by Large Language Models (LLMs). By applying Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) on large-scale real player behavioral data, we adapt LLMs from general priors to game-specific domains, enabling realistic and interpretable player decision-making. In parallel, a data-driven environment model trained on real gameplay logs reconstructs dynamic in-game systems. Experiments demonstrate strong consistency with real-world player behaviors and plausible causal responses under interventions, providing a reliable, interpretable, and cost-efficient framework for data-driven numerical design optimization.

Problem

• Simulate realistic player behavior in MMO games
• Optimize game design without costly live experiments
• Enhance simulation fidelity using generative AI agents

Innovation

• LLM-based generative agents simulate player decisions
• SFT and RL adapt models to game-specific domains
• Data-driven environment reconstructs dynamic game systems