Game-Playing Agents

Ask about this topic

Overview

Game-playing agents study how AI can perceive game states, plan actions, and adapt to opponents, providing testbeds for decision-making under competition, uncertainty, and long time horizons. Common approaches include tree search and Monte Carlo Tree Search, self-play reinforcement learning, regret minimization for imperfect-information games, and language- or vision-language models equipped with memory, tools, and learned world models. Evaluation spans bounded games such as chess and Go, poker and social deduction, real-time strategy, and interactive video games, using measures such as win rate, exploitability, task progress, and performance against people. Specialized systems can be highly capable in defined settings, while broad generalization remains difficult: agents still struggle with visual grounding, reliable memory, opponent modeling, and timely action in unfamiliar or long-running games.

Latest papers

CardsList

206 papers36 papers in the last four weeks, up 50% on the four weeks before. 0.2% of all new papers.

Jun 3, 2026

Episodic Memory Temporal Consistency for Cooperative Multi-Agent Reinforcement Learning

Cooperative Multi-Agent Reinforcement Learning (MARL) frequently suffers from severe reward sparsity and exploration bottlenecks. While episodic memory mechanisms mitigate these issues by reusing high-return trajectories, they often trap agents in local optima due to unconstrained incentive distribution and semantic representation collapse. To address this, we propose Episodic Memory Temporal Consistency (EMTC), a framework that robustly constructs and selectively leverages historical experiences. EMTC introduces two synergistic components: (1) a Temporally Consistent Semantic Embedder that integrates contrastive learning with time-conditioned state reconstruction, preventing representation collapse and enabling precise memory retrieval; and (2) a Temporal Consistency Gating Mechanism that dynamically modulates episodic incentives based on temporal consistency error. This adaptive gate filters misleading signals from pseudo-successful trajectories, effectively mitigating Q-value overestimation. We provide theoretical guarantees, establishing a strict error bound that directly links the observable temporal consistency error to the underlying trajectory optimality and representation quality. Extensive evaluations on the SMAC and GRF benchmarks demonstrate that EMTC consistently outperforms state-of-the-art baselines. Notably, compared to the strongest episodic baseline, EMTC achieves absolute win-rate improvements of up to 24% in super-hard SMAC scenarios and an average improvement of 28% across GRF tasks.

Zicheng Zhao, Yu Lan, Chengzhengxu Li +2

Xi’an Jiaotong University · Queen Mary University of London

Reinforcement LearningGame-Playing AgentsSparse-Reward RLCooperative MARL
AskPDF
Jun 3, 2026

ChessMimic: Per-Rating Transformer Models for Human Move, Clock, and Outcome Prediction in Online Blitz Chess

We present ChessMimic, a system of three small encoder-only transformers - for move, thinking-time, and outcome prediction - conditioned on the position, recent move history, player rating, and clock state. We fit a separate instance of each model per 100-Elo rating band, trading parameter efficiency for sharper per-skill calibration. On a held-out month-wide slice of Lichess Rated Blitz games ChessMimic's human move prediction accuracy outperforms Maia-2 in every Elo band. Compared to Maia-3, our 9M parameter model's accuracy sits between Maia-3-5M and Maia-3-23M without the additional complexity of Geometric Attention Bias. In addition to the move matching model, we also train a game outcome model that conditions not only on the position, but also player ratings, time control, and remaining clock times. The outcome model achieves an AUC of 0.78 out of sample, beating Maia-2 as well as logistic regressions based on material, ratings, and clock time. Finally, we train a clock model that predicts human thinking times. The clock model provides a usable but non-SOTA per-ply think-time signal under ALLIE-style filters (Pearson r = 0.41, Spearman rho = 0.50, MAE 4.10 s, against ALLIE's reported r = 0.70), with the residual gap concentrated in per-position bucket sharpness rather than bucket-marginal calibration. A public demo is at 1e4.ai and we release code, per-band weights, and the C++ data-filter pipeline code in GitHub.

Thomas Johnson

Nascent

Game-Playing AgentsTransformerChess
AskPDF
May 29, 2026

Safe Equilibrium Policy Optimization for Strategic Agent Policies

Language models fine-tuned with reinforcement learning typically optimize for task reward, ignoring multi-agent strategic structure. Because these agents condition on natural language game-state descriptions and emit actions through free-form generation, strategic failure modes -- exploiting weaker opponents, coordinating on harmful equilibria, and externalizing costs are inseparable from the language interface itself. We propose Safe Equilibrium Policy Optimization (\sepo{}), a training objective that augments expected payoff with explicit penalties for exploitability, collusion risk, and externality cost. We implement \sepo{} as a reward signal for Group Relative Policy Optimization (GRPO), applied to Gemma4 E4B-it and Qwen3.5-4B after supervised fine-tuning (SFT). Evaluated across five strategic domains: Iterated Prisoner's Dilemma, repeated auctions, two negotiation variants, and Kuhn Poker. \sepo{} achieves zero exploit-pool advantage in Kuhn Poker for both models, outperforms the base model on safety in four domains, and corrects the over-cooperative behavior introduced by SFT. In negotiation, \sepo{} achieves a positive-safety outcome and only the positive normalized relative advantage of any negotiation configuration. Ablation experiments confirm that per-rollout exploit computation is necessary: a shared constant penalty cancels in GRPO advantage normalization (constant control-variate property), producing zero gradient. To support further research in strategic safety for agents, we release our code and SFT datasets.

Karthika Arumugam, Kiran Kumar Manku, Amit Dhanda

†Amazon, USA

Multi-Agent Reinforcement LearningGame-Playing AgentsMulti-Agent System OptimizationSafe RLPolicy LearningReward Design for RL+2
AskPDF
May 28, 2026

PokerSkill: LLMs Can Play Expert-Level Poker without Training or Solvers

Poker is a landmark challenge for artificial intelligence. The dominant approach relies on equilibrium solvers built on counterfactual regret minimization, requiring millions of core-hours of training. Large Language Models (LLMs) possess extensive poker knowledge but perform far below solver-based agents when asked to play directly. Traditional rule-based poker agents are interpretable and training-free, but their strategic ceiling remains far below equilibrium play. We introduce PokerSkill, a training-free and solver-free framework that bridges this gap by using detailed rule-based poker skills as a structured action-grounding interface for LLMs. A deterministic context engine analyzes the current state and retrieves only the relevant fragments from a layered skill library, which is entirely designed by human poker experts, constraining the LLM's choice to reasonable actions. Against GTOWizard, a state-of-the-art GTO benchmark, GPT-5.5 XHigh with PokerSkill achieves −57±21-57 \pm 21 mbb/hand, Claude Opus 4.6 achieves −80±29-80 \pm 29 mbb/hand and Claude Opus 4.7 achieves −87±64-87\pm 64 mbb/hand, reducing losses by 49--61% compared to default-prompt baselines and outperforming the strong bot Slumbot. Our key finding is that rule-based skills alone do not constitute a strong strategy, and LLMs alone cannot play well, but their combination yields an agent that requires neither training nor solver access yet competes with systems built on millions of core-hours of computation. To our knowledge, this is the first demonstration of an LLM achieving competitive performance in a complex imperfect-information game without game-specific training or solver queries. Code is available at https://github.com/lbn187/PokerSkill.

Boning Li, Baoxiang Wang, Longbo Huang

IIIS, Tsinghua University · The Chinese University of Hong Kong, Shenzhen

Game-Playing AgentsImperfect-Information GamesLLM Agents
AskPDF
May 28, 2026

PTCG-Bench: Can LLM Agents Master Pokémon Trading Card Game?

Given a strategically complex board game, human players can quickly learn to devise strategies after playing a few rounds. Autonomous agents require similar capabilities in realistic interactive environments, yet existing agent benchmarks often fail to fully capture such strategic and evolving decision-making scenarios. We present PTCG-Bench, a benchmark built on the Pok'{e}mon Trading Card Game (PTCG) that evaluates LLM agents at two complementary levels: (1) their decision-making performance within a single complex environment, and (2) their ability to self-evolving through accumulated experience. We further include a modular harness ablation to better interpret agent performance without conflating it with model capability. Our experiments show that, although LLM agents can achieve non-trivial gameplay performance, sustained and stable self-evolution remains challenging, and performance is sensitive to harness design. We hope that PTCG-Bench will facilitate future research on harness-aware and self-evolving agents in realistic interactive environments.

Dongdong Hua, Yifei Sun, Renhong Huang +3

Zhejiang University · FinVolution Group

Game-Playing AgentsLLM Agent EvaluationAI Agent BenchmarksSelf-Evolving Agents
AskPDF
May 26, 2026

UniMaia: Steering Chess Policies with Language for Human-like Play

Recent advances in large language models have enabled natural language to serve as a flexible interface for controlling complex systems, but often at the cost of large-scale multimodal training or weakened domain-specific inductive biases. In structured decision-making domains such as chess, specialized policy networks achieve strong performance but lack semantic controllability, while prompt-conditioned language models are more flexible yet typically exhibit weaker domain grounding. We propose UniMaia\textbf{UniMaia}, a framework for prompt-conditioned policy modulation that adapts a frozen Lc0-based chess policy network using a parameter-efficient text encoder and a ControlNet-style conditioning mechanism. UniMaia enables semantic control over gameplay, including opening selection and player strength, while preserving the pretrained policy representations. We further introduce UniMaia-Aux\textbf{UniMaia-Aux}, which incorporates auxiliary temporal conditioning and behavioral prediction objectives. To support this work, we construct a large-scale metadata-augmented Lichess dataset, develop a semi-automated prompt-generation pipeline, and introduce benchmarks spanning both prompt-conditioned and metadata-conditioned settings. UniMaia achieves state-of-the-art expected accuracy on several prompt-conditioned benchmarks and competitive top-move accuracy on general instruction-following tasks, while remaining competitive with dedicated metadata-conditioned approaches on human move prediction benchmarks. UniMaia-Aux further improves expected accuracy and behavioral modeling across several evaluation settings, with modest trade-offs in top-move accuracy. Overall, our results demonstrate that prompt-conditioned control of domain-specific policy networks is feasible without end-to-end multimodal training, while highlighting trade-offs between controllability and predictive performance.

Sherman Siu, Lesley Istead

University of Waterloo · Carleton University

Game-Playing AgentsLanguage Model-Based ControlChess
AskPDF
May 25, 2026

Explore Before You Solve: The Speed--Depth Trade-off in Epistemic Agents for ARC-AGI-3

We systematically investigate all 25 public ARC-AGI-3 games and find that every one is reachable through non-intelligent strategies: 10 in a single blind step, 5 after one probing action, 1 via repeated ACTION1 presses, 1 via diverse exploration, and 8 via single repeated actions with sufficient budget (50-200 steps). A library-level null-coordinate vulnerability additionally bypasses 18 games in 1 step. This benchmark critique implies the public evaluation set cannot discriminate intelligent exploration from trivial heuristics - the private 55-game evaluation is the only genuine intelligence test. Against this backdrop, we present AERA (Adaptive Epistemic Reasoning Agent), a three-phase (EXPLORE / VERIFY / PLAN) agent achieving RHAE=0.2116 (4/25 solved) on these 25 games with Qwen2.5-0.5B, while random and no-explore baselines score 0.0000. We formalise AERA through a Speed--Depth trade-off framework: under a convexity assumption (proved for a class of environments in the Appendix), RHAE's quadratic form emerges as a second-order penalty for deviating from the Pareto frontier between action efficiency and information gain. Contributions: (i) a benchmark validity analysis showing that current interactive reasoning benchmarks fail to measure the exploration they claim to require, and (ii) the EXPLORE-before-PLAN framework and model-capability x exploration interaction. The linked code track entry achieves RHAE=0.30 on the full 55-game private evaluation. Code: CC0.

Liew Keong Han

Independent researcher.

Game-Playing AgentsAI Agent EvaluationARC-AGIAI Agent BenchmarksAgentic ReasoningBenchmark Validity
AskPDF
May 22, 2026

MAPLE: Multi-State Aggregated Policy Evaluation for AlphaZero in Imperfect-Information Games

Imperfect-information games (IIGs) are challenging, as players must make decisions without fully observing the true game state. While AlphaZero has achieved remarkable success in perfect-information games, extending it to IIGs remains difficult. Existing search-based approaches, such as Perfect Information Monte Carlo (PIMC), suffer from strategy fusion, while Information Set Monte Carlo Tree Search (IS-MCTS) incurs high computational cost when combined with neural networks. In this paper, we propose Multi-State Aggregated PoLicy Evaluation (MAPLE), a tree search method that aggregates policy and value evaluations from multiple sampled world states within a single search tree, combining the advantages of PIMC and IS-MCTS while maintaining a controllable computational cost. We further incorporate a Siamese-based sampling strategy to select informative world states from the information set. Experiments on Phantom Go and Dark Hex show that MAPLE significantly outperforms the PIMC-based AlphaZero baseline, achieving Elo improvements of 291 and 136, respectively. These results demonstrate that MAPLE is an effective approach for AlphaZero-style learning in imperfect-information games.

Qian-Rong Li, Hung Guei, I-Chen Wu +1

Department of Computer Science, National Yang Ming Chiao Tung University, Taiwan · Institute of Information Science, Academia Sinica, Taiwan

Game-Playing AgentsImperfect-Information GamesMonte Carlo Tree Search
AskPDF
May 22, 2026

One Policy, Infinite NPCs: Persona-Traceable Shared RL Policies for Scalable Game Agents

On a 300-persona life-simulation benchmark, pcsp achieves compositional zero-shot persona identification up to 17x above chance, Spearman rho approx 0.73 semantic-behavioral alignment, and 22x faster inference than an LLM-as-policy baseline. Life simulation games require hundreds to thousands of non-player characters (NPCs) that behave consistently with distinct personalities while remaining controllable through designer-authored natural language. Existing methods fail on constraints like persona consistency, controllability, or real-time inference. We introduce pcsp (Persona Conditioned Shared Policy), a single reinforcement learning policy conditioned on frozen LLM embeddings of free-form persona descriptions. pcsp combines once-per-NPC persona encoding, low-rank persona projection, neural persona conditioning, and a PPO + InfoNCE consistency + KL diversity training objective. Across three experimental settings, ablations show that the InfoNCE trajectory-consistency objective is load bearing: removing it collapses zero-shot persona identification to chance. External validation on Melting Pot 2.4.0 substrates confirms that our method produces persona-conditioned behavioral divergence in multi-agent strategic environments. We distinguish two senses of held-out evaluation: compositional zero-shot and vocabulary-expansion held-out. Finally, a UE5 deployment reproduces the in-engine persona-conditioning ablation at 64 agents with a low failure rate, showing that the sub-frame inference profile survives in a commercial game engine. These results prove that shared RL policies can support scalable, real-time, persona-conditioned NPC control.

Yoosung Hong

Independent Researcher

Reinforcement LearningGame-Playing AgentsHuman Behavior SimulationMulti-Agent Reinforcement Learning
AskPDF
May 22, 2026

Symbolic Reasoning Frameworks Trigger Memory-Mediated Ecosystem Dynamics in Multi-Agent LLM Systems

Large language models exhibit a risk-averse "turtle" bias as strategic agents. We show that injecting a symbolic reasoning framework as a per-round reflective prompt into one agent acts as a small perturbation whose consequences are not per-decision but emergent: the agent's risk posture is unchanged in isolation, yet over a campaign of accumulating memory and multi-agent interaction the conditions settle into distinct, condition-associated winner ecosystems. In a 7-player Warring States Diplomacy variant (61 games, 6 conditions), the winner distribution differs sharply across the four primary conditions (41 games; permutation omnibus p approximately 0.001): control -> Yan (7/11); I-Ching yarrow -> Yan/Chu co-dominance with Qin fully suppressed (0/10); Tarot -> Qin (5/10); scrambled-text ablation -> Qi (5/10). The scrambled->Qi attractor is robust (vs. pooled and control alone, p = 0.006 and 0.012); tarot->Qin is denominator-dependent (0.006 pooled, 0.064 vs. control). Han never wins and shows no survival difference (Fisher p = 1.0); neither framework's content predicts actions (chi-squared p = 0.95 hexagram, 0.69 Tarot). A memory-free decision-isolation probe (960 calls) shows the process does not change the agent's risk posture in isolation (Friedman p = 0.45; I-Ching p = 0.60; Tarot perturbs move content but not risk, p = 0.021). A 2x2 factorial separating yarrow's decision-time and learning-time components reveals a non-additive interaction: each alone freezes the board (50-60% stalemates), combined they produce zero (permutation p ~ 5e-5). Testing relocates Qin suppression to rival (Chu) expansion governed by campaign memory depth, not the oracle (p = 0.55). We present this as an observation paper: agent-level framework choice produces distinctive, non-additive system-level consequences, transmitted through emergent memory and multi-agent dynamics, not per-decision effects.

Augustin Chan

Multi-Agent LLM SystemsGame-Playing AgentsMulti-Agent SystemsEmergent Behavior in Multi-Agent SystemsMemory-Augmented Language Model Agents
AskPDF
May 21, 2026

Self-Play Reinforcement Learning under Imperfect Information in Big 2

Imperfect-information multiplayer games test whether agents can act under hidden information, sparse rewards, and non-stationary opponents. We study these challenges in Big 2, a four-player imperfect-information card game. We develop a self-play RL framework for Big 2 that enables controlled comparisons between policy-gradient and value-approximating agents. Under a common environment, input representation, training budget, and evaluation protocol, PPO outperforms Monte Carlo Q approximation, SARSA, and Q-learning against random, greedy, and heuristic Big 2 opponents. We further find that moderate entropy regularization improves PPO by preventing the policy from becoming overly deterministic, and that current-policy self-play provides a stronger finite-budget curriculum than checkpoint self-play or fixed-opponent training. Together, these results show that Big 2 is a useful controlled setting for studying deep RL under imperfect information, multiplayer interaction, delayed rewards, and variable action sets.

Aalok Patwa

University of Pennsylvania

Game-Playing AgentsSelf-Play RLImperfect-Information GamesProximal Policy OptimizationMulti-Agent Reinforcement Learning
AskPDF
May 21, 2026

Not Yet: Humans Outperform LLMs in a Colonel Blotto Tournament

The emergence of large language models (LLMs) has spurred economists to study how humans and LLMs behave in strategic settings. We organized a series of round-robin tournaments in the Colonel Blotto game. This game attracts game theorists' attention due to high-dimensional action space and the absence of pure strategy Nash equilibria. In the first tournament, more than 200 human participants competed against one another. In the second tournament, several popular LLMs were invited to submit strategies. In the third tournament, we matched the number of LLM strategies to the number submitted by humans. We find that humans more often employ better-calibrated intermediate-level allocation heuristics and outperform the simpler, more stereotyped strategies submitted by LLMs. Strategic sophistication is key to success if and only if the necessary level of reasoning depth is reached, while lower and higher levels of reasoning offer no clear advantage over the primitive strategies. Among humans, field of study weakly predicts success: participants with STEM backgrounds perform better in the first tournament. Surprisingly, humans almost do not adjust their strategies across tournaments with different sets of opponents. This result suggests that humans base their choices primarily on the game's rules rather than on the identity of their opponents, treating LLMs much like human competitors.

Dmitry Dagaev, Egor Ivanov, Petr Parshakov +2

HSE University · New Economic School · Central Economic Mathematical Institute, Russian Academy of Sciences +3

Game-Playing AgentsGame TheoryLLM Decision-MakingHuman-AI Decision Making
AskPDF
May 20, 2026

For How Long Should We Be Punching? Learning Action Duration in Fighting Games

Fighting games such as Street Fighter II present unique challenges to reinforcement learning (RL) agents due to their fast-paced, real-time nature. In most RL frameworks, agents are hard-coded to make decisions at a fixed interval, typically every frame or every N frames. Although this design ensures timely responses, it restricts the agent's ability to adjust its reaction timing. Acting every frame grants frame-perfect reflexes, which are unrealistic compared to human players, whereas longer fixed intervals reduce computational cost but hinder responsiveness. We consider an alternative decision-making framework in which the agent learns not only what action to take but also for how long to execute it. By jointly predicting both action and duration, the agent can dynamically adapt its responsiveness to different situations in the game. We implement this method using the open-source FightLadder environment with agents trained against scripted built-in bots, systematically testing different frame skip configurations to analyze their influence on performance, responsiveness, and learned behavior. Experiments show that learned timing can match the performance of well-chosen fixed frame skips and encourages repeatable action patterns, but does not ensure robustness on its own. In most cases, we see agents performing best with consistently high frame skip values (i.e., low responsiveness). This strategy makes it easier to learn exploitative strategies where the same action is repeated over and over, which the scripted bots appear to be susceptible to.

Hoang Hai Nguyen, Kurt Driessens, Dennis J. N. J. Soemers

Department of Advanced Computing Sciences, Maastricht University, Maastricht, the Netherlands

Reinforcement LearningGame-Playing Agents
AskPDF
May 20, 2026

Mahjax: A GPU-Accelerated Mahjong Simulator for Reinforcement Learning in JAX

Riichi Mahjong is a multi-player, imperfect-information game characterized by stochasticity and high-dimensional state spaces. These attributes present a unique combination of challenges that mirror complex real-world decision-making problems in reinforcement learning. While prior research has heavily relied on supervised learning from human play logs to pre-train the policy, algorithms capable of learning tabula rasa (from scratch) offer greater potential for general applicability, as evidenced by the AlphaZero lineage. To facilitate such research, we introduce Mahjax, a fully vectorized Riichi Mahjong environment implemented in JAX to enable large-scale rollout parallelization on Graphics Processing Units (GPUs). We also provide a high-quality visualization tool to streamline debugging and interaction with trained agents. Experimental results demonstrate that Mahjax achieves throughputs of up to 2 million and 1 million steps per second on eight NVIDIA A100 GPUs under the no-red and red rules, respectively. Furthermore, we validate the environment's utility for reinforcement learning by showing that agents can be trained effectively to improve their rank against baseline policies.

Soichiro Nishimori, Shinri Okano, Keigo Habara +3

The University of Tokyo, Tokyo, Japan. · Nara Institute of Science and Technology, Nara, Japan. · Independent Researcher. +5

RL BenchmarksGame-Playing AgentsImperfect-Information GamesMulti-Agent Reinforcement LearningRL for Game AI
AskPDF
May 19, 2026

Real-Time Parallel Counterfactual Regret Minimization

Counterfactual Regret Minimization (CFR) is the dominant algorithmic family for solving large imperfect-information games, underpinning breakthroughs such as Libratus and Pluribus in No-Limit Texas Hold'em poker. In real-time game-playing systems, the solver must compute a near-equilibrium strategy within a strict time budget of only a few seconds per decision, and the number of CFR iterations completed in this window directly determines play strength. We present Parallel CFR, the first parallelization framework for real-time depth-limited CFR solving that seamlessly integrates pruning, abstraction, and advanced CFR variants. We decompose each CFR iteration into a pipeline of seven stages and identify two orthogonal dimensions of parallelism: by information set and by tree node. Leaf node evaluation is offloaded to GPUs via batched neural network inference, creating a heterogeneous CPU--GPU pipeline. Experiments on Heads-Up No-Limit Texas Hold'em demonstrate that Parallel CFR achieves 3.33.3--3.4×3.4\times speedup over the single-threaded baseline on postflop streets, with per-iteration time of ∼47{\sim}47--5454~ms on a depth-limited game tree with over 11 billion histories. All experiments run on a single desktop-class device (NVIDIA DGX Spark), enabling hundreds of CFR iterations within a typical real-time decision budget without requiring datacenter-scale infrastructure.

Boning Li, Longbo Huang

IIIS, Tsinghua University

Game-Playing AgentsGPU AccelerationImperfect-Information GamesHigh-Performance ComputingCounterfactual Regret MinimizationRegret Minimization+1
AskPDF
May 18, 2026

Chessformer: A Unified Architecture for Chess Modeling

Chess has long served as a canonical testbed for artificial intelligence, but modeling approaches for its central tasks have diverged. Maximizing playing strength, predicting human play, and enabling interpretability are typically solved with disparate architectures, and these designs are often misaligned with the geometry of the domain. This raises the natural question of whether these objectives require separate modeling paradigms, or if there exists a single architecture that supports them simultaneously. We introduce Chessformer, a unified architecture that advances the state of the art on all three central goals in chess modeling. Chessformer is an encoder-only transformer that represents board squares as tokens, augments self-attention with a novel dynamic positional encoding called Geometric Attention Bias (GAB) that adapts to domain-specific geometry, and predicts actions with an attention-based source-destination policy head. We evaluate Chessformer on each front. First, we develop \maiathree, a family of models for human move prediction that reaches 57.1% move-matching accuracy, significantly surpassing the previous state of the art with fewer than a quarter of the parameters. Second, we integrate Chessformer into Leela Chess Zero, a leading open-source engine, adding over 100 Elo of playing strength and resulting in tournament victories over Stockfish in major computer chess competitions. Third, we show that Chessformer's square-token design makes attention patterns and activations directly attributable to board squares, enabling granular interpretability analyses that prior architectures do not naturally support. More broadly, our results demonstrate that aligning a model's tokenization, positional encoding, and output design with the underlying structure of a domain can yield simultaneous gains in performance, human compatibility, and interpretability.

Daniel Monroe, George Eilender, Philip Chalmers +2

University of Toronto · Williams College

Transformer InterpretabilityGame-Playing AgentsTransformerPositional EncodingChess
AskPDF
May 18, 2026

SPIKE: An Adaptive Dual Controller Framework for Cost-Efficient Long-Horizon Game Agents

Long-horizon multimodal agents in open-world games must stay goal-directed across many low-level interactions under tight token and latency budgets. Existing approaches often trade off costly per-step reasoning against reactive execution that can drift, repeat failures, and recover poorly. Our key idea is to reuse strategic reasoning across locally stable segments and reinvoke it at event boundaries. We present SPIKE, an adaptive dual controller framework for cost-efficient long-horizon game control. Its Strategic Controller performs low-frequency global planning, failure analysis, and recovery, while its Reactive Controller handles fast local execution under a strict token budget. An Event Trigger monitors visual change, task progress, repeated actions, and failure signals to decide when control should stay reactive or escalate to strategic reasoning. Hierarchical Memory separates short-term experience reuse in the State-Action Memory Bank (SA-MB) from structured evidence in the State Action Knowledge Graph (SA-KG), allowing each controller to retrieve the context it needs. This design reuses strategic proposals over multiple reactive steps, supports local override when plans become stale, and reserves expensive reasoning for moments where extra deliberation is useful. On the Lite-100 split of StarDojo, SPIKE improves Lite-100 success rate (SR) by 5.0 percentage points (38.5% relative) over the strongest Lite-100 baseline and Budgeted SR by 9.3 points (75.6% relative) over the strongest budgeted baseline. It also reduces token consumption by 54.9% and latency by 40.8%. Ablations show that event triggering, reactive override, and heterogeneous memory each contribute to success and recovery, supporting selective reasoning rather than reasoning at every step.

Wencan Jiang, Jiangning Zhang, Jianbiao Mei +6

1Zhejiang University · 2National University of Singapore · 3Nanyang Technological University

Game-Playing AgentsFeedback ControlEvent-Triggered ControlLong-Horizon Agent TasksMemory-Augmented AgentsAdaptive Control
AskPDF
May 16, 2026

From Imitation to Interaction: Mastering Game of Schnapsen with Shallow Reinforcement Learning

This paper investigates whether shallow neural network agents can master the card game Schnapsen and challenge a strong search-based baseline, RdeepBot, which uses Monte Carlo sampling and lookahead search. Guided by a progressively more complex experimental design, we first evaluate a supervised learning agent (MLPBot) trained on replay data and then a reinforcement learning agent (RLBot) with the same shallow architecture trained through asynchronous Monte Carlo updates and experience replay. The results show that supervised imitation does not generalize well enough to defeat strong RdeepBot opponents, whereas reinforcement learning produces substantially stronger agents. In the setting that focuses on the depth parameter of RdeepBot, the best performance is achieved when the learned value function is combined with deeper lookahead during gameplay, allowing RLBot to achieve statistically significant higher winning rates against the strongest evaluated RdeepBot baseline. In the sample-based setting, the gains are more conditional: the strongest performance appears at a relatively lower training num_samples parameter rather than increasing uniformly with stronger sampling.

Ján Klačan, Sizhong Zhang

Vrije Universiteit Amsterdam

Reinforcement LearningGame-Playing AgentsImperfect-Information GamesImitation LearningRL for Game AI
AskPDF
May 14, 2026

Cattle Trade: A Multi-Agent Benchmark for LLM Bluffing, Bidding, and Bargaining

We introduce \textsc{Cattle Trade, a multi-agent benchmark for evaluating large language models (LLMs) as agents in strategic reasoning under imperfect information, adversarial interaction, and resource constraints. The benchmark combines auctions, hidden-offer trade challenges (TCs), bargaining, bluffing, opponent modeling, and resource allocation within a single long-horizon game lasting 50--60 turns. Unlike prior agent benchmarks that test these abilities in isolation, Cattle Trade evaluates whether agents integrate them across a competitive, multi-agent economic game with conflicting incentives. The benchmark logs every bid, TC offer, counteroffer, and card selection, enabling behavioural analysis beyond final scores or win rates. We evaluate seven cost-efficient language models and three deterministic code agents across 242 games. Strategic coherence, in particular spending efficiency, resource discipline, and phase-adaptive bidding, is associated with rank more strongly than spending volume or any single subskill. Two heuristic code agents outperform most tested LLMs, and behavioural traces surface recurring LLM failure modes including overbidding, self-bidding, bankrupt TC initiation, and weak opponent-state adaptation. Evaluating agentic competence requires benchmarks that test the joint deployment of multiple capabilities in multi-agent environments with conflicting incentives, uncertainty, and economic dynamics.

Robert Müller, Clemens Müller

Multi-Agent NegotiationGame-Playing AgentsLLM Agent EvaluationLong-Horizon Agent EvaluationAI Agent Benchmarks
AskPDF
May 14, 2026

Watermarking Game-Playing Agents in Perfect-Information Extensive-Form Games

Watermarking techniques for large language models (LLMs), which encode hidden information in the output so its source can be verified, have gained significant attention in recent days, thanks to their potential capability to detect accidental or deliberate misuse. Similar challenges involving model misuse also exist in the context of game-playing, such as when detecting the unauthorized use of AI tools in gaming platforms (e.g., cheating in online chess). In this paper, we initiate the study of how game-playing strategies can be watermarked. We show how the KGW watermark for LLMs can be adapted to watermark game-playing agents in perfect-information extensive-form games. The watermark can then be detected using a statistical test. We show that the degradation in the quality of the watermarked strategy profile, quantified by the expected utility, can be bounded, but there is a tradeoff between detectability and quality. In our experiments, we bootstrap the watermarking framework to various chess engines and demonstrate that a) the impact of the watermark on the quality of the strategy is negligible and b) the watermark can be detected with just a handful of games.

Juho Kim, Fei Fang, Tuomas Sandholm

Computer Science Department Carnegie Mellon University · Software and Societal Systems Department Carnegie Mellon University · Computer Science Department, CMU +3

Game-Playing Agents
AskPDF
May 14, 2026

ASH: Agents that Self-Hone in Long-Horizon Worlds

Long-horizon visuomotor tasks remain a fundamental challenge in AI, as current methods rely on hand-engineered rewards or action-labeled demonstrations, neither of which scales. We introduce ASH, an agentic system that learns a long-horizon policy from unlabeled, noisy internet video, without reward shaping or expert annotation. ASH follows a self-improvement loop; when it gets stuck, ASH learns an Inverse Dynamics Model (IDM) from its own trajectories, and uses its IDM to extract supervision from relevant internet video. ASH uses unsupervised learning to identify key moments from large-scale internet video and retains them as long-term memory - allowing it to tackle long-horizon problems. We evaluate ASH on two complementary environments demanding multi-hour planning: Pokemon Emerald, a turn-based RPG, and The Legend of Zelda: The Minish Cap, a real-time action-adventure game. In both games, behavioral cloning, retrieval-augmented and zero-shot foundation-model baselines plateau, while ASH sustains progression across our 8-hour evaluation. ASH reaches an average of 11.2/12 milestones in Pokemon Emerald and 9.9/12 in Legend of Zelda, while the strongest baseline gets stuck in both environments at an average of 9.3/12 and 7.8/12 milestones, respectively. We demonstrate that self-improving agents are a scalable recipe for learning in long-horizon worlds.

Benjamin Schneider, Xavier Schneider, Victor Zhong +1

University of Waterloo · National Research Council Canada

Game-Playing AgentsSelf-Improving AgentsLong-Horizon Agent TasksMemory-Augmented Agents
AskPDF
May 12, 2026

Toward Modeling Player-Specific Chess Behaviors

While artificial intelligence has achieved superhuman performance in chess, developing models that accurately emulate the individualized decision-making styles of human players remains a significant challenge. Existing human-like chess models capture general population behaviors based on skill levels but fail to reproduce the behavioral characteristics of specific historical champions. Furthermore, the standard evaluation metric, move accuracy, inherently penalizes natural human variance and ignores long-term behavioral consistency, leading to an incomplete assessment of stylistic fidelity. To address these limitations, an architecture is proposed that adapts the unified Maia-2 model to champion-specific embeddings, further enhanced by the integration of a limited Monte Carlo Tree Search (MCTS) process to enrich tactical exploration during move selection. To robustly evaluate this approach, a novel behavioral metric based on the Jensen-Shannon divergence is introduced. By compressing high-dimensional board representations into a latent space using an AutoEncoder and Uniform Manifold Approximation and Projection (UMAP), move distributions are discretized on a common grid to compare behavioral similarities. Results across 16 historical world champions indicate that while integrating MCTS decreases standard move accuracy, it improves stylistic alignment according to the proposed metric, substantially reducing the average Jensen-Shannon divergence. Ultimately, the proposed metric successfully discriminates between individual players and provides promising evidence toward more comprehensive evaluations of behavioral alignment between players and AI models.

Loris Sogliuzzo, Aloïs Rautureau, Eric Piette

ICTEAM, UCLouvain · Louvain-la-Neuve, Belgium

Game-Playing AgentsAI Agent Evaluation
AskPDF
May 11, 2026

Continual Harness: Online Adaptation for Self-Improving Foundation Agents

Coding harnesses such as Claude Code and OpenHands wrap foundation models with tools, memory, and planning, but no equivalent exists for embodied agents' long-horizon partial-observability decision-making. We first report our Gemini Plays Pokemon (GPP) experiments. With iterative human-in-the-loop harness refinement, GPP became the first AI system to complete Pokemon Blue, Yellow Legacy on hard mode, and Crystal without a lost battle. In the hardest stages, the agent itself began iterating on its strategy through long-context memory, surfacing emergent self-improvement signals alongside human-in-the-loop refinement. Continual Harness removes the human fully from this loop: a reset-free self-improving harness for embodied agents that formalizes and automates what we observed. Starting from only a minimal environment interface, the agent alternates between acting and refining its own prompt, sub-agents, skills, and memory, drawing on any past trajectory data. Prompt-optimization methods require episode resets; Continual Harness adapts online within a single run. On Pokemon Red and Emerald across frontier models, Continual Harness starting from scratch substantially reduces button-press cost relative to the minimalist baseline and recovers a majority of the gap to a hand-engineered expert harness, with capability-dependent gains, despite starting from the same raw interface with no curated knowledge, no hand-crafted tools, and no domain scaffolding. We then close the loop with the model itself: an online process-reward co-learning loop, in which an open-source agent's rollouts through the refining harness are relabeled by a frontier teacher and used to update the model, drives sustained in-game milestone progress on Pokemon Red without resetting the environment between training iterations.

Seth Karten, Joel Zhang, Tersoo Upaa +5

Princeton University · ARISE Foundation · Google DeepMind

Continual Learning for LLM AgentsGame-Playing AgentsAgent Harness OptimizationSelf-Improving AgentsLong-Horizon LLM AgentsLLM Agent Harnesses
AskPDF
May 11, 2026

Towards Generalist Game Agents: A Survey of Foundation Models in Virtual Worlds

The real world unfolds along a single set of physics laws, yet human intelligence demonstrates a remarkable capacity to generalize experiences from this singular physical existence into a multiverse of games, each governed by entirely different rules, aesthetics, physics, and objectives. This omni-reality adaptability is a hallmark of general intelligence. As Artificial Intelligence progresses towards Artificial General Intelligence, the multiverse of games has evolved from mere entertainment into the ultimate ground for training and evaluating AGI. The pursuit of this generality has unfolded across four eras: from environment-specific symbolic and reinforcement learning agents, to current large foundation models acting as generalist players, and toward a future creator stage where agent both creates new game worlds and continually evolves within them. We trace the full lifecycle of a generalist game player along four interdependent pillars: Dataset, Model, Harness, and Benchmark. Every advance across these pillars can be read as an attempt to break one of five fundamental trade-offs that currently bound the whole system. Building on this end-to-end view, we chart a five-step roadmap, progressing from single-game mastery to the ultimate creator stage in which the agent simultaneously creates and evolves within theoretical game multiverse. Taken together, our work offers a unified lens onto a rapidly shifting field, and a principled path toward the omnipotent generalist agent capable of seamlessly mastering any challenge within the multiverse of games, thereby paving the way for AGI.

Kuan Zhang, Dongchen Liu, Qiyue Zhao +12

College of AI, Tsinghua University · MMLab, The University of Hong Kong · University of Chinese Academy of Sciences

Game-Playing AgentsAI Agent Benchmarks
AskPDF
May 9, 2026

AlphaExploitem: Going Beyond the Nash Equilibrium in Poker by Learning to Exploit Suboptimal Play

Poker is an imperfect information game that has served as a long-standing benchmark for decision-making under uncertainty. To maximize utility beyond the Nash equilibrium, an agent can deviate from Nash-equilibrium policies to exploit suboptimal play. We introduce AlphaExploitem, which extends the competitive RL poker agent AlphaHoldem by using a hierarchical transformer encoder that enables reasoning over previously played hands and modifying the training procedure with the inclusion of a diverse pool of exploitable opponents to facilitate learning to exploit. We train and evaluate AlphaExploitem on two standard benchmarks for imperfect-information games. Empirically, AlphaExploitem successfully exploits weak play by both in- and out-of-distribution opponents, without losing performance against NE opponents.

Vlad Murgoci, Matthijs Spaan, Yaniv Oren

Delft University of Technology · CD Delft, The Netherlands · Department of Intelligent Systems

Opponent ModelingGame-Playing AgentsImperfect-Information GamesGame Theory
AskPDF
May 8, 2026

Reason to Play: Behavioral and Brain Alignment Between Frontier LRMs and Human Game Learners

Humans rapidly learn abstract knowledge when encountering novel environments and flexibly deploy this knowledge to guide efficient and intelligent action. Can modern AI systems learn and plan in a similar way? We study this question using a dataset of complex human gameplay with concurrent fMRI recordings, in which participants learn novel video games that require rule discovery, hypothesis revision, and multi-step planning. We jointly evaluate models by their ability to play the games, match human learning behavior, and predict brain activity during the same task, comparing a suite of frontier Large Reasoning Models (LRMs) against model-free and model-based deep reinforcement learning agents and a Bayesian theory-based agent. We find that frontier LRMs most closely match human behavioral patterns during game discovery and predict brain activity an order of magnitude better than both reinforcement learning alternatives across cortical and subcortical regions, with effects robust to permutation controls. Through targeted manipulations, we further show that brain alignment reflects the model's in-context representation of the game state rather than its downstream planning or reasoning. Our results establish LRMs as compelling computational accounts of human learning and decision making in complex, naturalistic environments. Project page with interactive replays: https://botcs.github.io/reason-to-play/

Botos Csaba, Sreejan Kumar, Austin Tudor David Andrews +6

University of Oxford · Columbia University · New York University +2

Game-Playing AgentsHuman Behavior SimulationNeural EncodingLarge Reasoning ModelsHuman Behavior Modeling
AskPDF
May 7, 2026

Causal Reinforcement Learning for Complex Card Games: A Magic The Gathering Benchmark

Causal reinforcement learning (RL) lacks benchmarks for complex systems that combine sequential decision making, hidden information, large masked action spaces, and explicit causal structure. We introduce MTG-Causal-RL, a Gymnasium benchmark built on Magic: The Gathering with a 3,077-dimensional partial observation, a 478-action masked discrete action space, five competitive Standard archetypes, three reward schemes, and a hand-specified Structural Causal Model (SCM) over strategic variables. Every episode exposes causal variables, SCM-predicted intervention effects, and per-factor credit traces, making causal credit assignment, leave-one-out cross-archetype transfer, and policy auditability first-class metrics. We adapt a panel of reference baselines: random, heuristic, masked PPO, a causal-world-model PPO variant, and an architecture-matched scalar control. We propose Causal Graph-Factored Advantage PPO (CGFA-PPO) as a reference causal agent that uses SCM parents of win probability as factor-aligned critic targets with an intervention-calibration loss. All comparisons use paired seeds, paired-bootstrap confidence intervals, and Holm-Bonferroni correction within pre-registered families. Masked PPO and CGFA-PPO reach competitive in-distribution win rates and exceed the random baseline; per-factor calibration trajectories and leave-one-out transfer gaps expose diagnostic structure that scalar win rate alone cannot. We release the benchmark, reference-baseline results, and full evaluation protocol openly. By coupling a strategically rich, partially observed domain with an explicit causal interface and statistical protocol, MTG-Causal-RL gives causal-RL, world-model, and LLM-agent research a shared testbed for questions current benchmarks cannot pose together: causal credit assignment under masked action spaces, structural transfer across archetypes, and SCM-grounded policy auditability.

Cristiano da Costa Cunha, Ajmal Mian, Tim French +1

Department of Computer Science and Software Engineering, University of Western Australia, 6009 WA Australia

RL BenchmarksStructural Causal ModelsReinforcement LearningGame-Playing AgentsPartially Observable RLCausal RL
AskPDF
May 6, 2026

Strat-Reasoner: Reinforcing Strategic Reasoning of LLMs in Multi-Agent Games

While Large Language Models (LLMs) excel in certain reasoning tasks, they struggle in multi-agent games where the final outcome depends on the joint strategies of all agents. In multi-agent games, the non-stationarity of other agents brings significant challenges on the evaluation of the reasoning process and the credit assignment over multiple reasoning steps. Existing single-agent reinforcement learning (RL) approaches and their multi-agent extensions fail to address these challenges as they do not incorporate other agents in the reasoning process. In this work, we propose Strat-Reasoner, a novel RL-based framework that improves LLMs' strategic reasoning ability in multi-agent games. We introduce a novel recursive reasoning paradigm where an agent's reasoning also integrates other agents' reasoning processes. To provide effective reward signals for the intermediate reasoning sequences, we employ a centralized Chain-of-Thought (CoT) comparison module to evaluate the reasoning quality. Finally, we compute an accurate hybrid advantage and develop a group-relative RL approach to optimize the LLM policy. Experimental results show that Strat-Reasoner substantially improves strategic abilities of underlying LLMs, achieving 22.1% average performance improvements across various multi-agent games. Code is publicly available at https://github.com/ydhe1012/Strat-Reasoner.

Yidong He, Yutao Lai, Pengxu Yang +4

School of Software Engineering, South China University of Technology · Department of Computer Science, University of Oxford

Game-Playing AgentsRL for Language Model ReasoningMulti-Agent ReasoningMulti-Agent Reinforcement Learning
AskPDF
May 5, 2026

Agent Island: A Saturation- and Contamination-Resistant Benchmark from Multiagent Games

Static capabilities benchmarks suffer from saturation and contamination, making it difficult to track capabilities progress over time. We introduce Agent Island, a multiplayer simulation environment in which language-model agents compete in a game of interagent cooperation, conflict, and persuasion. The environment yields a dynamic benchmark designed to mitigate both saturation and contamination; new models can always outperform the current leading player in this winner-take-all game, and agents compete against other adaptive agents rather than face a fixed task set. We rank players with a Bayesian Plackett-Luce model, allowing us to quantify uncertainty in player skill. In 999 games involving 49 unique models, openai/gpt-5.5 dominates its peers with a posterior mean skill of 5.64, compared with 3.10 for the second-ranked model, openai/gpt-5.2, and 2.86 for the third-ranked model, openai/gpt-5.3-codex. We release the game logs as a dataset for analyses of model behavior. As an example, we investigate same-provider preference in final-round votes and find that models are 8.3 p.p. more likely to support a same-provider finalist than finalists from other providers. This preference is not uniform across providers: among separately estimated providers, the effect is strongest for OpenAI models and weakest for Anthropic models.

Connacher Murphy

Stanford University

Game-Playing AgentsBenchmark DesignLLM Agent EvaluationAI Agent BenchmarksMulti-Agent LLMs
AskPDF
May 4, 2026

A Structural Threshold in Decision Capacity Governs Collapse in Self-Play Reinforcement Learning

We show that a threshold in decision capacity determines whether self-play reinforcement learning agents collapse under asymmetric rule perturbations. Across poker variants, matrix games, a dice game, and multiple learning algorithms, eliminating all positive-reach contingent decisions causes rapid convergence to a deterministic exploitation attractor, a fixed point at near-maximal loss. Preserving even a single positive-reach contingent decision point prevents this collapse. A frozen baseline and fixed-opponent control confirm that the mechanism is co-adaptation under constraint, not the perturbation itself. The phenomenon is timing-invariant, fully reversible upon action restoration, and intensifies under function approximation. These results establish a sharp threshold at zero reach-weighted contingent action capacity, with severity scaling continuously via reach-weighted capacity in the tested domains.

Arahan Kujur

Independent Researcher

Game-Playing AgentsSelf-Play RLGame TheoryMulti-Agent Reinforcement LearningGame-Theoretic RL
AskPDF