How Far Are LLMs from Professional Poker Players? Revisiting Game-Theoretic Reasoning with Agentic Tool Use
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911412985528320 |
|---|---|
| author | Lin, Minhua Dai, Enyan Liu, Hui Tang, Xianfeng Yan, Yuliang Dai, Zhenwei Zeng, Jingying Zhang, Zhiwei Wang, Fali Gao, Hongcheng Luo, Chen Zhang, Xiang He, Qi Wang, Suhang |
| author_facet | Lin, Minhua Dai, Enyan Liu, Hui Tang, Xianfeng Yan, Yuliang Dai, Zhenwei Zeng, Jingying Zhang, Zhiwei Wang, Fali Gao, Hongcheng Luo, Chen Zhang, Xiang He, Qi Wang, Suhang |
| contents | As Large Language Models (LLMs) are increasingly applied in high-stakes domains, their ability to reason strategically under uncertainty becomes critical. Poker provides a rigorous testbed, requiring not only strong actions but also principled, game-theoretic reasoning. In this paper, we conduct a systematic study of LLMs in multiple realistic poker tasks, evaluating both gameplay outcomes and reasoning traces. Our analysis reveals LLMs fail to compete against traditional algorithms and identifies three recurring flaws: reliance on heuristics, factual misunderstandings, and a "knowing-doing" gap where actions diverge from reasoning. An initial attempt with behavior cloning and step-level reinforcement learning improves reasoning style but remains insufficient for accurate game-theoretic play. Motivated by these limitations, we propose ToolPoker, a tool-integrated reasoning framework that combines external solvers for GTO-consistent actions with more precise professional-style explanations. Experiments demonstrate that ToolPoker achieves state-of-the-art gameplay while producing reasoning traces that closely reflect game-theoretic principles. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_00528 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | How Far Are LLMs from Professional Poker Players? Revisiting Game-Theoretic Reasoning with Agentic Tool Use Lin, Minhua Dai, Enyan Liu, Hui Tang, Xianfeng Yan, Yuliang Dai, Zhenwei Zeng, Jingying Zhang, Zhiwei Wang, Fali Gao, Hongcheng Luo, Chen Zhang, Xiang He, Qi Wang, Suhang Artificial Intelligence As Large Language Models (LLMs) are increasingly applied in high-stakes domains, their ability to reason strategically under uncertainty becomes critical. Poker provides a rigorous testbed, requiring not only strong actions but also principled, game-theoretic reasoning. In this paper, we conduct a systematic study of LLMs in multiple realistic poker tasks, evaluating both gameplay outcomes and reasoning traces. Our analysis reveals LLMs fail to compete against traditional algorithms and identifies three recurring flaws: reliance on heuristics, factual misunderstandings, and a "knowing-doing" gap where actions diverge from reasoning. An initial attempt with behavior cloning and step-level reinforcement learning improves reasoning style but remains insufficient for accurate game-theoretic play. Motivated by these limitations, we propose ToolPoker, a tool-integrated reasoning framework that combines external solvers for GTO-consistent actions with more precise professional-style explanations. Experiments demonstrate that ToolPoker achieves state-of-the-art gameplay while producing reasoning traces that closely reflect game-theoretic principles. |
| title | How Far Are LLMs from Professional Poker Players? Revisiting Game-Theoretic Reasoning with Agentic Tool Use |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2602.00528 |