用人类专家设计的规则技能库，让大模型不训练也不调用求解器就能打职业级扑克。
PokerSkill: LLMs Can Play Expert-Level Poker without Training or Solvers
- 用分层规则技能库作为大模型的动作锚点，约束其决策范围。
- 在与顶级基准对战中，损失比默认提示降低49%至61%。
- 无需训练或求解器，适合想快速部署博弈智能体的研究者。
扑克是人工智能领域的标志性挑战。主流方法依赖基于反事实后悔最小化的均衡求解器，需数百万核时训练。大语言模型（LLMs）虽具备丰富扑克知识，但直接要求其下棋表现远逊于求解器驱动的代理。传统基于规则的扑克代理可解释且免训练，但策略上限远低于均衡水平。我们提出 extbf{PokerSkill}，一种免训练、免求解器的框架，通过人类专家设计的详细规则技能库作为结构化动作锚点，引导LLM做出合理决策。一个确定性上下文引擎分析当前状态，仅提取相关技能片段，从而限制LLM的选择空间。在与顶尖基准GTOWizard对抗中，GPT-5.5 XHigh使用PokerSkill取得 $-57 \pm 21$ mbb/hand，Claude Opus 4.6为 $-80 \pm 29$ mbb/hand，Claude Opus 4.7为 $-87\pm 64$ mbb/hand，相比默认提示基线减少49%–61%损失，优于强敌Slumbot。关键发现：仅靠规则技能无法构成强策略，仅靠大模型也无法下好牌，但二者结合可生成无需训练或求解器访问却能媲美数百万核时计算系统的智能体。据我们所知，这是首个展示大模型在复杂不完美信息博弈中无游戏特定训练或求解查询即达竞争力的实例。代码见 https://github.com/lbn187/PokerSkill。
原文摘要 · Abstract (English)
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 \textbf{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 $-57 \pm 21$ mbb/hand, Claude Opus 4.6 achieves $-80 \pm 29$ mbb/hand and Claude Opus 4.7 achieves $-87\pm 64$ 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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。