TradingAgents: Multi-Agents LLM Financial Trading Framework

2024-12-28

AI Summary

To address the limitations of single-model decision-making and inadequate risk management in financial trading, this paper proposes a multi-role large language model (LLM) agent collaboration framework. Inspired by real-world trading firms’ organizational structures, it designs specialized agents—including fundamental, sentiment, and technical analysts; risk-aware traders with heterogeneous risk preferences; and bull/bear researchers—coordinated via structured debate-and-negotiation mechanisms, a dynamic position-level risk control module, and multi-source market data fusion to enable closed-loop trading decisions. The framework introduces a novel collaborative paradigm that integrates hierarchical risk governance and real-time adversarial deliberation directly into the trading workflow. Empirical evaluation demonstrates substantial improvements over single-agent baselines: +23.6% cumulative return, +31.4% Sharpe ratio, and −18.9% reduction in maximum drawdown.

Technology Category

• Multiagent Systems: Adversarial Agents
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• 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: Architectures and workflows that use LLMs for crowd work
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Significant progress has been made in automated problem-solving using societies of agents powered by large language models (LLMs). In finance, efforts have largely focused on single-agent systems handling specific tasks or multi-agent frameworks independently gathering data. However, multi-agent systems' potential to replicate real-world trading firms' collaborative dynamics remains underexplored. TradingAgents proposes a novel stock trading framework inspired by trading firms, featuring LLM-powered agents in specialized roles such as fundamental analysts, sentiment analysts, technical analysts, and traders with varied risk profiles. The framework includes Bull and Bear researcher agents assessing market conditions, a risk management team monitoring exposure, and traders synthesizing insights from debates and historical data to make informed decisions. By simulating a dynamic, collaborative trading environment, this framework aims to improve trading performance. Detailed architecture and extensive experiments reveal its superiority over baseline models, with notable improvements in cumulative returns, Sharpe ratio, and maximum drawdown, highlighting the potential of multi-agent LLM frameworks in financial trading.

Problem

• Financial Transactions
• Decision Improvement
• Risk Reduction

Innovation

• TradingAgents
• Multi-agent Cooperation
• Financial Trading Optimization