Paper: arXiv 2408.06361

Abstract

Trading is a highly competitive task that requires a combination of strategy, knowledge, and psychological fortitude. With the recent success of large language models(LLMs), it is appealing to apply the emerging intelligence of LLM agents in this competitive arena and understanding if they can outperform professional traders. In this survey, we provide a comprehensive review of the current research on using LLMs as agents in financial trading. We summarize the common architecture used in the agent, the data inputs, and the performance of LLM trading agents in backtesting as well as the challenges presented in these research. This survey aims to provide insights into the current state of LLM-based financial trading agents and outline future research directions in this field.

Complexity vs Empirical Score

  • Math Complexity: 3.0/10
  • Empirical Rigor: 5.5/10
  • Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper is a survey discussing agent architectures and data inputs for trading, presenting low mathematical density but referencing backtesting results and implementation-heavy frameworks like reinforcement learning.

Research Flowchart

  flowchart TD
  A["Research Goal: Assessing LLM Agents<br>in Financial Trading"] --> B["Methodology: Literature Survey<br>Architecture & Data Synthesis"]
  B --> C["Data Inputs:<br>Market Data & News/Text"]
  C --> D["Computational Process:<br>LLM Agent Architecture<br>Analysis & Backtesting"]
  D --> E{"Outcome: Performance<br>in Backtesting"}
  E -->|Success/Failure| F["Key Findings:<br>Current State & Future Directions"]
  F --> G["Challenges:<br>Data, Regulation, &<br>Model Limitations"]