3S-Trader: A Multi-LLM Framework for Adaptive Stock Scoring, Strategy, and Selection in Portfolio Optimization

2025-10-20

AI Summary

Existing large language models (LLMs) for stock trading predominantly focus on single-stock decision-making, lacking capabilities for multi-asset portfolio construction and dynamic strategy adaptation, resulting in limited generalizability. This paper proposes a training-free, multi-LLM collaborative framework comprising three core modules: scoring, strategy generation, and stock selection. It achieves adaptive portfolio optimization and real-time strategy iteration through cross-dimensional signal fusion—including market regime inference, historical strategy analysis, and multimodal financial data summarization. Notably, this is the first framework to enable dynamic, cross-horizon strategy updates and joint evaluation of multiple candidate stocks without any fine-tuning. Empirical evaluation on Dow Jones Industrial Average constituents yields a cumulative return of 131.83%, a Sharpe ratio of 0.31, and a Calmar ratio of 11.84—substantially outperforming baseline methods.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Large Language Models (LLMs) have recently gained popularity in stock trading for their ability to process multimodal financial data. However, most existing methods focus on single-stock trading and lack the capacity to reason over multiple candidates for portfolio construction. Moreover, they typically lack the flexibility to revise their strategies in response to market shifts, limiting their adaptability in real-world trading. To address these challenges, we propose 3S-Trader, a training-free framework that incorporates scoring, strategy, and selection modules for stock portfolio construction. The scoring module summarizes each stock's recent signals into a concise report covering multiple scoring dimensions, enabling efficient comparison across candidates. The strategy module analyzes historical strategies and overall market conditions to iteratively generate an optimized selection strategy. Based on this strategy, the selection module identifies and assembles a portfolio by choosing stocks with higher scores in relevant dimensions. We evaluate our framework across four distinct stock universes, including the Dow Jones Industrial Average (DJIA) constituents and three sector-specific stock sets. Compared with existing multi-LLM frameworks and time-series-based baselines, 3S-Trader achieves the highest accumulated return of 131.83% on DJIA constituents with a Sharpe ratio of 0.31 and Calmar ratio of 11.84, while also delivering consistently strong results across other sectors.

Problem

• Addressing inflexibility in stock trading strategies during market shifts
• Overcoming single-stock focus limitations in portfolio construction
• Integrating multidimensional stock scoring with adaptive strategy generation

Innovation

• Multi-LLM framework integrates scoring, strategy, and selection modules
• Scoring module summarizes stock signals across multiple dimensions
• Strategy module iteratively generates optimized selection strategies