Abstract: Proposes a multi-agent LLM trading framework that decomposes investment analysis into fine-grained tasks. Evaluated on Japanese stock data, the framework shows that fine-grained task decomposition significantly improves risk-adjusted returns compared to coarse-grained instructions.
Key Insight: Alignment between analytical outputs and downstream decision preferences is a critical driver of system performance. Decomposing abstract roles (like "manager") into specific data-driven tasks reduces reasoning degradation and increases transparency.
Relevance to RSI: Demonstrates the "Structural RSI" approach—improving system capability by optimizing the coordination graph and task granularity rather than just increasing model scale. Crucial for agentic workflows in high-stakes domains.