Search papers, labs, and topics across Lattice.
This paper introduces Fin-Analyst, a hybrid trading agent that integrates eight specialist LLMs with rule-based signals to optimize trading strategies for Tesla (TSLA) and Bitcoin (BTC). The agent achieved a remarkable +13.51% return on TSLA, significantly outperforming a Buy-and-Hold strategy, while also revealing the limitations of fixed-threshold rules in volatile markets. Key findings include the dominance of event-driven signals from 8-K disclosures for TSLA and the need for a memory-aware successor to improve BTC trading performance.
Fin-Analyst's innovative LLM hybrid approach outperformed traditional strategies, achieving a +13.51% return on Tesla while exposing critical flaws in rule-based trading for Bitcoin.
Large language model (LLM) trading agents show promising performance in equity markets, yet remain narrowly focused on US equities with little evidence from live deployment. We present Fin-Analyst, a hybrid agent for FinMMEval 2026 Task 3: an eight-specialist LLM pipeline over news, SEC filings, fundamentals, analyst forecasts, technical indicators, and social sentiment, aggregated by a Meta-Agent for Tesla (TSLA), and a lightweight rule based three-signal vote for Bitcoin (BTC). On the final official leaderboard (accessed 2026-07-05), Fin-Analyst ranks first of all agents on TSLA with a +13.51% return, +28.33 points over Buy-and-Hold (Sharpe 4.10, 88% win rate), while the BTC vote ends flat yet well above a sharply falling baseline. Relative to the interim performance, the asset ranking reversed, indicating that short live windows yield volatility-sensitive rankings. Ablation identifies event-driven 8-K disclosures as the most influential TSLA signal. Error analysis shows that the memoryless agents repeat wrong calls for days at a time, and that the fixed-threshold BTC rules lost money by trading on noise in a sideways market while the LLM pipeline gained under similar conditions, motivating a memory-aware, LLM-based successor for both assets.
