Paper: arXiv 2603.16434

Authors: Haochen Luo, Zhengzhao Lai, Junjie Xu, Yifan Li, Tang Pok Hin, Yuan Zhang, Chen Liu

Abstract

Large Language Models (LLMs) excel at general code generation, yet translating natural-language trading intents into correct option strategies remains challenging. Real-world option design requires reasoning over massive, multi-dimensional option chain data with strict constraints, which often overwhelms direct generation methods. We introduce the Option Query Language (OQL), a domain-specific intermediate representation that abstracts option markets into high-level primitives under grammatical rules, enabling LLMs to function as reliable semantic parsers rather than free-form programmers. OQL queries are then validated and executed deterministically by an engine to instantiate executable strategies. We also present a new dataset for this task and demonstrate that our neuro-symbolic pipeline significantly improves execution accuracy and logical consistency over direct baselines.

Complexity vs Empirical Score

  • Math Complexity: 2.5/10
  • Empirical Rigor: 6.0/10 (capped from 6.5: transaction costs not modelled; see penalty flags)
  • Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper introduces a highly novel neuro-symbolic framework (OQL) bridging natural language and option strategy execution, offering strong practical utility. While heavy mathematical formalisms are minimal, the empirical evaluation using a specialized dataset and deterministic validation engine demonstrates solid rigor.

Research Flowchart

  flowchart TD
    A[Research Goal: Accurately Translate Natural-Language Options Strategy to Executable Code] --> B[Inputs: NL Trading Intent & Multi-Dimensional Option Chain Data]
    B --> C[LLM Neuro-Symbolic Parser: Generates OQL Intermediate Representation]
    C --> D[Deterministic Execution Engine: Validates OQL Rules & Queries Market Data]
    D --> E[Outcome: High Logical Consistency & Executable Option Strategy]
    D --> F[Outcome: Benchmark Dataset & Superior Accuracy over Direct Code Baselines]

Cited References (56)

  1. Fine-Tuning Large Language Models for Stock Return Prediction Using Newsflow — → 2407.18103
  2. DeepSeek-V3 Technical Report (2025) — link
  3. Chain-of-thought prompting elicits reasoning in large language models (2022)
  4. EvoAlpha: Evolutionary Alpha Factor Discovery with Large Language Models (2026) — link
  5. Codemirage: Hallucinations in code generated by large language models (2024) — link
  6. Qwen3 Technical Report (2025) — link
  7. Options, futures, and other derivatives (2016)
  8. The Llama 3 Herd of Models (2024) — link
and 48 more
  1. Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities (2025) — link
  2. GPT-4 Technical Report (2024) — link
  3. Finqa: A dataset of numerical reasoning over financial data (2021)
  4. Can large language models be query optimizer for relational databases? (2025)
  5. Cfinbench: A comprehensive chinese financial benchmark for large language models (2025)
  6. Finben: A holistic financial benchmark for large language models (2024)
  7. Pixiu: A comprehensive benchmark, instruction dataset and large language model for finance (2023)
  8. Learning to generate explainable stock predictions using self-reflective large language models — → 2402.03659
  9. AlphaBench: Benchmarking Large Language Models in Formulaic Alpha Factor Mining (2025) — link
  10. Investorbench: A benchmark for financial decision-making tasks with llm-based agent (2025)
  11. XF in B ench: Benchmarking LLM s in Complex Financial Problem Solving and Reasoning (2025) — link
  12. Can Large Language Models Mine Interpretable Financial Factors More Effectively? A Neural-Symbolic Factor Mining Agent Model (2024) — link
  13. Fingpt: Democratizing internet-scale data for financial large language models (2023) — link
  14. Bloomberggpt: A large language model for finance (2023) — link
  15. FinCon: A Synthesized LLM Multi-Agent System with Conceptual Verbal Reinforcement for Enhanced Financial Decision Making (2024) — link
  16. LLMF actor: Extracting Profitable Factors through Prompts for Explainable Stock Movement Prediction (2024) — link
  17. Finbert: A pre-trained financial language representation model for financial text mining (2021)
  18. Stock price prediction using llm-based sentiment analysis (2024)
  19. Deep Learning for Options Trading: An End-To-End Approach (2024)
  20. Finmem: A performance-enhanced llm trading agent with layered memory and character design (2025)
  21. TwinMarket: A Scalable Behavioral and Social Simulation for Financial Markets (2025) — link
  22. Low-Latency Machine Learning for Options Pricing: High-Speed Models and Trading Performance (2025)
  23. Options: trading strategy and risk management (2011)
  24. Deltahedge: A multi-agent framework for portfolio options optimization — → 2509.12753
  25. Machine learning for quantitative finance: fast derivative pricing, hedging and fitting (2018)
  26. Option pricing using machine learning (2021)
  27. Time Travel in LLM s: Tracing Data Contamination in Large Language Models (2024) — link
  28. Kelly-based options trading strategies on settlement date via supervised learning algorithms (2022)
  29. Machine learning in finance: the case of deep learning for option pricing (2017)
  30. Time Travel is Cheating: Going Live with DeepFund for Real-Time Fund Investment Benchmarking (2025) — link
  31. EFS: Evolutionary Factor Searching for Sparse Portfolio Optimization Using Large Language Models — → 2507.17211
  32. Navigating the Alpha Jungle: An LLM-Powered MCTS Framework for Formulaic Factor Mining (2025) — link
  33. Alpha-gpt: Human-ai interactive alpha mining for quantitative investment (2025)
  34. Alphaagent: Llm-driven alpha mining with regularized exploration to counteract alpha decay (2025)
  35. The pricing of options and corporate liabilities (1973)
  36. ST a R - SQL : Self-Taught Reasoner for Text-to- SQL (2025) — link
  37. Adaptive and Robust Translation from Natural Language to Multi-model Query Languages (2025) — link
  38. F inance R easoning: Benchmarking Financial Numerical Reasoning More Credible, Comprehensive and Challenging (2025) — link
  39. Next-generation database interfaces: A survey of llm-based text-to-sql (2025)
  40. DCG - SQL : Enhancing In-Context Learning for Text-to- SQL with Deep Contextual Schema Link Graph (2025) — link
  41. Decomposition for enhancing attention: Improving llm-based text-to-sql through workflow paradigm (2024)
  42. Benchmarking the text-to-sql capability of large language models: A comprehensive evaluation (2024) — link
  43. Ehrsql: A practical text-to-sql benchmark for electronic health records (2022)
  44. Text-to-sql empowered by large language models: A benchmark evaluation (2023) — link
  45. When is Tree Search Useful for LLM Planning? It Depends on the Discriminator (2024) — link
  46. Finbert: Financial sentiment analysis with pre-trained language models (2019) — link
  47. A Multi-Level Sentiment Analysis Framework for Financial Texts (2025) — link
  48. TradingAgents: Multi-Agents LLM Financial Trading Framework (2025) — link

4 of 56 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.