This paper proposes a hybrid model-first AI architecture for options trading, where LLMs are used to construct models rather than make trading decisions directly.

Applied to the options wheel strategy, the framework generates transparent signals based on 27 decision factors and delivers strong historical performance, with 15.3% annualized returns, a Sharpe ratio of 1.08, and a maximum drawdown of −8.2% over the period 2007–2025.

Comparisons with alternative approaches show that the hybrid architecture outperforms pure LLM strategies, static Bayesian networks, and rules-based systems.

Want more research breakdowns like this? Subscribe for weekly insights.