Options Selling Using Machine Learning

The goal of this paper is to develop a dynamic standalone option selling strategy using technical indicators, options Greeks and machine learning optimization. Options trading is hugely attractive due to the inherent leverage available to take view on market direction. We primarily focus on the options selling side, whose risk profile is akin to selling insurance. Options Sellers typically try to earn the “time value” of the options, i.e. the decay of option’s prices with time. We base our options selling strategy on the structure: “Strangle” . To optimize the strategy for the market trends and reversions, we make use of technical indicators, which have been shown to have significant predictive power with regards to market returns. Furthermore, we use options Greeks to take into account the risks associated with the options selling structure. A machine learning model is used to come up with an improved options selling strategy over the vanilla options selling strategy. The options selling strategy outperforms the baseline options selling strategy in both Sharpe ratio and monthly returns. In addition to the strategy implemented, this work provides a robust framework to develop robust machine learning optimized options trading strategies.