The model's investment mandate is exclusively ETFs, constrained to holding 10 to 40 instruments. It incorporates a risk management feature allowing a tactical shift to Money Market ETFs during bearish periods. The primary optimization goal is to Maximize the Sharpe Ratio. This system was deployed live with actual capital, not simulation, and is fully automated using a low-level TensorFlow AI model.
Two TensorFlow-based AI models trading S&P 500 stocks were paper-tested from 21 February to 19 April 2024, using only daily price data alongside a broad market signal. The models systematically executed trades 15 minutes before close, with short signals processed as a switch to a money market ETF. The experiment concluded that AI can successfully beat the market index, suggesting it will empower active management over passive strategies.
The Options Open Gap Model was built on the idea that when a stock opens at an extreme value relative to its previous close, it tends to close the price gap soon after. The model uses daily and intraday stock prices, fair value, and options trading history to predict the reversion level. Once the reversion level is predicted, a deep out-of-the-money contrarian option is bought to be sold when the anticipated price reversion is reached.
The journey to autonomous asset management started in February 2019 when I got access to my first GPU.