Low-latency liquidity inefficiency strategies

The vast amount of high-frequency data heralds the use of new methods in financial data analysis and quantitative trading. This study delivers a proof-of-concept for a high frequency-based trading system based on an evolutionary computation method. Motivated by a theoretical liquidity asymmetry theorem from the market microstructure literature, grammatical evolution is used to exploit volume inefficiencies at the bid–ask spread. Using NASDAQ Historical TotalView-ITCH level two limit order book data, execution volumes can be tracked. This allows for testing of the strategies with minimal assumptions. The system evolves profitable and robust strategies with high returns and low risk.