OverMind is a series of deep machine learning algorithms running on an artificial neural network built by Logiic's technical team on customized hardware. These algorithms were then “seeded” with over 40 years of historical price and market data, and provided with broad guidelines for finding profitable trading patterns that deliver a consistent volatility-adjusted return if consistently applied over a long term horizon.
After analyzing this vast corpus of data (more than 1 billion discrete data points), OverMind was able to identify specific, robust trading signals across different asset classes. Key features of these signals were their remarkable persistence over long periods of time across the economic cycle, and their liquidity – they were expressed in highly liquid asset prices where trades could be quickly made and profitably unwound in a short (often seconds’) time frame.
OverMind is currently connected to various trading platforms globally, and sends trade execution orders accordingly when the relevant signals appear in daily trading activity and market pricing. As markets/events move, the latest data is fed back into OverMind to constantly refine and backward test the validity of these signals, under human supervision in an iterative process.
Currently, OverMind takes in almost 250,000 data points a day, and is capable of executing up to 15,000 trades a second across multiple exchanges worldwide.