Pragma launches execution algorithms with deep-learning capabilities for equity trading
Pragma, an independent algorithmic trading technology provider, has launched a new generation of deep-learning enabled execution algorithms.
Pragma initiated the project in 2018 to see if deep neural networks could be applied to an execution algorithm’s micro-trading, engine–governing decisions such as the routing, sizing, pricing and timing of orders – and deal with complex multi-dimensional trading challenges more effectively.
Following a beta launch in 2020, Pragma managed a number of controlled trials with its clients. It observed a significant improvement to execution quality, with an average shortfall improvement of 33 per cent to 50 per cent across billions of traded shares.
“If you just want to add an ‘AI label’ to a product, it is easy to do so,” says David Mechner, Co-Founder and CEO of Pragma. “But getting a material, measurable benefit in average shortfall requires a lot of ingenuity. It took us a year and a half of intense research and development before we had our first version in production.
“The benefit of AI models is that it allows us to better tailor our algorithms’ use of dynamic real-time signals and market conditions to the complex, multi-dimensional interactions of stock characteristics and order requirements. There is no one-size-fits-all approach to trading, and simple stylised models just can't match the power of a deep neural network.
“2020 was a great proving ground for these algorithms. We have a number of enhancements planned in 2021, which we expect will bring even more benefits to our clients.”