Goldman Sachs is using neural networks and other AI tools to cut slippage in stock trading, aiming to better reflect intraday trends and correlations in global equities markets. Initial use of the technology has improved trade execution quality by 50% over the past six months, the bank claims. This improvement is measured against client benchmark expectations and single-stock alternatives.
The bank says the effort will help minimize slippage during completion of a trade. An unsupervised machine learning method called clustering is being used to create the models. “The models have improved trade execution by about 50% across all the trades that we’ve done in the last six months versus alternatives,” says Michael Steliaros, global head of quantitative execution services speaking to Risk.