Bank of America Merrill Lynch (BAML) has launched its flagship algorithmic trading strategy, Instinct – an implementation shortfall algo that uses real-time signals and a quantitative model to calculate order tradability.
Unlike many previous algos, Instinct supports trading across varying market caps rather than a specific segment.
It has been specifically designed to take into account how a stock trades by using historical data ranging from previous weeks to the last few seconds of activity. Using some measures traditionally used by high-frequency traders, Instinct draws on recent market activity to help predict what might happen to a stock in the next 10 seconds.
The algo also carries out market impact analysis, offering users five urgency levels to achieve various trading objectives. Parameters for the algo can be fine-tuned on an order-specific basis.
“What makes this unique is its ability to handle such a broad range of stocks,” said Brian Schwieger, head of EMEA algorithmic execution at BAML. “It combines market impact and stock analysis with liquidity signals to choose the best strategy to execute the trade effectively, based on the nature of the stock – for example how liquid it is.”
Instinct is intended to be a streamlined, simplified algo that combines overlapping strategies to work in diverse market conditions, such as periods of high volatility. It can be used for dark liquidity aggregation or instant market access, for example.
“We've taken the complexity out of the front-end and moved it to the back-end so entering trades is much more intuitive,” said Ashok Krishnan, head of execution services, EMEA at BAML. “Traders still have these very innovative and powerful tools at their fingertips, but we've made them easier to use.”
The algo is currently available in the US, EMEA and Asia, and has been in use internally at BAML since the beginning of this year. BAML publicly introduced its Delta Adjusted algo to its US equity options suite in July 2011. The firm also previously introduced a pairs trading algo to its European algo suite, in April 2011.