Agency brokerage Bloomberg Tradebook has enhanced B-SMART, its algorithm for placing passive orders at multiple trading venues, by adding a price predictor function that automatically controls the algo’s aggression level.
The new function, B-SMART AUTO, seeks to statistically mirror the actions of a trader by assessing a stock’s price dependent on factors such as activity on the tape, volatility of the stock, bid/offer spread, as well as peer group and sector analysis.
B-SMART AUTO consists of three underlying aggression levels: passive react mode; normal trading and aggressive liquidity capture, which are reliant on prevailing market conditions and predicted price movements. It also reassesses which venues are most active, and places lit and dark orders in those venues in an effort to optimise spread capture and to help clients avoid being traded around.
“The price predictor model seeks to create a clone of the trader by making statistically-based decisions of what the trader would likely do if he or she was trading the stock,” said Gary Stone, director of trading research at Bloomberg Tradebook, in a statement.