ITG has rolled out a new algorithm that makes dark pool liquidity available to portfolio traders.
The company unveiled its Dark List Algorithm, a next-generation algorithm which brings the firm’s dark trading capacity to portfolio traders. While this is not the company’s first portfolio algorithm, it is one that has been designed specifically to locate the most difficult stocks on a trader’s list and trade them first.
How does it work?
Using a proprietary optimization technique, the algorithm is able to efficiently trade a list in POSIT Marketplace and POSIT Alert without creating cash imbalances or over-trading liquid names.
“Portfolio traders are hesitant to leverage the cost-saving potential of dark pools at scale because of concerns about unpredictable execution rates that can create cash imbalances or increase the risk of residuals,” said Ben Polidore, head of ITG algorithms at ITG. “ITG Dark List is designed to address these issues with its unique approach to statistical cash management and its distinct optimization objectives.”
Dark List lets portfolio traders control the optimization process to suit their needs by offering two objective functions: maximize execution rates under a cash constraint or use the opportunistic nature of dark pools to clean up only the most costly parts of the list. These objectives allow traders to use ITG Dark List as their primary trading tool or as a complement to risk bids or non-discretionary algorithms like VWAP.
Polidore explained to Traders that the algo looks at all the stocks on a portfolio trader’s list, then selects the most illiquid names and seeks to execute them first off board. Then, once the more difficult names have been traded, then the software begins trading the more liquid names.
“If you can’t trade – a sell or a buy – of the most difficult stocks on a list then executing the contra trade or hedge becomes very difficult or impossible,” said Polidore. “These lists are quite tricky to trade manually and thus we use an optimizer to decide the speed and opportunity in which to trade a stock or stocks.”
Without an optimizer or algorithm, a human trader would be quite challenged to execute the myriad trades required by his list, which could have 600 or more names on it.
Polidore said that the optimizer looks at the historical liquidity of a stock to determine how to trade it. This differs from the company’s other portfolio algorithm, Dynamic Implementation Shortfall (DIS), which runs risk cost optimization and uses smooth estimates of cost and risk when determining to trade a name.
“Dark List uses a probability based approach to determine a name’s liquidity,” he said. “It looks at the probability of execution (the normal distribution) while other algos look at average liquidity to determine forecast execution.”
Dark List will try to execute block trades whenever it can, Polidore added. The algo looks at the whole list, scans for liquidity in the dark pools first where focusing on less liquid stocks can be more easily found and that have the highest expected costs. The algo, Polidore said, might not find all the liquidity in a name.
ITG did research for approximately six months to fine tune the optimization logic and then three months to actually build the algorithm. All in all, it took a year from logic research to actual algo creation.
Dark List is currently available only in the U.S. and ITG plans to roll it out in Europe in the latter half of 2015.
The algo joins POSIT, POSIT Alert and POSIT Marketplace, helping traders source liquidity at an optimum price by reducing the friction that prevents list and single-stock traders from crossing with one another directly.
