ITG Algo Finds Dark Canadian Liquidity06.21.2017 By John D'Antona
Trade inter-listed stocks?
That is the buying and selling of stocks that trade in both the U.S. and Canada. There is more to it than simply foreign exchange conversions – just ask any buy-side trader. Trading inter-listed stocks is fraught with unique complexities – different market structures in the US and Canada, different trading philosophies and venue discrepancies – but it can be done.
Enter ITG with its completely revamped and basically new-from -the-ground-up POSIT Marketplace Canada algorithm, which can be used to trade all 3000+ Canadian symbols as well as the almost 200 inter-listed stocks. Earl Cummings, head of Liquidity Management for ITG Canada spoke with Traders Magazine about the new electronic strategy that has incorporated machine learning into its approach. The goal – enhancing the buy-side trading experience – but how? As Cummings tells it, the new POSIT Marketplace reduces market footprint and adverse selection, increases transparency, finds alpha via new natural liquidity opportunities and helps efficiency, by looking at historical dark executions and extrapolating them into future trades.
First, the algorithm taps not just into the three existing Canadian dark venues – MATCHNow, Nasdaq CXD and Instinet CBX -but also routes to over 30 dark pools within the U.S. that trade inter-listed securities, including ITG’s own POSIT. The algo also accesses dark or “gray” order types on Canada’s lit exchanges – to find the natural contra.
“The idea is to maximize our liquidity seeking potential and minimize the parties that a buy-sider would rather not interact with,” Cummings said. “We want our clients to avoid negative outcomes and leverage dark marketplaces in an intelligent fashion to produce the best execution.”
The secret sauce here is in the usage of machine learning. Cummings explained that the algorithm now examines past dark executions– calculates the pros and cons of each, and then carefully looks at what a trader currently wants to achieve and apply what it has learned. In essence, the algorithm gets smarter with each trade as outcomes add to its depository of historical data from which it draws conclusions.
According to Canadian market data, TSX has just over 1500 listings and TSX Venture another 1600 listing, bringing the universe of tradable stocks to approximately 3100, of which approximately 190 are interlisted. These inter-listed stocks account for 41% of Canadian volume and 62% of value traded, Cummings added.
The POSIT Marketplace Canada routing also has anti-gaming logic built in as well as stock-specific minimum fill quantities based on advanced research.
And of course, it looks for blocks whenever it can, Cummings said.
The new algo took approximately six months to build, test and deploy. And, in order to help traders get the absolute best execution, it will automatically do all the necessary currency calculations on inter-listed orders to improve trading performance.
“The thinking here is to promote intelligent sourcing of dark liquidity in Canada since our market is somewhat behind the advancement in the U.S.,” Cummings said. “In the past, there has been little requirement for aggregation but now that is beginning to change with the two new dark venues and our exchanges adding dark order types. Our goal is to help traders get their orders filled using all accessible liquidity especially for inter-listed orders. .”
FXCM's leader looks back on FX market evolution and the changes ahead.
Buy-side says algo offerings determine order flow destination.
Exec discusses the fintech behind algos and 'best ex'.
IBM dips its toe into algorithmic trading.
Clients can lower funding requirement via initial margin optimization.