Algos Dip Into Multiple Dark Pools
Smarter Posting
Dark Inside also pursues a "smart posting" strategy - a concept outlined by Nenad Yashruti, head trader at Freestone Capital Management, in a related article, page 35 - that is similar to the strategy employed by EdgeTrade's FAN. Essentially, smart posting involves splitting an order into smaller "child" orders that match the best bid on each chosen venue or "hide" in reserve at the next-best price. If an order is executed, the algorithm is guaranteed at least to match the National Best Bid and Offer rather than force the algorithm to chase the best price by trying to cancel entire orders on one venue and move them to another. At that point, the time advantage starts to break down.
Smart posting protects the trader because each child order is small enough not to give away the size of the entire order, thereby concealing the trader's intentions and lowering the risk of resulting market movement. Additionally, smart posting also may reduce trading fees because ECNs charge for taking liquidity (i.e., buying shares), but not for posting liquidity (i.e., adding stocks for sale).
Yashruti relates that he is in the process of beta-testing Dark Inside. "When you think about posting more bids, if every algorithm did it, liquidity would be revitalized in the market," he says. "There is so much sitting in reserve. Other systems have sweeping routers, where they will go in and sequentially probe the spreads of all the exchanges one by one. But what they were failing to do was literally post bids and offers, to represent and protect me everywhere I trade." The issue of over-execution - getting stuck buying the same shares in two different markets at once - which was somewhat common with electronic orders in the past, largely is avoidable with this new generation of algorithms, so long as they both post bids and buy shares, Yashruti adds.
Venues that mire orders with slow response times or that simply do not turn up enough liquidity to satisfy customers eventually will be selectively skipped by this generation of smart algorithms, EdgeTrade's Zasky asserts. But in the meantime, FAN is fanning out to every venue that will accept it, he notes.
"Our job is to connect to all the relevant venues," Zasky says. "Some destinations may perform better in certain areas - say a certain exchange is terrible with exchange-traded funds and good with small caps. We database all this historical information so we can have a smarter approach in advance of the order. We don't say you have to be the best performer on all of the metrics."
Although the bombardment of new algorithm sales pitches makes it difficult to determine whether the latest algorithm really offers anything new, Freestone Capital's Yashruti says FAN and Dark Inside really do represent a step forward, ahead of the competition. "A lot of the bulge-bracket firms' next-generation algos are barely intelligent enough to trade on multiple exchanges at the same time," Yashruti contends. "They can sweep, but not really dance on multiple venues."
Still, like all other algorithms, FAN and Dark Inside are not likely to serve all of a trader's liquidity needs, for neither algorithm will be able to access both all public markets and all dark markets. In some cases, the dance hall doors simply are closed.
Tweet This