Pipeline Enhances Prediction Engine For Best Performing Algo
June 4, 2009
Pipeline Trading Systems has upgraded its Algorithm Switching Engine to give block traders a greater velocity range and more control over optimizing aggression levels as they execute their orders through direct market access platforms to over 40 different venues.
The enhancements to the Algorithm Switching Engine allow traders to take up to 50 percent of the liquidity of a given stock as opposed to 38 percent of typical liquidity allowing them to improve the execution quality of their more urgent trades, according to pipeline.
When executing a giant trade, capturing liquidity opportunities often takes precedence over short-term price improvement, said Henri Waelbroeck, Pipelines director of research. By combining predictive switching with access to high speed trading across all significant dark pools and displayed markets, the buy-side trader can maximize liquidity capture at significant points on the chart.
Algorithm Switching Engine was introduced in October 2007, six months after Pipeline Trading banned third-party algorithms from accessing its own electronic block trading market.
The Algorithm Switching Engine analyzes real-time and historical data to predict which of 120 algorithms is most effective for the trader in the next minute. If the algorithm is predicted to do poorly, the Algorithm Switching Engine will switch out of that algorithm beforehand into another one that will perform better in the next minute.
That means that Pipelines customers have the choice of waiting for match in Pipelines dark pool or using the Algorithm Switching Engine to send their order to other execution venues. Pipeline claims to be more effective than competitors in finding block matches because of its 50,000-share average execution size; large block orders are executed automatically without the possibility of sniffing out institutional interest with a small probing order.
According to Pipeline, the predictive switching offered by its Algorithmic Switching Engine reduces implementation shortfall a measure of total transaction cost -- by 30 percent from using a single algorithm to execute a large order.