Sandford Bernstein’s Shadow and Ninja strategies have the highest proportion of non-displayed fills. All of Bernstein’s algorithms have the ability to interact with non-displayed liquidity.
Bernstein maintains a connection to most major sources of dark liquidity regardless of their perceived level of toxicity. Venue quality is scored in real time selectively removing venues from the routing logic. Bernstein also recommends individual venues to be removed completely from the routing decision based on the style of trading.
Bernstein’s algorithms operate simultaneously across multiple non-displayed venues. The routing decision, venue selection and order type and size utilised is based on numerous factors. These include historic fill activity, recent inter-spread trading activity, short-term pricing models, short-term volatility, forecast slippage, current perceived gaming activity and chosen algorithmic strategy.
Bernstein algorithms have various types of anti-gaming techniques, including a fair price distribution and quality score for each child order based on recent price series to determine the price of the limit order. They also use pattern recognition to detect gaming activity and game theory to reduce risk. The user can control minimum fill in dark venues and relative price limits.
Many of Bernstein’s single stock algorithms, including Ninja and Shadow can be linked so that multiple orders can be traded while maintaining the cash ratio balance. Bernstein aims to provide full transparency on the execution and custom analytics. Complex customisation, auto switching between strategies and automatic filtering into strategies based on various criteria are available.