General overview:
What is the FX algo called?
Nightjar, Societe Generale’ flagship algo aiming at the highest level of discretion while mixing passive orders with hit orders.
What category does it fall into?
Opportunistic and participation-of-volume, for efficient cost of execution and dynamic impact management.
What does it attempt to do?
The Nightjar algorithm places passive, mid-book and aggressive orders with the objective to trade at mid-market in an undetectable fashion. One of the main objectives, being stealthy like the eponymous bird, is achieved by following live market volume curves to control the noise of the market participation, while the skew safeness of the child orders is tested as soon as practically possible.
Structure:
What is the algo’s software architecture?
Societe Generale uses low latency components written in C++ or C# that are deployed in the major colocation centers for the FX markets, to provide its customers the highest chances of getting filled. These software components can operate on a 24/5 mode, providing a full coverage of the FX markets.
Does it use proprietary modelling?
Yes. Like all the Societe Generale algos, the Nightjar uses proprietary modelling for pricing, order placement, pacing control signals, etc. Keeping the full control of the modelling enables the Nightjar to be very agile.
Does it use technology such as AI or ML? If so, how?
Yes, with moderation. Machine Learning is used to help the Nightjar guess what the market trading activity is, and is going to be, to control its market impact.
Functional aspects:
Does the algo adapt automatically to prevailing market conditions and if so how?
Yes, it does. The Nightjar has the capacity to detect slow or high market activity and adapt its target pacing accordingly, for any currency pair, at any time during the day.
Does it incorporate smart order routing?
Yes! Due to the highly fragmentation of the FX markets, it is logical to use some smart order routing tools. The Nightjar is able to skim cost efficient liquidity in all kinds of markets, by smartly interacting with a large and diversified pool of execution.
How does it minimise market footprint?
The Nightjar considers two aspects of the market footprint: the price impact generated by consuming liquidity and the information leaked by its child orders. FX is not different than the other asset classes, a large order execution usually results in a noticeable price movement, and the nightjar looks at minimizing this footprint by controlling its participation in the market. The child orders visibility is closely monitored and controlled to avoid adverse market movements.
What liquidity seeking and access capabilities does it deploy?
The Nightjar looks for liquidity in the primary and secondary venues, from Liquidity providers and from mid-market execution venues. When operating in the so-called lit venues, the Nightjar deploys some tricks to avoid being detected.
What operational risk management does it include?
The Nightjar, like all SG FX algorithmic orders, has several controls in place to protect its client electronic order flow from external negative events, including the following, but not exclusively:
a. Fat finger limit, pacing control,
b. Per exchange venue circuit breakers,
c. Per child order price and size controls,
d. Per venue aggregated notional value and throughput limits.
Parameters and Controls:
What client inputs are available in the algo?
Beside product (spot or NDF), direction, size, start and end times, liquidity pool, our customers can control these Nightjar specific parameters:
– Speed: slow, normal or aggressive.
– Alpha seeker: an option to speed-up or slow-down when the asset price over- or under-perform a recent average.
During the execution, our customers can change most of these parameters, pause and resume the Nightjar, and even switch to another algo type from SG’ algo suite.
How much real-time feedback does it provide?
The Nightjar reports in real-time its execution status (quantity and average price, running/paused, etc.), its child trades through the FIX API and the standard order management platforms. Its live orders that are also visible through our live TCA, a simple but efficient tool for visualizing the activity of the Nightjar.
Capabilities and Use:
What execution styles (e.g. passive/aggressive) does the algo support?
The Nightjar uses passive, aggressive and mid-market execution styles. For cost efficiency, it mainly posts passive or mid-book orders, and sometimes aggressive orders when skewed prices at mid-market or better are available.
How can it be integrated/called with/by higher-level workflows?
The Nightjar is available via FIX and FSS, so that it can be integrated in most systematic or semi-automated execution workflows.
What is the optimal scenario for its use?
The Nightjar is a stealthy execution strategy, recommended when discretion is more important than speed. It is a very versatile tool that is used for large executions over a day or a few hours, at best orders, and small to medium size orders over a few minutes.
Any other functionality worthy of note?
We like so much its ability to trade at mid-market that we have embedded the Nightjar in our TWAP+ algo, to improve spread capture, and it works! Our customers appreciate the capacity to control the end time of the TWAP+ order while benefitting from the agility of the Nightjar.
