The next generation of FX algos: pushing the limits of digital trading

This article was first published in FX Algo

There’s always a new frontier in FX algos, how do you stay ahead?

Part of what makes FX algo execution so exciting is that we’re always pushing the limits of the technology. We’re driven by advances in the tech, which create opportunities to redefine what’s possible, and at the same time our clients are becoming more and more sophisticated in the way they use algo execution, their expectations are higher all the time.

We’ve always looked to place ourselves at the cutting edge here. If you look back to 2013, algos were operating on a TWAP/VWAP basis and only the most advanced were able to manage fragmented liquidity. That’s when BNP Paribas came into the market with CORTEX iX, launching what we called the ‘3rd generation’ of adaptive FX algos. These actually read market conditions and adjust execution in real time, which was a game changer at the time.

One benefit of this was that, with intelligent execution we could maintain a simple offering of just three algos – Chameleon, Viper and Iguana. Rather than adding new algos for every market condition, we have just three, which can adapt their behaviour.

This was a practical application of AI which took algo execution in a whole new direction; rather than just picking up an algo and hoping for the best, clients could have confidence that the algo would respond to what was going on in the markets and manage their execution intelligently. It was a radical shift that made the world of algo execution much more accessible.

And what differentiates the latest generation of algos?

Control and interactivity are essential now and we’ve really driven the technology in this direction. As clients get more accustomed to algo execution, they are developing more sophisticated strategies which require a closer understanding of what drives the algo’s decision making, plus features to monitor and intervene during execution. As a response we’ve developed the 4th generation of interactive algos which adapt to markets like their predecessors, but bring new levels of control.

This capacity is really driven by data. Historically clients have lacked access to market data and
struggled to gain visibility on the workings of their algo execution, let alone interact with them. We countered this by providing data not only post-execution through TCA but in pre-trade via our ‘Insight’ tool, which lets clients simulate algo performance in different market conditions ahead
of execution, and adjust their strategy ahead of time.

But we didn’t stop there; what really differentiates our 4th generation, interactive algos is the ability to deliver data during execution. We achieve this with a new tool called ‘Insight Live’.

How does ‘Insight Live’ deliver control to clients?

‘Insight Live’ is a real time market intelligence portal offering true transparency and control of algo execution. It delivers analysis and interactivity through the whole trade lifecycle – before, after, and during the execution of a trade, which means clients can watch a live breakdown of the order, track execution performance against benchmarks, and understand the environment using live market data.

Monitoring liquidity in real time, they can track market conditions against historical values to fully
understand the liquidity regime in the moment – it’s like having an ‘algo-eye view’ of the market. Users can then leverage this information to track how an order is completed with live venue statistics, and use interactive tools to adjust an order mid-flight.

For example: having set up a strategy and deployed an algo, users can leverage Insight Live’s real time market intelligence to decide whether to use some of the interactive features such as ‘Get me out’, Iguana’s Speedometer, or Chameleon’s Rapid Fill, that let them adapt, adjust risk parameters, and even morph algorithm midexecution.

This way the user can choose to leave the algo running, applying its advanced AI to complete their
order, or they step in and apply their own market view, pursuing alpha by orientating the algo into specific venues, speeds, behaviours and more.

That sounds like a lot of data to manage, how do you help your clients handle this?

It’s a complex environment for sure and that’s why we have such a focus on simplicity. This was the case from the very beginning when we launch with only three algos. As we’ve built on the platform we’ve added features to them, to enhance their performance and interactivity, but the basic characteristics of each has stayed the same.

When it comes to interactivity we’ve kept this approach. FX markets are complex and fast moving, there are so many factors to consider, and the market is only getting more fragmented, so we knew that while clients wanted greater transparency and the ability to interact, this would have to be delivered in an accessible and user friendly way.

So we’ve taken a leaf from Silicon Valley’s book and have developed the first digital execution assistant, which we’ve called AliX. Working alongside ‘Insight Live’, AliX delivers a running commentary on execution, flagging progress, key events and so on, and responding to requests from the client, guiding them through the complex data landscape. This is the ultimate in interactivity, as the power of the data in ‘Insight Live’ is unlocked by the simplicity of AliX.

How do clients access these features?

We’ve built ‘Insight Live’ and AliX into CORTEX, which is being relaunched in September with a range of new features, so the best user experience is there. But in fact you can access them via the Bloomberg app store and the experience is still top-flight. We have a view to making the tools available on other MDPs, and are even planning the move to mobile. There’s always a new frontier in algos for FX, and we plan on staying right at the forefront of that.