After Refinitiv deal, LSEG Labs builds machine-learning market impact tool

Geoff Horrell talks about the Lab’s innovation strategy after the Refinitiv acquisition, deep learning, and the ideation process.

Podcast timestamps

5:00 – Geoff joins the podcast and gives an overview of his remit at LSEG.
7:00 – Then he explains what has changed since LSEG acquired Refinitiv and how that expands the coverage area of LSEG Labs.
9:00 – Geoff walks us through some examples of innovation in post-trade and capital markets space.
11:30 – They discuss the ideation process at LSEG Labs.
14:30 – It’s about identifying ‘quick wins.’
16:00 – Geoff talks about a recent project LSEG Labs built.
23:00 – Moving forward, LSEG Labs will prioritize projects in sustainable finance and digital assets.
27:00 – Fresh off the LSEG Labs AI/ML 2021 report, Geoff explains one of the key findings: deep learning is now the favored type of machine learning.
35:00 – They wrap up discussing innovation in Emea and Asia versus the US.

LSEG Labs—formerly Refinitiv Labs—has been working on a number of projects. One of them is a pre-trade market impact analysis tool for equity traders.

Geoff Horrell, head of innovation at the London Stock Exchange Group, said that traders choose which venue to use and what trade volume and size to put through when executing a trade. They also need to understand how a particular trade impacts the market so they can determine the cost analysis and performance of the trade.

LSEG Labs has developed a pre-trade market impact analysis model that will show the likely impact trades have on the market based on the stock for any particular volume size and time of day.

“So what the model does pre-trade, is tell you based on the current volume that we’re seeing in the stock historically, what we’re seeing in this particular time, what is the likely impact. And then you can decide your trading strategy accordingly, and you could perhaps change the order size, change the timings, change how you’re structuring, or laddering your trades. From the user point of view, it’s a very helpful piece of information for you to see that likely impact,” he said.

LSEG Labs used the I-Star model—a standard market impact model that estimates the instantaneous trading cost for orders—as a baseline and then introduced a machine learning-based model to predict the market impact.

For that, Horrell said the labs used six months of historical tick data from the S&P 500 and Russell 1000 indexes, as well as data from other markets.

“We did this for a number of markets. And you have all of the usual kinds of things, like market open, market closes, slight strangeness around how the exchanges report particular kinds of trades. You have to unpick all of that to understand and build your model,” he said.

LSEG Labs tried several different techniques but ended up using a neural network for the tool, as it performed best across different markets and conditions. Then the team built a user dashboard that visually shows traders what the model is predicting after consulting customers how they might want to interact with the tool.

LSEG Labs used Amazon’s SageMaker, a cloud-based machine-learning platform, for the heavy lifting when it comes to AI.

Horrell says he anticipates that the tool will be available to users within Refinitiv CodeBook (its cloud-hosted development environment for Python scripting) and Workspace (its next-generation workflow solution) in the first quarter of 2022, as part of a suite of trade performance analytics for equities.

“Rather than being a fixed analytic that a customer can see, it’s completely transparent to them in CodeBook in a Jupyter hub-type environment, and they can actually see the analytics, they can actually add their own features to those analytics. So it’s a more open way of delivering the analytics,” he said.

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact or view our subscription options here:

You are currently unable to copy this content. Please contact to find out more.

Systematic tools gain favor in fixed income

Automation is enabling systematic strategies in fixed income that were previously reserved for equities trading. The tech gap between the two may be closing, but differences remain.

Why recent failures are a catalyst for DLT’s success

Deutsche Bank’s Mathew Kathayanat and Jie Yi Lee argue that DLT's high-profile failures don't mean the technology is dead. Now that the hype has died down, the path is cleared for more measured decisions about DLT’s applications.

You need to sign in to use this feature. If you don’t have a WatersTechnology account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here