Live from the Equities Leaders Summit in Miami, the head of global systematic trading at Fidelity Investment talks exclusively to BEST EXECUTION about data science, digitisation and decision-making.
What were your top trading themes for 2023?
Top themes for 2023 were same as always: minimize execution cost in most efficient manner possible. Developing and delivering advanced execution capabilities through various forms of automation that are driven by advanced analytical efforts and that work in concert and collaboratively with traders is key to that objective.
How do you work with your desk to best locate liquidity in difficult markets?
The equity markets are continuously evolving, and there is constant innovation in the way liquidity is exchanged and moved around in the ecosystem. A focus of ours is to continually shape and curate the toolkit that allows the desk to have the greatest access to liquidity and flexibility in how to interact with it.
What was the most interesting/disruptive development you saw last year, and how did it change things?
Although generative AI/LLMs are nascent and they have yet to make their impact in the trading world, I would expect the impact of this capability/technology in our industry will be transformative. It’s just impossible to really know what that will look like at this point, but I think it will be really interesting to see how this unfolds.
How is AI driving progress in algorithmic trading?
I think that is difficult to say, as I think the answer to that can vary greatly depending on the firm. The amount of innovation in the industry seems to continue to accelerate. I guess what I think is a fairly common theme in advancement in trading seems to really come down to data. That is, making the most out of data and turning that from reporting purposes, to making data actionable. That’s where I think AI/ML can be very powerful. Trading is an endeavor that is about taking actions/making decisions in a data rich ecosystem with non-trivial, complex, nonlinear interrelationships. AI/ML can be instrumental in capturing and making sense of that complexity.
You’re speaking on a panel about the importance of data science – how is the intersection between data science and trading evolving, and how do you use data science on your desk?
I think the distinction of data science will eventually fade away, and we’ll just think of it as part of the general moniker of quant work—it’s the broadening out of the toolkit used for analysis, modeling, etc. Given that, we use data science heavily across the desk execution life-cycle: from pre-trade analytics, trade execution, and post-trade analysis. I think we’re still scratching the surface of it, particularly when we think about leveraging it to synthesize and deliver information and data to traders in real-time.
How can the use of data and digitisation help traders to capture new alpha?
I think this shows up in most clearly in a few ways: 1) embedded in the toolkit that the traders use to perform their job, 2) enabling traders to be more efficient and ability to focus on where they can add the most value, 3) provide new information that is more consumable and digestible in real-time, and 4) be part of the ongoing feedback loop to help make traders better informed and equipped to make better decisions.
How far can/will automation go, and how are the roles of trader and coder converging?
I think of there being two types of automation that we are typically talking about. The first is automation of (manual) processes. In trading, I think about that as button clicks—how can we get something down from N clicks to one (or none). I’m not sure there is an end to that process. The fewer the number of clicks, the better.
The second type of automation, and the more interesting one from my point of view, is automation of decision making. In the context of trading, an example of this can be “how to trade an order”— how fast or slow, what type of algo(s) and parameters, etc. Automation of this type can range along a spectrum from rule based to model driven predictive based—which can range from fairly simple traditional statistical models to more complex machine learning and AI based models. How far this goes is really dependent on many things, but really comes down to how intelligent is the automation.
I think all modes of decision making (human discretionary, rules based, and model based) can be successfully leveraged in concert with each other as they can be largely complimentary and the relative strengths of the three can be orthogonal to each other.
What are your top focus areas for 2024 and beyond?
Top focus for 2024 is the same as every year: minimize execution costs with the greatest amount of efficiency! I am a believer in the power of the complimentary efforts of human discretionary, rules based, and model driven decision making. A big focus will be on continuing to bring those capabilities together in a flexible manner within a common ecosystem—allowing for seamless collaboration across the three. The goal is simply to maximize the value added by the desk by utilizing the different capabilities collectively in the most efficient was possible.
What are you most looking forward to at ELS?
Connecting with peers and friends in the industry is always great. I also really enjoy hearing and discussing what people are focused on. We all are effectively trying to solve the same problem—the specifics can just look very different depending on where you are, but these different views and experiences are always useful in refining your own thinking about what you are trying to solve.
©Markets Media Europe 2023
