Data science for crude trading and risk management

This is a ZE Global Tech Summit Recap Blog.

Vlasios spoke about how Argus Media is leveraging data science to develop new products for its customers.

He explained how individual prices for commodities are a reflection of a large number of influences; these different influences can become more or less important over time. The challenge, he said, is to deploy data science to generate actionable outcomes from these different influences.

Through cooperation between data science and editorial teams, Argus can create a database of global drivers that influence prices. By overlaying financial data, physical prices (Argus assessments) and fundamentals (global drivers), Argus generates the “possibility curve”, a probabilistic forecast of prices for a specific commodity.

“The outcome does not what the price will be tomorrow or next month. It provides an estimate of the degree of uncertainty of the market. But most importantly, it provides an estimate of the upside risk versus the downside risk, that can be actually estimated,” Vlasios said.

“Clients get an overall estimate: not only a forward-looking estimate of the volatility, but quite importantly, the balance of risk, which identifies the biggest opportunity in terms of market dynamics.”

Argus has created a data science platform that presents its probabilistic forecasts, with the capability to add bespoke price curves and parameters to reflect clients’ own risk appetite or positions.

The aggregation of these probabilistic curves over time means that clients can also train machine learning models to work on both the forecast curves and the actual outcomes.

“That can give some sort of a signal for opening or closing a position,” Vlasios said. “And you can do all that without human intervention. Where the human comes in is in creating the model and training it to use a product like a possibility curve. But once you get that in place, then the whole process can become systematic.”

Presently, the possibility curves are being used in the crude oil market, but Argus plans to widen its application to other areas in its portfolio in the New Year.