CDN NEWS | 
US NEWS 
(Reuters) – The Iran conflict has put a spotlight on the somewhat murky world of refined fuel trading as the world battles constrained supply and record-high prices. But it is AI that may bring the trading of diesel, gasoline and other products out of the shadows for good.
While crude oil trading practices have been established for decades, attracting a wide range of market participants, the same cannot be said of refined products.
These markets, especially in the top-consuming region of Asia, have largely been the preserve of a relatively small group of traders who represent producers, refiners, shippers and large commodity houses such as Trafigura and Vitol.
But every now and again, the market can shift, and AI may provide the tools needed to democratise this still opaque segment of the commodities trading universe.
The idea sounds simple: use AI to collate all available data, add in the knowledge of specialised traders and analysts, and generate trading ideas tailored to specific users and situations.
One of the leaders in this process is Sparta Commodities, a Geneva-based firm that initially made its name by providing data that allowed traders to work out arbitrage opportunities when moving fuels from one region to another.
Earlier this month, it launched Leonidas AI, a decision-making tool that aims to speed up the process by which traders use data to arrive at trading calls.
It takes all the available data from Sparta’s own sources and from other providers, adds in news and the experience of what the company calls a dedicated agent in order to make recommendations.
In a demonstration for Reuters, Leonidas AI was asked about the implications of any ban on US diesel exports for various markets. Having come up with likely pricing moves, it was then asked to make trading recommendations.
It came up with five proposed trades, ranking them by conviction. The recommendations also covered various markets in different countries, and there was even a freight recommendation.
This all took a matter of seconds, whereas up until now, it would have taken even skilled and experienced trading analysts some time to work out likely price moves and how best to profit from them.
Competitors in this space, such as Kpler and Windward, are already rolling out or developing similar products.
Kpler, which started as a vessel-tracking service, uses AI to compress data into what it calls “actionable strategies” that take into account factors such as geopolitics.
Windward, another player in the tracking business, uses AI to analyse historical data and vessel movement patterns in order to highlight anomalies, such as illegal ship-to-ship transfers of cargoes.
Players such as hedge funds, which previously believed they lacked the knowledge to go up against established traders in an opaque fuel market, may now be able to get and process the necessary information just as fast.
RESISTING CHANGE
Are there any weaknesses in adopting AI to inform trading decisions?
The obvious question is whether everybody will just get the same recommendations, resulting in highly crowded trades that eliminate the potential for profits.
However, different users will be seeking different outcomes, which should mitigate this risk.
For example, a refiner may be willing to lock in a fixed price for some of its output of diesel and gasoline, while a trader may be willing to take the risk that prices will rise before taking physical delivery, thus creating the potential for profit.
As a result, the refiner and trader should get different recommendations from AI even if they are using the same model.
Another obvious question is whether the recommendations are any good — something that will only become clear with time.
Pushback could also come from established players who may resist newcomers entering their markets. A large trader could try to influence prices through a series of trades aimed at pushing smaller players into stop-loss situations.
But history suggests that over time the existing participants will have to adapt to changed circumstances — whether they like it or not.
An example of this type of change is vessel-tracking of physical cargoes in real time.
These services — from firms like Vortexa and London Stock Exchange Group — only started emerging about 15 years ago, but have since grown into household names and enjoy widespread penetration among commodity market participants.
NO RAGE AGAINST CHANGE
At the Asia Pacific Petroleum Conference (APPEC) earlier this month in Singapore, the main topic of conversation was how to resolve the Iran conflict and restore the flow of crude and products from the Middle East.
But the second most discussed issue was AI, with questions such as how companies are using it, and what the advantages and risks are.
Across numerous conversations at the event, it became clear that commodity executives expect AI to play an increasing role in data collection and analysis, as well as in operations such as shipping and inventory management.
The overall idea is that it’s pointless to rage against change. Rather, it’s preferable to be an early adopter and leader.
In a best-case scenario, AI will level the playing field in the trading of commodities that are currently the preserve of small groups of experienced participants.
As Sparta CEO Felipe Elink Schuurman put it, the trading desk that used to win was the one with the most information, but the new winners will be the ones that can turn data into trading moves fastest.
The views expressed here are those of Clyde Russell, a columnist for Reuters.
By Clyde Russell; Editing by Marguerita Choy and Anna Szymanski
Share This:
