Machine Learning enabled trading (gas/power) using price, LOB (Order book) and fundamental data

Offentlig prosjektsammendrag

The project is Machine Learning enabled trading (gas/power) using price, LOB (Order book) and fundamental data. Vega has developed models which have given encouraging results for 'Trading Signals' used to predict market price movements (>65% correct) and has developed a tool to backtest if this is profitable when combined with mechanistic trading strategies. Now, in a project with Refinitiv, Vega will add market fundamentals (demand, supply, flows, storage, etc) to improve the predictive power of the algorithms and test for corresponding improvements in trading results. Initially the focus is on UK and European gas (NBP / TTF) and then the same approach will be replicated on European power to see if the concept/model can travel across commodities.