🤖 AI Summary
To address price volatility and forecasting uncertainty induced by strategic trading in single-price imbalanced electricity markets, this paper proposes a fully automated cross-border intraday (XBID) trading strategy. Methodologically, we develop a hybrid probabilistic forecasting model to estimate system imbalance prices, integrate a Conditional Value-at-Risk (CVaR)-based adaptive risk-aversion mechanism, and introduce a novel sliding-window approach for real-time calibration of dynamic risk-measure parameters—thereby balancing forecasting accuracy and robustness. Evaluated on real-world Belgian market data within an XBID simulation environment, the strategy achieves significantly higher absolute profit and reduced trading frequency, demonstrating both economic efficiency and operational stability. Key contributions are: (1) the first real-time, risk-adaptive trading framework specifically designed for XBID; and (2) a sliding-window online parameter-tuning paradigm that effectively mitigates poor generalization under limited-sample conditions.
📝 Abstract
Efficient markets are characterised by profit-driven participants continuously refining their positions towards the latest insights. Margins for profit generation are generally small, shaping a difficult landscape for automated trading strategies. This paper introduces a novel, fully-automated cross-border intraday (XBID) trading strategy tailored for single-price imbalance energy markets. This strategy relies on a strategically devised mixture model to predict future system imbalance prices, which, upon benchmarking against several state-of-the-art models, outperforms its counterparts across every metric. However, these models were fit to a finite amount of training data typically causing them to perform worse on unseen data when compared to their training set. To address this issue, a coherent risk measure is added to the cost function to take additional uncertainties in the prediction model into account. This paper introduces a methodology to select the tuning parameter of this risk measure adaptively by continuously quantifying the model accuracy on a window of recently observed data. The performance of this strategy is validated with a simulation on the Belgian energy market using real-time market data. The adaptive tuning approach enables the strategy to achieve higher absolute profits with a reduced number of trades.