Issue: 2026/Vol.36/No.2, Pages 33-58
ARBITRAGE IN SHORT-TERM ELECTRICITY MARKETS: ECONOMIC VALUE OF MULTIDIMENSIONAL ENSEMBLE FORECASTS
Katarzyna Maciejowska
, Weronika Nitka
Cite as: K. Maciejowska, W. Nitka. Arbitrage in short-term electricity markets: Economic value of multidimensional ensemble forecasts. Operations Research and Decisions 2026: 36(2), 33-58. DOI 10.37190/ord/217347
Abstract
Generated quantity and market prices are major risk sources for electricity market participants, especially renewable producers with intermittent generation. Before delivery, wind and solar power plants must decide their day-ahead bids, which, through deviations from actual generation, determine the direction and volume of intraday trade. We identify an arbitrage opportunity between these short-term markets that increases expected profits and reduces trading risk. To exploit it, we propose a novel multiple-split approach for estimating multidimensional probabilistic forecasts of market fundamentals. This semi-parametric method produces ensembles preserving true error correlations and supports forecasting of linear and nonlinear functions such as price spreads, residual load, or trading income. Using German electricity market data, we demonstrate superior predictiv and economic performance over standard benchmarks
Keywords: electricity market, trading strategy, probabilistic forecasting, ensemble forecasting, multivariate forecasting
Received: 3 November 2025 Accepted: 25 January 2026
Published online: 25 January 2026