Economic Impact of Forecast Errors on BESS Participation in Day-Ahead electricity and Reserve Markets

Paper in proceeding, 2026

This paper quantifies the effect of price-forecast uncertainty on the profitability of a battery energy storage system (BESS) that participates simultaneously in the Nordic day-ahead (DA) market and the three frequency containment reserves: FCR-N, FCR-D up, and FCR-D down. A mixed-integer linear programming model has been developed, which co-optimizes DA arbitrage and multi-market reserve provision, subject to battery power and energy limits, ENTSO-E reserve technical requirements, and cycle- and calendar-aging costs. Price uncertainty is represented through forecast scenarios with a systematic bias and a zero-mean dispersion term. The forecast-based schedule is embedded in a forecast–decision–realization scheme, settled at realized prices, and compared against a perfect-information benchmark to determine the cost of uncertainty. Using 2025 DA and FCR markets and activation data from Nord Pool SE3 and Svenska kraftnät, respectively, for fictious 1 MW/1 MWh BESS, three results emerge. First, the revenue is earned almost entirely from the reserve capacity: FCR-N alone contributes to about 78% of the total revenue, while DA arbitrage contributes to only about 3%. Second, the mean cost of uncertainty stays below 1.5% of the total annual revenue even at a 50% forecast bias, but the loss is concentrated on a small number of high-cost days. Third, the loss depends substantially more on the coherence of forecast errors across markets than on their magnitude. Incoherent errors across the DA and reserve markets are an order of magnitude costlier than coherent errors of the same size.