New algorithm enables optimal bidding in European FCR markets

Marius Potfer, Cheng Wan, and Pierre Gruet have published a paper on arXiv (2605.31070) tackling the challenge of bidding in European Frequency Containment Reserve (FCR) markets. These markets are notoriously opaque: competing offers are hidden, and bidders only observe partial feedback like clearing price and awarded quantity. The authors show that for a participant active in a single country, the multi-country FCR clearing problem can be recast as a repeated multi-unit uniform-price auction against an endogenous vector of opposing bids. This reformulation turns the problem into an online learning task, allowing them to adapt a Best-of-Both-Worlds combinatorial semi-bandit algorithm that works with standard market feedback.

The resulting algorithm delivers strong theoretical guarantees: logarithmic pseudo-regret in stochastic environments and O(√T) regret under adversarial conditions. Synthetic experiments confirm the expected scaling, and backtests on historical European FCR data demonstrate competitive practical performance. The method performs especially well on stable products, while EXP3-type baselines can be safer under stronger non-stationarity. The authors conclude that learning-based bidding in FCR markets is both theoretically grounded and practically useful when the learning rule matches product-level market stability. Code and data have been released alongside the paper.