Probabilistic Electricity Forecasts and Battery Arbitrage Evaluation

Summary

The study examines whether better probabilistic day-ahead electricity price forecasts translate into better battery trading outcomes. It identifies two limitations of quantile-based trading strategies: they do not reward honest probabilistic forecasts, and they ignore the time dependence between electricity prices. As an alternative, it frames battery optimization as a stochastic program using full probabilistic forecasts, then considers how decision quality should be measured for both risk-neutral and risk-averse settings under different uncertainty models.

The work offers theoretical justification and empirical evidence from a case study in the German electricity market. It warns that battery trading strategy results can mislead when used to rank forecasting models and discusses how statistical forecast quality relates to decision quality and economic performance. The excerpt supplies no specific numerical results or details of the case study, so it does not support conclusions about which forecast model or battery strategy performs best. Its focus is evaluation practice, not a ready-to-deploy trading rule.

Key ideas

• Quantile-based battery strategies may fail to reward truthful probabilistic forecasts.
• Those strategies also omit intertemporal dependence in electricity prices.
• The study formulates battery optimization as a stochastic program using full predictive distributions.
• It evaluates decision quality across risk-neutral and risk-averse settings and different uncertainty models.
• The German market case study highlights pitfalls in ranking forecasts by battery trading results.

Tags

# 2604.19580


# Probabilistic Forecasting for Day-ahead Electricity Prices, Battery Trading Strategies and the Economic Evaluation of Predictive Accuracy









Electricity price forecasting supports decision-making in energy markets and asset operation. Probabilistic forecasts are increasingly adopted to explicitly quantify uncertainty, typically issued as quantile predictions or ensembles of the full predictive distribution. However, how improvements in statistical forecast quality translate into economic value remains unclear. Battery storage arbitrage in day-ahead markets is a popular application-based benchmark for this purpose. We analyze quantile-based trading strategies (QBTS) and identify two critical flaws: they do not incentivize honest probabilistic forecasting and they ignore the intertemporal dependence structure of electricity prices. We therefore frame battery optimization as a stochastic program based on fully probabilistic forecasts and examine decision quality measurement for risk-neutral and risk-averse settings under different uncertainty models. Our discussion touches both sides of the coin: How reliable is the economic evaluation of forecasting models though (simplified) application studies - and how do improvements in statistical forecast quality for stochastic programs relate to the decision-quality and economic performance? We provide theoretical justification and empirical evidence from a case study on the German electricity market. Our results highlight the pitfalls of ranking forecasting models through battery trading strategies. We conclude with implications for evaluation practice and directions for future research in application-based forecast assessment.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.