Probabilistic Forecasting for Day-ahead Electricity Prices, Battery Trading Strategies and the Economic Evaluation of Predictive Accuracy
Probabilistic Forecasting for Day-ahead Electricity Prices, Battery Trading Strategies and the Economic Evaluation of Predictive Accuracy
Authors: Simon Hirsch, Florian Ziel Date: 2026-04-21 Paper ID: openalex:2604.19580
Summary
This paper critically evaluates the relationship between statistical probabilistic forecast accuracy and economic decision-making in electricity markets. The authors analyze common quantile-based battery trading strategies, highlighting inherent failures in incentivizing honest forecasts and accounting for price interdependencies. By modeling battery arbitrage as a stochastic program, the study demonstrates the limitations of using simplified trading benchmarks for evaluating forecasting models. The work provides significant insights into the reliability of application-based forecast assessment and proposes more robust methodologies for energy market forecasting.
Key Contributions
- Identifies and formalizes flaws in current quantile-based trading strategies (QBTS) for battery arbitrage, specifically lack of incentive for honest forecasting and neglect of intertemporal price dependency.
- Proposes framing battery optimization as a stochastic programming problem utilizing full probabilistic forecasts rather than simple quantile approximations.
- Provides empirical evidence demonstrating that standard battery trading benchmarks can lead to misleading rankings of forecasting models due to sensitivity to model assumptions and risk profiles.
Open Questions & Future Work
Key Concepts
- quantile-based-trading-strategy: A trading heuristic that uses predictive quantiles to set limit orders, which can lead to biased model evaluation in energy market forecasting.
Archivist Review
I approved the term ‘Quantile-Based Trading Strategy’ as it is a specific, widely used heuristic in energy forecasting that researchers should critically evaluate. The open question regarding the alignment between economic performance and statistical forecasting accuracy is highly relevant and captures a core methodological tension in energy time-series research. I adhered to the policy of being selective, approving only these two items while rejecting other general concepts.
Approved Concepts
- Quantile-Based Trading Strategy: This highlights a common but flawed practice in energy forecasting evaluation, making it a valuable term for future reference in discussions about model evaluation.
Approved Open Questions
- Economic vs Statistical Forecast Evaluation: Understanding this relationship is fundamental for the field of electricity price forecasting, as it determines whether researchers should rely on statistical scores or application-specific outcomes when selecting and validating forecasting models.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2604.19580
- paper_id
- 2604.19580
- paper_source
- openalex
- domain
- time-series
- tags
- time-seriesforecastingenergy-markets
- architectures
-
- datasets
-
- concept_slugs
- quantile-based-trading-strategy
- dataset_slugs
-
- skill
- TimeSeriesSkill
- created_at
- 2026-04-24T05:51:07Z