Combining predictive distributions of electricity prices: Does minimizing the CRPS lead to optimal decisions in day-ahead bidding?

ArXiv ID: 2308.15443 “View on arXiv”

Authors: Unknown

Abstract

Probabilistic price forecasting has recently gained attention in power trading because decisions based on such predictions can yield significantly higher profits than those made with point forecasts alone. At the same time, methods are being developed to combine predictive distributions, since no model is perfect and averaging generally improves forecasting performance. In this article we address the question of whether using CRPS learning, a novel weighting technique minimizing the continuous ranked probability score (CRPS), leads to optimal decisions in day-ahead bidding. To this end, we conduct an empirical study using hourly day-ahead electricity prices from the German EPEX market. We find that increasing the diversity of an ensemble can have a positive impact on accuracy. At the same time, the higher computational cost of using CRPS learning compared to an equal-weighted aggregation of distributions is not offset by higher profits, despite significantly more accurate predictions.

Keywords: probabilistic forecasting, day-ahead bidding, electricity markets, ensemble learning, CRPS, Commodities

Complexity vs Empirical Score

• Math Complexity: 7.0/10
• Empirical Rigor: 8.5/10
• Quadrant: Holy Grail
• Why: The paper involves advanced statistical concepts like proper scoring rules (CRPS) and complex ensemble weighting techniques, indicating high math complexity. It also conducts a rigorous empirical study with six years of real market data, state-of-the-art neural network models, and direct profit calculations for a trading strategy, showing high empirical rigor.

  flowchart TD
    A["Research Goal<br>Does CRPS-minimizing ensemble<br>lead to optimal day-ahead bidding?"]
    B["Data Source<br>Hourly German EPEX<br>day-ahead prices"]
    C["Methodology<br>Ensemble prob. forecasting<br>CRPS vs. Equal-weight aggregation"]
    D["Computational Process<br>Hourly bidding decisions<br>based on prob. forecasts"]
    E["Outcome 1<br>CRPS yields<br>more accurate predictions"]
    F["Outcome 2<br>Higher accuracy does NOT<br>translate to higher profits"]
    G["Key Finding<br>Computational cost of CRPS<br>not justified by profit increase"]

    A --> B
    B --> C
    C --> D
    D --> E
    D --> F
    E --> G
    F --> G