Paper: arXiv 2306.13419

Authors: Simon Hirsch, Florian Ziel

Abstract

Intraday electricity markets play an increasingly important role in balancing the intermittent generation of renewable energy resources, which creates a need for accurate probabilistic price forecasts. However, research to date has focused on univariate approaches, while in many European intraday electricity markets all delivery periods are traded in parallel. Thus, the dependency structure between different traded products and the corresponding cross-product effects cannot be ignored. We aim to fill this gap in the literature by using copulas to model the high-dimensional intraday price return vector. We model the marginal distribution as a zero-inflated Johnson’s \(S_U\) distribution with location, scale and shape parameters that depend on market and fundamental data. The dependence structure is modelled using latent beta regression to account for the particular market structure of the intraday electricity market, such as overlapping but independent trading sessions for different delivery days. We allow the dependence parameter to be time-varying. We validate our approach in a simulation study for the German intraday electricity market and find that modelling the dependence structure improves the forecasting performance. Additionally, we shed light on the impact of the single intraday coupling (SIDC) on the trading activity and price distribution and interpret our results in light of the market efficiency hypothesis. The approach is directly applicable to other European electricity markets.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 7.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced statistical methods including copulas, zero-inflated Johnson’s SU distribution, and latent beta regression, indicating high mathematical complexity. It is empirically rigorous with a detailed simulation study on German intraday electricity market data, analyzing cross-product effects and market structure impacts, making it backtest-ready.

Research Flowchart

  flowchart TD
  A["Research Goal: Multivariate Probabilistic Forecasting for Intraday Power Markets"] --> B["Model Marginals: Zero-Inflated Johnson's S_U Distribution<br>(Market & Fundamental Data)"]
  B --> C["Model Dependence: Copulas with Time-Varying Latent Beta Regression<br>(SIDC & Market Structure)"]
  C --> D["Computational Process: Simulation-based Approach"]
  D --> E["Outcome: Improved Forecasting Accuracy<br>(vs. Univariate Approaches)"]
  E --> F["Outcome: Insights on Market Efficiency & SIDC Impact"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a multivariate, simulation‑based probabilistic forecasting model for intraday electricity markets using copulas to capture cross‑product price effects.

Source sentence
As novelty, we analyse the correlation and dependence structure in the German intraday electricity market and develop a multivariate, probabilistic forecasting model for the German intraday electricity markets to take cross‑product effects into account.

Assumes 6

  • The marginal distribution of intraday price returns follows a zero‑inflated Johnson's S_U distribution with parameters depending on market and fundamental data. model
    Source sentence
    We model the marginal distribution as a zero‑inflated Johnson's $S_U$ distribution with location, scale and shape parameters that depend on market and fundamental data.
  • The dependence structure is captured by a Gaussian copula (Gaussian dependence structures). model
    Source sentence
    We restrict ourselves to Gaussian dependence structures.
  • The dependence parameter may be time‑varying, but time‑dependent modelling adds little value compared to a constant dependence. model
    Source sentence
    we find that time‑dependent modelling of the dependence structure is of little added value compared to constant dependence.
  • Latent beta regression is used to model dependence, reflecting overlapping but independent trading sessions. model
    Source sentence
    The dependence structure is modelled using latent beta regression to account for the particular market structure of the intraday electricity market, such as overlapping but independent trading sessions for different delivery days.
Show 2 more
  • Benchmark models assume independence between delivery hours when sampling each hour individually. model
    Source sentence
    first assuming independence between different delivery hours (i.e sampling each delivery hour h = 0, ..., H individually)
  • The random walk benchmark (RW.Emp) assumes the price process is a martingale (mean‑zero innovations). model
    Source sentence
    The second term centers all residuals to mean zero, ensuring the price process to be a martingale.

Data

  • German intraday electricity market data market-prices unknown
    2020-01-01 to 2023-01-01 · German intraday electricity market
    Source sentence
    Our data comprises of the three years between 2020-01-01 and 2023-01-01.

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • Modeling the dependence structure leads to significantly improved forecasting performance compared to univariate approaches. performance
    Source sentence
    Our results show that modelling the dependence structure leads to significantly improved forecasting performance compared to univariate approaches.
  • Time‑dependent modelling of the dependence structure provides little added value over a constant dependence. negative
    Source sentence
    we find that time‑dependent modelling of the dependence structure is of little added value compared to constant dependence.
  • The RW.Emp benchmark yields the best point‑forecasting performance (lowest MAE and RMSE). performance
    Source sentence
    On an aggregate level, we see that the RW.Emp yields the best point forecasting performance (MAE and RMSE)
  • The Naive.Dep benchmark yields the best probabilistic forecasting performance, confirmed by Diebold‑Mariano tests. predictive-accuracy
    Source sentence
    the Naive.Dep yields the best probabilistic forecasting performance. The Diebold‑Mariano test confirms the statistical significance of superior probabilistic forecasting forecasts for the Naive.Dep.
  • Mixture models that incorporate dependence (Mix.CD, Mix.TD) perform worse than simple benchmarks in both point and probabilistic metrics. performance
    Source sentence
    the forecasting performance of the mixture models is not as good as the rather simple benchmark models.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 31 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 1 paper in the archive (1 by independent authors)

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.