Paper: arXiv 2508.11372
Authors: Arkadiusz Lipiecki, Kaja Bilinska, Nicolaos Kourentzes, Rafal Weron
Abstract
We introduce the concept of temporal hierarchy forecasting (THieF) in predicting day-ahead electricity prices and show that reconciling forecasts for hourly products, 2- to 12-hour blocks, and baseload contracts significantly (up to 13%) improves accuracy at all levels. These results remain consistent throughout a challenging 4-year test period (2021-2024) in the German power market and across model architectures, including linear regression, a shallow neural network, gradient boosting, and a state-of-the-art transformer. Given that (i) trading of block products is becoming more common and (ii) the computational cost of reconciliation is comparable to that of predicting hourly prices alone, we recommend using it in daily forecasting practice.
Complexity vs Empirical Score
- Math Complexity: 5.0/10
- Empirical Rigor: 9.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs formal hierarchical reconciliation mathematics (e.g., summing matrices, covariance shrinkage) but is heavily grounded in a comprehensive, multi-year backtest across multiple model architectures, using real market data and detailed error metrics.
Research Flowchart
flowchart TD
A["Research Goal: Improve Day-ahead<br>Electricity Price Prediction"] --> B["Methodology: Temporal Hierarchy<br>Forecasting (THieF) with Reconciliation"]
B --> C{"Data Inputs"}
C --> D["German Power Market 2021-2024"]
C --> E["Hourly Prices, Block &<br>Baseload Products"]
D & E --> F["Computational Process"]
F --> G["Model Training:<br>Linear Reg, NN, GBM, Transformer"]
F --> H["Forecast Reconciliation<br>across Temporal Levels"]
G & H --> I["Key Findings"]
I --> J["Accuracy Improved up to 13%<br>at all temporal levels"]
I --> K["Consistent results across<br>4-year test period & all models"]
I --> L["Reconciliation cost<br>comparable to hourly prediction"]
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 temporal hierarchy forecasting (THieF) with reconciliation of hourly and block forecasts to improve day-ahead electricity price prediction.
Source sentence
We introduce the concept of temporal hierarchy forecasting (THieF) in predicting day-ahead electricity prices and show that reconciling forecasts for hourly products and 2- to 24-hour blocks can significantly (up to 13%) improve accuracy at all levels.
Assumes 5
- Temporal aggregation levels are non‑overlapping. model
Source sentence
The temporal aggregation levels prescribed by k are non-overlapping for two reasons.
- Base models may be misspecified, which could complicate reconciliation. model
Source sentence
under the reasonable assumption that the base models may be misspecified, such serial autocorrelation can further complicate reconciliation in THieF.
- Variance scaling sets off‑diagonal elements of the error covariance to zero (WLS estimator). statistical
Source sentence
the so-called ‘variance scaling’ (Athanasopoulos et al., 2017) that sets all off-diagonal elements to zero and corresponds to the Weighted Least Squares (WLS) estimator.
- The training window rolls forward by 24 hours each day. execution
Source sentence
Each day the training window is rolled forward by 24 hours.
Show 1 more
- Publicly available market data (prices, load, wind forecasts) are accurate representations of the underlying market. data
Source sentence
three series specific to each market (hourly resolution) – day-ahead prices pd,h, day-ahead load ˆLd,h and wind ˆWd,h generation forecasts
Data
- EPEX-DE (Germany) day‑ahead market market-prices public
Source sentence
To ensure a sound assessment of the THieF approach, we consider two major European power markets: EPEX-DE (Germany) and OMIE (Spain). Both datasets span seven years (05.01.2018‑31.12.2024, see Figure 3) and include publicly available data
- OMIE (Spain) day‑ahead market market-prices public
Source sentence
To ensure a sound assessment of the THieF approach, we consider two major European power markets: EPEX-DE (Germany) and OMIE (Spain). Both datasets span seven years (05.01.2018‑31.12.2024, see Figure 3) and include publicly available data
Costs
Not applicable
Finds 5
- Reconciliation improves forecast accuracy by up to 13% across all hierarchy levels. predictive-accuracy
Source sentence
reconciling forecasts for hourly products and 2- to 24-hour blocks can significantly (up to 13%) improve accuracy at all levels.
- Hourly MAE improves up to 5.5% and baseload MAE up to 13.4% after reconciliation. predictive-accuracy
Source sentence
We find that temporal reconciliation leads to significant improvements in terms of the mean absolute error (MAE) and the root mean squared error (RMSE) across all classes of models and hierarchy levels considered: up to 5.5% for hourly and 13.4% for baseload prices.
- XGB model achieves the largest average gains: 6.7% MAE improvement in Germany and 4.8% in Spain. performance
Source sentence
The highest accuracy improvements in MAE are obtained for the XGB model, with a 6.7% average gain in Germany and a 4.8% in Spain.
- All reconciled forecasts are statistically significantly more accurate (Diebold‑Mariano test at 1% level). statistical
Source sentence
Importantly, all reconciled forecasts are significantly more accurate than the corresponding unreconciled ones, as measured by the multivariate variant of the Diebold‑Mariano test (Ziel and Weron, 2018; Lago et al., 2021) based on the L2-norm of forecast errors.
- Gains increase with larger block sizes, supporting the relevance of block trading. predictive-accuracy
Source sentence
quadrupling of the number of products in the European day-ahead market and the resulting importance of block trading.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 56 related papers). Coverage is partial, so this is neither replication nor refutation.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (39)
- Stochastic coherency in forecast reconciliation (2021) — link
- Cross-temporal coherent forecasts for Australian tourism (2019) — link
- LASSO principal component averaging: A fully automated approach for point forecast pooling (2023) — link
- From day-ahead to mid and long-term horizons with econometric electricity price forecasting models (2025) — link
- Forecasting: Theory and practice (2022) — link
- Electricity price forecasting: A review of the state-of-the-art with a look into the future (2014) — link
- Short- and Mid-Term Forecasting of Baseload Electricity Prices in the U.K. : The Impact of Intra-Day Price Relationships and Market Fundamentals (2016) — link
- Forecast reconciliation: A review (2024) — link
and 31 more
- Forecasting with temporal hierarchies (2017) — link
- Temporal hierarchies with autocorrelation for load forecasting (2020) — link
- Likelihood-based inference in temporal hierarchies (2024) — link
- Probabilistic forecast reconciliation with applications to wind power and electric load (2019) — link
- Improving the forecast accuracy of wind power by leveraging multiple hierarchical structure (2024) — link
- Optimal reconciliation of hierarchical wind energy forecasts utilizing temporal correlation (2024) — link
- Reconciling solar forecasts: Temporal hierarchy (2017) — link
- Spatio-temporal reconciliation of solar forecasts (2023) — link
- Hierarchical forecasting for aggregated curves with an application to day-ahead electricity price auctions — → 2305.16255
- Day-ahead electricity price forecasting with high-dimensional structures: U nivariate vs. multivariate modeling frameworks (2018) — link
- A Shrinkage Approach to Large-Scale Covariance Matrix Estimation and Implications for Functional Genomics (2005) — link
- Loss functions in regression models: Impact on profits and risk in day-ahead electricity trading (2025) — link
- Mitra: Mixed synthetic priors for enhancing tabular foundation models (2025) — [link](https://www.amazon.science/blog/mitra-mixed-synthetic- priors-for-enhancing-tabular-foundation-models)
- Improving forecasting by estimating time series structural components across multiple frequencies (2014) — link
- Honey, I shrunk the sample covariance matrix (2004) — link
- Variance Stabilizing Transformations for Electricity Spot Price Forecasting (2018) — link
- A note on averaging day-ahead electricity price forecasts across calibration windows (2019) — link
- Enhancing load, wind and solar generation for day-ahead forecasting of electricity prices (2021) — link
- Forecasting electricity prices with expert, linear, and nonlinear models (2023) — link
- Forecasting with trees (2022) — link
- Interpretable Transformer Model for Capturing Regime Switching Effects of Real-Time Electricity Prices (2023) — link
- Forecasting day-ahead electricity prices: A review of state-of-the-art algorithms, best practices and an open-access benchmark (2021) — link
- Dynamic short-term risk management strategies for the choice of electricity market based on probabilistic forecasts of profit and risk measures. T he G erman and the P olish market case study (2022) — link
- Time series prediction based on NARX neural networks: An advanced approach (2009)
- On the importance of the long-term seasonal component in day-ahead electricity price forecasting with NARX neural networks (2019) — link
- Distributional neural networks for electricity price forecasting (2023) — link
- The Elements of Statistical Learning: Data Mining, Inference, and Prediction (2009)
- XGBoost : A Scalable Tree Boosting System (2016) — link
- Optuna: A Next-generation Hyperparameter Optimization Framework (2019) — link
- EU electricity trading in the day-ahead markets becomes more dynamic (2025)
- Optimal Forecast Reconciliation for Hierarchical and Grouped Time Series Through Trace Minimization (2019) — link
1 of 39 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.