{"paper_id":"2508.11372","title":"Stealing Accuracy: Predicting Day-ahead Electricity Prices with Temporal Hierarchy Forecasting (THieF)","authors":["Arkadiusz Lipiecki","Kaja Bilinska","Nicolaos Kourentzes","Rafal Weron"],"paper_date":"2025-08-15","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2508.11372v1","page_url":"https://thequant.space/flowcharts/stealing-accuracy-predicting-day-ahead-electricity-prices-w/","scores":{"empirical_rigor":9,"math_complexity":5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes temporal hierarchy forecasting (THieF) with reconciliation of hourly and block forecasts to improve day-ahead electricity price prediction.","quote":"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.","summary":"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…","strategy":"Temporal Hierarchy Forecasting (THieF)"},"assumptions":{"status":"available","items":[{"assumption":"Temporal aggregation levels are non‑overlapping.","type":"model","stated":"explicit","quote":"The temporal aggregation levels prescribed by k are non-overlapping for two reasons."},{"assumption":"Base models may be misspecified, which could complicate reconciliation.","type":"model","stated":"explicit","quote":"under the reasonable assumption that the base models may be misspecified, such serial autocorrelation can further complicate reconciliation in THieF."},{"assumption":"Variance scaling sets off‑diagonal elements of the error covariance to zero (WLS estimator).","type":"statistical","stated":"explicit","quote":"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."},{"assumption":"The training window rolls forward by 24 hours each day.","type":"execution","stated":"explicit","quote":"Each day the training window is rolled forward by 24 hours."},{"assumption":"Publicly available market data (prices, load, wind forecasts) are accurate representations of the underlying market.","type":"data","stated":"implied","quote":"three series specific to each market (hourly resolution) – day-ahead prices pd,h, day-ahead load ˆLd,h and wind ˆWd,h generation forecasts"}]},"datasets":{"status":"available","items":[{"name":"EPEX-DE (Germany) day‑ahead market","provider":"EPEX Spot / ENTSO‑E transparency platform","kind":"market-prices","access":"public","period":"05.01.2018‑31.12.2024","universe":null,"quote":"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"},{"name":"OMIE (Spain) day‑ahead market","provider":"OMIE / ENTSO‑E transparency platform","kind":"market-prices","access":"public","period":"05.01.2018‑31.12.2024","universe":null,"quote":"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"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"intraday-bars"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"forecast-accuracy","findings":[{"finding":"Reconciliation improves forecast accuracy by up to 13% across all hierarchy levels.","kind":"predictive-accuracy","quote":"reconciling forecasts for hourly products and 2- to 24-hour blocks can significantly (up to 13%) improve accuracy at all levels."},{"finding":"Hourly MAE improves up to 5.5% and baseload MAE up to 13.4% after reconciliation.","kind":"predictive-accuracy","quote":"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."},{"finding":"XGB model achieves the largest average gains: 6.7% MAE improvement in Germany and 4.8% in Spain.","kind":"performance","quote":"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."},{"finding":"All reconciled forecasts are statistically significantly more accurate (Diebold‑Mariano test at 1% level).","kind":"statistical","quote":"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."},{"finding":"Gains increase with larger block sizes, supporting the relevance of block trading.","kind":"predictive-accuracy","quote":"quadrupling of the number of products in the European day-ahead market and the resulting importance of block trading."}],"reported_performance":{"strategy":"Temporal Hierarchy Forecasting (THieF)","sharpe":null,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"walk-forward","headline_out_of_sample":true}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":56,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":true,"code_url":"https://github.com/lipiecki/thief","note":null},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T17:29:55Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}