{"paper_id":"2509.04452","title":"Directional Price Forecasting in the Continuous Intraday Market under Consideration of Neighboring Products and Limit Order Books","authors":["Timothée Hornek","Sergio Potenciano Menci","Ivan Pavić"],"paper_date":"2025-08-20","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2509.04452v1","page_url":"https://thequant.space/flowcharts/directional-price-forecasting-in-the-continuous-intraday-mar/","scores":{"empirical_rigor":7,"math_complexity":6,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a novel directional electricity price forecasting method for hourly products in the European continuous intraday market.","quote":"In turn, we propose a novel directional electricity price forecasting method for hourly products in the European continuous intraday market.","summary":"The increasing penetration of variable renewable energy and flexible demand technologies, such as electric vehicles and heat pumps, introduces significant uncertainty in power systems, resulting in greater imbalance; defined as the deviation between scheduled and actual supply or demand. Short-term…","strategy":"directional electricity price forecasting model"},"assumptions":{"status":"available","items":[{"assumption":"Reference price is the average VWAP of the four most recent transactions before the forecast time.","type":"data","stated":"explicit","quote":"To determine the reference price, we average the VWAPs of the four most recent transactions occurring before the forecast time following the authors of [22]."},{"assumption":"Market direction is defined by whether the future price exceeds the reference price.","type":"model","stated":"explicit","quote":"We then define the market direction by comparing these two values: if the future price exceeds the reference price, the market is rising; otherwise it is falling."},{"assumption":"Logistic regression uses L2 regularization to mitigate multicollinearity.","type":"model","stated":"explicit","quote":"For logistic regression, we apply l2 regularization, which distributes weights across correlated variables and reduces the adverse effects of multicollinearity."},{"assumption":"The empirical evaluation uses data from April 14, 2024 to April 13, 2025.","type":"data","stated":"explicit","quote":"We demonstrate and evaluate our method using real data from the EPEX for the German CID market, with the test period spanning from April 14, 2024, to April 13, 2025."},{"assumption":"The analysis focuses exclusively on the ID3 trading period.","type":"market","stated":"explicit","quote":"Our analysis focuses exclusively on the ID3 trading period, which may limit the generalizability of our findings to other lead times."},{"assumption":"Gradient boosting (LightGBM) is used to capture non‑linear relationships while remaining computationally practical.","type":"model","stated":"explicit","quote":"We choose logistic regression for its simplicity as a linear model and computational efficiency, and LightGBM to capture non‑linear relationships while maintaining practical computational requirements."}]},"datasets":{"status":"available","items":[{"name":"German EPEX CID market data","provider":"European Power Exchange (EPEX)","kind":"market-prices","access":"commercial","period":"2024-2025","universe":"German continuous intraday market","quote":"We demonstrate and evaluate our method using real data from the EPEX for the German CID market, with the test period spanning from April 14, 2024, to April 13, 2025."},{"name":"EPEX limit order book data","provider":"European Power Exchange (EPEX)","kind":"order-book","access":"commercial","period":"2024-2025","universe":"German continuous intraday market","quote":"The results indicate that features derived from the limit order book are the most influential exogenous variables."}],"known":[],"sample":{"start":"2024-04","end":"2025-04","market":"European continuous intraday power market (German CID)","asset_class":"commodities","frequency":"intraday"},"data_resolution":"l2-book"},"transaction_costs":{"status":"available","modeled":"unclear","headline_net_of_costs":false,"detail":null,"quote":null,"source":"conflict","rule":"R4b","candidates":{"front_matter":"no","evidence_card":"n/a","evidence_card_quoted":true},"conflict_kind":"applicability","card_detail":null,"card_quote":"Finally, although our directional forecasts show promise, we did not evaluate their profitability within realistic trading strategies.","meaning":"the platform classifier and the paper-text extraction disagree, so no value is asserted"},"results":{"status":"available","evaluation_type":"unclear","findings":[{"finding":"Features derived from the limit order book are the most influential exogenous variables for forecasting.","kind":"predictive-accuracy","quote":"The results indicate that features derived from the limit order book are the most influential exogenous variables."},{"finding":"Including features from neighboring products, especially those with overlapping delivery start times, improves forecast accuracy.","kind":"predictive-accuracy","quote":"In addition, features from neighboring products—especially those with delivery start times that overlap with the trading period of the target product—improve forecast accuracy."},{"finding":"Gradient boosting model accuracy ranges from 51.73% to 53.00% across different lead‑times, with the best performance (53.00%) using LOB top‑MW features for the 1h→h2 horizon.","kind":"performance","quote":"Feature Set Period 3ℎ→2ℎ 2ℎ→1ℎ 1ℎ→ℎ2 Current 51.73 51.26 53.00"},{"finding":"The robustness and high accuracy observed in the shortest lead‑time period (1h→h2) highlight its suitability for fast trading strategies.","kind":"performance","quote":"The robustness and high accuracy observed in the shortest lead‑time period (1ℎ→ℎ2 ) highlight its particular suitability for trading strategies that rely on fast and reliable market signals."}],"reported_performance":{"strategy":"directional electricity price forecasting model","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":17,"citations":{"cited_by_in_corpus":3,"by_independent_authors":3,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":3,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"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":"2026-11-14T17:29:43Z"},"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/"}