{"paper_id":"2408.03594","title":"Forecasting High Frequency Order Flow Imbalance","authors":["Aditya Nittur Anantha","Shashi Jain"],"paper_date":"2024-08-07","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2408.03594v1","page_url":"https://thequant.space/flowcharts/forecasting-high-frequency-order-flow-imbalance/","scores":{"empirical_rigor":8,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes using Hawkes processes with a Sum of Exponentials kernel to forecast the near‑term distribution of Order Flow Imbalance and a framework to compare many forecasting models.","quote":"observe that the Hawkes process modeled with a Sum of Exponential’s kernel gives the best forecast among all competing models.","summary":"Market information events are generated intermittently and disseminated at high speeds in real-time. Market participants consume this high-frequency data to build limit order books, representing the current bids and offers for a given asset. The arrival processes, or the order flow of bid and offer…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Trades can be deterministically classified as BUY or SELL using the two exchange order numbers in the NSE tick data.","type":"data","stated":"explicit","quote":"This allows for a deterministic classification of every trade."},{"assumption":"The cross‑effect between BUY and SELL market order flows is essential and must be captured by the model.","type":"model","stated":"explicit","quote":"Note that accurately capturing the cross effect of the buy and the sell signed trades on each other is essential for the latter."},{"assumption":"Order flow arrivals can be modeled as a Hawkes point process.","type":"model","stated":"explicit","quote":"Bacry et al. [2015] and Fosset et al. [2022] model the market order point process using a Hawkes process."},{"assumption":"The loss function used for model comparison is correctly specified and can be aggregated across models.","type":"statistical","stated":"explicit","quote":"We develop a general framework for computing the loss function."},{"assumption":"The tick data from a single day (September 19th 2018) is representative for evaluating the forecasting methods.","type":"data","stated":"implied","quote":"We use processed tick data recorded on September 19th 2018 from the National Stock Exchange of Nifty futures expiring on September 27th 2018."},{"assumption":"Inter‑trade arrival times contain information about the dependency structure of the limit order book.","type":"model","stated":"explicit","quote":"Our results show that adequately modeling the inter-arrival time of trades and the dependency structure of the limit order book, as detailed previously, leads to a significant improvement in forecast quality."}]},"datasets":{"status":"available","items":[{"name":"NSE Nifty futures tick data (Sept 2018)","provider":"National Stock Exchange","kind":"order-book","access":"unknown","period":"September 19th 2018","universe":"NIFTY futures expiring September 27th 2018","quote":"We use processed tick data recorded on September 19th 2018 from the National Stock Exchange of Nifty futures expiring on September 27th 2018 (current month futures contract)."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"tick"},"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":"The Hawkes process with a Sum of Exponential kernel provides the best OFI forecast among all competing models.","kind":"predictive-accuracy","quote":"observe that the Hawkes process modeled with a Sum of Exponential’s kernel gives the best forecast among all competing models."},{"finding":"Incorporating inter‑trade arrival times and the dependency structure of the limit order book significantly improves forecast quality.","kind":"predictive-accuracy","quote":"We observe a significant enhancement in forecast quality when considering the dependency structure of the limit order book through inter-trade arrival times."},{"finding":"A method to forecast the near‑term distribution of OFI is presented.","kind":"theoretical","quote":"In this paper, we present a method to forecast a near-term distribution of the OFI."},{"finding":"A general framework for computing the loss function enables comparison of multiple forecasting models.","kind":"theoretical","quote":"We develop a general framework for computing the loss function."},{"finding":"Modeling the time distribution of market order events and the temporal dependence between BUY and SELL flows enhances confidence in selecting a benchmark model in high‑frequency settings.","kind":"statistical","quote":"By modeling the time distribution of market order events and the temporal dependence between ‘BUY’ and ‘SELL’ market order flows, this paper demonstrates that incorporating the inter-arrival time of trades significantly enhances the confidence with which a benchmark model can be chosen in high-frequency settings."}],"reported_performance":null,"out_of_sample":{"method":"walk-forward","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[],"contradicts":[],"qualifies":[],"compared_with":20,"citations":{"cited_by_in_corpus":2,"by_independent_authors":2,"how_cited":{"used_as_baseline":0,"uses_method_or_data":2,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[{"paper":{"paper_id":"2411.08382","title":"Hybrid Vector Auto Regression and Neural Network Model for Order Flow Imbalance Prediction in High Frequency Trading","paper_date":"2024-11-13","authors":["Abdul Rahman","Neelesh Upadhye"],"empirical_rigor":7.5,"page_url":"https://thequant.space/flowcharts/hybrid-vector-auto-regression-and-neural-network-model-for-o/"},"how":"uses_method_or_data","quote":null},{"paper":{"paper_id":"2510.22685","title":"TABL-ABM: A Hybrid Framework for Synthetic LOB Generation","paper_date":"2025-10-26","authors":["Ollie Olby","Rory Baggott","Namid Stillman"],"empirical_rigor":6,"page_url":"https://thequant.space/flowcharts/tabl-abm-a-hybrid-framework-for-synthetic-lob-generation/"},"how":"uses_method_or_data","quote":null}],"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":"2026-11-14T18:04:39Z"},"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/"}