{"paper_id":"2503.18005","title":"A Simple Strategy to Deal with Toxic Flow","authors":["Álvaro Cartea","Leandro Sánchez-Betancourt"],"paper_date":"2025-03-23","paper_type":"theoretical","paper_url":"https://arxiv.org/pdf/2503.18005v1","page_url":"https://thequant.space/flowcharts/a-simple-strategy-to-deal-with-toxic-flow/","scores":{"empirical_rigor":2,"math_complexity":9,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes an internalisation‑externalisation strategy for a broker trading with informed and uninformed counterparties, derives closed‑form optimal strategies, and provides an algorithm that bypasses the need to calibrate individual model parameters.","quote":"The broker’s strategy optimally balances three key objectives: (i) internalise and externalise client flow, (ii) manage inventory risk, and (iii) infer market signals to execute speculative trades. 4 Conclusions We developed an internalisation-externalisation strategy for a broker who trades with informed and uniformed counterparties.","summary":"We model the trading activity between a broker and her clients (informed and uninformed traders) as an infinite-horizon stochastic control problem. We derive the broker’s optimal dealing strategy in closed form and use this to introduce an algorithm that bypasses the need to calibrate individual…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The broker and the informed trader use the same discount rate.","type":"model","stated":"explicit","quote":"3For simplicity, we assume that the broker and the informed trader use the same discount rate.","support":"full"},{"assumption":"The discount rate β is large enough so that the expression inside the square root is non‑negative.","type":"model","stated":"explicit","quote":"Here, we assume that the value of the discount rate β is large enough so that the expression inside of the square root is non-negative – note that if 2 aB > b the expression 4 aB β kB −2 b β kB + β2 (kB)2 + 2 kB ϕB is non-negative for β > 0.","support":"full"},{"assumption":"Parameter condition 2 aB > b holds.","type":"model","stated":"explicit","quote":"In what follows, we assume that 2 aB > b.","support":"full"},{"assumption":"When the discount rate β is large, the sign of G is positive.","type":"model","stated":"explicit","quote":"We remark that when the value of the discount rate β is large, the sign of G is positive.","support":"full"},{"assumption":"The broker’s optimal strategy is a linear combination of four state variables: broker inventory, informed trader inventory, volume traded by informed traders, and volume traded by uninformed traders.","type":"model","stated":"explicit","quote":"Specifically, the broker’s strategy is a linear combination of four state variables: broker’s inventory, informed trader’s inventory, vol-ume traded by informed traders, and volume traded by uniformed traders.","support":"full"},{"assumption":"The broker classifies clients into informed and uninformed groups for parameter learning.","type":"data","stated":"explicit","quote":"First, classify clients into informed and uninformed.","support":"full"}]},"datasets":{"status":"none_found","items":[],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"unclear"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R4a","candidates":{"front_matter":"n/a","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"kI = kU = 1.0 × 10⁻³, kB = 1.2 × 10⁻³","card_quote":"The price impact and penalty parameters are kI = kU = 1.0 × 10−3, kB = 1.2 × 10−3, aI = 1, aB = 1, and ϕB = ϕI = 10−2.","meaning":"the paper reports no trading or execution result; the cost the extractor found is a model parameter"},"results":{"status":"available","evaluation_type":"none","findings":[{"finding":"The broker’s optimal dealing strategy can be expressed in closed form and implemented without calibrating individual model parameters.","kind":"theoretical","quote":"We derive the broker's optimal dealing strategy in closed form and use this to introduce an algorithm that bypasses the need to calibrate individual parameters, so the dealing strategy can be executed in real-world trading environments.","support":"full"},{"finding":"The broker trades at a loss to the informed trader, and these losses increase when the informed trader penalises inventory holdings less.","kind":"performance","quote":"As discussed above, the broker trades at a loss to the informed trader. These losses are exacerbated when the informed trader hardly penalises inventory holdings because his strategy becomes more responsive to the signal.","support":"full"},{"finding":"Parameter combinations that yield high value for the informed trader correspond to low value for the broker.","kind":"statistical","quote":"We observe that parameter combinations of κ and ϕ that result in high values of the informed trader’s value function, are those that result in low values for the broker.","support":"full"},{"finding":"When the discount rate β is large, the sign of the coefficient G in the broker’s strategy is positive, influencing the strategy’s behavior.","kind":"theoretical","quote":"We remark that when the value of the discount rate β is large, the sign of G is positive.","support":"full"},{"finding":"The algorithm for learning the broker’s strategy parameters (c∞*, g∞*, h∞*, f∞*) is interpretable and easy to implement.","kind":"other","quote":"The outcome of an optimisation of the function above is an interpretable set of parameters (c∞∗, g∞∗, h∞∗, f ∞∗) that characterises the trading strategy of the broker.","support":"full"}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":19,"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":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":2,"extracted_at":"2026-11-14T19:51:18Z"},"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/"}