{"paper_id":"2412.17712","title":"Broker-Trader Partial Information Nash-Equilibria","authors":["Xuchen Wu","Sebastian Jaimungal"],"paper_date":"2024-12-23","paper_type":"theoretical","paper_url":"https://arxiv.org/pdf/2412.17712v2","page_url":"https://thequant.space/flowcharts/broker-trader-partial-information-nash-equilibria/","scores":{"empirical_rigor":1,"math_complexity":9.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a multi‑asset broker‑trader model with partial information where the informed trader holds a trading signal and the broker off‑loads inventory, and characterises the Nash equilibrium as the solution of a filtered system of forward‑backward stochastic differential equations (FBSDEs","quote":"We study partial information Nash equilibrium between a broker and an informed trader. Furthermore, we establish that this equilibrium corresponds to the solution of a nonstandard system of forward-backward stochastic differential equations (FBSDEs) that involves the two differing filtrations.","summary":"We study partial information Nash equilibrium between a broker and an informed trader. In this setting, the informed trader, who possesses knowledge of a trading signal, trades multiple assets with the broker in a dealer market. Simultaneously, the broker offloads these assets in a lit exchange…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The broker only observes aggregate asset prices and cannot separate underlying trends from volatility.","type":"market","stated":"explicit","quote":"The broker, however, only observes aggregate prices and cannot distinguish between underlying trends and volatility.","support":"full"},{"assumption":"The broker and the informed trader have different filtrations: the broker’s strategies are G‑adapted while the trader’s are F‑adapted.","type":"model","stated":"explicit","quote":"This leads to a partial information setting for the Broker, but a full information setting for the informed trader. Of particular note is the differing ﬁltrations that the two agents have.","support":"full"},{"assumption":"Transient price impact is linear with constant positive semi‑definite matrices p and h that commute.","type":"execution","stated":"explicit","quote":"Transient price impact model: The broker’s trading activity in the lit exchange has transient impact on the asset price, which we denote Y and whose dynamics is assumed3 to be: ( dY ν t = (h νt −p Y ν t ) dt , Y ν 0 = y , where y ∈L2(G0), p and h are K × K constant positive semi-deﬁnite matrices representing decay coefﬁcients and instantaneous impact parameters, respectively. We further assume p and h commute.","support":"full"},{"assumption":"The time horizon is short enough so that a unique solution to the FBSDE system exists.","type":"other","stated":"explicit","quote":"For short enough time horizons, we prove that a unique solution of this system exists.","support":"full"},{"assumption":"The performance criteria for both agents are strictly concave functionals of their trading strategies.","type":"statistical","stated":"explicit","quote":"Speciﬁcally, we show JI and JB are strictly concave and Gâuteaux-differentiable in the component over which the corresponding agent maximizes. Note that, unlike in [9], we obtain the following result without any further restriction on a, p, h, and T . Proposition 3.2.","support":"full"},{"assumption":"Matrices p and h are simultaneously diagonalizable (they commute and are positive semi‑definite).","type":"model","stated":"explicit","quote":"re positive semi-deﬁnite and commute, they are simultaneously diagonalizable (see, for example, Theorem 5.76 of [4]), that is, there is an invertible K × K matrix O such that p = OˆpO−1 and h = OˆhO−1, where ˆp and ˆh are diagonal matrices whose diagonals consist of eigenvalues of p and h, which are nonnegative, respectively.","support":"full"}]},"datasets":{"status":"none_found","items":[],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"none"},"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":"Transient price impact is modelled via linear dynamics dY = (h ν – p Y) dt with p and h positive semi‑definite matrices.","card_quote":"Transient price impact model: The broker’s trading activity in the lit exchange has transient impact on the asset price, which we denote Y and whose dynamics is assumed3 to be: ( dY ν t = (h νt −p Y ν t ) dt , Y ν 0 = y , where y ∈L2(G0), p and h are K × K constant positive semi-deﬁnite matrices representing decay coefﬁcients and instantaneous impact parameters, respectively.","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":"A Nash equilibrium exists and is unique for sufficiently short time horizons.","kind":"theoretical","quote":"Using convex analysis, we characterize the Nash equilibrium and demonstrate its existence and uniqueness. For short enough time horizons, we prove that a unique solution of this system exists.","support":"full"},{"finding":"The Nash equilibrium is exactly characterised by a filtered system of FBSDEs involving the two differing filtrations.","kind":"theoretical","quote":"Furthermore, we establish that this equilibrium corresponds to the solution of a nonstandard system of forward-backward stochastic differential equations (FBSDEs) that involves the two differing filtrations. As a result, we obtain a ﬁltered system of forward-backward stochastic differential equations (FBSDEs) that characterizes the Nash equilibrium.","support":"full"},{"finding":"The solution to the FBSDE system admits a polynomial (asymptotic) approximation in the strength of the transient impact parameter to arbitrary order, with controlled error.","kind":"theoretical","quote":"Finally, under quite general assumptions, we show that the solution to the FBSDE system admits a polynomial approximation in the strength of the transient impact to arbitrary order, and prove that the error is controlled. The approximation we develop is represented as an asymptotic series, where each term in the series admits a closed form solution in terms of the previous coefﬁcients in the expansion.","support":"full"},{"finding":"An explicit perturbation analysis yields closed‑form expressions for each term of the asymptotic series representing the equilibrium strategies.","kind":"theoretical","quote":"Speciﬁcally we obtain a representation of the form ηt = M X m=0 hmηm t + Rη,M,h t , and νt = M X m=0 hmνm t + Rν,M,h t , where we provide explicit representations of (ηm t , νm t )m. To this end, the following lemma provides the FBSDEs that each term in the series satisﬁes, and proves that a unique solution exists, as well as provides the explicit form of the solution. Lemma 5.1.","support":"full"},{"finding":"The remainder terms from truncating the asymptotic series at order M are o(M).","kind":"theoretical","quote":"Moreover, we prove that the error terms, Rη,M,h t and Rν,M,h t , from truncating the series at a given order M, are both o(M). Moreover, we prove the remainder terms from truncating the series at order M are o(M).","support":"full"}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[],"contradicts":[],"qualifies":[],"compared_with":34,"citations":{"cited_by_in_corpus":6,"by_independent_authors":5,"how_cited":{"used_as_baseline":0,"uses_method_or_data":1,"extends":0,"discusses_findings":0,"mentions":4,"unclassified":1},"independent_papers":[{"paper":{"paper_id":"2610.03598","title":"When a Correct Reward Is Not Enough: Diagnosing and Guiding PPO in an Analytically Solved Broker-Trader Game","paper_date":"2026-10-02","authors":["Siu Tung Wong","Carlo Campajola"],"empirical_rigor":7,"page_url":"https://thequant.space/flowcharts/when-a-correct-reward-is-not-enough-diagnosing-and-guiding-ppo-in-an/"},"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":2,"extracted_at":"2026-11-14T19:53:26Z"},"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/"}