{"paper_id":"2405.18936","title":"Optimizing Broker Performance Evaluation through Intraday Modeling of Execution Cost","authors":["Zoltan Eisler","Johannes Muhle-Karbe"],"paper_date":"2024-05-29","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2405.18936v2","page_url":"https://thequant.space/flowcharts/optimizing-broker-performance-evaluation-through-intraday-mo/","scores":{"empirical_rigor":3,"math_complexity":7.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a methodology for evaluating broker execution algorithms by estimating linear and quadratic (impact) costs using intraday price and trading data, with novel weighting of price changes to improve estimation accuracy.","quote":"This paper presents a methodology for evaluating the effectiveness of broker execution algorithms using trading data. We focus on two primary cost components: a linear cost that quantifies short-term execution quality and a quadratic cost associated with the price impact of trades.","summary":"Minimizing execution costs for large orders is a fundamental challenge in finance. Firms often depend on brokers to manage their trades due to limited internal resources for optimizing trading strategies. This paper presents a methodology for evaluating the effectiveness of broker execution…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The broker executes orders over a fixed time interval [0,T] with a deterministic trading rate process qt.","type":"execution","stated":"explicit","quote":"2.1 Broker Trading We assume that the broker executes the orders over some fixed time period t ∈[0, T] with a trading rate qt, representing the broker’s adjustments of the position Qt = R t 0 qsds.","support":"full"},{"assumption":"Price impact follows the transient Obizhaeva‑Wang model with exponential decay at rate 1/τM.","type":"model","stated":"explicit","quote":"We describe this by the standard price impact model of Obizhaeva and Wang (2013):5 Mt −M0 = λ Z t 0 e−(t−t′)/τM qt′dt′ + σMW M t . Impact subsequently decays at rate 1/τM > 0, i.e., the half-life of impact is proportional to τM.","support":"full"},{"assumption":"The unaffected mid price follows a Brownian motion WM with volatility σM, independent of the Brownian motion driving the broker’s trading rate.","type":"market","stated":"explicit","quote":"(2.2) This means that the unaffected price follows a Brownian motion W M t with volatility σM > 0. We assume it is independent from the Brownian motion W q t driving the fluctuations of the broker’s trading rate.","support":"full"},{"assumption":"Impact decay timescale τM and spread s are known constants.","type":"model","stated":"explicit","quote":"(2.3) To simplify the analysis, we assume that both the impact decay timescale τM and the spread s are known constants.","support":"full"},{"assumption":"The execution price includes a fraction a of the bid‑ask spread, with a ≤ 1/2 in realistic cases.","type":"execution","stated":"explicit","quote":"The execution price Pt achieved by the broker additionally includes a fraction a of the bid-ask spread s > 0: Pt = Mt + as × sign(qt). 4 if a broker always aggressively crosses the spread to take liquidity, and has no skill in timing the orders to before favorable price moves, one expects a = 1/2.","support":"full"},{"assumption":"Linear cost is captured by slippage relative to continuously sampled mid prices, which filters out most impact cost and market noise.","type":"statistical","stated":"explicit","quote":"To address this, our study outlines how to exploit the transient nature of price impact to improve the signal-to-noise ratios considerably.1 To estimate linear costs, this can be achieved by replacing slippage relative to the “arrival price” with slippage relative to mid prices continuously sampled during the execution. We show that by doing so, one filters out a large part of impact cost and market noise, while","support":"full"}]},"datasets":{"status":"available","items":[{"name":"E‑mini S&P futures trading data (simulated numerical examples)","provider":null,"kind":"market-prices","access":"unknown","period":null,"universe":"E‑mini S&P futures contracts","quote":"2.4 Parameter Values In our numerical examples, we select sensible parameter values for trading E-mini S&P futures, see Table 1. To wit, the target trade size Q is taken to be roughly 0.2% of the real-life trading volume which is of the order of 106 contracts.","support":"full"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"synthetic"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"Linear cost = a·s·∫|qt|dt (spread component); impact cost = λ·∫∫ e^{-(t‑t')/τM} q_{t'} dt'·qt dt; market noise term also modeled.","quote":"We focus on two primary cost components: a linear cost that quantifies short-term execution quality and a quadratic cost associated with the price impact of trades. Using a model with transient price impact, we derive analytical formulas for estimating these costs.","source":"evidence_card","rule":"R2","candidates":{"front_matter":"unclear","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Linear cost = a·s·∫|qt|dt (spread component); impact cost = λ·∫∫ e^{-(t‑t')/τM} q_{t'} dt'·qt dt; market noise term also modeled.","card_quote":"We focus on two primary cost components: a linear cost that quantifies short-term execution quality and a quadratic cost associated with the price impact of trades. Using a model with transient price impact, we derive analytical formulas for estimating these costs.","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"none","findings":[{"finding":"The proposed estimators increase the signal‑to‑noise ratio for both linear and impact cost estimation by a factor of 6 to 7 for typical E‑mini S&P futures parameters.","kind":"statistical","quote":"For model parameters typical for the trading of E-mini S&P futures, we find that our simple alternative estimators increase the signal-to-noise ratio by a factor of 6 to 7. Such an improvement can significantly speed up the convergence of broker selection procedures that are ubiquitous in the industry.","support":"full"},{"finding":"Using slippage relative to continuously sampled mid prices filters out most impact cost and market noise while preserving the linear cost term.","kind":"theoretical","quote":"To address this, our study outlines how to exploit the transient nature of price impact to improve the signal-to-noise ratios considerably.1 To estimate linear costs, this can be achieved by replacing slippage relative to the “arrival price” with slippage relative to mid prices continuously sampled during the execution. We show that by doing so, one filters out a large part of impact cost and market noise, while","support":"full"},{"finding":"Weighting earlier price changes more heavily yields substantial improvements in impact‑cost estimation, matching recent findings on early‑trade information content.","kind":"theoretical","quote":"We show that by doing so, one filters out a large part of impact cost and market noise, while preserving the linear cost term.2 For impact costs, suitably re-weighting mid price changes with a greater emphasis on earlier rather than later trades leads to similar improvements.3 The intuition here is that price impact has a much stronger effect on prices at the beginning of execution, but then gradually saturates","support":"full"},{"finding":"The methodology provides a robust and efficient framework for selecting the most cost‑effective brokers.","kind":"performance","quote":"Our results demonstrate substantial improvements in estimating both linear and impact costs, providing a robust and efficient framework for selecting the most cost-effective brokers. While these specific improvements are derived in a particular model, the mechanisms underlying them are simple and robust.","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":17,"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":"2026-11-14T19:57:07Z"},"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/"}