{"paper_id":"2411.13403","title":"Path weighting sensitivities","authors":["Liu Xuan","Gauthier Michel"],"paper_date":"2024-11-20","paper_type":"methodological","paper_url":"https://arxiv.org/pdf/2411.13403v1","page_url":"https://thequant.space/flowcharts/path-weighting-sensitivities/","scores":{"empirical_rigor":4,"math_complexity":9,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes an explicit path weighting formula for spot sensitivities of path‑dependent derivatives, a variance‑reduction adjustment for the Monte‑Carlo estimator, and a covariance‑inflation technique to handle singular covariance matrices.","quote":"We propose explicit path weighting formula and variance reduction adjustment in order to address the large variance happening when the first simulation time step is small. For the second problem, we propose a variance reduction adjustment, which improves the convergence significantly at the cost of one more pricing for full Delta vector and two more pricing for full Gamma matrix.","summary":"In this paper, we study the computation of sensitivities with respect to spot of path dependent financial derivatives by means of path weighting. We propose explicit path weighting formula and variance reduction adjustment in order to address the large variance happening when the first simulation…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Stock dynamics follow a Markovian diffusion model dSt = a(t,St)dt + b(t,St)dWt with a correlation matrix Σ.","type":"model","stated":"explicit","quote":"2 The Path Weighting Formula 2.1 General Case Denote the stock prices at time t by a column vector St = \u0000S(1) t ,...,S(n) t \u0001T. Consider a Markovian diffusion model for stock dynamics dSt = a(t,St)dt +b(t,St)dWt, d⟨W (i),W (j)⟩= Σijdt, where a(t,s), b(t,s) are Rn×n-valued functions, and Σ is the correlation matrix.","support":"full"},{"assumption":"The functionals Hi have continuous second‑order derivatives with at most polynomial growth (regularity assumption for variance analysis).","type":"statistical","stated":"explicit","quote":"The functionals Hi have continuous second order derivatives with at most polynomial growth at infinity.","support":"full"},{"assumption":"The payoff function F(S) does not contain the first simulation time t1 as an event time (or an extra time point is inserted).","type":"model","stated":"explicit","quote":"We should point out that the assumption t1 not in event times of a Bermudan type payoff F can be satisfied by inserting an extra simulation time point. Suppose that the payoff function F(S) does not contain t1 as its event time.","support":"full"}]},"datasets":{"status":"available","items":[{"name":"test_spot_ladder","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"Results shown in Firgure 3.2, Figure 3.3, and Figure 3.4 are generated by the test case test_spot_ladder in pwsen.","support":"full"},{"name":"test_variance_reduction","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"Numerical test results in this example are generated by the test case test_variance_reduction in pwsen.","support":"full"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"synthetic"},"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":"unclear","findings":[{"finding":"For unsmoothed payoffs, the path‑weighting (PW) estimator has slightly higher standard deviation than finite differences (FD) for most cases, with the gap widening when correlation approaches ±1.","kind":"statistical","quote":"As shown by Table 1, for Delta of unsmoothed payoff, PW has slightly higher standard deviation than FD for most of the cases, while the difference becomes more significant when correlation is close to 1 (similar result holds for correlation −1).","support":"full"},{"finding":"Applying the variance‑reduction adjustment (RPW) yields a huge variance reduction compared to the raw PW estimator.","kind":"statistical","quote":"For all cases, there is huge variance reduction from RPW to PW.","support":"full"},{"finding":"For Gamma, PW generally shows smaller standard deviation than FD, except when correlation is near ±1 where FD performs better.","kind":"statistical","quote":"As shown by Table 2, for Gamma of unsmoothed payoff, PW has smaller standard deviation than FD for most of the cases, while PW has larger standard deviation than FD when correlation is close to 1 (similar result holds for correlation −1).","support":"full"},{"finding":"The covariance‑inflation technique further reduces Monte‑Carlo variance and enables handling of near‑singular covariance matrices.","kind":"statistical","quote":"lation is close to being singular or volatility is close to zero) is to insert an extra small simulation time point (1/360 for example), and then apply Theorem 3.3 to the covariance inflated process Sε1,ε2 t , by which we sacrifice the property of being unbiased for the ability to handle singular covariance matrix and also a smaller MC variance.","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":47,"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":"2026-11-14T19:53:50Z"},"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/"}