{"paper_id":"2501.11164","title":"A statistical technique for cleaning option price data","authors":["Jaco Visagie"],"paper_date":"2025-01-19","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2501.11164v1","page_url":"https://thequant.space/flowcharts/a-statistical-technique-for-cleaning-option-price-data/","scores":{"empirical_rigor":7,"math_complexity":4,"quadrant":"Street Traders"},"proposes":{"status":"available","statement":"The paper proposes a statistical technique for cleaning option price data that identifies and removes arbitrage violations and outliers without relying on option pricing model assumptions.","quote":"In this paper, we list three distinct problems often found in recorded option price datasets alongside means of addressing these. The methods used are justified using sound statistical reasoning and remove option prices violating the standard assumption of no arbitrage.","summary":"Recorded option pricing datasets are not always freely available. Additionally, these datasets often contain numerous prices which are either higher or lower than can reasonably be expected. Various reasons for these unexpected observations are possible, including human error in the recording of…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Option prices must satisfy no‑arbitrage bounds for European calls and puts.","type":"model","stated":"explicit","quote":"The methods used are justified using sound statistical reasoning and remove option prices violating the standard assumption of no arbitrage.","support":"full"},{"assumption":"All datasets share the same risk‑free interest rate, taken as the six‑month Treasury bill rate.","type":"market","stated":"explicit","quote":"Since all of these datasets are obtained at the same time and in the same market, the same risk free interest rate will be used for each dataset. We use the discount rate of Treasury bills in the secondary market as the risk free interest rate.","support":"full"},{"assumption":"Dividend yields for all underlying assets are zero.","type":"market","stated":"explicit","quote":"As a result, the dividend rate is 0 for all underlying assets considered.","support":"full"},{"assumption":"Residuals from the polynomial fit are i.i.d. normal with mean zero.","type":"statistical","stated":"explicit","quote":"Assume that the residuals are independently and identically 3 A statistical technique for cleaning option price data 450 500 550 600 650 700 750 0 50 100 150 Strike price Option price Figure 1: Google call option prices with the ﬁtted second degree polynomial. distributed normal random variables with a mean of 0 and standard deviation of σ.","support":"full"},{"assumption":"Options are actively traded, so any arbitrage opportunity would be exploited immediately.","type":"market","stated":"implied","quote":"Since the options considered are actively traded, it is reasonable to assume that these arbitrage opportu-nities would be exploited immediately.","support":"full"},{"assumption":"The analysis is restricted to European call and put options.","type":"other","stated":"explicit","quote":"Regarding the type of option under consideration, we restrict our attention to European call and put options below.","support":"full"}]},"datasets":{"status":"available","items":[{"name":"S&P 500 options dataset","provider":"Yahoo Finance","kind":"options","access":"public","period":"11 May 2012","universe":"S&P 500 index call and put options","quote":"The data used consist of European call and put option prices on the S&P 500 index, the PowerShares index and Google shares. On 11 May 2012 the S&P 500 index closed at $1353.39.","support":"full","known_id":"yahoo-finance"},{"name":"PowerShares index options dataset","provider":"Yahoo Finance","kind":"options","access":"public","period":"11 May 2012","universe":"PowerShares index call and put options","quote":"The PowerShares index is an index weighted by market capatilisation which consists of the largest non-ﬁnancial companies the world over. On this day the PowerShares index closed at $64.18.","support":"partial","known_id":"yahoo-finance"},{"name":"Google Inc. options dataset","provider":"Yahoo Finance","kind":"options","access":"public","period":"11 May 2012","universe":"Google stock call and put options","quote":"On the same date the share price of Google Inc. closed at $605.23. We obtained the prices of 545 call options and 532 put options on this stock from [3].","support":"partial","known_id":"yahoo-finance"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"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 the S&P 500 dataset, 146 call options (25%) and 129 put options (17%) were removed during cleaning.","kind":"statistical","quote":"This means that 146 call options and 129 put options were removed during the cleaning process, which corresponds to 25% and 17% of the original datasets, respectively.","support":"partial"},{"finding":"For the PowerShares dataset, 29% of call options were removed.","kind":"statistical","quote":"Turning our attention to the PowerShares dataset, we see that the data cleaning process reduces the number of call options from 413 to 293, which means that 29% of the options were removed.","support":"full"},{"finding":"For the PowerShares put options, 41% of the recorded options were removed.","kind":"statistical","quote":"In this case 41% of the option prices are removed by cleaning.","support":"full"},{"finding":"For the Google dataset, call options were reduced by 5% and put options by 16% after cleaning.","kind":"statistical","quote":"For the Google Inc. dataset, the number of call options is reduced from 545 to 519, constituting a 5% reduction in the number of options in the dataset. The number of options used is reduced by 16% in this instance.","support":"full"},{"finding":"The most extreme case required removal of 41% of recorded options to ensure realistic prices.","kind":"statistical","quote":"In the most extreme case (that of the PowerShares put options), it was necessary to remove 41% of the recorded options in order to ensure that the prices are realistic.","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":11,"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":true,"code_url":"https://github.com/JacoVisagieNWU/Option-pricing-data","note":null},"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":"2026-11-14T19:53:02Z"},"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/"}