{"paper_id":"2410.00854","title":"Impermanent loss and loss-vs-rebalancing I: some statistical properties","authors":["Abe Alexander","Lars Fritz"],"paper_date":"2024-10-01","paper_type":"methodological","paper_url":"https://arxiv.org/pdf/2410.00854v2","page_url":"https://thequant.space/flowcharts/impermanent-loss-and-loss-vs-rebalancing-i-some-statistical/","scores":{"empirical_rigor":2,"math_complexity":7.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper analyzes the statistical properties of impermanent loss (IL) and loss‑versus‑rebalancing (LVR) and shows that, for Brownian motion with a given volatility, they have identical expectation values while their distribution functions differ markedly.","quote":"In this short paper we shed light on the statistical aspects of both concepts and show that they are more similar than conventionally appreciated. Our main finding can be summarized in one sentence: For Brownian motion with a given volatility, IL and LVR have identical expectation values but vastly differing distribution functions.","summary":"There are two predominant metrics to assess the performance of automated market makers and their profitability for liquidity providers: ‘impermanent loss’ (IL) and ’loss-versus-rebalance’ (LVR). In this short paper we shed light on the statistical aspects of both concepts and show that they are…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The analysis assumes non‑toxic or unspecific trading.","type":"market","stated":"explicit","quote":"We consider non-toxic or rather unspecific trading in this paper.","support":"full"},{"assumption":"Price dynamics follow a Brownian motion (or geometric Brownian motion) with a given volatility.","type":"statistical","stated":"explicit","quote":"Our main finding can be summarized in one sentence: For Brownian motion with a given volatility, IL and LVR have identical expectation values but vastly differing distribution functions. This leads to a price that follows Brownian motion.","support":"full"},{"assumption":"The random‑walk simulation uses price updates pi+1 = pi + σ0 ξ with ξ = ±1 drawn with equal probability.","type":"model","stated":"explicit","quote":"We define it such that the price at time t+1 obtains from the price at time t according to pi+1 = pi + σ0ξ (4) where σ0 is the variance and ξ is ±1, both drawn with equal probability.","support":"full"},{"assumption":"Price changes are small such that dp ≪ p, allowing a second‑order expansion.","type":"model","stated":"explicit","quote":"We start with assuming that the changes in price dp ≪p and expand the expression for LVR according to ∆LV R(p, p + dp) = L √p \u0012 1 − r p p + dp \u00132 ≈L 4 dp2 p5/2 For Brownian motion it is well known that dp2 = σ0dt meaning we find ∆LVR(p, ∆t) = L 4 σ2 0dt p5/2 .","support":"full"},{"assumption":"The constant‑product AMM satisfies the invariant xy = L² with a starting condition x0y0 = L².","type":"model","stated":"explicit","quote":"(1) 2 We assume from now on that there is a starting condition at time t = 0 defined by x0y0 = L2 (2) and price p0 = y0/x0.","support":"full"},{"assumption":"For Brownian motion, the infinitesimal variance satisfies dp² = σ0 dt.","type":"statistical","stated":"explicit","quote":"We start with assuming that the changes in price dp ≪p and expand the expression for LVR according to ∆LV R(p, p + dp) = L √p \u0012 1 − r p p + dp \u00132 ≈L 4 dp2 p5/2 For Brownian motion it is well known that dp2 = σ0dt meaning we find ∆LVR(p, ∆t) = L 4 σ2 0dt p5/2 .","support":"full"}]},"datasets":{"status":"available","items":[{"name":"Synthetic random‑walk simulations of price paths","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"For numerical simulations, we simulate the price movement with a simple random walk. A characteristic single run for a starting price of p0 = 100, σ0 = 0.001, and 5000 (blue) and 20000 (orange) steps is shown in Fig. 1.","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":"R1","candidates":{"front_matter":"n/a","evidence_card":"no","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"none","findings":[{"finding":"For Brownian motion with a given volatility, IL and LVR have identical expectation values.","kind":"theoretical","quote":"Our main finding can be summarized in one sentence: For Brownian motion with a given volatility, IL and LVR have identical expectation values but vastly differing distribution functions.","support":"full"},{"finding":"For small price movements, IL and LVR yield identical results despite different interpretations and distributions.","kind":"theoretical","quote":"In this paper, we find that for small price movements, impermanent loss (IL) and loss-versus-rebalancing (LVR) yield identical results, despite their different interpretations and distributions.","support":"full"},{"finding":"Numerical simulations show that IL and LVR have the same average value across many parameter settings.","kind":"statistical","quote":"We find, maybe surprisingly at this point, that while both quantities have very different distribution functions, they possess the same average.","support":"full"},{"finding":"Analytical calculations confirm ⟨IL(t)⟩ = ⟨LVR(t)⟩.","kind":"theoretical","quote":"(11) We thus conclude that we manged to show that ⟨IL(t)⟩= ⟨LVR(t)⟩, as we already observed from the numerical simulation.","support":"full"},{"finding":"For 5000 simulation steps the average of both IL and LVR is 0.00125.","kind":"statistical","quote":"The ana-lytical prediction for the plots shown in Fig. 4 and Fig. 5 is ⟨LVR(5000)⟩= ⟨IL(5000)⟩= 0.00125 for both averages.","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":52,"citations":{"cited_by_in_corpus":1,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":1},"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":"2026-11-14T19:54:57Z"},"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/"}