{"paper_id":"2406.16600","title":"Profit Maximization In Arbitrage Loops","authors":["Yu Zhang","Zichen Li","Tao Yan","Qianyu Liu","Nicolo Vallarano","Claudio Tessone"],"paper_date":"2024-06-24","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2406.16600v1","page_url":"https://thequant.space/flowcharts/profit-maximization-in-arbitrage-loops/","scores":{"empirical_rigor":6,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes three strategies—MaxPrice, MaxMax, and a Convex Optimization strategy—to maximize monetized arbitrage profit in DEX arbitrage loops.","quote":"we put forward three different strategies to maximize the monetized arbitrage profit for each arbitrage loop.","summary":"Cyclic arbitrage chances exist abundantly among decentralized exchanges (DEXs), like Uniswap V2. For an arbitrage cycle (loop), researchers or practitioners usually choose a specific token, such as Ether as input, and optimize their input amount to get the net maximal amount of the specific token…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The price of each token in a CEX is given (e.g., Px = 2$, Py = 10.2$, Pz = 20$).","type":"data","stated":"explicit","quote":"We assume the price of each token in a CEX is given, namely, Px = 2$, Py = 10.2$, Pz = 20$"},{"assumption":"Uniswap V2 liquidity pools follow the constant product market maker equation with a transaction tax rate λ.","type":"model","stated":"explicit","quote":"The constant AMM equation in any liquidity pool of Uniswap V2 is in the form of [x + (1 −λ)∆x](y −∆y) = xy = k, where ... λ"},{"assumption":"Arbitrage profit is maximal when the marginal output equals marginal input (d∆xout/d∆xin = 1).","type":"model","stated":"explicit","quote":"The arbitrage profit is maximal when d∆xout d∆xin = 1."},{"assumption":"Traders concentrate only on arbitrage loops starting and ending with a single token (e.g., X→Y→Z→X) and ignore other permutations.","type":"execution","stated":"explicit","quote":"We assume traders only concentrate on the arbitrage starting from token X and backing to token X again (X →Y → Z →X), but ignore the other two ways of arbitrage"},{"assumption":"A token loop is an arbitrage loop if the product of relative prices along the loop exceeds 1 (pij·pjk·pki > 1).","type":"market","stated":"explicit","quote":"A token loop is called an arbitrage loop when the product of all relative prices along the token loop is larger than 1."},{"assumption":"If log(p*ij)+log(p*jk)+log(p*ki)=0 after optimization, no arbitrage opportunities remain for any token in the loop.","type":"other","stated":"explicit","quote":"Because of log(p∗ij) + log(p∗jk) + log(p∗ki) = 0, there are also no arbitrage chances starting from other tokens in the loop"}]},"datasets":{"status":"available","items":[{"name":"Token graph constructed from on‑chain data","provider":"on‑chain data (Uniswap V2) and Binance prices via CoinGecko ","kind":"crypto-onchain","access":"public","period":null,"universe":"51 tokens, 208 liquidity pools","quote":"The token graph we got had fifty‑one nodes (tokens) and two hundred and eight edges (liquidity pools). The token’s price from CEX (Binance) was downloaded on CoinGecko by API.","known_id":"binance"}],"known":[{"id":"binance","name":"Binance","provider":"Binance","kind":"market-prices","access":"public","quote":"The token’s price from CEX (Binance) was downloaded on CoinGecko by API.","mentions":1,"used_in":1},{"id":"coingecko","name":"CoinGecko","provider":"CoinGecko","kind":"market-prices","access":"public","quote":"The token’s price from CEX (Binance) was downloaded on CoinGecko by API.","mentions":1,"used_in":1}],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"unclear"},"transaction_costs":{"status":"not_extracted","modeled":"unclear","headline_net_of_costs":null,"detail":null,"quote":null,"source":"none","rule":"R0","candidates":{"front_matter":"unclear","evidence_card":"unclear","evidence_card_quoted":false},"meaning":"the classifier was not confident"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"Empirically, the Convex Optimization strategy yields monetized profit almost equal to the MaxMax strategy.","kind":"performance","quote":"all points are almost on the 45◦line, which means that the profit from the Conevx Optimization strategy is almost equal to that from the MaxMax strategy."},{"finding":"The MaxPrice strategy is not reliable for obtaining maximal monetized arbitrage profit compared to MaxMax.","kind":"performance","quote":"the MaxPrice strategy is not reliable in getting the maximal monetized arbitrage profit compared to the MaxMax strategy."},{"finding":"Theoretically, the Convex Optimization strategy can obtain more monetized arbitrage profit than the MaxMax strategy.","kind":"theoretical","quote":"proved in theory that more monetized arbitrage profit can be obtained by the Conevx Optimization strategy than the MaxMax strategy."},{"finding":"If no arbitrage profit is found by the MaxMax strategy, the Convex Optimization strategy also cannot detect any arbitrage profit.","kind":"theoretical","quote":"if no arbitrage profit can be found by the traditional arbitrage strategies or the MaxMax strategy, then no arbitrage profit can be detected by the ConvexOptimization strategy, either."},{"finding":"Both traditional strategies and the MaxPrice strategy are not reliable for achieving maximal monetized arbitrage profit.","kind":"negative","quote":"both the tradi‑tional strategies and the MaxPrice strategy are not reliable if we focus on getting the maximal monetized arbitrage profit."}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":26,"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":1,"extracted_at":"2026-11-14T18:07:22Z"},"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/"}