{"paper_id":"2508.03217","title":"Measuring DEX Efficiency and The Effect of an Enhanced Routing Method on Both DEX Efficiency and Stakeholders' Benefits","authors":["Yu Zhang","Claudio J. Tessone"],"paper_date":"2025-08-05","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2508.03217v1","page_url":"https://thequant.space/flowcharts/measuring-dex-efficiency-and-the-effect-of-an-enhanced-routi/","scores":{"empirical_rigor":7.5,"math_complexity":8.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes the Standardized Total Arbitrage Profit (STAP) as a systematic measure of DEX efficiency and evaluates the impact of two token routing algorithms (a line‑graph‑based method and a depth‑first search algorithm) on DEX efficiency and stakeholder benefits.","quote":"This paper introduces the concept of Standardized Total Arbitrage Profit (STAP), computed via convex optimization, as a systematic measure of DEX efficiency. Simulations comparing two routing algorithms-Yu Zhang et al.'s line-graph-based method and the depth-first search (DFS) algorithm-show that employing more profitable routing improves DEX efficiency and trader returns over time.","summary":"The efficiency of decentralized exchanges (DEXs) and the influence of token routing algorithms on market performance and stakeholder outcomes remain underexplored. This paper introduces the concept of Standardized Total Arbitrage Profit (STAP), computed via convex optimization, as a systematic…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"We assume that \\(\\Delta a_1^{out} = \\Delta b_1^{in} = 0\\), i.e., the trade sells token A to buy token B.","type":"model","stated":"explicit","quote":"We assume that ∆a1 out = ∆b1 in = 0, then ∆a1 in ≥0 and ∆b1 out ≥0, which means we sell token A to buy token B. Then, the above equation 4 will be simplified as the following equation 5: ∂F ∂∆a1 in = −PA + L1γ(b1 −∆b1 out) = 0 ∂F ∂∆b1 out = PB −L1(a1 + γ∆a1 in) = 0 · · · (5) Then we have PA PB = γ(b1−∆b1∗ out) a1+γ∆a1∗ in .","support":"full"},{"assumption":"We assume token prices from centralized exchanges (CEX) remain unchanged during the simulated token trading sequences.","type":"market","stated":"explicit","quote":"When we research token routing algorithms’ effect on the DEX efficiency and stakeholders’ benefits, we simulate a sequence of random token trading and keep the tokens’ CEX prices unchanged.","support":"full"},{"assumption":"Only liquidity pools with total value locked (TVL) larger than ten thousand dollars are selected for the token graph.","type":"data","stated":"explicit","quote":"During the construction of the token graph with Uniswap V2 pool reserve data, we first select pools whose total value locked (TVL) is larger than ten thousand dollars.","support":"full"},{"assumption":"The token graph is limited to 11 tokens and 18 liquidity pools because of the computational limitation of the cvxpy package.","type":"data","stated":"explicit","quote":"After token filtering, there are 11 tokens and 18 liquidity pools in total in the token graph, which is shown in Fig.3 1. cvxpy is a Python package for solving convex optimization problems WETH WBTC SYLO USDC USDT DAI DAO BLZ RACA MKR AGIX Figure 3: Token graph.","support":"full"},{"assumption":"The convex optimization problem used to compute maximal arbitrage profit is solved to optimality, implying the model accurately captures the AMM pricing and fee structure.","type":"model","stated":"implied","quote":"ponds to the liquidity pool containing a1 units of token A and b1 units of token B. Then we calculate the derivative of F to all variables, including ∆ai out, ∆bi in, etc.. ∂F ∂∆a1 in = −PA + L1γ(b1 + γ∆b1 in −∆b1 out) = 0 ∂F ∂∆b1 out = PB −L1(a1 + γ∆a1 in −∆a1 out) = 0 · · · (4) Because this is a convex optimization problem, the optimal solution must be on the boundary of the constraints. Namely (a1 + γ∆a1 in −∆a1","support":"full"},{"assumption":"Liquidity providers receive the retained transaction fee (\\(\\lambda\\Delta a\\) or \\(\\lambda\\Delta b\\)) as a reward after each trade.","type":"execution","stated":"explicit","quote":"After calculating the number of target tokens based on the CPMM rule as shown above in equation 1 (that are ∆b in the first equation and ∆a in the second equation), the retained transaction fee (that are λ∆a in the first equation and λ∆b in the second equation) will go to the liquidity pool again as a reward to liquidity providers. In the first phase, the transaction fee is retained, the number of tokens to be","support":"full"}]},"datasets":{"status":"available","items":[{"name":"Uniswap V2 pool reserve data","provider":"Uniswap V2","kind":"crypto-onchain","access":"public","period":"June 21, 2024 to November 8, 2024","universe":"11 tokens and 18 liquidity pools","quote":"Data Description The pool reserves’ data is from Uniswap V2 on the Ethereum blockchain, ranging from June 21, 2024, to Figure 2: The trading interface of Sushiswap. After token filtering, there are 11 tokens and 18 liquidity pools in total in the token graph, which is shown in Fig.3 1. cvxpy is a Python package for solving convex optimization problems WETH WBTC SYLO USDC USDT DAI DAO BLZ RACA MKR AGIX Figure 3:","support":"full"},{"name":"CoinGecko token price time series","provider":"CoinGecko","kind":"market-prices","access":"public","period":"June 21, 2024 to June 20, 2025","universe":"tokens present in the Uniswap V2 token graph","quote":"Cryptocurrencies’ daily price time se-ries data is from CoinGecko.","support":"partial","known_id":"coingecko"}],"known":[{"id":"coingecko","name":"CoinGecko","provider":"CoinGecko","kind":"market-prices","access":"public","quote":"Secondly, we select pools where both tokens’ price time series data can be downloaded from CoinGecko.","mentions":5,"used_in":2}],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"unclear"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"A retained transaction fee (λ·Δa or λ·Δb) is taken from each trade and added back to the liquidity pool as a reward to liquidity providers.","quote":"After calculating the number of target tokens based on the CPMM rule as shown above in equation 1 (that are ∆b in the first equation and ∆a in the second equation), the retained transaction fee (that are λ∆a in the first equation and λ∆b in the second equation) will go to the liquidity pool again as a reward to liquidity providers. In the first phase, the transaction fee is retained, the number of tokens to be","source":"evidence_card","rule":"R2","candidates":{"front_matter":"unclear","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"A retained transaction fee (λ·Δa or λ·Δb) is taken from each trade and added back to the liquidity pool as a reward to liquidity providers.","card_quote":"After calculating the number of target tokens based on the CPMM rule as shown above in equation 1 (that are ∆b in the first equation and ∆a in the second equation), the retained transaction fee (that are λ∆a in the first equation and λ∆b in the second equation) will go to the liquidity pool again as a reward to liquidity providers. In the first phase, the transaction fee is retained, the number of tokens to be","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"unclear","findings":[{"finding":"DEX efficiency, measured by STAP, declines from June 21, 2024 to November 8, 2024.","kind":"theoretical","quote":"Using a token graph comprising 11 tokens and 18 liquidity pools based on Uniswap V2 data, we observe a decline in DEX efficiency between June 21 and November 8, 2024.","support":"full"},{"finding":"Using the line‑graph‑based routing algorithm improves DEX efficiency more than the depth‑first search (DFS) algorithm.","kind":"performance","quote":"Simulations comparing two routing algorithms-Yu Zhang et al.'s line-graph-based method and the depth-first search (DFS) algorithm-show that employing more profitable routing improves DEX efficiency and trader returns over time. We find that the line-graph-based method can help to improve the efficiency of the DEX more than the DFS algorithm.","support":"full"},{"finding":"Total value locked (TVL) remains almost constant when the line‑graph method is used, but increases almost linearly under the DFS algorithm.","kind":"performance","quote":"Moreover, while total value locked (TVL) remains stable with the line-graph method, it increases under the DFS algorithm, indicating greater aggregate benefits for liquidity providers. The TVL increases almost linearly with more token trading if the DFS algorithm is used for token routing, which means that the liquidity pool providers will benefit as a whole sector from a poor routing algorithm.","support":"full"},{"finding":"When the line‑graph routing is applied, the cumulative monetized value of target tokens received by traders is almost equal to the value of source tokens invested; under DFS, the cumulative received value is much lower.","kind":"performance","quote":"If we take all traders as a whole sector, we find that the accumulative monetized value of the target tokens that traders receive is almost equal to the monetized value of the source tokens that traders input if the line-graph-based routing method is applied. However, this accumulative monetized value of traders received by using the DFS algorithm is much lower comparatively.","support":"full"},{"finding":"Executing the trade order that maximizes STAP eliminates all cyclic arbitrage opportunities within DEXs and arbitrage between DEXs and CEXs.","kind":"theoretical","quote":"We prove that executing the trade order maximizing STAP and reintegrating the resulting transaction fees eliminates all arbitrage opportunities-both cyclic arbitrage within DEXs and between DEXs and centralized exchanges (CEXs).","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":25,"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":"2026-11-14T19:48:55Z"},"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/"}