Optimistic MEV in Ethereum Layer 2s: Why Blockspace Is Always in Demand

Lioba Heimbach; Ozan Solmaz; Roger Wattenhofer; Yann Vonlanthen

REVIEW 3 major objections 6 minor 15 references

Optimistic MEV in Ethereum Layer 2s: Why Blockspace Is Always in Demand

T0 review · 3 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Cyclic arbitrage on Ethereum Layer 2s has become speculative on-chain probing that dominates blockspace.

desk verdict First large-scale measurement of optimistic MEV on L2s is valuable and mostly convincing, but its central gas-share claims depend on an unvalidated classification of no-trade DEX interactions as on-chain probing. read the letter →

arxiv 2506.14768 v2 pith:DJEHUSAY submitted 2025-06-17 cs.CE

Ozan Solmaz , Lioba Heimbach , Yann Vonlanthen , Roger Wattenhofer This is my paper

classification cs.CE

keywords blockchainMEVoptimisticcyclicarbitrageLayer2EthereumDEXrollups

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper defines and quantifies a class of blockchain value extraction it calls optimistic MEV, in which a bot decides inside the transaction itself whether an arbitrage opportunity exists, rather than checking off-chain before submitting. Its central finding is that on the rollups Base and Optimism, transactions from cyclic-arbitrage bots accounted for about 51% and 55% of all on-chain gas in Q1 2025, and that most of those transactions were interaction probes that touch DEX pools but never execute a swap. The paper reads this as evidence that cheap Layer 2 blockspace is being consumed by speculative machine activity, which is why blocks on those networks stay persistently full even when organic user demand is low. It further shows that the same behavior is far weaker on Arbitrum, and argues that design choices such as fee floors, transaction ordering, and block interval explain the difference. If true, the result changes how one thinks about Layer 2 capacity, fee markets, and MEV mitigation.

What carries the argument

The central object is optimistic MEV, defined as MEV whose opportunity existence and execution parameters are resolved by on-chain computation at execution time rather than verified off-chain before submission; by construction such strategies produce many speculative transactions that terminate without trading. The argument is carried by a three-stage classification pipeline: a cyclic-arbitrage detector that reconstructs token paths from swap events, drops router and aggregator calls, and keeps only cyclic, non-negative-profit transactions to seed a validated set of bot contracts; a per-transaction classifier that marks the first callee as cyclic-arb or other, the DEX activity as trade, interaction, or residual, and the outcome as success or revert; and an inference layer that uses calldata-reuse ratios, median calldata lengths, bytecode similarity, and regression to show that parameters are not precomputed off-chain.

What would settle it

Pick a random sample of cyclic-arbitrage interaction transactions on Base and Optimism, replay their traces, and count how many use verbose calldata that encodes a specific path or amount or appear in bursts immediately after a single user trade; if a substantial share of so-called probes look like off-chain-planned back-run attempts that simply failed, the optimistic-MEV interpretation of the numbers would be falsified.

Watch

Extended reading notes

Core claim

On its own terms, the paper claims that cyclic arbitrage on Layer 2 rollups has flipped from a precomputed, off-chain-verified strategy into an optimistic one: bots submit high-frequency transactions that begin with read-only calls into DEX pools, decide inside the Ethereum Virtual Machine whether a profitable cycle exists, and usually abort without any swap. The supporting measurements are that on Base and Optimism only 6.3% and 12% of cyclic-arbitrage DEX-touching transactions produce a trade, that the vast majority of these transactions terminate without a swap and without reverting, that top bots reuse minimal calldata tens to millions of times so paths and amounts cannot be precomputed off-chain, and that this activity is what fills post-Dencun blocks after Base and Optimism raised their gas targets. The paper contrasts this with Ethereum Layer 1, where 95.8% of cyclic-arbitrage DEX-touching transactions execute trades, and with Arbitrum, where a 0.01 GWei fee floor, first-come-first-served ordering, and roughly 250ms blocks coincide with a 7% gas share and a 52.6% trade share.

Load-bearing premise

The headline gas-share figures assume that every no-trade transaction sent to a validated cyclic-arbitrage contract is an on-chain probe, even though that label is assigned by contract address and DEX-pool touch rather than verified per transaction.

Editorial extensions

If this is right

A network that cuts costs and raises its gas target, as Base and Optimism did after the Dencun upgrade, hands most of the new capacity to speculative probes rather than to user traffic.
Fee floors or minimum transaction costs become the most direct policy lever: Arbitrum's 0.01 GWei floor and short blocks coincide with a 7% gas share, versus over half on the OP-Stack chains.
MEV auctions and revert protection do not transplant cleanly to Layer 2s, because high-frequency probing could flood an auction with noise and short block intervals leave little scheduling room.
Capacity planning and gas-target setting for rollups must separate this speculative probe baseline from genuine user demand, or risk inflating the apparent organic load.
Searcher economics change: bots compete by the cost of repeated probes rather than by priority-fee bidding, which is why these transactions pay less than a quarter of fees despite consuming most of the gas.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A testable extension would apply the same classification to other low-fee rollups such as zkSync, Scroll, or Unichain; the paper's mechanism predicts the highest probe shares on chains that combine very low fees with multi-second block intervals.
The calldata-reuse statistics hint at a concentration the aggregate figures understate: 50 identical bytecode clones on Base suggest that a handful of operators, not a competitive crowd, may be generating a large fraction of the activity, which would also change how the regression drivers should be read.
A way to stress-test the optimistic mechanism is to raise fees on a test network while holding block time fixed; if probe volume stays high, the cheap-fee explanation would be incomplete and the probes would be carrying some other value, such as information about pending pool states.
The results imply that any rollup introducing an express lane or priority ordering should expect the speculative probe pattern to adapt to that ordering rather than disappear, since deterministic inclusion lowers the risk that a probe arrives after the opportunity has faded.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague, here's my read on arXiv:2506.14768.

The paper does two genuinely new things: it formally defines optimistic MEV and it measures it at scale on the three big L2s. The headline finding—that cyclic arbitrage contracts account for over half of gas on Base and Optimism in Q1 2025, mostly via no-swap "interaction" transactions—is striking and, if right, important for fee market and sequencer design. The authors also do real work on the pipeline: the cyclic arb detector, the contract validation in Appendix E, the bytecode similarity analysis, and the cross-network contrasts all show care. The open repo and Dune queries are a plus, and the regression section is honestly labeled exploratory.

The soft spot is the load-bearing classification assumption. Section 4.2 labels any successful transaction from a known cyclic arb contract that touches a DEX pool without emitting a swap as "interaction," and the paper then treats that as evidence of on-chain opportunity probing. That is a reasonable inference, but it is not per-transaction verification. The aggregate calldata reuse and short calldata evidence is consistent with probing, but it is also consistent with bots resubmitting fixed calldata and letting the contract terminate when a precomputed path is no longer profitable, or with parameters loaded from storage. The stress test is right that this is not a minor boundary case: on Base and Optimism, the interaction-success category dominates the gas share, so the central claim rests almost entirely on this label. The paper gives no precision/recall for the classification and no trace-level validation of a random sample.

I would not call this fatal. The definition is useful, the measurement is the first of its kind, and the cross-network differences (Arbitrum's lower share, the revert patterns, the fee dynamics) hang together and support the qualitative story. But the specific >50% optimistic MEV numbers should be read as upper-bound-ish until the authors quantify the classifier's error or validate a sample manually. I'd recommend they release per-transaction labels and do exactly that.

This paper deserves a serious referee. It is a strong measurement contribution with a central interpretation that needs tightening, not a desk reject. I'd send it out and ask the referees to focus on the classification validity.

For me: I'd bring it to reading group, and I'd probably cite it for the definition and the measurement, though with a caveat on the headline numbers.

Referee Report

3 major / 6 minor

Summary. This paper defines "optimistic MEV" as MEV in which opportunity discovery and execution parameters are determined on-chain at execution time, and measures its prevalence among cyclic arbitrage on Arbitrum, Base, Optimism, and Ethereum L1 using Dune data from August 2023 to May 2025. A multi-stage pipeline detects profitable cyclic swaps, validates the involved contracts (Cbot) through manual and heuristic audits (Appendix E), and labels each transaction by purpose, DEX involvement, and execution outcome. The headline results are that in Q1 2025 cyclic-arbitrage contracts account for 51% of gas on Base, 55% on Optimism, and 7% on Arbitrum, with the Base and Optimism shares dominated by successful no-trade "interaction" transactions, which the paper interprets as on-chain opportunity probes. The paper further compares top bot contracts across chains, reports fee and revert dynamics, and uses OLS regressions to associate cyclic-arbitrage activity with ETH volatility, retail trading, and DEX aggregator usage.

Significance. If the interpretation holds, this is the first systematic measurement of a distinct L2 MEV strategy and a valuable contribution to the MEV and rollup-design literature. The paper ships an open implementation and classification artifact, validates contracts through a multi-stage audit, and produces internally consistent cross-network contrasts—trade vs. interaction shares, revert behavior, calldata reuse, and fee levels—that are informative for sequencer and fee-market design. The finding that a small set of arbitrage contracts can consume the majority of blockspace at very low fees is policy-relevant. The main caveat is that the optimistic label is an interpretation of an aggregate proxy rather than a per-transaction verified property; the significance of the paper therefore depends on the trace-level evidence requested below.

major comments (3)

[§4.2, §5.2, Appendix E] The headline gas-share figures—51% on Base and 55% on Optimism in Q1 2025, with 48% and 52% respectively attributed to cyclicArb-interaction-success—depend on classifying every successful, no-swap transaction that invokes a validated cyclic-arbitrage contract and touches a DEX pool as an on-chain opportunity probe. The assignment rule uses only the first callee and pool contact; it does not verify that the execution path actually performed an opportunity search, such as reading multiple distinct pool states before terminating. The validation in Appendix E is contract-level (majority of no-trade transactions touch DEX contracts, caller diversity, swap percentiles) and is compatible with a transaction being a failed off-chain-planned attempt, a fixed-calldata resubmission, or a maintenance call that happens to read a pool. Because the interaction-success category is the dominant component of the gas share, the central claim that cyclic arbitrage on L2s is "predominantly executed as optimistic MEV" rests entirely on this assumption. Please add per-transaction trace evidence, for example the distribution of the number and diversity of pool reads in interaction-success transactions or path variation under identical calldata, or weaken the claim to what the classification supports.
[§5.5, Table 1] The calldata-reuse and short-calldata statistics are presented as the main evidence that "key parameters, i.e., the path of the arbitrage transaction and the swap amounts, are determined during execution rather than precomputed off-chain" (Section 3). However, high transactions-per-unique-calldata ratios (e.g., 7,979,548 on Optimism for 0xabf4...017c) and median calldata lengths of 2–6 hex characters are also consistent with a bot resubmitting a fixed trigger while the contract reads parameters from storage updated by earlier transactions, or with repeated attempts at a precomputed path that is no longer profitable. The defining property in Definition 1 is that path and amount decisions are made within the transaction itself; that property should be demonstrated on traces, for instance by showing that identical calldata yields different multi-pool paths across executions, or by reporting what fraction of interaction-success transactions issue multiple distinct DEX staticcalls before terminating.
[Appendix E] The validation procedure reports a gas-coverage stopping rule (more than 80% of gas from candidate-contract transactions originates from validated contracts) but does not report the number of candidates inspected and removed at each step, nor any precision/recall estimate against a labeled ground truth. Since Cbot is the denominator for all gas-share percentages and for the regression dependent variables, the absence of quantified accuracy leaves the central numerical claims without an error bound. Please report candidate counts and removal counts per validation step and, where possible, compare the final Cbot on each chain with an independent source such as EigenPhi [33] or Entropy Advisors [34,35] rather than using those sources only as a consistency check for revert behavior.

minor comments (6)

[§4.1] In item 3 of the cyclic arbitrage detection description, "the the token bought" should read "the token bought".
[§1, contribution iii] The phrase "out analysis reveals" should read "our analysis reveals".
[Appendix D] The sentence "This classification stage aimes to identify" contains a typo; "aimes" should be "aims".
[Appendix E, Swaps per Transaction] The validation text says "a cyclic arbitrage requires at least two transactions," but the surrounding discussion is about swaps per transaction; this should read "at least two swaps."
[§5.5, Table 1] The denominator of the "cum. MEV bot gas (%)" column is described as "all identified MEV bots within the dataset C′bots,chain," but C′bots,chain is not defined in the main text; please clarify its relation to Cbot.
[§6, Table 2] For Base and Optimism, Model 1 has adjusted R² values of 0.0736 and 0.0240, indicating very limited explanatory power for the overall cyclicArb transaction count; the text should avoid presenting these models as establishing drivers of overall cyclicArb activity and should emphasize the trade-count models instead.

Circularity Check

2 steps flagged · score 4.0 of 10

Interaction category is defined by the same no-swap/DEX-touch observable that is then cited as evidence of optimistic probing; calldata and trace evidence provide partial independent support.

self definitional [Section 5.2, Figure 3 caption and following paragraph]
"interaction (on-chain probing or contract calls without token transfers) ... The share of gas attributed to cyclicArb - interaction transactions rises sharply over time, indicating a surge in speculative, non-token-transfer, DEX contract probing activity associated with optimistic MEV."

Section 4.2 defines 'interaction' purely by the observable that no swap is emitted while the trace touches a recognized DEX pool. The Figure 3 legend renames the same category 'on-chain probing,' and the accompanying text treats the rising interaction gas share as evidence of 'speculative ... probing activity.' Thus the conclusion that these transactions are probes is entailed by the label, not established by independent per-transaction evidence. Since cyclicArb-interaction-success drives the >50% gas-share headline on Base and Optimism, the headline inherits this definitional equation.
self definitional [Section 5.1, paragraph after Figure 2]
"This pattern of profitability being resolved only at execution and the prevalence of transactions that terminate without trading is exactly what defines optimistic MEV."

Definition 1 defines optimistic MEV as strategies whose 'profitability is only resolved during execution, resulting in a significant share of transactions that terminate without performing a trade.' The paper observes the no-trade outcome and then asserts the unobserved mechanism (profitability resolved at execution) as part of the observed pattern. Because the no-trade symptom is written into the definition, using it as confirmation that the activity is optimistic is confirmatory by construction; the independent content comes from ancillary calldata and STATICCALL evidence, not from the aggregate gas-share classification itself.

full rationale

The paper's central empirical claim—that cyclic arbitrage on Base and Optimism is predominantly optimistic MEV—rests on the cyclicArb-interaction-success category. Section 4.2 defines interaction solely by the absence of swaps plus trace contact with a DEX pool; Figure 3's legend renames this 'on-chain probing,' and Section 5.2 then reads the interaction gas share as evidence of 'speculative ... probing activity.' This is a definitional conflation: the category label already asserts what the analysis claims to show. Similarly, Section 5.1 uses the no-trade prevalence, a symptom written into Definition 1, as 'exactly what defines optimistic MEV,' making the observable and the construct coincide. However, the paper does provide independent evidence beyond the label: the case study's STATICCALL trace (Example 1), calldata reuse statistics (e.g., one Optimism bot reusing identical calldata over 7 million times), and validation against Entropy Advisors' dataset. Therefore the circularity is partial, not total: the >50% gas-share figure is a measurement, not a fit, and the optimistic interpretation is supported by ancillary evidence that could in principle falsify it. The main caveat is validity—the interaction label may include off-chain-planned stale attempts—rather than a closed definitional loop. Score 4 reflects that the headline conclusion partially reduces to the operational definition, while retaining independent content.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central measurement rests on a pipeline with several hand-chosen thresholds and on the interpretive assumption that no-trade DEX interactions from cyclic arbitrage contracts are speculative probes. No new physical or formal entities are introduced beyond the conceptual category 'optimistic MEV', which is defined and operationalized in the paper.

free parameters (5)

Minimum swaps for cyclic detection = 2
A transaction is considered cyclic arbitrage only if it forms a sequence of at least two swaps (Section 4.1 and Appendix D).
Profit filter thresholds = non-negative balance changes with at least one strictly positive gain
Defines profitable cyclic arbitrage; any negative token balance disqualifies the transaction (Section 4.1).
Validation gas coverage target = 80%
Manual validation continues until more than 80% of gas used in transactions involving candidate contracts originates from validated contracts (Appendix E).
DEX-interaction majority threshold = 60%
Candidate contracts must have at least 60% DEX-related interaction in non-trading transactions to be kept in the final set (Appendix E).
Caller diversity threshold = 3 distinct unrelated EOAs
Contracts with more than 3 distinct and unrelated EOAs interacting with human-like frequency are excluded as likely non-MEV (Appendix E).

assumptions (5)

domain assumption Dune Analytics tables (dex.trades, transactions, traces, raw_pools) provide complete and accurate records of all swaps, calls, traces, and gas usage on the studied chains.
All measurements rely on Dune data; the authors acknowledge some missing DEXs and freshness but assert they do not affect MEV results (Section C).
domain assumption The first callee of a transaction is the originator of the MEV activity, so labeling by the 'to' address is sufficient.
Transaction-level classification tags a transaction as cyclicArb when the first callee is in Cbot (Section 4.2).
ad hoc to paper Transactions from validated cyclic arbitrage contracts that touch a DEX pool but emit no swap event are on-chain opportunity probes (optimistic MEV), not other contract behavior.
This interpretive assumption underpins the central claim that interaction gas is optimistic MEV; supported only by aggregate calldata and STATICCALL examples (Sections 3, 5.1).
domain assumption The set of cyclic arbitrage contracts Cbot is stable over the study period, so contract identity can be applied to all transactions in the window.
Contracts are detected from transactions in the same window and then used to classify all transactions, implying no major behavioral or ownership changes (Section 4).
domain assumption The EIP-1559 fee model and Arbitrum's fee-floor dynamics are correctly represented in the cross-network fee analysis.
Used to explain why Base and Optimism blocks remain full while Arbitrum has unused headroom (Section 5.4).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Optimistic MEV in Ethereum Layer 2s: Why Blockspace Is Always in Demand." pith.science (2026). https://pith.science/paper/DJEHUSAY

@misc{pith2026250614768,
  author       = {Pith},
  title        = {Pith review of: Optimistic MEV in Ethereum Layer 2s: Why Blockspace Is Always in Demand},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DJEHUSAY}},
  note         = {Machine review of arXiv:2506.14768}
}

abstract

Layer 2 rollups are rapidly absorbing DeFi activity, securing over $40 billion and accounting for nearly half of Ethereum's DEX volume by Q1 2025, yet their MEV dynamics remain understudied. We address this gap by defining and quantifying optimistic MEV, a form of speculative, on-chain MEV whose detection and execution logic reside largely on-chain in smart contracts. As a result of their speculative nature and lack of off-chain opportunity verification, optimistic MEV transactions frequently decide not to execute any trades. In this work, we focus on cyclic arbitrage, which we find is predominantly executed as optimistic MEV on Layer 2s. Using our multi-stage identification pipeline on Arbitrum, Base, and Optimism, we show that in Q1 2025, transactions from cyclic arbitrage contracts account for over 50% of on-chain gas on Base and Optimism and 7% on Arbitrum, driven mainly by "interaction" probes (on-chain computations searching for arbitrage). This speculative probing indicates that cyclic arbitrage on Layer 2s is predominantly executed as optimistic MEV and contributes to generally keeping blocks on Base and Optimism persistently full. Despite consuming over half of on-chain gas, these optimistic MEV transactions pay less than one quarter of total gas fees. Cross-network comparison reveals divergent success rates, differing patterns of code reuse, and sensitivity to varying sequencer ordering and block production times. Finally, OLS regressions link optimistic MEV trade count to ETH volatility, retail trading activity, and DEX aggregator usage. Together, these findings show that optimistic MEV has become a major source of persistent spam-like transaction activity on Layer 2s, dominating blockspace with low-value probes and reshaping the composition of on-chain activity.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

15 extracted references · 14 canonical work pages

[1]

Crucially, we also collect detailed transaction traces, which include internal transactions, emitted logs (events), and relevant state changes

Transactional Data: We retrieve high-level transaction details including, but not limited to, sender and recipient addresses, gas utilized, transaction value, input data, block numbers, and timestamps. Crucially, we also collect detailed transaction traces, which include internal transactions, emitted logs (events), and relevant state changes. This granul...

work page
[2]

This includes detailed swap event data and liquidity pool contract addresses, leveraging curated datasets such as those described in [32]

Decentralized Exchange Activity Data:To analyze MEV transactions, we collect extensive data on trades and liquidity pools from DEX protocols operating on the target Layer 2s. This includes detailed swap event data and liquidity pool contract addresses, leveraging curated datasets such as those described in [32]

work page
[3]

69 Samczsun

Accessed: 2025-05-25. 69 Samczsun. Escaping the dark forest.https://samczsun.com/escaping-the-dark-forest/,

work page 2025
[4]

We run our own archive nodes for Base, Arbitrum, and Optimism networks to fetch the bytecodes using JSON-RPC method eth_getCode

Contract Bytecodes: For the contracts that we marked on Layer 2s, we fetch their bytecodes to be utilized in future steps to measure contract similarity (see Appendix A). We run our own archive nodes for Base, Arbitrum, and Optimism networks to fetch the bytecodes using JSON-RPC method eth_getCode

work page
[5]

Price Data: We additionally obtain daily Open-High-Low-Close (OHLC) data for the ETH price on Ethereum Layer 1 using Dune Analytics [27], which we use for regression analysis and volatility calculation. This multi-faceted data collection strategy was designed to provide the necessary inputs for our MEV and transaction classification pipeline, enabling a r...

work page
[8]

Contract Address Identification:A curated list of smart contract addresses relevant to DEX activities, such as prominent DEX routers and aggregator contracts on the studied Layer 2s, is compiled. This is achieved through a combination of leveraging existing tagged address lists within Dune Analytics [27], cross-referencing them with on-chain explorers [1,...

work page
[11]

This is an ordered sequence of token addresses reflecting the actual path of token conversions in the transaction

Π =⟨tokensold,1,tokenbought,1,..., tokensold,k,tokenbought,k⟩. This is an ordered sequence of token addresses reflecting the actual path of token conversions in the transaction

work page
[12]

This is a map representing the net balance changes after the transaction from the initiator’s perspective across all tokens involved in the trade tr

∆B ={(κ,δκ)|κ∈ K,δκ∈ R}. This is a map representing the net balance changes after the transaction from the initiator’s perspective across all tokens involved in the trade tr. Specifically, for each tokenκ, δκ = ∑k j=1(amountbought,j· I(tokenbought,j = κ)−amountsold,j· I(tokensold,j =κ)), where I(·) is the indicator function. LetT′ be the set of transactio...

work page

Show all 15 references

[13]

This filter removes transactions directly interacting with these intermediary contracts

Filter α1 (Router/Aggregator Exclusion): LetRcontracts andAcontracts be the sets of known router and aggregator contract addresses, respectively. This filter removes transactions directly interacting with these intermediary contracts. T (1) =α1(T′) ={(hash,to,from, (Π, ∆B))∈T ...
[14]

The predicateisCyclic(Π) holds true if: k≥ 1 (i.e., there is at least one swap)

Filter α2 (Cyclic Swap Detection):This filter identifies transactions whose sequence of token swapsΠ =⟨π1,π 2,...,π 2k⟩ forms a cycle. The predicateisCyclic(Π) holds true if: k≥ 1 (i.e., there is at least one swap). π1 =π2k (the first token sold is the same as the last token b...
[15]

It evaluates the balance changes map ∆B

Filter α3 (Profitability Assessment):This filter retains transactions that result in a net profit for a token without incurring losses in any other token. It evaluates the balance changes map ∆B. The predicate isProfitable(∆B) holds true if: There exists at least one tokenκ su...
[2020]

70 Ozan Solmaz

Accessed: 2025-05-25. 70 Ozan Solmaz. Dune analytics query: 5534551.https://dune.com/queries/5534551, 2024. Accessed on 2024-07-25. 71 Sorella Labs. Brontes: A blockchain analysis tool. https://github.com/SorellaLabs/ brontes, 2025. Accessed: 2025-05-18. 72 ThogardPvP. Tweet p...

2025
[2021]

60 Optimism

URL: https://arxiv.org/abs/2112.01472, arXiv:2112.01472. 60 Optimism. Optimism. https://www.optimism.io. Accessed: 2025-05-18. 61 Ethereum Optimism. Add initial devnet op-geth changes. https://github.com/ ethereum-optimism/op-geth/pull/118, 2023. Accessed: 2025-05-12. 62 Burak...

2025 arXiv
[2022]

liquidity-sniping

GitHub repository, accessed 2025-05-09. 78 Uniswap Labs. Universal router – overview. URL:https://docs.uniswap.org/contracts/ universal-router/overview. 79 Virtuals.io. Virtuals app, 2025. Accessed: 2025-05-29. URL:https://app.virtuals.io. 80 Ye Wang, Yan Chen, Haotian Wu, Liy...

2025 arXiv
[2023]

@uniswap what you doing here? poor man just wants to plow 0.36 eth into some shitcoin

Accessed: 2025-05-25. 23 Tarun Chordia, Avanidhar Subrahmanyam, and Richard W. Roll. Market liquidity and trading activity. SSRN Electronic Journal , July 2000. URL: https://ssrn.com/abstract=237674, doi:10.2139/ssrn.237674. 24 Clanker. Clanker world, 2025. Accessed: 2025-05-2...

2025

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Crucially, we also collect detailed transaction traces, which include internal transactions, emitted logs (events), and relevant state changes

Transactional Data: We retrieve high-level transaction details including, but not limited to, sender and recipient addresses, gas utilized, transaction value, input data, block numbers, and timestamps. Crucially, we also collect detailed transaction traces, which include internal transactions, emitted logs (events), and relevant state changes. This granul...

work page

[2] [2]

This includes detailed swap event data and liquidity pool contract addresses, leveraging curated datasets such as those described in [32]

Decentralized Exchange Activity Data:To analyze MEV transactions, we collect extensive data on trades and liquidity pools from DEX protocols operating on the target Layer 2s. This includes detailed swap event data and liquidity pool contract addresses, leveraging curated datasets such as those described in [32]

work page

[3] [3]

69 Samczsun

Accessed: 2025-05-25. 69 Samczsun. Escaping the dark forest.https://samczsun.com/escaping-the-dark-forest/,

work page 2025

[4] [4]

We run our own archive nodes for Base, Arbitrum, and Optimism networks to fetch the bytecodes using JSON-RPC method eth_getCode

Contract Bytecodes: For the contracts that we marked on Layer 2s, we fetch their bytecodes to be utilized in future steps to measure contract similarity (see Appendix A). We run our own archive nodes for Base, Arbitrum, and Optimism networks to fetch the bytecodes using JSON-RPC method eth_getCode

work page

[5] [5]

Price Data: We additionally obtain daily Open-High-Low-Close (OHLC) data for the ETH price on Ethereum Layer 1 using Dune Analytics [27], which we use for regression analysis and volatility calculation. This multi-faceted data collection strategy was designed to provide the necessary inputs for our MEV and transaction classification pipeline, enabling a r...

work page

[6] [8]

Contract Address Identification:A curated list of smart contract addresses relevant to DEX activities, such as prominent DEX routers and aggregator contracts on the studied Layer 2s, is compiled. This is achieved through a combination of leveraging existing tagged address lists within Dune Analytics [27], cross-referencing them with on-chain explorers [1,...

work page

[7] [11]

This is an ordered sequence of token addresses reflecting the actual path of token conversions in the transaction

Π =⟨tokensold,1,tokenbought,1,..., tokensold,k,tokenbought,k⟩. This is an ordered sequence of token addresses reflecting the actual path of token conversions in the transaction

work page

[8] [12]

This is a map representing the net balance changes after the transaction from the initiator’s perspective across all tokens involved in the trade tr

∆B ={(κ,δκ)|κ∈ K,δκ∈ R}. This is a map representing the net balance changes after the transaction from the initiator’s perspective across all tokens involved in the trade tr. Specifically, for each tokenκ, δκ = ∑k j=1(amountbought,j· I(tokenbought,j = κ)−amountsold,j· I(tokensold,j =κ)), where I(·) is the indicator function. LetT′ be the set of transactio...

work page