ANTHILLAsk your own question
Answer · confidence: partial

What do studies say about sandwich attacks and front-running, and who wins?

Research question: What does academic research say about MEV (sandwich attacks, front-running) and who captures that value?

Research finds sandwich attacks are rare and unprofitable on rollups with private mempools, unlike on Ethereum L1 where they are endemic and profitable 12. In CEX-DEX arbitrage, value concentrates in a few firms: Wintermute, SCP and Kayle together captured about 73% of cumulative extracted value, and around 90% by Q1 2025 13. Another study found that eleven searchers account for more than 80% of identified arbitrage volume 4. Builders and validators also take a share, since searchers tied to a single builder often share most of their value with it 6, and validators capture more MEV when searcher competition is high 8. The most recent dated fact is from March 2026 21, about seven months before October 2026, and the evidence covers specific chains and periods rather than MEV overall 39.

Checked 2026-10-18. Only documented and supported facts are used.
Dig deeper on this question

Sources

[1] documented
The median net return for sandwich attacks on Ethereum rollups with private mempools is negative.
"We find that the majority of flagged patterns are false positives and that the median net return for these attacks is negative."
How to Serve Your Sandwich? MEV Attacks in Private L2 Mempools
[2] documented
Sandwich attacks are rare and unprofitable in rollups with private mempools, unlike on Ethereum L1 where they are endemic and profitable.
"Our results suggest that sandwiching, while endemic and profitable on Ethereum L1, is rare, unprofitable, and largely absent in rollups with private mempools."
How to Serve Your Sandwich? MEV Attacks in Private L2 Mempools
[3] supported
Nineteen major searchers extracted about 233.8M USD from roughly 7.2 million CEX-DEX arbitrages, and the three leading searchers captured about three-quarters of the volume and extracted value.
"We estimate that 19 major searchers extracted a total value of 233.8M USD from 7,203,560 CEX-DEX arbitrages during the observed period and uncover an increasing centralization trend, with the three leading searchers capturing approximately three-quarters of the total arbitrage volume and extracted value."
Measuring CEX-DEX Extracted Value and Searcher Profitability: The Darkest of the MEV Dark Forest
[4] supported
Heimbach et al. found that eleven searchers account for more than 80% of identified arbitrage volume, with the two largest searchers alone accounting for 49.2%.
"Heimbach et al. [ 9 ] find that eleven searchers account for more than 80% of identified arbitrage volume, with the two largest searchers alone accounting for 49.2%."
Towards Decentralized Searcher Competition in MEV Markets
[5] supported
Long-tail sandwiching is profitable in aggregate: 1,162 sandwiches over 2,550 blocks were net-positive ex post, totaling about 35.08 BNB (roughly $20,200).
"Long-tail sandwiching is not: 1,162 ex-post net-positive sandwiches over 2,550 blocks total 35.08 BNB (∼$20,200)—a ∼91× gap in count."
Who Captures Atomic MEV after PBS? Receipt-exact Evidence of Builder Centralization and a Vanishing Independent-searcher Edge on BNB Smart Chain (2026)
[6] supported
Searchers that spread their flow across several builders keep higher profit margins, while searchers exclusive to one builder keep lower margins and often share most of their value with that builder, sometimes running at a net loss.
"Neutral searchers, distributing flow among multiple builders, retain higher profit margins. Conversely, exclusive searchers maintain lower margins, sharing most of their extracted value with affiliated builders, sometimes even operating at negative net profit."
Measuring CEX-DEX Extracted Value and Searcher Profitability: The Darkest of the MEV Dark Forest
[7] supported
In the paper's threat model, attackers insert transactions at specific positions to front-run or back-run a victim, and build a bundle from them.
"to front-run or back-run a victim [12] to build a bundle."
MEVisor: High-Throughput MEV Discovery in DEXs with GPU Parallelism Weimin Chen
[8] supported
Mamageishvili et al. showed that validators capture more MEV value when searcher competition is high, while searchers keep more value when competition is weak.
"Mamageishvili et al. [ 11 ] showed that validator revenue increases with searcher competition: when many searchers identify similar opportunities, competition pushes most of the value to the validator, while searchers retain more value when competition is weak."
Towards Decentralized Searcher Competition in MEV Markets
[9] supported
Realized sandwich capture by independent searchers is tiny: 45 of 96,064 opportunities (0.0468% by count, 0.0078% by value) and none captured by builders, so the inferred independent searcher edge is vanishingly small.
"by an in-block counterfactual, is tiny: 45 of 96,064 (0.0468% by count, 0.0078% by value; 0 by builder)"
Who Captures Atomic MEV after PBS? Receipt-exact Evidence of Builder Centralization and a Vanishing Independent-searcher Edge on BNB Smart Chain (2026)
[10] supported
Wintermute and SCP together consistently took over 50% of total CEX-DEX revenue throughout the observation period.
"The largest two searchers, Wintermute and SCP consistently extracted over 50% of total revenue throughout the entire observation period."
Measuring CEX-DEX Extracted Value and Searcher Profitability: The Darkest of the MEV Dark Forest
[11] supported
Sandwich attacks typically involve only two trades, but choosing the most profitable pair is computationally expensive, with quadratic complexity in the number of pools.
"For sandwich attacks, although the attacker typically searches only two trades, selecting the optimal pair incurs quadratic complexity."
MEVisor: High-Throughput MEV Discovery in DEXs with GPU Parallelism Weimin Chen
[12] supported
Empirical studies found that the searcher market is highly skewed toward large or specialized actors.
"Empirical studies further confirmed that the searcher market is highly skewed toward large or specialized actors."
Towards Decentralized Searcher Competition in MEV Markets
[13] supported
Wintermute, SCP, and Kayle together captured about 170.8M USD, roughly 73% of total cumulative extracted value, and by Q1 2025 they accounted for around 90%.
"The market thus resembles a clear trident, with three searchers— Wintermute , SCP , and Kayle —collectively capturing 170.8M USD, representing approximately 73% of the total cumulative extracted value."
Measuring CEX-DEX Extracted Value and Searcher Profitability: The Darkest of the MEV Dark Forest
[14] supported
Prior research found that optimizing the starting capital for arbitrage on Uniswap V2 is a convex problem, so it can be solved reliably and efficiently.
"Prior work [15] shows that the capital optimization for Uniswap V2 DEXs [2] is a convex"
MEVisor: High-Throughput MEV Discovery in DEXs with GPU Parallelism Weimin Chen
[15] supported
The builder Titan earned the highest total profit despite having no vertical integration with any searcher, which the authors take as evidence it rebates part of its profit to high-value orderflow providers.
"Finally, Titan , despite not being vertically integrated with any searcher, earns the highest total profit. This is strong evidence that they have exclusive deals with high-value orderflow providers"
Measuring CEX-DEX Extracted Value and Searcher Profitability: The Darkest of the MEV Dark Forest
[16] supported
In a head-to-head comparison, the Lanturn tool found only 4 new sandwich opportunities, with a total profit of 0.048 ETH after 0.003 ETH gas cost.
"As for the sandwich attacks, Lanturn identifies 4 new opportunities, yielding a total profit of 0.048 ETH with 0.051 ETH revenue and 0.003 ETH gas cost."
MEVisor: High-Throughput MEV Discovery in DEXs with GPU Parallelism Weimin Chen
[17] supported
Transactions waiting in the Ethereum mempool are publicly visible, which makes them vulnerable to front-running.
"While waiting in the mempool, these ‘in-flight’ transactions are publicly visible and vulnerable to front-running attacks, such that malicious parties use information in the transaction for their own gain and at a direct cost to the transaction owner."
Simulation of Front-Running Attacks and Privacy Mitigations in Ethereum Blockchain (2022)
[18] supported
Front-running attackers capture value for themselves by using information from a pending transaction, at a direct cost to the transaction's owner.
"such that malicious parties use information in the transaction for their own gain and at a direct cost to the transaction owner."
Simulation of Front-Running Attacks and Privacy Mitigations in Ethereum Blockchain (2022)
[19] supported
The term Miner Extractable Value was used in earlier work; as more protocols moved to Proof-of-stake consensus, it broadened into the more general term Maximal Extractable Value.
"в роботі [6] вжито термін Miner Extractable Value"
Maximal extractable value (mev) in blockchain networks and its impact on blockchain ecosystem (2023)
[20] supported
The MEV-geth protocol is proposed as a way to mitigate these attacks by keeping in-flight transactions private.
"mitigate attacks using the MEV-geth protocol for in-flight transaction privacy."
Simulation of Front-Running Attacks and Privacy Mitigations in Ethereum Blockchain (2022)
[21] supported
The sandwich attack works by placing one transaction before and one after the victim's trade, so the bot profits from both sides of the victim's trade.
"The technique is called a “sandwich attack” because the bot places one transaction before and one after the victim’s trade. The user gets squeezed in the middle, paying inflated prices while the bot extracts value from both sides of the trade."
MEV Bot Sandwich Attack Drains $50M Swap To $36K