[Paper Review] How to Serve Your Sandwich? MEV Attacks in Private L2 Mempools

Krzysztof Gogol, Manvir Schneider+2 more|arXiv (Cornell University)|Jan 27, 2026

Auction Theory and Applications0 citations

TL;DR

The paper analyzes sandwich attacks on Ethereum rollups with private mempools, develops economic and execution feasibility models, and empirically shows such attacks are rare, often unprofitable, and largely absent in current private-mempool L2s.

ABSTRACT

We study the feasibility, profitability, and prevalence of sandwich attacks on Ethereum rollups with private mempools. First, we extend a formal model of optimal front- and back-run sizing, relating attack profitability to victim trade volume, liquidity depth, and slippage bounds. We complement it with an execution-feasibility model that quantifies co-inclusion constraints under private mempools. Second, we examine execution constraints in the absence of builder markets: without guaranteed atomic inclusion, attackers must rely on sequencer ordering, redundant submissions, and priority fee placement, which renders sandwiching probabilistic rather than deterministic. Third, using transaction-level data from major rollups, we show that naive heuristics overstate sandwich activity. We find that the majority of flagged patterns are false positives and that the median net return for these attacks is negative. Our results suggest that sandwiching, while endemic and profitable on Ethereum L1, is rare, unprofitable, and largely absent in rollups with private mempools. These findings challenge prevailing assumptions, refine measurement of MEV in L2s, and inform the design of sequencing policies.

Motivation & Objective

• Motivate understanding of MEV sandwich feasibility on rollups with private mempools.
• Extend an optimal frontrun/backrun sizing model to CPMMs and CLMMs.
• Develop an execution-feasibility framework for private mempools and probabilistic inclusion.
• Empirically measure sandwich activity across major L2s and assess economic viability.

Proposed method

• Develop an economic model for optimal frontrun size V_f relative to victim input V_v under CPMMs and CLMMs.
• Derive propositions for optimal frontrun sizing inside a tick (CPMM) and across CLMM ticks.
• Formulate an execution-feasibility model with co-inclusion probability under FCFS and PGA sequencing.
• Create an empirical pipeline to identify sandwich patterns with actor attribution and ordering constraints.
• Compute net PnL with and without gas costs, assessing profitability.

Figure 1: Density heatmaps of swap sizes (USD) for front, victim, and back legs. The top row plots back-run versus front-run swaps; a genuine sandwich would concentrate mass along the diagonal $b=f$ , yet observed densities scatter widely away from it. The bottom row plots victim versus front-run sw

Experimental results

Research questions

• RQ1 What is the profitability of sandwich attacks on rollups with private mempools under CPMM and CLMM models?
• RQ2 How do private mempool sequencing policies (FCFS vs PGA) affect the probability of successful co-inclusion?
• RQ3 Are observed sandwich-like patterns economically consistent and profitable in major L2 rollups?
• RQ4 What are the typical victim trade sizes and attacker-backrun relationships observed in L2 data?

Key findings

• Economic: optimal frontrun size is approximately half the victim trade inside a tick, with CLMMs producing piecewise quadratic profits and boundaries when pushing into thinner liquidity.
• Execution: co-inclusion probabilities on private mempools are low (roughly 5%–20%), making same-block sandwiches probabilistic rather than atomic.
• Empirical: naive heuristics overstate sandwich activity; most flagged triples fail economic consistency checks and median net PnL is negative.
• Profitability: median gross PnL is near zero and net PnL remains negative after slippage and gas costs across studied L2s.
• Bot activity: sandwich attackers are few and concentrated; many bots show low sandwich efficiency, suggesting other arbitrage or market-making activities dominate.
• Overall conclusion: sandwiches are rare, unprofitable, and largely absent in rollups with private mempools, challenging L1 MEV generalizations.

Figure 2: Violin distribution of estimated net PnL in USD. The distributions are tightly centered around zero, with heavy tails, confirming that median profits are negligible and most candidate sandwiches are not profitable.