On Ethereum, trades are not executed instantly. When a user submits a transaction, it enters a waiting area, and whoever builds the block gets to decide which transactions are included and in what order. This creates opportunities for profit through transaction reordering, a phenomenon known as Maximal Extractable Value (MEV).
MEV is not inherently malicious, but some forms of it directly harm users. A typical example of MEV is the sandwich attack. A sandwich attack occurs when an attacker sees a user’s trade before it is finalized and places one trade immediately before it and another immediately after it. The attacker profits from the price movement caused by the victim’s trade, while the victim receives a worse execution price. Economically, this behaves like a hidden fee on traders: one that varies by token liquidity, market volatility, and where and how the trade is ordered.
In this project, we will identify sandwich attacks directly from on-chain data, estimate how much value is extracted from users, and analyze when sandwich attacks are most severe (trader characteristics, platform liquidity, exchange protocols, builder concentration, etc.). We will then focus on mitigation. In recent years, several approaches have been proposed to reduce sandwich attacks, including private transaction submission, alternative auction rules, and mechanisms that redistribute MEV back to users.
Throughout the semester, we will engage with recent research on decentralized exchanges and sandwich attacks, collect and analyze on-chain transaction data, identify attack patterns, and evaluate the strengths and potential vulnerabilities of different decentralized exchange designs and operating mechanisms.
Supervisor: Xiaochen Jing
Graduate Supervisor: Zhonghe Wan