Equilibrium: Preventing Arbitrage Attacks in Optimistic Rollups

DOI: 10.1109/ICBC67748.2026.11575540

One of the most prominent applications of blockchains is decentralized finance, like Automated Market Makers (AMMs), which are exchange houses without a central authority. AMMs currently handle large amounts of tokens and are subject to vulnerabilities. One attack, called arbitrage, involves trading with collections of AMMs in a way that the attacker ends up with more tokens than before trading, thus gaining tokens from the AMMs. Moreover, arbitrage attackers compete by offering higher transaction fees, which affects the whole blockchain ecosystem.

In this paper, we present Equilibrium, a novel solution to prevent arbitrage attacks in Layer-2 optimistic rollup blockchains (L2) where many AMMs currently execute. In our solution, L2 miners re-establish the balance by removing large arbitrage opportunities when computing new blocks. The re-balance is computed offchain as a solution to an optimization problem. To ensure the correct evolution of Equilibrium, we provide novel fraud-proof games that guarantee the correctness of the re-balancing process and, consequently, of the blockchain. Our preliminary empirical evidence suggests that re-balancing reduces arbitrage profits while redirecting part of them to AMMs, and is inexpensive to compute and infrequently required.