Modeling game-theoretic corruptive majority attack on ethereum

Abstract

This paper studies the economic feasibility of coordinated majority attacks on the Ethereum blockchain under its Proof-of-Stake consensus mechanism. Focusing on Ethereum’s validator-based finality rules, the paper models a corruptive attack in which validators are induced to deviate from the protocol in exchange for off-chain transfers. Under perfect information, the model derives a bribery cost schedule implying that an attacker must effectively finance just over half of total staked ether to assemble two-thirds of finality power, despite only needing to corrupt a subset of validators at each margin. The paper further shows that, under risk-neutral Bayesian behavior, uncertainty about the attacker’s stake weakly reduces expected bribery costs due to the concavity of the compensation function. The results highlight how Ethereum’s economic security depends critically on slashing design, information structure, and validator fragmentation.