Vitalik proposes anti-correlation incentives to promote decentralization of staking protocols, punish misbehaving actors, and implement them in various scenarios based on empirical data.
Vitalik Buterin recently published a thought-provoking article exploring the concept of using anti-correlation incentives to support distributed staking. This article, written by Vitalik Buterin, presents a preliminary study and recommends independent replication attempts to validate the proposed ideas.
The main focus of this article is to address the challenge of encouraging better decentralization within staking protocols. The authors suggest that penalizing correlations between actors could be an effective mechanism to encourage more distributed and resilient ecosystems.
The current approach of the Ethereum slashing mechanism already includes an element of anti-correlation incentives. However, this article argues that relying solely on extreme incentives, which can arise only in very exceptional attack situations, may not be enough to drive decentralization.
To further strengthen the anti-correlation incentive, this article proposes to extend this concept to address more general failures such as missing proofs. It is assumed that large stakers, including wealthy individuals and staking pools, often run multiple validators on the same internet connection or physical computer, resulting in correlated failures. The article acknowledges that expecting these stakeholders to set up independent physical setups for each validator eliminates the economies of scale in staking.
To test the hypothesis, the authors combine proof data from recent eras with information mapping validator IDs to publicly known clusters. This article provides empirical evidence against excessively correlated failures within clusters by analyzing the occurrence of co-failures (when two validators within the same cluster fail during the same slot). This data supports the notion that validators from the same cluster are more likely to miss attestations simultaneously than validators from different clusters.
Based on this analysis, this article proposes a penalty mechanism that bases the number of currently missing slots on the average of the last 32 slots. This mechanism ensures that the penalty for missing proofs is proportional to the number of validators that fail in a particular slot compared to recent slots. This article highlights the resilience of this mechanism because it cannot be easily manipulated and does not provide incentives for actors to intentionally fail.
The research presented in this article contributes to the ongoing discourse on distributed staking and provides insight into the potential benefits of anti-correlated incentives. By encouraging decentralization and mitigating correlated failures, staking protocols can become more robust and more resistant to attacks.
It is important to note that the study presented in the article is preliminary and the authors encourage independent replication attempts to support the results. The code used in the analysis is available on GitHub.
In conclusion, supporting decentralized staking through anti-correlated incentives offers a promising way to improve protocol decentralization. Staking protocols can create a stronger and more resilient ecosystem by penalizing the correlation between misbehaving actors. Further research and experiments in this area will contribute to the development of decentralized blockchain networks such as Ethereum.
Image source: Shutterstock