Ethereum Reorgs "Extremely Unlikely" After POS Merge: Research Paper

A new research paper by Ethereum co-founder Vitalik Buterin and blockchain researcher Georgios Konstantopoulos suggests that after “the Merge” or when Ethereum shifts from Proof of Work (POW) to Proof of Stake (POS), the recently discussed possibility of miners running custom clients (with existence not established with certainty) to accept bribes to execute Ethereum reorgs and extract harmful MEV would become “extremely unlikely”.

The paper titled “Ethereum Reorgs After The Merge” published on July 20 describes the various “finality” parameters for different consensus mechanisms, highlights the danger of Ethereum reorgs, delves into the miner’s likelihood of doing that, explains why the Merge is the preferred solution and complete activation of Proof of Stake (POS) is the block-chain’s best bet to combat such issues.

Ethereum and Bitcoin rely on the “longest chain” to define the fork choice rule – the determination of the chain to follow when multiple valid chains exist. Any reorg means that a block which was previously part of the chain is discarded in favor of a competing block. Now, short reorgs are common in Proof of Work (POW) consensus mechanism and typically arise from latency – that is different participants in the network see different valid blocks at the same time and have to define the valid one.

There’s little doubt that “intentional” Ethereum reorgs are detrimental to the health of the blockchain network. They can increase node operation costs, cause user experience to degrade by requiring more confirmations for certainty, introducing transaction context uncertainty by introducing execution vulnerabilities – degraded DeFi trade results and increased vulnerability to 51% attacks.

Ethereum reorgs are “solved” or made “highly unlikely” by the Ethereum shift to Proof of Stake (POS) implementation Gasper, which introduces two roles during block production – the validators proposing the block and attesters determining the valid block or giving weight to it. Here’s why it makes reorgs unfeasible.

The opportunity to propose a block arises every 12 sec. A random committee of ~1/32 validators is selected, where one is a proposer and the rest are the attesters. Hence, any concentration of “malicious attesters” is extremely unlikely, since an attacker has to concentrate validators greatly and employ difficult hard coordination to have such an opportunity.

Vitalik and Konstantopoulos assert that “Today, the Beacon Chain has ~196k validators, meaning every slot has a committee with a size of ~6125. As a result, even single-block reorgs are extremely difficult, because an attacker controlling only a few validators has no way to beat the honest majority of thousands of attesters. Reorg-attacking successfully requires solving the hard coordination problem of getting a large portion of validators onboard at the same time”. Ethereum reorgs might become a problem closer to the Merge, which is why the Merge should be sped up and the ability to deploy emergency merge should be developed to tackle any misadventure from the miners.