An uncle block, also called an ommer block, is a valid block mined concurrently with another block but does not become part of the canonical blockchain. It meets all technical requirements but arrives slightly after a competing block at the same height has been accepted, so it is excluded from the main chain.
The term "uncle" comes from the Merkle tree data structure used to organize blockchain data. In this model, blocks have family-like relationships: parent, child, and sibling. A block that shares a parent with the current chain tip but is not on the main chain occupies a position like an uncle in a family tree. Ethereum's documentation later adopted the gender-neutral term "ommer" instead of "uncle," though both remain widely used in the blockchain community.
Uncle blocks should not be confused with orphan or stale blocks, though the terms are used interchangeably in informal settings. Orphan blocks lack a known parent and are invalid. Stale blocks, often referenced in Bitcoin, are well-formed but no longer part of the longest valid chain. Uncle blocks, by contrast, are fully valid with a known parent; they simply lost a propagation race to another block at the same height.
Uncle blocks arise from the decentralized and asynchronous nature of blockchain networks. When multiple miners work simultaneously, two or more can find a valid block nearly at the same time. Each miner broadcasts their block across the network, but because nodes receive information at different speeds due to latency, some accept one block while others accept the competing block first.
Ethereum's proof-of-work system targeted a block time of about 13 to 15 seconds. This short interval made it likely that two miners would occasionally solve their puzzles in close succession. When the network resolved the tie by selecting the block with the most accumulated mining difficulty, the losing block became an uncle. Transactions in that uncle block could be rolled back but remained visible in the chain's underlying data structure.
Unlike Bitcoin, which discards losing blocks without compensation, Ethereum rewarded miners whose blocks became uncles. Two types of rewards were involved. The miner whose block became an uncle received a partial block reward, reduced because the block did not join the main chain. The miner of the subsequent "nephew" block who referenced the uncle received a small inclusion reward.
Ethereum allowed a block to reference up to two uncle blocks, which had to fall within seven blocks of the current chain tip. This gave miners a reasonable window to have their uncle referenced and compensated while keeping the protocol predictable.
Uncle blocks serve functions beyond being excluded alternatives. By incorporating uncle block references into the main chain, Ethereum increased the total mining work acknowledged by the network. This raised the effective cost of a 51% attack, since an attacker would need to overpower not only the main chain's hash rate but also the computing power reflected in uncle references.
The reward mechanism also addressed a subtle centralization risk. Large mining pools benefit from faster internal communication and can propagate new blocks more quickly than solo miners or smaller operations. Without uncle rewards, smaller miners would suffer disproportionately from propagation delays, pushing the network toward pool dominance. Compensating miners for uncle blocks reduced this disadvantage and kept mining participation more distributed.
Ethereum transitioned from proof-of-work to proof-of-stake in September 2022 through an upgrade called the Merge. Under proof-of-stake, validators replace miners and block production follows a deterministic schedule instead of a computational race. This change eliminated the conditions for uncle blocks, and ommer rewards were removed from the protocol.
Uncle blocks remain a feature of Ethereum Classic, the network that split from Ethereum following the DAO hack in 2016. Ethereum Classic has retained proof-of-work consensus and continues to generate and reward uncle blocks as part of its standard operation.
Bitcoin handles competing blocks differently. When two miners produce valid blocks at the same height, Bitcoin's longest-chain rule selects the winner and discards the losing block without compensation or lasting reference. Bitcoin's ten-minute block target also makes simultaneous discoveries less frequent, reducing the problem's impact. The tradeoff is a harsher environment for miners who produce a valid block but lose the propagation race through no fault of their own.
.webp)