Competing Chain

Overview

When two blocks extend the same parent, each can gain descendants and form a competing chain. Different nodes may temporarily prefer different branches because messages arrive in different orders or because they are partitioned. In proof-of-work, the branch with greater accumulated work typically becomes preferred. In proof-of-stake and BFT systems, votes, locks, checkpoints, and finality rules determine whether competition can continue.

A competing branch can contain transactions that conflict with the current Canonical Chain. The same funds may be spent differently, or a transaction may appear only on one side. If fork choice changes, a reorganization replaces part of the previous history. This is why transaction inclusion in a recent block is provisional and why the Best Chain can change before stronger settlement evidence accumulates.

Not every competing chain is malicious. Short forks are a normal consequence of distributed timing in some networks. Persistent or deep competition can indicate network partition, client incompatibility, validator equivocation, or an attack. A chain can also be structurally valid but lack enough consensus weight to become canonical. Operators should investigate depth, duration, affected providers, and whether finality has stalled or been violated.

Payment systems should detect disagreement by comparing block IDs at the same number and following parent links. They should avoid releasing irreversible value based on a branch observed by only one provider. If a processed block becomes noncanonical, internal records need reversal and replay logic. Cross-chain protocols require particular caution because acting on a competing source chain can create destination assets or messages that cannot be safely undone.