A chain reorganization, often called a reorg, is when a blockchain node switches from its current view of the chain to a different valid chain that has more accumulated work or stake. In practice, this means one or more of the most recent blocks are discarded and replaced, so the network converges on a single agreed history.
How a reorg happens
Reorgs usually occur because block production is probabilistic. Two miners (or validators) can produce different blocks at nearly the same time, creating a temporary split where some nodes see one tip and others see another. Both branches can be valid, but consensus rules instruct nodes to follow the chain with the most cumulative proof of work, or the heaviest chain under the protocol’s weighting.
Once a new block is added to one branch, making it longer or heavier, nodes on the other branch will reorganize, rolling back the blocks that are no longer on the winning chain. The transactions from those discarded blocks are not necessarily lost, they typically return to the mempool to be re-included in a future block, unless they conflict with transactions already confirmed on the winning branch.
Finality, confirmations, and risk
Reorgs are most common at the “tip” of the chain and usually affect only a small number of blocks. For example, a merchant might see a payment confirmed in a block, but if that block is later orphaned during a reorg, the payment can become unconfirmed again. This is why many systems wait for multiple confirmations before treating a transaction as settled.
While many reorgs are benign and arise from normal network latency, deeper reorgs can increase the risk of double-spending and operational issues for exchanges, wallets, and DeFi apps. Understanding reorgs matters because it explains why blockchains offer probabilistic finality, and why confirmation practices are essential for secure crypto transactions.