Finality

In practice, finality is what separates a payment that merely looks confirmed from one a merchant, exchange, or another protocol can safely treat as settled. Every blockchain reaches this point differently, and the design choice shapes how long users must wait before trusting a transfer.

Proof of Work chains like Bitcoin rely on probabilistic finality: each new block mined on top of a transaction makes reversing it exponentially more expensive, but the guarantee is never absolute. The convention of waiting for six confirmations, roughly an hour, traces back to Satoshi Nakamoto's original probability calculations, and it drops the chance of a successful reorg to a fraction of a percent under normal conditions, though it cannot rule out a deep reorganization by an attacker with enough hashing power.

Many Proof of Stake networks instead aim for deterministic, or absolute, finality. Ethereum's Casper FFG mechanism groups blocks into epochs and finalizes a checkpoint once two thirds of staked ETH has attested to it across two consecutive epochs, typically around 13 minutes. Reversing a finalized block would require attackers to forfeit a huge share of their staked ETH, a penalty severe enough that the outcome is often called "economic finality" rather than pure mathematical certainty.

• Fast custom chains built on Tendermint style Byzantine fault tolerant consensus can finalize blocks within a few seconds.
• Exchanges and payment processors set their own confirmation thresholds based on a chain's finality model and the size of a transaction.
• Layer 2 rollups add another wrinkle, since their transactions often only inherit full finality once posted to and finalized on the underlying base layer.

Understanding a chain's finality model matters for anyone accepting large payments, bridging assets, or building applications that assume a settled transaction cannot later disappear.