Proof of Work, Explained

A costly signal

Proof of work requires repeated hashing but carries no shortcut that guarantees success. A valid result demonstrates that computation was expended, allowing independent nodes to compare competing histories using an objective accumulated-work rule. <strong>Security through incentives.</strong> An attacker attempting to rewrite settled history must catch up with honest mining while forgoing or risking legitimate rewards.

The exact security budget changes with miner revenue, hardware and energy economics. <strong>What proof of work cannot do.</strong> It does not decide whether a transaction or block follows consensus rules; nodes do. It also cannot make fraud outside the protocol - such as a stolen exchange password - valid or safe.

What this establishesProof of work is expensive to produce and cheap to verify.

Why finding is hard and checking is easy

A miner changes values in a candidate header and hashes it repeatedly, looking for a result below the current target. There is no known shortcut that selects a winning nonce. Success is probabilistic, so more hashes create more chances. Once a result is found, every node can hash that header once, compare the result with the target, and verify the block's transactions.

The network therefore spends substantial work to propose history but little work to reject a false claim. If someone changes an earlier transaction, the affected header and every descendant header must be rebuilt. Honest miners continue extending the accepted chain during that attempt.

Proof of work orders valid histories by accumulated work; it does not excuse invalid signatures, excessive issuance, or broken scripts, which nodes reject separately.

What this establishesNodes enforce validity while work orders valid history.

The puzzle, precisely

A block header is hashed with SHA-256 twice. The result is a 256-bit number, and the block is valid only if that number is below a target set by the current difficulty. Because a hash is unpredictable, the only strategy is to try candidates - changing the nonce, the timestamp within limits, or the coinbase data - until one succeeds.

Checking a candidate takes microseconds; finding one takes the whole network about ten minutes. That asymmetry is the point: producing a block is hard, verifying one is trivial, so anyone can confirm that work was done without redoing it.

What this establishesNetwork security does not automatically secure wallets or exchanges.

Difficulty, targets and the two-week clock

Every 2,016 blocks each node compares the time those blocks took with the intended 20,160 minutes and scales the target so the next 2,016 land closer to ten-minute intervals, within a factor of four per adjustment. Difficulty is simply the ratio of the easiest possible target to the current one.

When hashrate grows, blocks arrive early and difficulty rises; when miners leave, blocks slow until the next adjustment lowers it. The mechanism needs no coordinator and no clock beyond the block timestamps themselves, and the network dashboard on this site shows where the current period stands.

Proof of work compared with the alternatives

Proof of stake, used by Ethereum since 2022, replaces external cost with internal cost: validators lock coins and lose them for misbehaviour. It uses far less energy and settles faster, but it ties block production to existing holdings, requires the chain to define and punish misbehaviour, and makes 'who owns the coins' the security question.

Proof of work ties production to a resource anyone can buy anywhere, and makes rewriting history cost real electricity that cannot be recovered. Reasonable people weigh these differently; Bitcoin's choice is deliberate, and its developers have declined to change it.