The puzzle is a search through unpredictable answers
In proof-of-work systems, computers repeatedly run transaction data through a hash function—a calculation that turns input into a short, fixed-length string. Miners change a small piece of the input, often called a nonce, and calculate again. Because a hash’s output is effectively unpredictable, there is no shortcut for finding one that meets the required condition; miners mostly have to try again and again.
The condition is usually that the hash fall below a target value. A lower target means fewer outputs qualify, so each attempt has a smaller chance of success. The system can make the search harder or easier by changing that target, while checking a proposed answer takes only one calculation.
Why the search consumes so much power
Many miners search at once, each making enormous numbers of guesses. Only one qualifying result is needed to propose the next block, so the work done by the other machines does not add more transactions to that block. Their competition is what makes the result costly to produce: a miner that wants to win must devote computing hardware and electricity to the search.
The computation is not solving a useful scientific problem. Its purpose is to make block production require a real, measurable expenditure. That cost helps deter attempts to rewrite transaction history. An attacker trying to replace older blocks would need to perform the work for those blocks again and compete against the honest network’s continuing work. The faster the rest of the network adds work, the harder it is to catch up.
The power use is therefore the security tradeoff built into proof of work: verification is cheap, but gaining influence over the ledger requires sustained computation. That asymmetry is why a tiny answer can be easy to check yet expensive to find.
