If you want a group of strangers scattered around the world to agree on a single, shared record of who sent what to whom, you run into a problem: who decides which version of the record is correct? Traditional systems solve this with a trusted authority, like a bank. Proof of work solves it differently, by making the act of proposing new entries costly, and by making it cheap for everyone else to check that the cost was really paid.
At the heart of proof of work is a cryptographic hash function. A hash function takes any input, of any length, and produces a fixed-length string of characters that looks like random gibberish. The same input always produces the same output, but changing even one character of the input produces a completely different result. Crucially, the function only runs in one direction. Given an output, there is no known shortcut to work backward and find an input that produces it. The only approach is to guess inputs and check what comes out.
Miners bundle pending transactions into a candidate block. That block contains the transaction data, a reference to the previous block, a timestamp, and a changeable number called a nonce. The miner hashes the whole block and looks at the result. The network's rules say the hash must fall below a certain target value, meaning it must start with a long run of zeros. Almost every guess fails. So the miner changes the nonce and hashes again, and again, trillions of times per second, until one attempt happens to produce a hash under the target.
This is why mining is often described as a lottery rather than a puzzle with a clever solution. There is no skill involved, only raw attempts. Your chance of finding a valid block is roughly proportional to your share of the total computing power on the network. What makes the system work is the asymmetry: finding a valid hash takes enormous effort, but anyone can verify it by running the hash function once. Every other node on the network checks the new block in a fraction of a second, confirming both that the hash meets the target and that all the transactions inside are valid.
When a miner finds a valid block, they broadcast it and other nodes add it to their copy of the chain. The winning miner collects a reward, which typically consists of newly issued coins created by the protocol plus the fees attached to the transactions they included. This reward is what pays for the electricity and hardware, and it is what makes honest participation economically sensible.
Each block references the hash of the block before it, which is what turns a pile of blocks into a chain. If someone wanted to alter an old transaction, they would change that block's contents, which would change its hash, which would break the reference in the next block, and every block after it. Rewriting history therefore means redoing all that work, while the rest of the network keeps extending the honest chain. Nodes follow the chain with the most accumulated work, so an attacker would need to out-compute everyone else combined to catch up and overtake it. That requirement is the security model.
Networks do not want blocks arriving too fast or too slow, so the target automatically adjusts. If blocks are being found more quickly than intended, the target tightens and mining gets harder. If miners leave and blocks slow down, the target loosens. This difficulty adjustment keeps the average pace of new blocks roughly steady no matter how much computing power joins or exits.
Because the odds favor whoever can hash fastest, mining hardware evolved from ordinary computer processors to graphics cards to purpose-built chips designed to do nothing but compute one hash function. Individual miners with modest equipment may go a very long time without ever winning a block, so many join mining pools, where participants combine their hashing power and split rewards in proportion to the work each contributed.
This competition is also the source of proof of work's most discussed tradeoff: the security comes directly from real-world energy expenditure. That cost is the feature, not a bug in the design, since it is precisely what makes attacking the network expensive. It is also why alternative consensus mechanisms, which secure networks through staked collateral rather than computation, were developed as a different approach to the same underlying problem.
This article is for general education only — not financial advice, and nothing here is a recommendation to buy, sell, or hold any asset. Cryptocurrency carries real risk of loss; always do your own research before making a financial decision.
