Proof of work explained: hard to find, one hash to check
Proof of work is a search that can only be done by trying, over and over, until a hash comes out below a target. Finding one costs a fortune in guesses and electricity. Checking it takes a single hash, which is why cheating is expensive and catching a cheat is cheap.
In a gold rush, anyone could walk into the assay office with a nugget and claim it came from his own claim. The assayer didn't need to know how hard the digging had been. He weighed it and tested it. Real gold or not, the answer took minutes.
Proof of work turns that idea inside out and makes it stricter. The nugget is a number. It can only be found by digging, guess after guess, for an enormous amount of time. And anyone can test it, in a fraction of a second, with one calculation.
That gap is the whole trick. Finding is expensive, checking is cheap. So lying about your work costs you, and catching a liar costs everyone else almost nothing.
What it is
Proof of work is a puzzle with no shortcut. The miner must find a block header whose SHA-256 fingerprint, taken twice, is lower than a target number the network sets. The only way to find one is to try: change the nonce, hash, compare, repeat.
The white paper puts it in two lines. The proof of work involves "scanning for a value that when hashed, such as with SHA-256," starts with a number of zero bits, and "The average work required is exponential in the number of zero bits required and can be verified by executing a single hash" (Nakamoto, 2008). The idea goes back to Hashcash, covered in "b-money, Bit Gold and Hashcash". How the hash itself works is in "SHA-256 explained".
How it works: a toy search, then a real one
The toy. For this article, on 19 November 2026, a short Python script hashed the text "bestcrypto.guide block" followed by a counter: 0, 1, 2 and so on. It stopped when the SHA-256 hash started with a given number of zeros in hexadecimal. Each hexadecimal zero makes the search 16 times harder on average.
- 4 zeros: found at counter 13,446, so 13,447 tries. The average to expect is 65,536 (16 to the power 4). This search got lucky.
- 5 zeros: 646,954 tries, against an average of 1,048,576.
- 6 zeros: 41,655,884 tries, against an average of 16,777,216. This one was unlucky, and took the script about 18 seconds.
Two things to notice. One extra zero means 16 times more work on average. And single searches swing far above or below that average, because every try is a fresh lottery ticket.
Checking any of those results takes one hash. Type the text with the winning counter, hash it once, look at the start. Done.
The real one. Bitcoin block 900,000, mined on 6 June 2025, has a block hash starting with 19 hexadecimal zeros: 000000000000000000010538... (mempool.space). The rule wasn't "19 zeros", though. The rule was "below the target", and that block's target, decoded from its header, starts with 19 zeros followed by 237740.
The same script took the 80-byte raw header that mempool.space publishes for block 900,000, ran SHA-256 on it twice, and got exactly that hash, below the target. One check, under a millisecond. Bitcoin Core's function for this is called CheckProofOfWork (pow.cpp).
How much digging did it take to find? The block's difficulty was 126,982,285,146,989. Multiplied by 2 to the power 32, that is about 545,000,000,000,000,000,000,000 hashes expected on average, roughly 5.45 times 10 to the power 23. The toy script managed about 1.9 million double hashes a second. At that pace, the search for one Bitcoin block would take about 9 billion years. The checking took the same script one try.
Why it matters
- Cheating costs real money. A fake block needs the same search as an honest one. Its electricity is spent whether the network accepts the block or not.
- Checking is democratic. Any full node can verify every block since 2009 on ordinary hardware. Nobody has to trust the miners' word.
- The heaviest chain wins. The white paper: "The majority decision is represented by the longest chain, which has the greatest proof-of-work effort invested in it" (Nakamoto, 2008). To rewrite history, you have to outdig everyone else. What that would take is in "The 51% attack explained".
- Who pays. Miners pay in hardware and power. The network pays them back in new coins and fees, as long as those are worth the digging.
Where people get it wrong
- "The work is useful computation." It isn't, and that's the point. The work has to be pure cost so it can't be faked or recycled.
- "Bitcoin requires a fixed number of zeros." It requires a hash below a target. Counting leading zeros is a rough way of reading that target.
- "A faster computer finds the answer by being smarter." It just guesses faster. No guess is better than any other.
- "Proof of work means the block is correct." It only proves effort. Every node still checks the transactions, and a block full of invalid ones is rejected, however much work it carries.
- "Proof of work is just a way to waste energy." The energy use is real, and the honest numbers are in "Bitcoin's energy use: the honest numbers". But the cost is what makes rewriting history expensive.
Test one nugget yourself
Open a block explorer, the assay office, and look at any block hash. Count the zeros at the front. Then look up the block's difficulty. Multiply it by about 4.3 billion and you have the average number of guesses a whole industry spent, in roughly ten minutes, on the zeros you just counted.
Then remember that you checked it in a second, for free. That's the deal proof of work offers: the miners dig, and you just look.
Frequently asked questions
What is proof of work in simple terms?
Proof of work is a search for a number that makes a block's hash come out below a target. It can only be found by trying again and again, but anyone can check the answer with a single hash. That makes cheating expensive and checking cheap.
Why is proof of work hard to fake?
Because there is no shortcut. Every guess is equally likely to win, so the only way to find a valid block is to spend the guesses, and the electricity behind them.
How long does it take to check proof of work?
One double SHA-256 hash of the 80-byte block header, a tiny fraction of a second on an ordinary computer. The node then compares the result with the target written in the header.
What does the number of leading zeros in a block hash mean?
It is a rough sign of how low the target was. Each extra hexadecimal zero means about 16 times more guesses on average. The real rule is that the hash must be below the target, not a fixed count of zeros.
Who invented proof of work?
Bitcoin's proof of work builds on Hashcash, which Adam Back proposed in May 1997 as a way to make spam costly (Back, 2002). Satoshi Nakamoto's white paper cites Back's paper among its references.

