What is inside a block: header, height, size and subsidy

What is a block made of?

A block has two parts, a header and a body. Mining hashes only the header.

• The body holds the transactions. It can be a few hundred bytes or several megabytes, depending on how busy the chain is.
• The header is a fixed summary of the block. On most chains it is exactly 80 bytes.
• Mining hashes the header and nothing else.

That last point is the most useful fact in this article. A miner tries billions of variations per second. It could not do that if each attempt meant rehashing a block full of transactions. It only rehashes 80 bytes.

What are the six fields in the header?

The header holds six fields. Each one is there for a reason, and two of them do most of the work.

• The version.
• The hash of the previous block.
• The merkle root, one hash that sums up every transaction.
• A timestamp.
• The difficulty bits.
• The nonce, a counter the miner can change.

The previous block hash makes it a chain. Each header names the block before it. Rewriting an old block would change its hash. That would break the next header, and the next, all the way to the tip. This one field is the whole security argument for proof of work.

The merkle root summarises every transaction in the body. Change one transaction by a single byte, and the merkle root changes. Then the header changes, and every hash tried so far is worthless. This is how 80 bytes can honestly stand in for a block of any size.

The difficulty bits are the target the hash must come in under, in a compact form. The timestamp is roughly when the block was made, and it may be approximate. The nonce is the field the miner is free to change. That freedom is what makes the search possible at all.

What happens when the nonce runs out?

The nonce is a small counter. A modern miner uses up every value it can hold almost instantly. That is not a dead end. The nonce is only the easiest thing to change, never the only one.

When it runs out, the miner changes something else that feeds the header.

• The timestamp can move.
• The coinbase transaction (the first transaction in the block, explained below) can be altered slightly. That changes the merkle root, which gives a completely fresh header to search.

On a pool, this is what the extra nonce is for. The pool gives every miner a different starting point. So no two miners grind the same numbers.

What is block height?

Block height is how many blocks came before a given block. The genesis block, the very first one, is height zero. The next is one, and so on. It is the number beside every entry on the pool blocks page and on any explorer.

Height is not really stored in the header. It comes from where the block sits in the chain.

On most modern chains, the height is also written into the coinbase transaction. That rule was added later, so two blocks at different heights can never be byte identical. Still, the authoritative answer to a block's height is its position.

Height is also the natural way to count confirmations. A transaction buried six blocks deep sits in a block six below the tip. Each further block makes rewriting it more expensive.

What does block size measure?

Block size is essentially the size of the body. The header is a rounding error next to it. A block size limit is about how many transactions fit, not about the header.

A full block and an empty block take exactly the same mining work. Difficulty decides how hard a block is to find. Size decides how much the block carries. The two are unrelated, and mixing them up causes a lot of bad intuition about fees.

Size does affect the fee market. When more transactions want in than fit, those offering higher fees get in first. The total fees in the block rise, and that flows straight into what the block pays.

What is the coinbase transaction?

The first transaction in every block is special. It has no sender. It creates coins out of nothing and pays them to whoever made the block.

It is called the coinbase transaction. It has nothing to do with the exchange of the same name.

It pays the block reward, which has two parts.

• The subsidy, the newly issued coins. The chain's emission schedule sets it, and on most coins it shrinks over time.
• The fees of every other transaction in the block. Users attach them to get their transactions included.

On a pool, the coinbase pays the pool. The pool then shares the reward among everyone who sent shares in that round. This is not a detour. It is how a pool pays many people from a reward that arrives as one indivisible payment.

Why does this matter while you mine?

Once the shape of a block is clear, three things become obvious. All three come up constantly.

• A new block makes your work worthless instantly. The previous block hash changed, so the header you were searching no longer exists. Anything you submit for it afterwards is a stale share, and no amount of speed fixes that.
• The pool sends you fresh work often. Every new transaction the pool wants to include changes the merkle root. Fresher work carries more fees, which means more reward for the same effort.
• Block size never appears in an earnings estimate. What you earn depends on difficulty, the subsidy and the fees. It does not depend on the bytes in the body.

The pool blocks page shows each block's height, its reward and when it was found. That row is the useful part of all of the above. The explorer shows the rest, including the coinbase transaction that paid it.