Bitcoin in Practice: Network Policies

By Kurt Wuckert Jr.

Network Policies

Bitcoin’s protocol defines what the node network is, and how it is supposed to behave. At the center of that is a simple idea: a node is an actor that adds new blocks to the blockchain. In plain terms, a node is a miner, and mining is not just “running software.” It is producing blocks and participating in the network’s enforcement of the rules.

What nodes do

Nodes are characterized by a few concrete capabilities:

  • Produce blocks. They gather transactions and assemble them into a candidate block.

  • Validate and relay blocks. They receive blocks from peers, check them against the rules, and share valid blocks onward.

  • Share found blocks efficiently. When a node finds a block, it announces it to peers, but the full contents are usually already known to other nodes because the transactions have already been propagated across the network.

  • Validate transactions. They check that transactions follow the protocol rules before accepting them into their candidate block.

The protocol’s operational rule for progress is also straightforward: the longest chain of blocks is the chain the network treats as correct, and that is the chain nodes use as the basis for building the next blocks.

Nodes enforce the rules, but they do not invent them. A useful analogy is law enforcement: police enforce laws; they do not get to rewrite the law because it is inconvenient that day.

Block subsidy and obligations

New coins enter circulation through the block subsidy, which is part of the block reward created in the first transaction of each block, the coinbase transaction. In this description, miners are treated as being bound by a unilateral contract from the issuer to perform the duties of a node, and failure to do so is framed as a breach that could create legal liability.

Beyond that, nodes are described as having network-facing obligations:

  • Include every transaction they see in the block they are proposing, including transactions with no fee, up until they hit a real capacity constraint. At that point, if they must prioritize, they can choose fee-paying transactions over free ones.

  • Protect the network from malicious activity by users, such as double spend attempts or movement of stolen funds, and also from dishonest mining nodes attempting to attack the network.

A key distinction is made here: an entity that runs node software but never produces a block is not really participating in block production, and in that sense is not recognized as a true node in the system. In other words, “node” is being defined by economic function and output, not by simply having a copy of the software installed.

Difficulty adjustment and the practical number of miners

The protocol also specifies that mining difficulty is readjusted on a schedule. In the original protocol description, difficulty adjusts every two weeks. With an average target of 10 minutes per block, that implies up to 2016 blocks in a two-week window.

This text then uses that 2016 number as an upper bound on the number of possible nodes at a point in time, under the thought experiment that every block in that window could be found by a different node. In practice, the text states the network tends to have a small number of large nodes and a handful of smaller ones, roughly three to five large nodes plus some small miners, and treats that as a practical picture of decentralization in this model.

It also notes that when multiple forks exist using the same hashing algorithm, difficulty behavior can change. In the specific case described here, difficulty is adjusted after every block.

Longest Chain

Once you understand what a node is, the next key concept is what the protocol means by “longest chain,” because that defines the network’s current state at any moment.

The whitepaper describes it like this: proof of work is essentially one CPU, one vote. The “majority decision” is represented by the longest chain, meaning the chain with the most proof of work invested. If most of the computational power is controlled by honest nodes, that honest chain grows faster and outpaces competing chains.

What “majority” actually means here

This has sometimes been misread as “democracy for users.” The clarification in the text is that it is not democratic in the sense of one person, one vote. It is not one vote per IP address or per machine either. It is a majority decision among commercial block producers, weighted by investment. You could summarize it as: one vote per unit of sustained work.

Automation is not the whole story

Another common mistake is assuming the system is fully autonomous in a way that removes human judgment. The longest chain represents the network’s majority decision, but the text argues it is an error to think every node must blindly follow it.

The reason is that nodes are not just passive robots; node operators can decide a chain is invalid and choose to override the automatic choice, selecting an alternative chain they believe contains the longest chain of valid transactions. That choice can cost money in the short term, but it is framed as a deliberate tradeoff to preserve network honesty, which supports profitability over the long term.

The bottom line in this model is simple: nodes do not merely drift with whatever chain happens to be longest. They make a conscious decision about what they consider the honest chain, then build on it.

ELI5(ish...)

Think of bitcoin like a public record book that gets updated in pages. The people who get to write the next page are miners, and in this explanation, those miners are the “nodes.”

A real node is not just someone who downloaded an app. A node is someone who actually writes pages, meaning they build blocks, share them with other miners, and check that new blocks and transactions follow the rules.

The “longest chain” is basically the version of the record book that has the most work behind it. Work is like receipts for electricity and machines burned to earn the right to write those pages. More work means a heavier, harder-to-fake history.

But miners are not supposed to follow the longest chain blindly if they think it is cheating. If they believe a chain is invalid, they can choose to build on a different one, even if that costs them money at first. In this view, keeping the record book honest is part of what keeps the whole system valuable.