Mining pool

Pools solve a statistics problem rather than a technical one. Miners submit near misses called shares to prove how much work they are doing, and the pool pays out on that evidence whether or not it finds a block that day. The first pool opened in November 2010; the two largest now mine close to half of all blocks between them, which is the uncomfortable part.

How it works

A pool turns one improbable jackpot into a predictable wage, and it does that by measuring work that would otherwise be invisible.

The measuring device is the share. Alongside the real network target, the pool sets a much easier target of its own, and every time a miner finds a header below that easier threshold it submits the result. Shares are worthless to the network and perfect as evidence: they cannot be forged, they cannot be reused at another pool because the payout address inside the block is the pool's own, and their arrival rate is a direct sample of a miner's hash rate. Occasionally a share happens to clear the real target too, and that is the pool's block.

Payment schemes differ in who carries the bad luck. Under full pay per share, the pool pays a fixed amount for each share regardless of whether it finds blocks, absorbing all the variance itself and charging for the privilege, commonly 2 to 4 percent. Under pay per last N shares, miners are paid only out of blocks actually found, so the pool takes less, often between 0 and 2 percent, and quiet weeks hit the miner directly. Home miners generally prefer the smoother version; large operators with financing to service almost always do.

Two details surprise people. First, the coinbase output pays the pool, not you, so a pool is a custodian holding your earnings until payout. Second, under the long standing Stratum V1 protocol the pool builds the block template, meaning it, not its thousands of members, chooses which transactions are included. Braiins Pool, the descendant of that first 2010 pool, and newer pools built on Stratum V2 are working to hand that choice back.

Why this matters when you buy bitcoin

Pool concentration is the closest thing bitcoin has to a control point, and it decides whether your transaction gets into a block at all.

The concentration is real and worth knowing about. Foundry USA has run around a third of network hash rate, with AntPool next, and a handful of pools account for most blocks. Nothing about that lets them steal coins or change the supply, because nodes reject invalid blocks whoever produced them. What a pool can do is refuse to include specific transactions.

This has been tested once already. In May 2021 Marathon Digital launched a pool that filtered transactions touching addresses on the United States sanctions list and mined a block that excluded them. The backlash was immediate and the company abandoned the policy within weeks, publicly committing to stop censoring transactions. The episode is the reason transaction filtering is discussed seriously rather than theoretically, and it matters most to people in the jurisdictions this site covers where sanctions screening is aggressive.

For an ordinary buyer the practical effect is mild but real. Filtering by one pool only delays a transaction, since the next pool includes it, and the cost is measured in blocks rather than in coins. It is still the strongest argument for the protocol work aimed at letting individual miners choose their own transactions.

The variance a pool removes

A single machine's odds explain why pools took over within two years of bitcoin's launch.

One 200 terahash machine against a network of roughly 850 exahashes per second holds about 0.000024 percent of the total, which works out to a 1 in 29,500 chance per day of finding a block. That is one block every 81 years on average, with no guarantee it arrives in your lifetime. Join a pool and the same machine earns roughly 10,600 satoshis a day, every day, minus the pool fee. The expected value is nearly identical; what changes is that you can pay an electricity bill with it. That trade, predictable income for a lottery ticket, is the entire product a pool sells.

Mining pool vs solo mining

A mining pool spreads one block reward across many miners; solo mining keeps the whole thing and usually keeps nothing. The confusing part is that services called solo pools exist, such as solo.ckpool.org, which provide pool infrastructure but pay winner takes all: you use their servers and connection, and if your machine finds the block you receive the entire reward minus a small fee. That is solo mining with the plumbing outsourced, not pooled mining. A real pool splits every reward it wins in proportion to shares submitted.

Mining pool vs 51 percent attack

A large pool is not the same thing as an attacker with a majority of hash rate. A pool with 40 percent of blocks does not own that hardware; it rents the loyalty of thousands of operators who can and do repoint their machines within minutes, as happened when GHash.io briefly passed half the network in June 2014, then pledged publicly to stay under 40 percent while its members left for rivals. An attacker would need the machines themselves, or a pool willing to burn its business overnight. The concentration is a genuine governance problem and a fragile foundation for an attack, which is why the industry treats it as a decentralization issue rather than an imminent theft risk.