The Lottery Nobody Wants to Play Alone
You've been running your ASIC for six months. Your colleague started the same day, same hardware, same hashrate. She's found two blocks. You've found zero. Statistically, that's entirely plausible, and there's no one to complain to about it: it's the direct consequence of how proof-of-work block discovery works. The math simply doesn't promise you anything on a six-month horizon when you control 0.001% of global hashrate.
Mining pools exist specifically to kill that variance. Not the expected value. The variance. That distinction is the whole point, and most explainers bury it.
What Variance Actually Means Here
In probability terms, variance measures how widely your outcomes scatter around the average. A solo miner with a small rig has a high expected time-to-block (potentially decades) and enormous variance around that expectation. The pool doesn't change your fair share of total rewards. It changes how reliably you receive it.
Here's the core mechanism. A Bitcoin block requires a hash below a target value, currently demanding something in the vicinity of 10^21 attempts per block network-wide. Your individual miner produces, say, 100 terahashes per second. The probability of you personally finding the next block is roughly your hashrate divided by total network hashrate. Small number. Very small.
A pool aggregates thousands of miners. Together they might control 20% of network hashrate, which means the pool finds a block roughly every five blocks on average rather than one in every hundred thousand. The pool's variance collapses toward a smooth, predictable income stream. Your share of that stream is proportional to your contribution. That's the trade.
Shares: The Accounting Unit That Makes It Work
Here's the part most guides skip entirely.
The pool can't just trust miners to report how much work they've done. Instead, it issues a parallel proof-of-work system with a much easier target, called a share. Where a real block might require a hash starting with 20 leading zeros, the pool's share target might only demand 10. Miners submit these easier hashes continuously, the pool logs each one, and occasionally a submitted share also happens to meet the real block target. Block found, everyone gets paid.
Shares are tamper-resistant work receipts. You can't fake them without actually doing the hashing, and the rate at which a miner submits valid shares is a statistically reliable proxy for their true hashrate over any window longer than a few minutes. They function less like a clock-in card and more like a scale: you can't bluff your weight.
Under a basic proportional scheme, the payout math looks like this: if the pool earns 3.125 BTC from a block (the post-fourth-halving subsidy), and your shares represent 2% of all shares submitted during the round, you receive 2% of 3.125 BTC, minus the pool's fee. Simple division. No trust required beyond the pool's accounting software.
A Worked Scenario: Two Miners, One Pool
Take Marcus and Priya. Both join the same pool on the same day. Marcus runs a single ASIC producing 100 TH/s. Priya runs a small farm at 1,000 TH/s. The pool's combined hashrate is 10,000 TH/s.
Over a 24-hour window, the pool finds eight blocks, which is plausible for a pool this size. Marcus's share of total submitted shares comes out to 1% (100/10,000). Priya's is 10%. Marcus receives 1% of eight blocks' worth of rewards. Priya receives 10%.
Now compare this to solo mining. In solo, Marcus has a 1-in-10,000 chance of finding any given block. In an eight-block day, his expected solo earnings are 0.0008 blocks. His actual solo earnings that day: almost certainly zero. The pool gave him a small, reliable slice instead of a lottery ticket. Priya, at 10% of hashrate, would expect roughly one solo block per day anyway, so the variance reduction matters less to her, though it still smooths things out considerably.
This is why variance reduction matters more the smaller you are relative to the network. It is not a uniform benefit.
Fee Structures and Payout Schemes
Different pools use different payout schemes, and the differences aren't cosmetic. The proportional model described above has a known vulnerability called pool-hopping: a miner who joins at the start of a round, when expected payout per share is high, and leaves mid-round, when it drops, can extract more than their fair share over time.
This pushed pools toward more sophisticated schemes. Pay-Per-Share (PPS) eliminates round variance entirely for miners by paying a fixed rate per valid share regardless of whether the pool finds a block. The pool absorbs all variance itself and prices that risk into a higher fee. Pay-Per-Share-Plus (PPS+) adds transaction fees into the calculation. Full-Pay-Per-Share (FPPS) is a common contemporary variant doing the same.
Then there's PPLNS (Pay Per Last N Shares), which pays based on shares submitted in a rolling window rather than discrete rounds, making pool-hopping much less profitable. Most of the large pools you'd recognise today use some variant of PPLNS or PPS+.
The catch: no scheme eliminates variance for the pool operator. PPS just shifts it onto them. Smaller pools running PPS can and do face extended losing streaks that stress their reserves. It's happened, and it will happen again.
What People Get Wrong About Pool Mining
The most persistent misconception is that joining a bigger pool gives you better expected returns. It doesn't. A pool with 30% of network hashrate doesn't pay you more per unit of work than one with 5%. The expected value is identical: your hashrate times the block reward divided by network hashrate, minus fees.
What the bigger pool gives you is lower variance around that expectation. Payments come more frequently, in smaller increments, with less scatter. For a solo miner waiting potentially years for a block, that's genuinely valuable. For someone running a large operation with enough hashrate to expect frequent solo blocks anyway, the marginal variance reduction from joining the biggest pool versus a mid-sized one is nearly irrelevant.
So why do so many experienced miners still treat pool size as a proxy for pool quality? Inertia, mostly, dressed up as due diligence.
Still, pool fees aren't free money for the operator. Running reliable payout infrastructure, absorbing variance in PPS schemes, and maintaining uptime is a real cost. A 1-2% fee on a PPS pool is not obviously overpriced, and anyone who dismisses it as a skim hasn't thought through the risk the operator is carrying.
The Real Trade
Mining pools are, at their core, a financial smoothing instrument. They convert a high-variance, lumpy income stream into something that behaves more like a salary. The math is transparent: shares measure work, shares determine proportion, proportion determines payout. No mysticism.
What you give up is self-sovereignty over your block rewards and a small percentage to fees. What you get is predictability. For most miners without enormous hashrate, that is not a close call. The lottery model sounds exciting until month eighteen of zero blocks.
Know your pool's payout scheme before you point your ASICs at it. The difference between PPLNS and PPS+ isn't jargon; it's a genuine difference in who absorbs risk and when you get paid. That's worth ten minutes of reading. The hardware commitment you're making deserves at least that much.