You're three weeks into pool mining and the daily payout numbers are bouncing around like pressure readings on a pipe with a slow leak somewhere upstream. Some days the pool finds three blocks. Some days, zero. Your earnings swing 40% and you haven't touched a single setting. That's not a bug. That's the payout scheme doing exactly what it was designed to do.
The two dominant models, Pay-Per-Share (PPS) and Pay-Per-Last-N-Shares (PPLNS), are not interchangeable. They distribute risk differently, reward loyalty differently, and attract fundamentally different types of miners. Understanding the mechanics takes about ten minutes. Ignoring them can cost you real money.
A Fixed Wage vs. a Piece of the Hunt
PPS works like an hourly wage. Every valid share your miner submits gets credited at a pre-calculated rate, regardless of whether the pool actually finds a block that day. The pool operator absorbs the variance. You submit a share, you earn a fraction of a coin, end of story.
That rate is derived from the current network difficulty and the block reward. If a block is worth 3.125 BTC and the network expects one solution every 700 billion hashes (a rough illustrative figure, not a live number), the pool prices each share accordingly, then shaves off a fee, typically 2-4%, to cover the risk it's carrying. The math holds whether the pool runs hot or cold on block luck that week.
PPLNS works like a hunting party's split. The pool waits until it actually finds a block, then looks backward at the last N shares submitted across all miners, and divides the block reward proportionally among whoever contributed to that window. N is usually defined as a multiple of the pool's difficulty, often somewhere between 2x and 10x the expected shares per block.
Notice what that means: your payout depends not just on your work, but on when you worked.
The Luck Window, Illustrated
Take two miners, call them Priya and Marcus. Both run identical 100 TH/s rigs on a PPLNS pool. Priya has been mining continuously for two weeks. Marcus just plugged in this morning.
The pool goes eight hours without finding a block, a bad-luck streak, then suddenly finds two in quick succession. When those blocks land, the reward window looks backward at the last N shares. Priya's contributions are woven densely through that window. Marcus joined four hours ago, so his shares make up maybe 20% of his expected slice of N.
Priya gets paid close to her full proportional share of both blocks. Marcus gets paid for roughly what he submitted since joining, which is less than his steady-state rate would suggest. Over the next 48 hours, as Marcus's shares build up in the window and the pool finds blocks at a normal rate, his payouts normalize.
The PPLNS model is self-correcting. It also punishes pool-hopping hard, and that is entirely intentional.
Why Pool-Hopping Was a Real Problem
Before PPLNS became standard, some miners ran scripts that would join a pool right after it found a block, when the expected-value math on the next block was highest, mine until the pool's luck started going negative, then jump to a fresher pool. Pure PPS makes hopping pointless since every share pays the same fixed rate. But proportional schemes without the N-window were genuinely vulnerable.
PPLNS closed that hole by making your effective payout rate depend on your historical presence in the window. A hopper who arrives late contributes shares that may fall outside the reward window if a block lands quickly. The longer you stay, the more densely your shares pack into N, and the closer your effective rate approaches your true proportional contribution.
For a solo miner running a single ASIC, this is mostly academic. For an operation running 500 machines across multiple pools, it used to matter quite a lot.
Variance: The Number That Should Drive Your Decision
Here's the honest tradeoff that gets glossed over in most pool comparisons.
PPS has near-zero payout variance for the miner. You can model your monthly revenue from your hashrate and the published share price with reasonable confidence, predictable enough to plan around. The pool operator takes all the variance risk, which is why PPS fees run higher.
PPLNS has high short-term variance that converges to the same expected value over time. A PPLNS pool with good luck in a given month will pay out more than PPS would have. A pool running below expected luck will pay out less. The law of large numbers eventually smooths this out, but "eventually" might mean three months, not three days.
Are you running two or three ASICs in a garage? That variance can make budgeting genuinely awkward. A bad-luck week on PPLNS looks like your hardware broke. It didn't. The pool just hasn't found its expected share of blocks yet, the same way a stretch of highway can go weeks without a traffic incident and then have three in a day.
Larger miners, particularly those running hundreds of machines with strong cash flows, often prefer PPLNS precisely because the lower fees compound meaningfully at scale. A 1% fee difference on 10 PH/s of hashrate is not a rounding error.
One Honest Caveat About Fee Comparisons
Pool fee percentages are not directly comparable across PPS and PPLNS without adjusting for variance. A PPS pool charging 3% and a PPLNS pool charging 1% are not straightforwardly 2% apart in cost. The PPS pool is selling you insurance against variance. The PPLNS pool is not.
Whether that insurance is worth 2% depends entirely on your operation's size, cash flow needs, and time horizon. A miner who needs predictable weekly revenue to cover electricity bills on net-30 terms is buying something real from a PPS pool. A miner who can absorb a lumpy month without stress is paying for something they don't need. Those are genuinely different products, and conflating them because the fee line looks smaller on one side is a mistake.
There's also a third model worth knowing exists: FPPS (Full Pay-Per-Share), which folds transaction fees from each block into the PPS calculation rather than keeping them for the pool. It's PPS with a slightly better share price, and it's become common on larger pools as transaction fee revenue has grown more significant relative to the block subsidy.
The model you choose doesn't change the physics of your miner or the network difficulty you're working against. It changes who holds the variance, and for how long. Figure out which side of that trade fits your operation. The choice gets simple after that.