The real cost of switching mining pools: what you lose in the move

TL;DR

Switching pools costs you in three places: downtime while you reconfigure, an income shortfall while the PPLNS window fills at the new pool, and the wait for a payout threshold you have just reset. At home scale, 100 TH/s, an hour of downtime runs about 200 satoshi. Entering a PPLNS pool costs roughly half a window's worth of income; for Ocean, whose window is eight network difficulty blocks, that lands in the range of a few days of income, with the exact figure set by the pool's share of network hashrate. Moving from a 4% fee to a 1% fee saves 9 cents a day on the same 100 TH/s, so a fee driven move takes months to pay back and sometimes never does. At petahash scale the arithmetic flips.

Our pool by farm size guide promised this calculation. Here it is.

Where the numbers come from

Every figure below traces back to the network snapshot of 2026-08-14 used in our August 2026 pool comparison: network hashrate 933.99 EH/s, difficulty 127,479,855,693,691, block subsidy 3.125 BTC, BTC at 64,558 USD. At 100 TH/s after a 1% fee that comparison landed on 0.00004803 BTC a day, about $3.10. In 4% pools the same 100 TH/s returns $3.01.

The PPLNS mechanics and Ocean's window size come from our payout schemes breakdown: Ocean documents a window of eight network difficulty blocks. Other large pools do not publish their N, so for those we give a range and say so plainly.

Payout thresholds come from the same August comparison table. Nothing beyond those sources enters the math.

What does an hour of downtime cost during a pool switch?

An hour of full downtime at 100 TH/s costs roughly 200 satoshi, about 13 cents at the snapshot price. The math is one line: daily income of 0.00004803 BTC equals 4,803 satoshi, divided by 24. Fifteen minutes costs around 50 satoshi. That is an upper bound, since part of the setup happens while the miner keeps hashing.

The 15 to 60 minute range is my assumption, not a measurement. Fifteen minutes covers changing pool details in one ASIC's web UI with the login and payout address already prepared. An hour covers several machines, a fresh account, an email confirmation, and one worker authorization error to untangle. With twenty machines on mixed firmware, plug in your own number: income scales linearly, so an hour of downtime across 1 PH/s runs about 2,000 satoshi.

One cost I am not pricing here: a move that trips an ISP stratum block pushes downtime well past the top of that range. Diagnosing and working around it is covered in our piece on a pool blocked by your ISP or router.

What happens to your money when you leave a PPLNS pool?

Nothing burns. Your shares stay in the window and keep earning until they age past N. But the money arrives late, and only when the pool finds a block. The genuine loss shows up in two situations: the tail leaves you stranded below the payout threshold, or you closed the account.

The symmetry matters. On the way in you earn below expectation while the window fills with your shares, because the first block is split over a history that barely contains you. On the way out you collect the mirror image of that. Add both ends of one complete stay and they cancel. The switching cost appears when the cycle breaks: you move, skip the tail, never reach the threshold, and leave a balance parked at a pool you no longer use.

A lower bound on the entry shortfall, assuming your weight in the window grows linearly:

```

shortfall ≈ 0.5 × T_window × daily income

```

The linear assumption is deliberately crude. The real curve depends on how the pool weights shares and how steady your hashrate was.

Converting a PPLNS window into hours

Ocean states its window in difficulty blocks, which converts to time in one step. The network produces a block every 600 seconds on average, so work equal to one difficulty block accumulates at a pool in 600 seconds divided by its share of network hashrate:

```

T_window = 8 × 600 s / s, where s is the pool's share of network hashrate

```

Which gives this table without inventing a single number:

Income is taken at $3.10 a day per 100 TH/s. I am not printing Ocean's own share here: pool shares move week to week, and any number written into an article ages faster than the article does. Current shares for any pool are easy to read off miningpoolstats.stream, and from there you put your own s into the formula and take the matching row.

For AntPool, F2Pool or ViaBTC in PPLNS mode, N is not published, so the honest estimate is that same span: hours at a large pool with a short window, days at a pool with a long one. Without a published N nobody can narrow it further, and no calculator closes that gap.

One boundary condition worth stating: all of this applies to PPLNS, TIDES and relatives only. Under FPPS and PPS+ the bulk of your income accrues per accepted share whether or not the pool found a block, so there is no entry shortfall and no exit tail. Moving between two FPPS pools costs exactly zero on this line.

How long until the first payout at the new pool?

Anywhere from two days to seven months, depending on the threshold. At 100 TH/s a 0.001 BTC threshold fills in roughly 21 days, 0.005 BTC in roughly 107 days, and Foundry USA's 0.01 BTC in roughly 215 days. The 0.0001 BTC threshold at EMCD or Promminer clears in about two days. The money is yours throughout, it just sits in the pool balance.

That is lost access rather than lost income, but it is the line that bites hardest in practice. Someone leaves a pool with a 0.0001 BTC threshold for one at 0.005 BTC to chase a fee percentage, then watches three months pass with nothing hitting the wallet, while a leftover balance sits at the old pool.

Table: what you lose and how much

All amounts at home scale, 100 TH/s, $3.10 a day.

Against all that sits the saving. Dropping from a 4% fee to 1% at 100 TH/s is 9 cents a day, about $33 a year. If the move costs $8.7 in shortfall plus 13 cents of downtime, payback lands around 98 days. At 1 PH/s the same saving is $0.94 a day and every line above clears in a week or so. That is why your farm profile decides more than the fee table does.

A checklist for moving without losses

1. Check what you are moving for. If it is a fee percentage at home scale, price the saving in the profitability calculator and compare it against the table above. The answer is often "not worth it".
2. Identify the old pool's scheme. FPPS or PPS+ means no window and no tail, so the move costs downtime and threshold time only. PPLNS or TIDES means steps 3 and 4 apply.
3. Under PPLNS, leave late in the cycle, once most of your contribution has already paid out through found blocks. Nobody can name the exact date; use the window length from the table, measured from the pool's last block.
4. Do not delete the old account or change its payout address until the final payment lands. The window tail arrives after you have already unplugged.
5. Check whether you can still reach the threshold at the old pool. Walking away halfway to a 0.005 BTC threshold parks that balance indefinitely. Mining to the threshold is sometimes cheaper than saving 9 cents a day.
6. Create the new account and set the payout address before the miner goes down. Email confirmation and 2FA are done on a running farm.
7. Use all three stratum slots in firmware. New pool first, old pool second, a backup address for the new pool third. A failed worker authorization then does not leave the farm idle.
8. Switch one machine, not the whole farm. The first unit proves the login, port and connection stability; the other nineteen follow with no surprises.
9. Reconcile the credits after 24 hours. Expected daily income follows from your hashrate, your share of the network and the new pool's fee, with the formulas laid out in our calculation methodology.
10. Do not move back a week later. Every PPLNS entry is paid for again, and over short spans luck variance swamps any real difference between pools.

When the move is clearly worth it

Some cases skip the arithmetic entirely. The pool stopped paying or delays payouts. The pool does not disclose its fee, so there is nothing to reconcile credits against. The stratum drops often enough that you lose hours instead of minutes. The payout scheme does not suit your hashrate, PPLNS on a single home ASIC being the classic example, where ordinary variance reads like a hardware fault. In each of those the cost of moving is smaller than the cost of staying.

The reverse case is just as clear. A 4% fee against 1% at 100 TH/s is 9 cents a day, measured against every line in the table. If the pool pays reliably, those nine cents are not a reason to touch anything.

FAQ

How many satoshi do I lose per hour of downtime when switching pools?

About 200 satoshi an hour at 100 TH/s, roughly 13 cents at the snapshot price: 4,803 satoshi a day divided by 24. At 1 PH/s it is about 2,000 satoshi an hour. A single machine switched with the account already prepared usually costs closer to 15 minutes, so about 50 satoshi.

Do my PPLNS shares burn if I leave the pool?

No. They stay in the window and count toward the pool's next blocks until they age past N. The risk is collection, not destruction: the tail pays out after you have left, and if it does not carry the balance over the payout threshold, the money sits at the old pool indefinitely.

How long does a switch take to pay for itself?

Divide the one-off cost by the daily saving. At 100 TH/s the gap between 4% and 1% is 9 cents a day, so an $8.7 move pays back in roughly 98 days. At 1 PH/s the saving is $0.94 a day and the same move clears in about ten days.

How do I leave a PPLNS pool without losing anything?

Wait until the pool finds a block with your contribution fully weighted, then stop hashing but leave the account and payout address alone. The tail keeps crediting for roughly one more window length. If the balance still sits below the threshold after that, the only ways back to it are mining more or waiting for the pool to change its payout rules.