Noob question: why do low-fee transactions ever get confirmed?

[StackGambler]

I've been into cryptocurrencies for several years, but I think it's kind of embarassing that I don't know the answer to this, so I thought I might as well ask.

In Bitcoin Core, there's an option to send transactions that will confirm in 20 minutes, 7 days, 2 weeks, and so on, based on the current sat/byte rate. By the looks of it, there are always transactions that are paying a fee. People say that if you set a low fee, then your transaction will take a long time to confirm. But, let's say that there is consistent volume on the bitcoin network. Wouldn't miners only include transactions with a decent feerate in their blocks, and by that logic, wouldn't the lower fee transactions never get mined? I'm not sure how the estimated time for a confirmation is deduced, because if a miner doesn't want to include a certain transaction in the block today because the fee is too low, then if the transaction volume stays similar, why would a miner want to include the very same transaction in a block 2 weeks from now?

Apologies if the answer is obvious. I'm a bit of a stupid person. Thanks for reading.

[1713515845]

1713515845

[1713515845]

1713515845

[1713515845]

1713515845

[mocacinno]

You're not stupid, it's a legit question...
The fact of the matter is that the volume doesn't always stay the same... When the mempool contains 10000's of transactions with a fee over 100 sat/byte, the odds of broadcasting a 10 sat/byte transaction and seeing it end up in a block BEFORE those higher paying transactions were either pruned from the mempool, double spent or confirmed is very small... But if you wait long enough and keep rebroadcasting your 10 sat/byte transaction, the odds are good that some day there will be a mining node whose mempool only contains transactions in the same feerange as yours (either because all other transactions were mined, or because the node pruned higher paying unconfirmed transactions, or maybe because the node just started out and had an empty mempool when you re-broadcasted your transaction).

There are also some nodes that let you prioritise transactions (either for free, or paying), and sometimes the CPFP mechanism kicks in when you use the unspent output created by the unconfirmed transaction as input for a new transaction with a higher fee.

Segwit can also influence which tx's are included into blocks... A transaction with a smaller fee (in sat/byte) but with a lot of witness data can be more interesting than a non segwit transaction with a bigger fee (in sat/byte)... But it still boils down to the same principle: a miner wants to maximise his income.

Last but not least: there *might* still be miners honouring high priority transactions... Altough i doubt there are a lot of those miners still hanging around.

But, in general, you are correct, as long as a miner can fill his block with transactions with a higher fee (in sat/byte), they won't include transactions with a lower fee...

[StackGambler]

You're not stupid, it's a legit question...
The fact of the matter is that the volume doesn't always stay the same... When the mempool contains 10000's of transactions with a fee over 100 sat/byte, the odds of broadcasting a 10 sat/byte transaction and seeing it end up in a block BEFORE those higher paying transactions were either pruned from the mempool, double spent or confirmed is very small... But if you wait long enough and keep rebroadcasting your 10 sat/byte transaction, the odds are good that some day there will be a mining node whose mempool only contains transactions in the same feerange as yours (either because all other transactions were mined, or because the node pruned higher paying unconfirmed transactions, or maybe because the node just started out and had an empty mempool when you re-broadcasted your transaction).

There are also some nodes that let you prioritise transactions (either for free, or paying), and sometimes the CPFP mechanism kicks in when you use the unspent output created by the unconfirmed transaction as input for a new transaction with a higher fee.

Segwit can also influence which tx's are included into blocks... A transaction with a smaller fee (in sat/byte) but with a lot of witness data can be more interesting than a non segwit transaction with a bigger fee (in sat/byte)... But it still boils down to the same principle: a miner wants to maximise his income.

Last but not least: there *might* still be miners honouring high priority transactions... Altough i doubt there are a lot of those miners still hanging around.

But, in general, you are correct, as long as a miner can fill his block with transactions with a higher fee (in sat/byte), they won't include transactions with a lower fee...

Enlightening answer. Thank you for that.

[buwaytress]

I wonder also, given that I know of quite a few miners (who are practically breaking even or even slightly in loss) doing this as a hobby or even as a voluntary form of supporting the network, if they have a role in helping confirm these ultra low fee txs.

I know that the age of corporate pools make all these volunteers redundant (plus they are probably also mining as part of a pool) but perhaps there will again be a time when miners are adding txs also, as mocca suggests, trying to determine priority. Maybe one day, txs could be identified to be... For example, a charitable transaction.

Am I naive to think altruism still exists among miners? Conversations within this forum community suggests people still genuinely want to help.

[o_e_l_e_o]

By the looks of it, there are always transactions that are paying a fee.

Your assumption here is mistaken.

You can use this website (https://jochen-hoenicke.de/queue/#1,2d) to see how full the mempool currently is. As you can see, over the past several days (or weeks, or months), the mempool pretty regularly drops to near empty, allowing all these very low fee transaction to get mined. This wasn't always the case - for example, if you zoom out by clicking on "all", you'll see a period of around 3 months from December 2017 - February 2018 at the height of the bull run where anything below 40 sats/byte was hardly getting touched. However, even during this time, some lower fee transactions were being mined, for the reasons stated by mocacinno above.

[StackGambler]

By the looks of it, there are always transactions that are paying a fee.

Your assumption here is mistaken.

You can use this website (https://jochen-hoenicke.de/queue/#1,2d) to see how full the mempool currently is. As you can see, over the past several days (or weeks, or months), the mempool pretty regularly drops to near empty, allowing all these very low fee transaction to get mined. This wasn't always the case - for example, if you zoom out by clicking on "all", you'll see a period of around 3 months from December 2017 - February 2018 at the height of the bull run where anything below 40 sats/byte was hardly getting touched. However, even during this time, some lower fee transactions were being mined, for the reasons stated by mocaccino above.

Oh, interesting. Thank you. So, let's say that interest in bitcoin piques again... is it safe to assume that in that case (like in December 2017), only transactions with a medium to high fee will be confirmed? And in that time range, low fee transactions will be in limbo forever or returned to the sender? I know it's a hypothetical, but if bitcoin is to become as big as we all think it'll become, then it's an eventuality.

[CryptopreneurBrainboss]

Am I naive to think altruism still exists among miners? Conversations within this forum community suggests people still genuinely want to help.

No you aren't, it still happens. We still have miners genuinely willing to help. Early last year during the time of traffic in the bitcoin blockchain neti I mistakenly sent a transaction with very low fee and started a thread asking for help from forum members and someone recommended this thread FREE BTC Tx Accelerators. Stop paying for something that is free! I used either of 2-3 of those services listed and my transaction got confirmation in no time.

[o_e_l_e_o]

Oh, interesting. Thank you. So, let's say that interest in bitcoin piques again... is it safe to assume that in that case (like in December 2017), only transactions with a medium to high fee will be confirmed? And in that time range, low fee transactions will be in limbo forever or returned to the sender? I know it's a hypothetical, but if bitcoin is to become as big as we all think it'll become, then it's an eventuality.

Theoretically, yes. If the mempool became that full again, then low fee transactions would be unlikely to confirm. However, there are several ongoing changes to bitcoin which are intended to address this problem.

SegWit (Segregated Witness) transactions currently make up about 45% of all transactions. As the name suggests, SegWit takes the witness data of each transaction, and moves it to the end where it takes up less weight units, effectively making the transaction smaller and therefore allowing more transactions to fit in each block. This helps to mitigate the mempool filling up, as more transactions can be mined in the same amount of time. Also, as the transactions are smaller, it also means the transactions are cheaper.

Lightning Network has the potential to completely solve this issue. LN is what is known as a second layer protocol, which functions on top of the blockchain. Very basically, it involves opening channels with nodes (which involves a transaction on the blockchain), but then once the channel is open, you can make as many near-instant transactions as you like without any further transactions (until you close the channel) appearing on the blockchain. LN is still in fairly early stages - currently there are around 4000 active LN nodes with a capacity of around 1000 BTC - but these numbers are on a fairly steady upward trend.

[ABCbits]

However, there are several ongoing changes to bitcoin which are intended to address this problem.
--snip--

Don't forget there are 2 more changes :

Schnorr MuSig which reduce transaction size by aggregate multiple signature into one signature. It also improves user privacy since people can't see who are and how many the signer.

MAST (Merklized Abstract Syntax Tree). Basically it's improvement where you only submit part of script which is needed when you  want to spend your coins. It also improve user privacy since not all condition to spend the coins is revealed.

Lightning Network has the potential to completely solve this issue.

I disagree, LN isn't suitable for big transaction since it's less secure than on-chain transaction and most LN nodes/channel don't store big amount of Bitcoin.

[o_e_l_e_o]

I disagree, LN isn't suitable for big transaction since it's less secure than on-chain transaction and most LN nodes/channel don't store big amount of Bitcoin.

True, but the median USD value of bitcoin transactions fluctuates around the $50-$100 mark. If we are assuming that LN reaches a point where all these small transactions are done via LN, especially considering that LN is particularly suited to paying for relatively inexpensive things in person (such as a meal, groceries, general shopping, etc) considering how fast it is, then that leaves a much smaller number of transactions which have to be made via the blockchain. Hopefully we could reach a point where the number of transaction on the blockchain is never enough to overwhelm the mempool.

[joniboini]

No you aren't, it still happens. We still have miners genuinely willing to help. Early last year during the time of traffic in the bitcoin blockchain neti I mistakenly sent a transaction with very low fee and started a thread asking for help from forum members and someone recommended this thread FREE BTC Tx Accelerators. Stop paying for something that is free! I used either of 2-3 of those services listed and my transaction got confirmation in no time.

I think those are accelerator services, or at least rebroadcast services which sometimes can help your transaction to get confirmed faster according to some users. I don't think it implies miners taking zero fees or something similar, but those who want to help accelerate broadcast. Sure, some 'miners' might have this service such as ViaBTC but I've never found those who accept very low fees transaction on their services.

CMIIW.