Insert again the option of doing transaction less than 1 sat/byte · Issue #12536

[Ayanamirs]

https://github.com/bitcoin/bitcoin/issues/12536

[ranochigo]

It's probably not going to be done. Other than the obvious minimum relay fee rule, the fee is designed to make spamming transactions more expensive.

With the current volume of transactions, each transaction requires at least 10 satoshi/byte and that is already 10x of the fee that you've suggested. By lowering the fees, this just gives spammers an easier way to spam up the mempool and doesn't benefit anyone else in any way; miners aren't going to accept transactions with fees that low.

Damn. I'm on different frequencies.

[LoyceV]

It's probably not going to be done. Other than the obvious minimum relay fee rule, the fee is designed to make spamming transactions more expensive.

Spamming doesn't rely on the minimum fee set in Bitcoin Core. I just ran into this limitation yesterday, I couldn't set less than 1 sat/byte, while blocks weren't even full.

With the current volume of transactions, each transaction requires at least 10 satoshi/byte and that is already 10x of the fee that you've suggested.

Both https://btc.com/stats/unconfirmed-tx and https://coinb.in/#fees recommend 1 sat/byte fee for a fast transaction. Just a few minutes ago, this transaction with 0.774 sat/B fee confirmed in 12 minutes.

By lowering the fees, this just gives spammers an easier way to spam up the mempool and doesn't benefit anyone else in any way; miners aren't going to accept transactions with fees that low.

The moment spam happens, fees go up. There's no need to artificially raise fees, let's leave that up to the market.

[Thirdspace]

I just ran into this limitation yesterday, I couldn't set less than 1 sat/byte, while blocks weren't even full.

With the current volume of transactions, each transaction requires at least 10 satoshi/byte and that is already 10x of the fee that you've suggested.

Both https://btc.com/stats/unconfirmed-tx and https://coinb.in/#fees recommend 1 sat/byte fee for a fast transaction. Just a few minutes ago, this transaction with 0.774 sat/B fee confirmed in 12 minutes.

[tx1] https://btc.com/17af9a5502f4778dfd2ce6b319518848d52f82b50904ee62c140172b653208a7
if you read the tx data closely, the result is from 175 sats / 226 Bytes = 0.774 s/B (bc.info: based on Size)
but it seems that the tx was accepted because of 175 sats / 144 Bytes = 1.215 s/VB (btc.com: based on Virtual Size)

same thing with [tx2] https://btc.com/748a9298545c0e120b99abafcdb0ac149dba4c637000926ab6493cabcdf05af0
282 sats / 424 Bytes = 0.665 s/B (bc.info: based on Size)
282 sats / 261 Bytes = 1.08 s/VB (btc.com: based on Virtual Size)

so essentially I think both txs are treated as having higher than 1 s/VB tx fee
while bc.info is displaying them (inaccurately) having less than 1 s/B

[NeuroticFish]

Even if you want to pay less than 1 sat/byte, you may have to pay the relay fee of 500 sat, or was it removed? My memories are somehow mixed up.
All in all I have the feeling that removing the limitation won't be of much help.
And let's be fair, the miners have to buy an extra coffee too, don't they?

[achow101]

so essentially I think both txs are treated as having higher than 1 s/VB tx fee
while bc.info is displaying them (inaccurately) having less than 1 s/B

That's correct. Bitcoin Core does things in sat/vByte, not sat/byte (although things are still named the old way).