On 8 August 2015 at 00:57, Gavin Andresen via bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org> wrote:
I answered: 
> 1. If you are willing to wait an infinite amount of time, I think the
> minimum fee will always be zero or very close to zero, so I think it's a
> silly question.
That's not what I'm thinking. It is just an observation based on the fact that blocks are found at random intervals. 
Every once in a while the network will get lucky and we'll find six blocks in ten minutes. If you are deciding what transaction fee to put on your transaction, and you're willing to wait until that six-blocks-in-ten-minutes once-a-week event, submit your transaction with a low fee. 
All the higher-fee transactions waiting to be confirmed will get confirmed in the first five blocks and, if miners don't have any floor on the fee they'll accept (they will, but lets pretend they won't) then your very-low-fee transaction will get confirmed.

​That depends a bit on how ra​tional miners are, doesn't it? Once the block subsidy is retired, hashpower is only paid for by fees -- and if there's no fee paying transactions in the queue, then there's no reward for applying hashpower, so mining a block won't even pay for your costs. In that case, better to switch to hatching something else (an altcoin with less fees than bitcoin has on average but more than nothing, eg), or put your hashing hardward into a low power mode so you at least cut costs.

That will only be needed for a short while though -- presumably enough transactions will come in in the next five or ten minutes for a block to be worth mining again, so maybe implementing that decision process is more costly than the money you'd save.

​(C​
onversely, when the queue is over-full because there's been no blocks found for a while, that should mean you can fill a block with higher-than-average fee transactions, so I'd expect some miners to switch hashpower from altcoins and sidechains to catch the temporary chance of higher revenue blocks. 
​Both tendencies would help reduce the variance in block time, compared to a steady hashrate, which would probably be a good thing for the network as a whole)​


I think the same incentives apply with mining being paid for by assurance contracts rather than directly by transaction fees -- if you get a bunch of blocks done quickly, the existing assurance contracts are dealt with just as well as if it had taken longer; so you want to wait until new ones come in rather than spend your hashpower for no return.

​All of this only applies once fees make up a significant portion of the payment for mining a block, though.​

Cheers,
aj

--
Anthony Towns <aj@erisian.com.au>