If we cannot trust that the majority of the mining power will do what is best for Bitcoin then Bitcoin has already fundamentally failed which I do not think is presently the case. In order to orphan all non-abiding blocks a single entity would need to be able exert control over more then fifty one percent of the mining power, if this happens Bitcoin has already been undermined anyway. I do not think it will be possible to enforce such policy across every jurisdiction in the world, this is however more of a question for geopolitics.
My point still stands: larger blocks increase mining centralization pressures. Truth is, even with a single miner you can trust that it will do what's best for Bitcoin. But it doesn't have to. It's all based on incentives. Mining centralization is hard to prevent, as the past tells us, and even harder to undo. You likely won't notice until the bad happens.
I agree with you that mining centralization is difficult to prevent, because of centralization of manufacture and economies of scale among other reasons, to a certain extend I have accepted this reality however. I still do not think that increasing the blocksize would increase mining centralization since miners do not run full nodes.
In my position for instance I just point my hashing power towards a pool of my choice that I think reflects my beliefs well and acts responsible, increasing the block size does not effect my mining operation whatsoever and the majority of the hashing power is under the control of people that are in a similar position to myself.
Increasing the blocksize does introduce centralization pressures but not in regards to mining. Keeping the blocksize at one megabyte also introduces centralization in the form of an increased reliance on third parties. So for me considering this balancing act, increasing the blocksize is the most decentralized option with everything considered.
You don't say anything new; in fact, you're reiterating your claims as if I never challenged them.
I'll try again, the last time: there are network topology issues which result in an uneven block propagation. Uneven propagation alters the orphan race outcomes for different players (mostly pools), making some of them more profitable. The larger blocks get, the larger is the financial difference. This is what alters the incentive structure.
This is exactly my point, it effects the pools not the miners. This is a very important distinction, since if miners are not effected it does not increase mining centralization, it could effect pool centralization if increased to much however for pools it is trivial to setup a full node inside of a data center, which is why pools can support much larger blocks without increasing centralization. It is important to point out the distinction between mining centralization and pool centralization, what you are discussing effects pool centralization, therefore increasing the blocksize does not increase mining centralization.
You are claiming that you can vote with your feet, and I hate to repeat that a lot of hashing power is industrial scale, and is attached to a particular pool. Whether there is a million 1 GH/s miners going back and forth is irrelevant, their combined hashpower is what matters. I don't have any statistical data, but I suspect the majority of hashpower is already industrial-scale and is rigid w/r/t to pools.
I do not think that hashpower is rigid with pools, after all over seventy percent of hashpower is currently concentrated in public pools:
https://blockchain.info/pools?timespan=4daysThe majority of hashpower is already industrial scale however that does not mean that this hashpower is exclusive to particular pools, it is after all very easy for a miner to switch between pools. If we had one hundred industrial miners responsible for the majority of the hashpower spread over 20 pools this would still be very decentralized, there are small miners or "home" miners like myself but I suspect that the hashpower that these types of miners represent would not be much more then one of these larger industrial mines. This is the reality today, it would be better if there where more small miners like myself.
I can tell you with absolute confidence that an increase in the blocksize would not effect my ability to compete with the larger industrial mines. It is very simple really, as a miner I do not run a full node, therefore increasing the difficulty of running a full node does not effect my mining operation.
In response to the incident of Antpool SPV mining and causing a fork a few months back, since then Antpool has lost a lot of its mining power which proves my point exactly, such a pool is incentivized to attract more miners by acting responsibly and when they act irresponsibly they lose support, proving the points that I have been making. I very much doubt that Antpool would make such a mistake again since they lost a lot of money because of the reward and reputation that was lost, if miners understand one thing it is profit.
I believe this is a common logical fallacy known as
post hoc ergo propter hoc. Moreover, I couldn't find anything that can support your claim. According to
organofcorti, there were no significant plunges in AntPool hashrate during June-August, and it has risen over that two-fold in the period. And has risen since then as well.
Fair enough on Post hoc ergo propter hoc, I can not prove with absolute certainty that the decrease in the hashrate of Antpool was because of this incident. However since I tend to keep an eye on pool distribution I do distinctly remember Antpool having over thirty percent of the hashpower before the time of this incident and now they only have sixteen percent. One aspect of that has been the recent increase in the overall hashpower since I am looking at the percentage not the hashrate. I do share your concerns about mining centralization, however I do not think that increasing the blocksize would effect mining centralization for the reasons I have explained here. That is a good website by the way, thank you for linking it, there is some very good analysis on there.
