Quote from: smooth on January 06, 2015, 09:26:39 PM

Quote from: JorgeStolfi on January 06, 2015, 09:16:00 PM

The size of the user base and the price are not important for the experiment. But the distribution of mining power is.

Only if: 1. The mining power is abused. That hasn't happened yet in any significant way, so Peter R's argument applies equally. Every month that passes without abuse increases the chances for continued success, albeit slightly.

I understand "success" as "the bitcoin protocol provides a P2P payment system that does not require trust in a third-party authority".

We now know that the bitcon protocol does not provide that. The protocol requires users to trust the miners; which is OK, as long as mining power is sufficiently well distributed over thousands of independent entities, benign or short-term greedy. However, the protocol cannot prevent concentration of mining, and in fact forsters it. Once mining is concentrated in a small number of companies, having to trust them is no longer OK.

It does not matter whether those companies ever did something bad, the problem is that users must trust that they will not do it. That is supposed to be the bad feature of banks and credit cards that bitcoin was designed to avoid. (Personally, I would more easily trust my money to a a bank cartel than to those 4 top mining companies.)

You seem to be saying that FooBank's saferoom is secure because in the past 2 years no one has tried to force it open; and gets more secure at each month that passes without anyone trying to force it. Sorry...

You seem to define "success" as "the current bitcoin network will keep working and people will keep using it".

Bitcoin may still be successful in that sense; but what would be the point, if it is no longer "trustless"?

Quote

2. The distribution of mining power is stable. The real world evidence suggests the contrary. Yet you ignore it.

Quote

those 4 companies (which may actually be only 3, or 1, or may become 1 next year)

Or may become 10. Funny how you didn't include that possibility in your rhetoric. I wonder why.

The distribution of power does not have to be stable for Satoshi's goal to be frustrated.

The protocol asks us to trust that, at any time during the life of the network, no subset of miners that comprises 51% of the total power will agree to do something that violates the protocol, and will maintain that agreement long enough to make the damage permanent.

Again, one can take for granted that the members of that susbet will not agree to do that, if the subset has thousands of independent miners. But if the subset has only a few members, one cannot make that assumption. There is significant probability that they will agree and maintain their agreement for the time needed, especially if the violation of the protocol is profitable for that subset. The smaller that subset, the more liikely it is that they will cooperate in such "attack". But the, it is highly probable that such thing will happen at some point during the life of the network.

We had already a situation when GHash.io controlled by itself more than 51% of the network's total hashpower, IIRC for several days. During that time, everybody had to trust that they would not use their power to violate the protocol; in particular, that they would not decide to change the protocol by threatening users with DoS.

Right now, users have to trust that Discus, GHash, AntPool and BTCChina will not join forces to violate the protocol. They have to trust that Discus, GHash, AntPool and KnCMiner will not do it either. And that Slush, pool34, BTCMiner, KnCMiner, BTCChina, GHash, and Discus will not do it either. And ...

You see, if power was distributed over many thousands of independent miners, all such lists would have thousads names, and users would not have to worry about the chance of all the members in one list agreing to play foul. But right now there are two sets of 4 companies that would have the power to do that, and several sets of 8 companies or less. The chances that all members of one of these list will form a cartel and violate the protocol is not longer negligible. Users will have to trust that none of those potential cartels will exploit that opportunity.

The risk of protocol violation by a 51% cartel is even greater if two or more of those big companies are owned by the same people. It is quite possible that GHash still controls more than 51%, but made an arrangement with other pools for them to nominally "host" part of GHash's power.

Even if the pie chart changes completely by next month, it is a safe bet that power will remain quite concentrated (as it has been for many months). Again, the economics and nature of PoW mining fosters concentration, overt or disguised.