Copying this from another thread, since it continues some discussion that started over here:
If you wanted to upend Core, then you should have more competent people who would have advised you that unbounded block size doesn't have an equilibrium.
If you have successfully demonstrated that unbounded block size cannot reach an equilibrium
outside a majority-collusion environment, I have missed it.
Yes you missed it. ...
Once we do model differing orphan rates for different miners, then the optimal strategies for mining come into play. And if you work out the game theory of that, you realize that
collusion and centralization are the only possible outcome.
So you seem to be acknowledging that I am correct above...
There is no outcome that is
outside a majority-collusion environment. I explained that unbounded block size cannot reach an equilibrium
outside a majority-collusion environment. You said you must have missed it and I explained you did miss it.
You are conflating that the fact that the equilibrium is reached inside majority-collusion with your thought that I haven't stated what occurs outside of majority-collusion. But I have stated what happens, which it is always devolves to majority-collusion.
You are making a similar error as those two others did upthread. A 51% (or even 33% selfish mining) attack is not a change in protocol. In other words, in BTC the miners can't make huge blocks, because it violates the protocol limit of 1MB.
Collusion is collusion, irrespective of the protocol. Nakamoto consensus is only possible when a majority of participants are 'honest' as per the whitepaper terminology. Unbounded blocks does nothing to change this.
Conflation is conflation. (Meaning you apparently entirely missed the relevance of the point)
I'm trying to be respectful, but please don't waste my time. You see I have too many messages to reply to.
And as a practical matter, Bitcoin operated just fine for multiple halvings with no practical bound on blocksize.
There was minimum advised fee and there were pools doing anti-spam such as I think I've read that Luke Jr's pool rejected dust transactions.
Yes, minimum
advised fee. 'Advised', as not encoded within the protocol. The fact that this worked up to the point that the production quota was finally persistently hit forms an existence proof that the system
can work. The fact that it
did work may or may not have something to do with all players having beneficial intent, but there it is. Indeed a populist sentiment includes the notion that it is against the best interests of all participants to do anything that kills the system. Which probably explains why our past known-majority miner (Discus Fish?) turned back from their position of mining majority without ever forming an attack from their assuredly-successful posture.
Everyone was incentivized by the fact that once the 1MB limit was reached, then the destruction of Bitcoin would ensue as is currently happening with the battle between the miner and codester cartels.
It was the 1MB protocol limit that provided the barrier that everyone had to try to swim far from. Also miners had an incentive to get minimum level of fees and they didn't yet have enough centralization to extract higher fees. And the decentralized miners at that time before ASICs also had an incentive to keep spam low since as I explained to @dinofelis today that it wasn't a fully connected mesh so propagation time was a bigger deal than he realized. Also the decentralization at that time when people were still mining on GPUs, meant there was more of an altruistically driven Nash equilibrium than now.
That not at all like the cut throat, big money economics situation now. As @dinofelis pointed out, the only altruism (and internal discord) from miners now is probably all faked to make us think there isn't a cartel.