I used to think PoW was far superior to PoS, but I'm currently rethinking this belief, at least regarding minority chains such as Bitcoin Cash.
Marc De Mesel suggested in a recent video of his to consider forking BCH to PoS and this is what finally triggered me to think about this option more seriously. He may be right in saying "maybe it's a little bit too much", but I'm certainly in favor of considering the option seriously.
I need help in thinking this through and that's why I'm writing this article. It will be biased towards doing a fork towards Proof of Stake in BCH and that's not because I strongly support it at this point. It's because it hopefully helps me think through the option and maybe it inspires discussion more readily than some unbiased blablabla would.
"good guardians": Miners are incentivized to take good care of the chain. This assumption is based on there being nothing else that could be done with mining hardware (or rather a mining operation in more general terms) that would be more valuable than mining Bitcoin. This was true in GPU/FPGA days and that assumption got even stronger when sha256 ASICS entered the market.
"no barrier to entry": everybody can mine, no permission necessary. There is nothing keeping anyone from installing a publicly available opensource mining node software onto some cumputer and connect it to the Bitcoin network in order to participate in mining Bitcoin. You don't need to register an account with any central company or institution. You don't need to be human, you don't need to be from earth, you just need computing hardware and a network connection. You simply don't need permission.
Note that it seems Satoshi was making both of these assumptions in the white paper and I used to make these assumptions, too. In fact over time they turned into beliefs I didn't question much any more. "I've thought it through and that's whats true. Period"
"good guardians": I don't believe this is true for minority chains. Miners can always switch their equipment to another chain, so there always is something equally valuable to use the operation for in case something bad happens to the coin in question.
"no barrier to entry": From the wording I used ("no barrier" instead of "permissionless participiation") it already becomes clear that this is not the case. Not everybody can mine. Everybody might have "permission", yes, but that doesn't mean it's feasable: successfully mining SHA256 in an economically sustainable way requires access to cheap enough electricity and access to efficient enough mining equipment. It also requires the operation to have a certain scale and thus require a certain amount of capital investment. If these conditions aren't met, you're not going to be competitive and soon enough you're not going to be a miner, because you're simply going to be priced out. Your operation will be too inefficient either due to too large operational overhead cost (scale too low), due to too large operational cost (electricity cost too high), or due to too large capital cost (mining hardware to inefficent in terms of hash per dollar).
When the "no barrier to entry"-assumption crumbled (I noticed first-hand, because I had to quit mining in the "fall" of 2011. First by mining moving to places with cheap electricity and then by the introduction of ASICS, I brushed it off and told myself: "it's not so bad... these guys will be good guardians of the coin, we can trust them to that extent because they are incentivized in that way." So basically I fell back to the "good guardians" assumption.
When the "good guardians" assumption crumbled, I didn't notice. It was post-BCH split and I was all entangled in hoping BCH would flip BTC soon so I really didn't see it clearly since I believed ("hoped", rather) the minority chain situation would be temporary. I told myself: Jihan has a lot of BCH and he loves BCH and so do at least some of the other BCH miners. "Look: they're throwing hash at a huge cost to get the coin started, so obviously they're believers, holders and good guardians of BCH." Well, this might be true coincidentally, but it's not due to any built-in incentives for miners. Rather, it's because they happen to also be holders and they happen to also be believers. The emphasis here is on "happen to" and "coincidental". This means we cannot be sure this remains to be the case... it's not "built-in".
Let's assume the following happens (which I don't think is unlikely): IFP activates through BIP-9 voting and is actually supported by majority hash after activation. 5% of the block reward (probably has to be renamed at that point) flows to Bitcoin ABC (or other projects) for development.
Let's be frank: I want no piece of that coin (at least in the long run) and I'll probably detach emotionally from it (it's already starting). I'm sure I'm not alone. At that point I'll support a split. I don't want a split, but having the IFP pill forced down my throat is just not "an option" for me so I'll "vote" with what I have available to me: my feet.
But how could we possibly fork off? Certainly just making a non-IFP version of the node and getting some miners to use it is going to result in a minority-of-a-minority chain, because if majority of the hash had been against the IFP, it wouldn't have been activated and supported by majority hash. So really the only option is to switch the mining algo.
Considering other PoW algorithms (or multi-PoW or whatever there is) is one thing, but why not go the full mile and consider PoS? The coin is distributed enough already so maybe a "transitional" approach (first PoW, then a mix and then transition to PoS) can be skipped in the case of BCH?
Proof of Stake, at least on naive inspection (which is what I'm doing here), brings the two assumptions back to the table:
"no barrier to entry": clearly the barrier is much lower than with PoW: no need for cheap electricity or hard-to-get specialized hardware: A simple VPS will do.
"good guardians": if implemented right, those that hold coins are the ones who can mint blocks and those that hold coins are naturally incentivized to be good guardians because they are holders and hurting the coin value will hurt them. To make this true, there would probably have to be a mechanism that ensures those minters don't turn around and sell their coins right away after minting. They may have to be "locked for staking" for a period of time. Also to avoid people just buying up a bunch of coins right before some important voting takes place it might be good to require a certain coin age for staking. But these are all points I'd like to discuss because frankly: I haven't thought this through.
The problems with PoS
well, see the comment section, I hope.
love you guys, molec