So you think it's easy to tell thousands or millions of "average Joes" to send coins to themselves?
The only event I can imagine where massive sending of currency at the same time could occur is a generalized panic on exchanges (sending to other addresses should have the same effect, as the coins would have to "mature" there).

But for 51% to send coins the panic would have to be really massive, and these 51% holders would have to be online in a relatively short timeframe for the "blockchain stop" to occur.

I think it's an extremely difficult attack, even if you consider a short seller that wants to profit from it. And it doesn't fit really in the "Nothing at stake" category. A PoW/PoS hybrid coin, for example, would not be affected.

Another one that is only feasible if there are few holders and distribution is very uneven.

When it comes to uneven distributions, you shouldn't forget that majority of worlds money is in hands of top 1% (https://en.wikipedia.org/wiki/Distribution_of_wealth#Data_about_global_distribution_of_wealth).
It shouldn't be unreasonable to assume that distribution would be quite uneven. Plus we are talking about control of the coins here, not just ownership, which is even more centralized since you have banks and investment funds, where few control a lot of money.

This world is quite old and still the distribution of wealth isn't even at all. This isn't due to any particular event either, the wealthiest man at the moment got rich relatively quite recently, so we shouldn't assume that coins will be more evenly distributed over time, maybe the near future, but in the very long run I wouldn't bet on it. Bitcoin should be designed to be here and stay for as long as it can.

Only small or very badly designed Proof of Stake currencies would have these weaknesses.

Small PoW coins have a similar problem - they are also weak and can be attacked easily via 51%ing them because mining power typically is weak and very unevenly distributed. Even in stronger currencies this problem persists - there is often a pool or miner group that in theory could 51% the currency. I've followed the Segwit adoption in Vertcoin and Litecoin and there were often situations when a single address (pool or miner) mined more than 50% of the blocks of a day.

I don't think so.Aiming at a series of problems, the future should have a solution.

Well the real world doesn't really have PoS, it is actually more like PoW. My point was that distribution probably will always tend to be uneven. Uneven distribution in PoS system adds a new vector of fewer parties controlling the system. That new vector is time of adoption of that stake, it this case adoption of cryptocurrency. If distributions always tend to be uneven then PoW and PoS will have that problem over time for whatever the reason that is occurring (might be natural selection at play or something else), the problem of centralization. However PoS will have it occurring for one more reason than PoW and that is the early adoption, like I said in my first post quoted below.


I don't think Bitcoin is ideal crypto utopia, there are plenty of good advancements made in many altcoins, Bitcoin just takes a slow,stable and secure approach in it's development compared to other cryptos. Centralization of mining power is a very real problem in Bitcoin and that might be inevitable, however we should fight for the chance that it isn't. One of the biggest issues there is mining pools and some cryptos saw that and did their best to stop it happening it their own coin, using smaller block time target of couple of seconds and making it more accessible for more people to start mining, like trying to stop ASIC mining with new hashing algorithms. A PoS coin has no way to battle the initial distribution, it is a more likely a lost cause.

The weaknesses of small PoS and PoW coins are not the same ones. But botnets and cloud mining are two forms an attacker can use to get easily a majority for a short time in a small PoW currency. This kind of attack isn't free, but should be easier to perform than a history/long range attack on PoS currencies. However, I think you are right that a _very_ small PoS currency is probably weaker than a _very_ small PoW currency, if the PoW currency is using an ASIC/GPU-unfriendly algorithm.

As far as I know only some PoS algorithms have the problem to come to a halt after a large "stake rate" drop. Even then, a hard fork or a snapshot-based restart can save the currency. That would need a collective effort of the "validators", but it is comparable with the collective action Bitcoin miners took in March of 2013 to kill the fork that had been produced by BTC 0.8.

By the way: with Bitcoin Cash we saw that a extremely large hashrate drop in a PoW currency also can result in a "semi-permanent halt" - one block per 6 hours for a whole difficulty period, like at the beginning of the BCC existence would have made the currency almost unusable for a long time and it would probably have died. I know that is an extreme situation, but in my opinion we here are debating about gradual differences and not about a fatal flaw.

I get that point. It's a theoretical disadvantage of PoS currencies, but in my opinion not a fatal flaw because currencies without a working market practically are not currencies, and so the option to "buy in" is as easy (in most currencies much easier!) than to mine a PoW currency.

When choosing between both systems, it has to be analyzed if this disadvantage is more important than the generally much higher cost of a PoW cryptocurrency node network because of the higher energy consumption.

In the context of one of these attacks it is very different. How will you buy PoS coins from an exchange when the network isn't producing blocks? In this case, there is no way for the network to recover barring a hard fork, which is very damaging for a currency which is supposed to be decentralised.

You comparison to BCC is actually a very revealing example of how a PoW coin deals with a sudden and catastrophic drop in hash rate. Less blocks, slow recovery. If a PoS coin suddenly lost 95% of the staking power, a hard fork would have been the only way to recover.

I would say that Primecoin might be a better example of a useful secure PoW coin. You get big primes as a result, which is always useful in cryptography. Since prime numbers are used in public/private key cryptography, it would be kind of like it is feeding it self Who knows, maybe one day we figure out how to make a nice switch in Bitcoin's PoW algorithm, that might take some time tho.

Early adopters control the system forever. All the problems of centralization apply. One point of failure and all that.