As for the insiders killing it, it is irrelevant. What is relevant is the cost to attack. Because an outsider can just buy enough stake at X cost and become an insider. As for Proof of Work, I believe it is more costly.
Of course not. On the contrary. if you want to do a 51% attack on consensus decisions, you need to own 51% of the market cap. That's MUCH MUCH more of a cost (in bitcoin, it would amount to having to buy up 20 billion $ of coins, but your cornering of the market would make it still much, much more expensive).... in order to destroy the system for which you just paid 20 billion ! No miner is ever going to invest 20 billion in mining equipment, because he will not mine 50% of the stake in block rewards and fees !
Also, the cost of a 51% attack in PoW is just the cost of 51% of the mining rewards, and you don't need to be stake holder. From the outside, you simply cannot attack a PoS system, because it is based upon much harder to fake digital signatures.
Currently, ALL of the proof of work ever delivered to bitcoin has a security level of 90 bits. A simple 256 bit key signature has a bit security level of 128 bits, and only costs a few milliwatts of power and an old laptop (or even a mobile phone).
Cryptographically, PoW is pure BS as a protection. Any digital signature beats it with tens of orders of magnitude in "efficiency" (that is, spent resources versus security obtained).
Your arguments are compelling. I do not possess the Cryptography expertise to dismount any of the technical arguments.
I will say this:
51% attack in PoW does not only cost 51% of mining rewards. It is not immediate, and it costs all the energy and capital that was put onto it before any ROI.
Furthermore if you consider it is possible to hold a 51% hashpower, if we estimate like a 40% ROI for mining (and I think this is too much considering mining alone), according to current market cap it should cost total about 13.007.185.339,7 USD to achieve a 51% Hashrate. the actual loss of revenue from an attack today would be 2.410.118.399,27 USD Bringing the total to about 15.5 B.
For a 51% PoS, the cost would be 21.678.642.232,83This is where your reasoning ultimately fails
Basically, you roughly divide the current market cap of Bitcoin by 2 and get the cost at today's price (since this is what the market cap shows), implicitly assuming that the price will remain the same when you actually start buying 51% of all Bitcoin monetary supply. Nothing could be more false and farther from reality than that. First, not all coins are being traded, I guess, it is somewhere in the range of a few millions (maybe, 3-4 at most), and when you have bought your first million, you will have to spend like 10 times more money to buy the next million of coins (due to prices flying to the moon). Further, you chose to completely ignore my argument that you may never get there at all, no matter how much money you could have since some stake holders may not be going to sell their stakes at any price
First of all, you mentioned a PoS attack from inside.
This would cost exactly 21.678.642.232,83. A PoS attack from the outside, would cost alot more obviously and might be arguably impossible.
For an outside POW You would need to double the existing Hashrate to acquire 51% of Market Cap.
This would cost 1.793.266,75 USD a day to run based on maintenance rate of Hashing24 or Genesis Mining. The cost to setup, let's consider Antminers S9 selling at 2500 USD on Amazon a piece. Since we would need 5.434.141.693 GH/s, this would mean approximately 388.152 Antminers S9, 970.380.000 USD.
Ignoring Logistics and setup costs this would cost around 1 Billion if we could do this in a day. With nothing to gain. Considering that existing miners already have existing infrastructure, they would also easily increase Hashrate to defend (as the market would for PoS buy-in). An attacker would have not only more cost, but a logistically difficult problem to resolve - immediate supply of Hashing Power. Furthermore, each day it took to get the necessary hashpower would arguably cost around 1M USD considering a 40% ROI.
As opposed, a 51% stake can be acquired virtually hassle free and even "anonimously" using multiple wallets during the course of time.
This considering all participants act in their best interest. Otherwise, a 51% stake always costs market cost. Whereas a 51% hashrate can possibly be bribed for 100 USD depending on the stupidity of the actors involved.
So I grant you are correct in a doomsday scenario where incurring cost to no gain happens. PoS is better security from the outside.
As for the inside all the incentives are there and the Miners can easily defend an outside attack, furthermore, they are logistically invested. It is not easy to divest resources. And their cost to defend is a lot less than that of an outside attacker. Whereas Capital, goes where its most valued pretty easily.
I am still inclined to maintain that I think Proof of Work is economically more sound than Proof of Stake.