Yes, that's all taken into account in my simple formula. That's why I admitted that the security would be comparably to 2017 and not 2018/19.

Of course if hashrate dropped now to 50 EH/s, 95% of all miners went bankrupt and sold their equipment potentially to attackers, then the network would be seriously in danger. I don't dispute that. But we're here talking about a scenario for 2140. The process thus will be very gradual. Even if the equation for miners becomes already problematic in 10 years like MeGold666 feared, this would also not cause the hashrate to drop instantly to dangerous levels. We would probably have 10-20 years more where one after another miner would go bankrupt, and the end result would be a hashrate with a significantly cheaper attack cost. But probably the hashrate would still increase or stagnate, only that due to Moore's law and used equipment the attack cost would sink.


I actually agree here at least partly. The attacker's incentive could be higher if he can short enough coins to be able to profit from the resulting crash, so the market cap can have a small influence on incentives if we assume that "higher market cap" means "more value available for short selling on BTC lending markets", although the correlation is of course not direct. For double spends, as I wrote in the last post, the incentive should be linked to the transaction volume in a stable currency, i.e. how much he's able to steal with this double spend if he's able to rollback a block of transactions.

But in general what is most important in my opinion is the pure attack cost, i.e. the cost to acquire the hardware or hashrate to perform the attack.

If Bitcoin had a sidechain with Monero-style ring signatures (and other privacy features) and small transaction fees, would this make it digital cash for you too?