You're assuming that attackers are always profit-motivated, which would not be true in the case of a government protecting the sovereignty of their national currency, for instance.
If the thing protecting the network is the self interest and good behavior of the miners then people should stop saying that the thing protecting the network is the awesome mining power of the network.
This kind of analysis are vulnerabilities in the consensus sense. You can think of a state actors not caring about the investment done but about the harm they can do.
Almost any state is capable of spending a few billion dollars and making a >50% attack (assuming they can buy enough ASICs), no matter how good your crypto is. There are much cheaper ways to bring something like Bitcoin to its knees. DDoS the nodes for a year, for instance.
So, are you saying that this is not part of the fundamental analysis you should do if you have money at stake? You have made such analysis to argue about the threshold numbers.
Where did I say "this is not part of the fundamental analysis you should do"?
My only point is that >50% attacks by the people involved are purely theoretical. Attacks by almost any state are the end of your project. A billion dollars is enough to DDoS pretty much everything.
My point was that we need to analyze consensus models (and new protocols), and that is very complex (and few people in the world have the skills needed). I gave the example of Bitcoin where the cryptography is perfect but the model has some security bounds.
Then you argued about the bounds and if they were theoretical or not which was not the central point of the argument and we can choose another issue to illustrate our central point. I think we can argue if it is theoretical or not ad infinitum.
So, to push forward my central argument I will again say: we need to check beyond the cryptography. Not only that, I will tell you: stay tuned because a new security finding with Bitcoin will be published soon.