I think that the size of the bet is depending of my estimation of the probabilities of different outcomes (may be not in this case) and so we can't completely exclude probabilities estimation.

But in general I agree this your theory. It is useful in estimation of the x-risks. We don't need exact probabilities of different risks. We need information how to use our limited resources to prevent them. This is our bets on our ability to prevent them. But to make such bets we need some idea about order of magnitude of risks and their order in time.