Toward a Fundamental Decision Theoretic Grid of Risk and Cyber Threat

What is the greater probality that a cyber attack will occur versus that it will not occur as a random event which cannot be predicted, outside of fully secured intelligence data (that is "bullet-fool" proof, parlance.) This cannot be done outside of cyber-intrusion, and detection of assessing the the threat level to which there may be an attack. 1-->0--->1, in binary, odds/even probalistic ratio on a Baysian subjective logical view that an attack will occur or not-- in contradistinction to its converse relationships. This is the limits to which we can order any estimation of a decision theoretic model than any attack can happen or not happen, at any time, any place, any where. Can we prepare for such a assault on the any type of epistemological-logical reasoning in relation to cyber-Intelligence, and surveillance, if we are under cyber-attack?