Information Entropy: Managing Unknown Risks, Disasters and Losses

We establish the probability and risk for the next disaster or loss outcome. We select proactive indicators of progress to support risk management and cultural analysis, and hence make predictions of the next outcome probability based on data. We answer the questions asked by Bernoulli and Bayes so long ago: knowing what we know has already happened, how we can predict what will happen next. With limited prior knowledge, the fundamental paradox is that we cannot learn other than by making those very errors, which we seek to avoid. We show how we must follow the Universal Learning Curve, even if we have had no outcomes and may face an unknown risk. Assuming that the future probability, rate and number are based solely on the past experience, we have shown formally the prediction of the future outcome probability and number. Future rate and number reductions will depend on our sustaining a true learning environment. For rare events and risks, we also determine the probability of having (or observing) the very next unknown. We show that it does not matter if we do not know the exact numbers: the learning trend is the key for decision making and risk taking. The rational choice and implication is to trust our experience and not to be afraid of the risk due to the perceived unknown.