The expected value of perfect information in unrepeatable decision-making

Abstract

This paper reflects on the concept of the 'Expected Value of Perfect Information' (EVPI) and the procedure used to determine it. It is widely accepted that this value is the difference between the expected value when we have perfect information and the best expected value provided by alternatives. However, this difference often results in values that no rational decision-maker would accept. Here, we overcome this difficulty by defining the 'Value of Perfect Information for the Problem' (VPIP) where we consider not only the price of perfect information (EVPI) but also two additional parameters: the 'Loss to be Avoided' and 'The Most Favourable Payoff in the Worst Scenario'. In this way, we are able to obtain a more accurate value of the amount a decision-maker might be willing to pay for perfect information. We also seek to show that the indiscriminate employment of probability theory, based by definition on the repetition of the experiment, can be misleading in the case of decisions which, owing to the very nature of the problem, are unrepeatable.