When decisions are made without perfect knowledge of the results that may ensue, an unavoidable exposure to risk occurs. Petroleum industry business decisions often carry substantial economic consequences, thus it is critical to understand the risk involved. Exploration provides a suitable context for examining risk analysis issues. A common means of handling risk is to quantify it numerically in the form of probabilities. Conventional probabilistic risk analysis is appealing in that a complex situation can be reduced to a mathematical problem subject to computation. However, a critical examination of the extension of probabilistic techniques to typical business situations raises concerns. Logical inconsistencies are encountered when the theory of probability is applied in practice due to the limited number of opportunities normally available to evaluate the outcome of decisions. Specific concerns with probability based techniques are discussed as grounding for an alternative approach that seeks to avoid their limitations. A different paradigm is suggested to comprehend and manage risk, including a rationale for decision making without recourse to numerical probabilities. Risk is expressed in terms of a reasonable range of possible outcomes and described by linguistic statements. The decision process encourages quality technical analysis to reduce uncertainty, contingency planning, and mitigation measures that may allow acceptance of the risk that remains.
The purpose of this paper is to provoke contemplation of probabilistic risk analysis methods by examining them from a skeptical viewpoint. Methods based on probability are the currently accepted foundation for risk analysis throughout industry. Techniques utilizing numerical probabilities are highly compatible with electronic computation, hence are increasingly easy to use. There are situations where the use of probabilistic analysis is quite appropriate, but under certain conditions the results are questionable. Examples set in a petroleum industry context will be used to explain the causes for concern.
Risk is most simply defined as exposure to loss. The exposure occurs when decisions are made, influencing the prospects for the future. Risk is always present in the absence of perfect knowledge of the consequences of the chosen course of action. Mitigation of loss, or maximization of benefit, depends upon correct decisions founded on reliable assessment of the nature and magnitude of the risks that must be taken. The accuracy of the decision and reliability of the risk assessment can only be judged in retrospect, after the passage of time reveals the actual outcome.
Conventional risk analysis attempts to numerically quantify the unknown future using probability. Probabilities, expressed as decimal fractions or percentages, are established for specific outcomes either statistically-based on existing data, or subjectively-based on opinion. The values can then be utilized in mathematical equations or decision trees. The approach permits complex and sometimes intangible risk to be scaled, manipulated, and expressed in simple terms. Arriving at a single value representing risk perhaps instills unjustified confidence that the issue has been addressed, and further thought may tend to center on the number itself rather than the underlying risk factors.
Probability theory originated from the mathematical analysis of situations characterized by well defined rules, outcomes having equal probability, and the potential for unlimited repetition. Games of chance such as dice meet these criteria. In this domain the answers provided by probabilistic analysis are precise and irrefutable, but situations encountered in activities such as petroleum exploration do not necessarily share these characteristics. As a result, application of probability theory can produce answers that are demonstrably incorrect or incapable of being verified.