Abstract: Some Simple Statistical Tools to Aid in Risk Analysis

James H. Hartsock

Classical decision theory separates decision making into three groups, namely, certainty, risk and uncertainty. This paper defines the differences. In exploration and production we deal mostly with uncertainty, yet we persist in treating problems as through they are related to risk. Risk analysis can be formulated in many ways such as decision trees, decision matrices, Monte Carlo simulation, etc. We shall focus on Monte Carlo simulation because it is a simple and attractive tool for treating risk and the results are easily understood. However, the underlying assumptions are not as well known and are often misunderstood. This paper clarifies these assumptions as well as identifies some of the pitfalls. Finally we present several simple statistical techniques that can be u ed prior to applying Monte Carlo simulation to:

-- better define the density of each variable

-- establish the independence of two random variables

-- determine whether a sample is truly random.

AAPG Search and Discovery Article #90960©1995 AAPG Southwest Section Meeting, Dallas, Texas