Determinism and Statistics in Prospect Appraisal

D. Sluijk, J. R. Parker

The many different approaches to prospect appraisal can be grouped into two broad categories: those that emphasize a deterministic approach and those that follow statistical principles. However, these approaches are by no means mutually exclusive.

The necessity for prospect appraisal is undisputed, both to translate geology into figures (to arrive at monetary values) and to express the uncertainty implicit in oil and gas exploration. The first stage of any prospect appraisal is developing a play concept: collecting and analyzing data to understand the processes that lead to the formation and preservation of oil and gas accumulations in the subsurface. Such studies have a theoretical basis in geological and geochemical research, and thus a deterministic approach is probably the first choice.

However, a deterministic model also contains the limits of the analysis; it would be unrealistic to assume that every detail of every process is described. Some processes are poorly understood, or if they are understood in principle, they may be controlled by factors that are difficult (or impossible) to predict before drilling.

A statistical approach can help overcome these limitations in a purely deterministic method, providing that the statistics are applied in a way that uses the benefits of a deterministic understanding. One approach should supplement the other. Such statistical techniques require values for several independent variables (i.e., geologic parameters) that help explain the parameter to be predicted--the dependent variable. Such geologic parameters can best be derived from a deterministic model for the process in question.

The task of the designer of a prospect appraisal system is thus defined. The basis for the system should be the best (deterministic) model available, with data collected from several case histories representing the "world" to be evaluated. Statistical analysis can then test the relevance of the various geologic parameters, and define the limits of understanding and ability to describe the relevant geologic parameters. The resulting formulas, together with the residual errors to account for incompleteness and inaccuracy, can be built into a simulation model for prospect appraisal. If the two approaches are carefully combined, the resulting prospect appraisal system will combine the benefits and avoid the shortcomings of the two techniques.

Also, an important aspect of prospect appraisal is largely neglected by the purely deterministic appraiser: the aspect of chance. In an ideal, full-fledged deterministic system, chance (or risk) does not play a role; it is either "yes" or "no." However, such an ideal system does not exist, and we must acknowledge our inability to describe all conditions and configurations in detail. In this situation, the statistical procedures provide a tool to quantify the risks that are a major element in hydrocarbon exploration.

AAPG Search and Discovery Article #91043©1986 AAPG Annual Convention, Atlanta, Georgia, June 15-18, 1986.