A Comparison of Probabilistic and Deterministic Reserve Estimates: A Case Study

Abstract

Summary Twelve practicing petroleum engineers participated in a study comparing deterministic with probabilistic reserve estimates. Production data for five wells were provided sequentially in four phases covering the first 5 years of production history. A deterministic and probabilistic reserve estimate was made by each evaluator for each phase. The deterministic reserve estimates were based on Securities & Exchange Commission (SEC) definitions. For the probabilistic reserve estimates, the evaluators were asked to supply three values: P90, P50, and P10. P90 is the value where there is a 90% probability that the actual reserves are greater than this value; P50 is the value where there is a 50% probability that the actual reserves are greater than this value; and P10 is the value where there is a 10% probability that the actual reserves are greater than this value. Most evaluators used their deterministic estimate as the basis for the probabilistic estimates. In these cases, subjective estimates were made rather than using a statistical approach. The evaluators have more difficulty in estimating the P90 value when the decline exponent b is low. The evaluators have the same difficulty in estimating the P10 value when the decline exponent b is high. In comparing the individual evaluators, there is considerable variability in deterministic, P90, P50, and P10 values. The most variability is in the P10 values. Also, it was observed that over time, as more production history is recorded, P90 values do not approach truth from the low side.