In an exploratory search for oil and gas in any basin, it is desirable to evaluate from time to time the remaining hydrocarbon potential of the basin. From a statistical point of view, the productivity of the search area and the efficiency (or skill) of exploration are the parameters which are embodied in the actual history of exploration. A careful analysis of the past history can enable us to (1) extract an index for the efficiency of the past exploration, (2) obtain objective estimates of remaining reserves, and (3) predict the probability of success in the next few drillings.Here, we propose a model for which the input is the past chronological record of exploration in the search area. Efficiency is quantified by comparing the observed history with that which would have resulted had the search been random, other parameters remaining the same. In this respect, our method is qualitatively akin to that proposed recently by Cozzolino (1977). However, there are major conceptual differences between the two models resulting in different quantitative predictions for the same data. These are discussed and illustrated by applying the two methods to exploration in the southwestern portion of the Williston basin in the United States. It is shown that our approach yields parameters which are more stable with respect to variation in the length of the input history; in this sense our model is more reliable.