Decision Analytics, Value of Information, and Piloting in Unconventional Reservoirs

Abstract

Abstract Suppose we plan to develop a field by drilling up to N wells. Each well i (= 1 to N) will produce an uncertain quantity of reserves Xi. We assume that the Xi are identically distributed with common mean m and variance s2 This is often reasonable, as companies view the development of unconventional reservoirs as a statistical plays; some wells will be profitable and others will not. Lacking an ability to identify the wells that fall into each category, a package of wells is drilled with the hope that the average reserves m will exceed some minimum threshold t. We further assume that given m and s2 are known, the wells are probabilistically independent. Therefore, we assume the wells are independent and identically distributed (iid). In practice, the iid assumption means that (1) we do not have any reason to believe one location is likely to be more profitable than any other and (2) the drilling results at one location do not change our beliefs about the potential profitability of another location, given that m and s2 are known. If either m or s2 are uncertain then prior drilling results at one or more locations will tell us something about the future prospects for the field.