This paper considers sampling proportional to expected size from a partly unknown distribution. The applied context is the exploration for undiscovered resources, like oil accumulations in different deposits, where the most promising deposits are likely to be drilled first, based on some geologic size indicators (“creaming”). A size distribution within the Beta-class turns out to have nice analytical features in this context, and fits available data reasonably well, after rescaling. The theoretical and practical consequences for the accumulation of knowledge on the underlying distribution based on this scheme, named Beta-creaming, are explored in some detail.