Abstract
Many individual oil fields have multiple producing horizons. A question of some interest in developing an oil field is when does it make sense to drill for further horizons, based on the knowledge one has already obtained from the known horizons. Such an assessment must incorporate the economic chances of successfully finding another horizon and also the likely gains and costs involved. This paper shows how that question can be answered using Bayesian up-dating methods for improving success probabilities based on the information of known producing horizons, and incorporating both uncorrelated and correlated reservoir information to improve the chances of success. A case history is provided to illustrate how this objective procedure operates in practice.