A probabilistic discovery model modified after Kaufman's earlier model is simplified to reduce computational demands and to reduce the sensitivity of the resulting estimates. The model is applied to the North Sea to estimate remaining oil reserves and forecast future discoveries. The simplification jeopardizes some informational detail, but the errors and approximations inherent in historical data sometimes overvalue the available information. A broader categorization scheme helps to control errors in this case. The model uses a stochastic production function based on a timing relationship between exploratory efforts and reservoir discovery and on a dynamic relationship of productivity and resource depletion. 21 references, 4 tables. (DCK)