--> Production Forecasting Optimism – Impact of Workflows, Sparse Data, and Human Bias

Southwest Section AAPG Annual Convention

Datapages, Inc.Print this page

Production Forecasting Optimism – Impact of Workflows, Sparse Data, and Human Bias


Accurate oil and gas production forecasting is one of the keys to the industry maximizing return on invested capital. Detailed geological and dynamic reservoir models are used along with sophisticated economic models to generate probabilistic production forecasts that enable evaluation of development options and to justify sanction of a preferred development alternative. A major question for the industry is how accurate are these forecasts. Although there have been few published studies, a recent SPE conference paper by Nandurdikar and Wallace (2011) based on a study of nearly 150 industry major capital projects concluded that the industry is delivering only 75% of forecast production overall and projects with identified subsurface issues delivered only 55% of the production volumes forecast at time of project sanction.

So why are production forecasts generally optimistic? Among the contributors to forecast optimism, the following appear to most significant: (1) human factors or biases that favor project development includes “misuse” of analog data; (2) use of well placement optimization workflows without full appreciation and understanding of possible technical limitations; (3) failure to realistically capture the full range of uncertainty due to sparse subsurface data; and, (4) failure to mitigate the various ways in which optimistic forecasts are favored in response to by common workflow choices.

Studies completed as well those still in progress suggest that the major contributors to the observed shortfall in actual production compared to pre-sanction are in order of decreasing importance (1) human biases, both technical and management, (2) sparse data bias, (3) well placement optimization workflows, and (4) workflow and model parameter selections such as grid size or upscaling.