810-1 Of P99's and Other Risk Management Urban Legends

See more from this Division: Gulf Coast Association of Geological Societies
See more from this Session: Risk Management and Uncertainty

Tuesday, 7 October 2008: 1:30 PM
George R. Brown Convention Center, 310CF

Dean R. Hennings1, Tom Griffith2 and Kevin Martindale2, (1)Anadarko Petroleum Corporation, The Woodlands, TX
(2)Risk Consistency Team, Anadarko Petroleum Corporation, The Woodlands
Abstract:
We present several “urban legends” we have encountered in our role as explorers and internal risk consistency consultants to make some observations about some of the pressure points we see in the world of probabilistic estimation of resource potential.

We have all heard that explorationists tend to over-estimate the resource potential of their exploration prospects and portfolios. We present several years of exploration drilling results from a predecessor company to support this view. In the context of probabilistic estimating methods, we propose that this is primarily the result of overestimating the small end of the lognormal resource distribution. And, we make suggestions on what can be done to improve the accuracy of our pre-drill resource estimates.

Practitioners of probabilistic resource estimation understand that the resource potential of a given prospect is represented as a lognormally distributed range of possible outcomes, from P99 (small) to P1 (large). On a probit scale, this distribution is represented by a straight line. We have found that most estimating errors are created by over-estimating the P99 end of the resource range of possible outcomes. We present some data to support this contention and offer a theoretical model that explains why the P99 should be a very small number.

Finally, we discuss former Secretary of Defense Donald Rumsfeld's infamous “known knowns” quote and suggest that the statement perfectly encapsulates the world of risk and uncertainty in which we work.

See more from this Division: Gulf Coast Association of Geological Societies
See more from this Session: Risk Management and Uncertainty

Previous Abstract | Next Abstract >>