In this brief video, Dr. Chris Albright demonstrates how to build and interpret a decision tree for a classical oil drilling problem. New features in PrecisionTree 6.0 are introduced, including Appending a Symmetric Subtree, Bayesian Revision, and Adding a Node anywhere in an existing tree. These new features make it easier than ever to perform decision analysis in PrecisionTree using decision trees and influence diagrams.
The problem itself is straight-forward: An oil company must decide whether to drill in a given region. There is uncertainty about the amount of oil in the ground — it can be dry, wet, or soaking. What will be the payoff for drilling? Should the oil company make a decision right away, or run a seismic test? How does one account for the conditional probabilities of the seismic test result? The goal is to find the strategy with the largest expected net gain. Dr. Albright sets up a model in PrecisionTree to address all of these questions.