Many times we do geological models and we trust in them blindly, forgetting that a model is just a conceptual simplified version of what we think can be the real stuff. As the paper I link today, quoting Loizou (2002), "Even in highly explored areas such as the North Atlantic margin of the UK, only one in five wells is successful".
"Structural models: optimising risk analysis by understanding conceptual uncertainty", published in First Break in June 2008 (volume 26) by Clare Bond, Zoe Shipton, Alan Gibbs and Serena Jones reviews the fundament of geology as an interpretation based science, how dependant it is on uncertainty and what does it mean in terms of risk analysis.
Very worthy to take a look and stop to think... "How confident I am with my model"?