The most challenging part in the model is extracting expert opinion. Some ways can be tested or validated through drilling team: (1) Using a set of questions to build a probability distribution about cost and duration for every well type, (2) Letting the expert state the minimum, maximum and most likely or center of the data point, or (3) Providing different probability distribution of beta family given the expert minimum and maximum for particular well type. The third method could be more convenient, but other ways might emerge as development of the estimation model goes forward. Consider the expert opinion about the cost distribution in the well in figure 7. Obviously the expert(s) thinks that the actual cost will be more toward less values i.e. less than 2 million. Then after observing data of 2008-2012, getting the posterior distribution using Bayes theory, the cost distribution (the third bottom distribution) changes accordingly and centers towards values between 1.8-2.0 million. Note that the variance is lowered once the expert observed the data. Also the model is dynamic i.e. it can update itself once the expert receives new information where the final cost distribution centers around 2 million (o green line).