- Probability definitions, objective and subjective probabilities, calculation rules, expected value
- Simple exploration economics
- Decision trees
- Bayes formula and the Value of Information (VOI)
Participants will learn about simple exploration economics, decision trees, and the value of information. A case study will combine decision trees and Bayesian inversion to show how the probability of geologic success should be revised to incorporate new seismic information and how the quality of this information will impact its value. The psychological aspects of decision making, as well as utility curves, attitudes towards risk, and cognitive biases, will be discussed.
- Deterministic versus probabilistic: what is the difference?
- Describing uncertainty: population, samples, histogram, pdf, cdf, percentiles
- Mean, mode, median; variance, standard deviation, coefficient of variation
- Standard error of the mean
- Random and systematic errors and how to deal with them
- Winner's curse
Day 2 will include a discussion of estimation in general, confidence intervals, coverage probability, and sample size. A case study on the selection of blocks will show the economic implications of making decisions based on estimates rather than true values. Distributions commonly used in geosciences such as normal, lognormal, triangular, uniform, stretched beta, and binomial will be presented.
Monte Carlo and Applications
- The Monte Carlo approach
- Presentation of @RISK™, a Monte Carlo add-in for Excel™
- Computing uncertainty on STOIIP
- Prospect evaluation
- Modeling dependencies between risks
Day 3 will be devoted to Monte Carlo simulation and hands-on applications. The estimation of the production increase expected from a workover program will demonstrate the importance of scale in assessing uncertainty. A comprehensive case study of a prospect with multiple dependent objects will have the participants model shared risks, build the probability tree of the different cases, derive the distribution of aggregate resources, and select development scenarios.
- Probabilistic resources aggregation
- Efficient Frontier
- The correlation coefficient and its pitfalls; Simpson's paradox.
- Effect and interpretation of correlation
Day 4 will focus on portfolio issues, notably probabilistic aggregation, put in perspective with the PRMS resources classification system. An exercise will illustrate the effects of diversification, and an exploration and appraisal approach for a cluster development will be proposed. The day will close with a presentation and discussion of correlation.
Trend Lines and Review
- Least squares
- The predictive value of a trend line
- General wrap-up of the week
Uncertainty on fluid densitites computed from pressure data will take participants through the various steps involved in establishing and checking a regression model and deriving prediction uncertainty. The session will end with a review of the various concepts presented during the week.
Anyone interested in learning more about the approach and techniques of risk, uncertainty, and decisions in E&P projects.
Some basic statistics knowledge would be a plus but is not required.