Dr. Frank Peel:  Course creator and presenter. Previously responsible for global prospect risk consistency in his role as Chief Geologist at BHP Billiton Petroleum.

Why this is important:

In order to decide whether or not to drill a prospect, we need to know how much it is worth if it is successful, how much it will cost to drill, and an estimate of the chance of success (or “prospect risk”). All three factors are equally important to the decision, but while there are many resources available to help us estimate the potential size and the drill cost, there is very little guidance available to help us estimate the chance of success. This course will help us to understand what risk is, and will provide a set of tools and methods that enable us to obtain an estimate logically, consistently, and efficiently.

We provide a process that can be used by a team of people to obtain a good estimate, through a session of guided discussion, and we provide a simple-to-use software application that delivers a single usable number, while also capturing the range of opinions of the participants.

In addition to traditional teaching material, this course involves a series of interactive exercises that will help the participant understand probability, and how we can estimate the chance of a future outcome (such as an exploration well being a geological success) using the information that we commonly have at hand. Many of these methods have general applicability to business decisions in other aspects of business and life.

Summary:  

This practical course explains the science of subsurface geological risk, supplying practical methods to estimate it logically, consistently, and efficiently. We show why risk is important, and how it is applied in decision-making. We use a set of realistic exercises to address key questions (e.g. how do we apply success/failure statistics? How do we apply seismic attributes?) and provide practical methods for obtaining a risk estimate. The methods used are applicable throughout the upstream business, from wildcat exploration wells to appraisal and development wells that test new field compartments.

Learning outcomes. On completion of this course, you will be able to:

• Understand the meaning and basis of prospect risk. Relate prospect risk to probabilistic ranges for volume and other parameters, and understand how it is part of the process used in probabilistic volume calculations using tools such as GeoX. Understand the difference between prospect risk, play risk, shared risk, CRS (Common Risk Segment) bands, the chance of a well encountering hydrocarbons, the chance of commercial success, etc…
• Understand how geological risk is used to make decisions and evaluate opportunities in the exploration business
• Know what information we need to collect to move the evaluation of a prospect to the risking stage; know when to stop interpretation and start making a decision.
• Use prospect mapping, well data, local and global success rate statistics, etc. to estimate a “prior” risk value.
• Use seismic attributes and apparent DHIs (direct hydrocarbon indicators) to boost or reduce the “prior” value using Bayes Theorem and other methods: know how this works, when to use it (and when not to), and recognize some of the common pitfalls.
• All prospects are different. You will be able to devise methods and strategies for obtaining a valid estimate that address the nature of your own prospect, using a range of different techniques.
• Plan and execute a successful session to review the information and estimate the risk numbers. Design a process to fit the needs of your own organisation. Combine different components (such as reservoir, source etc.) to obtain an overall number.

Who should attend:

Any geoscientist involved in the identification and evaluation of prospects, and undrilled segments of discovered and producing fields. It will be applicable to geoscientists at any career stage.

Course outline

Estimating the chance of success is a critical part of decision making, but there are few available resources that describe how to do this in a real-world setting using incomplete or equivocal data. The presenter has applied 30 years of prospect review experience in the oil industry to create a practical course, focussing on how we can use the type of data commonly available to us, especially seismic and well data and regional knowledge, to come up with a rational and defensible risk estimate. Exercises based on realistic information illustrate each component of the process.

We will explain the definition of prospect risk, and how it relates to the volume and other geological parameters that go into the volume calculation. We will show how this number is used in the critical business decisions. We will introduce lessons from other risk related activities; for example a crucial concept (Pot Odds) borrowed from high-stakes poker players may be used to inform us when we know enough to make efficient decisions.

Every prospect is different, and each may require a different approach: we will teach a range of common-sense methods that may be used to come up with the best estimate. We will show how to use well success statistics to inform your estimate, and deal with the question of what to do when local well data is sparse or equivocal.

The use of seismic anomalies and other geophysical attributes is powerful but can be difficult to understand. We do not teach the geophysics, but we will show how such information can be used appropriately to raise or lower the risk, including a discussion of the power and pitfalls of using Bayesian updating.

We will outline an effective method for risk collection by peer review: participants will be given an Excel application that they can modify for use by their own organisation.

Participants are encouraged to bring their own examples of risking problems that may be addressed using the methods set out in the course.

Course Agenda

Why is risk important? Use in decision making. The concept of ENPV, rate of return, etc., and how these depend on risk. What do we need to know to make the right decision? When is it important to obtain a fine-tuned risk estimate? The concept of Pot Odds – what the exploration business can learn from high-stakes poker players; knowing when you know enough. Exercises: 1. Valuing risk opportunities in exploration and business. 2. Pot Odds – can we make a quick decision, or do we need to do more work?

What does prospect risk mean (chance of geological success, Pg)?  Understanding what it is makes the task of estimating it easier. There is not much guidance in existing literature, and much of that is dated, confusing, or not applicable to modern probabilistic methods used to estimate hydrocarbon volumes. We explain precisely what we mean by prospect risk, and how it differs from other related concepts.  How risk and volume ranges are related – you can’t define one without the other. How risk relates to probabilistic volume ranges, P90-P50-P10 values, etc. How we can use Pg to derive the chance of an exploration well finding hydrocarbons, the chance of commercial success, etc.  Why the chance of geological success for a given prospect  is NOT the same as chance of the exploration well being a success, or the chance of a commercial discovery; and how we can estimate those.

Risk components – multiplication of independent factors to obtain overall risk. What are the “right” components? How many factors do we need, and how flexible can we be?  How to avoid “double-dinging”. Why multiplying components gives prospect volume curves that are not lognormal – and why this matters.

Characterisation of the prospect: What do we need to know in order to risk a prospect? e.g. Trap – we must define all its components and all the edges. We need to specify all the parameter ranges associated with success case. We must define the likely failure mode.

Methods for estimating risk: Basic logic. We can often frame apparently difficult problems in a simple way which allows us to use available information.  Exercises. 1. chance of death by meteorite. 2. Chance of source rock maturity in an undrilled basin. 3. Chance of reservoir presence in a deepwater basin.

Sherman Kent’s Scale: translating words into numbers. What it is, how it saved the world, how to use it. Cultural context – why the set of numbers defined for the words that were used by 1950s US spies may not apply in all cultures today; designing a Sherman Kent table for your culture and your language. Exercises: 1. prevent World War III. 2. Use S-K to estimate geological risk

 

Success/failure statistics. What is a valid statistic? When are past rates of success a valid guide to future expectation, and when are they not? How to obtain an estimate from a small data set? How to apply imperfect statistics when we have other information? What do you do when there are no local statistics – does ignorance mean it’s 50:50?  Exercises: 1 Strat trap – combining local and global statistics. 2. Farm out vs. new acreage

Complexity and risk (Murphy’s law). The more ways it can fail, the more likely it is to fail.

Look-up tables: strengths and weaknesses. The process of assigning defined risk numbers to specific prospect component types – is it valid? Can we use it?

The principal of reasonability and the coefficient of ridiculosity. To estimate the chance of failure, we need to define a failure mode. How reasonable does the success case look? How hard do you have to strain the available data to make a component break? If you can’t devise a failure mode, is success proven? Exercise: create success and failure case trap maps from posted well and seismic data, judge relative reasonability.

Common Risk Segment (CRS) vs. prospect risk: Can you use CRS maps as an input to prospect risking? Should they be hard-linked?

Evidence Theory and Bayes Theorem: How do we properly use independent supporting evidence, that is consistent with our model but which does not by itself prove it? How do you combine this with a “prior” estimate of probability? Basic Bayes’s theorem and its simple application.  Pitfalls – The importance of estimating reliability, the danger of too much certainty. Have we identified all the possible outcomes? How to incorporate the possibility of the unknown. Exercise – the bell curve and the unknown outcome.  Exercises: 1. The Bloody Glove – is the defendant guilty? 2. Detection of FTL particle velocities? 3. Medical diagnosis. 4. The Monte Hall Problem.

“Direct hydrocarbon indicators”: their use and abuse.  In some circumstances, we are able to see features on seismic data that look like oil or gas accumulations – for example bright amplitudes, AVO anomalies, etc. These may be DHIs or ” Direct Hydrocarbon Indications”  – or they may be other things. Such information can be very valuable, but how do we incorporate it into prospect risk estimation? What are the potential pitfalls? The proper use of seismic attributes can be a powerful aid to risking; but if used improperly it can give wildly wrong estimates.

Topics covered: The importance of separating out attribute-influenced factors from non-influenced factors. Which risk components can attributes modify – and which ones can they not? Recognising that what the attributes respond to is not the same as simple success/failure. “Not everything that glows, flows” Not all porosity is created equal. Permeability is invisible and compartmentalisation may be invisible. Not all situations with hydrocarbons in porosity equate to a success case – the importance of hydrocarbon-bearing failure cases.  Conformance is vital. An apparent contact that does not conform to structure is a very strong negative. What else glows? Recognizing alternative sources of apparent DHIs: diagenesis, residual, low saturation hydrocarbons. Rules of common sense when applying seismic attributes. Exercise: using attributes to risk a prospect

Non-Bayesian DHI boosting

The Wisdom of Crowds : the power and pitfalls of crowdsourcing. Advantages of peer review. Are explorers all incorrigible optimists? Why it is loved by some, and loathed by others.  Avoiding leading, groupthink and cognitive bias. Exercise: Use crowdsourcing to estimate an unknown quantity.

Prospect risking by Peer Review:  How to design a peer review for risk estimation: ideal size and  composition of the group. Creation of a simple computational tool for risk assessment – attendees will be given the basic software application, and advised how to modify it for their own company’s preferences. The importance of recording and reporting. Exercise: risking a prospect by group voting.

Lookback analyses: how to use past results to calibrate your method of estimation. Methods and pitfalls. Component lookback – how to maximize the information from lookbacks. Why lookbacks inevitably give the apparent message that explorers are too optimistic – even when we get it right!

Prospects, plays and dependency: What is shared risk? How does this change? What is play risk?

 

What to expect when you’re prospecting: What happens to risk of an individual prospect, or of a portfolio of prospects, as you gain new information? Can you expect to de-risk a prospect by acquiring new data? If not, how does acquiring new information add value?