Calculating Expected Monetary Value by using Decision Trees is a recommended Tool and Technique for Quantitative Risk Analysis. The value of seismic in these new resource plays is dependent on the probability distribution of and economic differences in the spectrum of outcomes available. A decision is being contemplated as to drill a well into a welldefined, thick, and broad, Mannville channel system in West Central Alberta. Let us work through the solutions to Bayes’ theorem for the three interpreted seismic cases corresponding to each of the three events we defined; that our interpretation could indicate a poor well, and average well, or a superior well. The expected NPV at chance node B, the seismic interprets an average well is: = 0.026 * (-$2,550,000) + 0.947 (+ $1,450,000) + 0.027 * (+$4,950,000) = +$1,436,842. It is: NPV (E1) * P(E1) + NPV (E2) * P(E2) = -$3,000,000 * 0.20 + $1,500,000 * 0.80 = +$600,000. These drilling experiences include heavy oil, shallow gas, deep carbonate exploration, deep basin, Peace River Arch, Saskatchewan and Manitoba oil, and include vertical as well as horizontal drilling. This is the probability, accuracy, or reliability of the interpreted information being correct. The seismic further discriminates the likely quality of the well by identifying undesirable stress regimes with azimuthal amplitude versus offset analysis and converted wave analysis of shear wave splitting (Close et al, 2012). This has long been intuitively understood, which is why geophysical data has been most frequently used in exploration. Lee Hunt was the 2011/2012 CSEG Distinguished Lecturer. If you notice any problems with an article (examples: incorrect or missing figures, issue with rendering of formulas etc.) Hunt et al (2012) also showed that improved steering accuracy generated economic improvements as measured by modeled barrels of oil per day in a Saskatchewan and Manitoba Devonian horizontal oil play. The probability of the interpretation being correct in either case is 0.9. There are two possible (states of nature) outcomes for the well’s production and consequent present value: the well misses the channel system, or the well encounters the channel system and hits statistically average rock for the system. Business or project decisions vary with situations, which in-turn are fraught with threats and opportunities. These are the likelihood of each of the Ei events occurring. Decision node D follows the “interpreted no channel” branch and explores the possibility that a decision is then made to either not drill the well, or to drill the well. The original, or initial, probability of each possible outcome is also an important aspect of this analysis. A large volume of data is being converted to make this online archive. This means the no channel interpretation is now understood to mean no well is drilled, yielding an NPV of -$100,000. NPV = -$4,500,000. We will show that if we understand these probabilities and their economic differences, we can quickly assess how valuable seismic data is. NPV = +$5,000,000. The key with decision tree analysis is that before we can assess the decisions that we face soonest, we must first reconcile the decisions that we would make in the future depending upon the circumstances we find ourselves in. Gray (2011) cites other case study based examples, and the annual CSEG Symposium typically contains several value oriented case studies. Let us look at another example involving a true resource play. A scenario approach to capacity planning. The revised probability of the average well interpretation has the highest probability for any event, the average well event of 0.947, which we would also expect given the original probabilities. In the case of seismic, B might be an interpretation from the data that a channel is present. This new scenario describes a resource play where the average outcome is much more likely than either of the other events, but that the poor event is more economically damaging as compared to example II. The value of geophysical data may be estimated in a variety of ways, none of which are materially different from the way of determining the value of other kinds of information such as log data. Column three in Table 1 shows the calculation of the numerator of Bayes’ Theorem. Given a change in these values, we can further explore the required reliability of the imperfect seismic information. In Chart 1, we see that we have several choices: first, at decision node F we choose to either gamble and drill using our original estimate of probabilities, or do we choose to purchase the seismic. The expected NPV at chance node C, the seismic interprets a superior well is: = 0.048 * (-$2,550,000) + 0.857 (+ $1,450,000) + 0.095 * (+$4,950,000) = +$4,259,524. The valuation of the processing technique in the Viking case specifically measures the economic effect of the reliability of the seismic data, and is unusual in the literature. This cost is minor, but fully accountable, and is $100,000 per well deferred. Scott Hadley, David Gray, John Duhault, George Fairs, Ron Larson, and Satinder Chopra all helped make this paper better through their comments, criticism, and support. Gray, D., 2011, Quantify the Economic Value of Geophysical Information: CSEG RECORDER, 36, 3, 29-32. The decision is made to defer drilling the well. Given the high chance (80%) of encountering the channel, this approach may be tempting, and many operators would choose it. Decision Trees in financial analysis are a Net Present Value (NPV) calculation that incorporates different future scenarios based on how likely they are to occur. We examine the sensitivity to the accuracy of the seismic interpretation and to the spread of expected values of the drilling being considered. This is no surprise; we should not invest in anything that lacks a value argument, otherwise it is charity. The change in original probabilities was not as dominant as the change in poor well outcome. The information that we bring in (say geophysical data such as seismic) defines event B. Let us go through the solution at each node. Chance node C explores the possibility that despite the “interpreted no channel” outcome from the seismic, the decision was made to drill the well anyway. in geophysics in 1990, after which he started his career working for PanCanadian Petroleum Ltd. At PanCanadian, he was mentored by Bill Goodway. Interested in starting, or contributing to a conversation about an article or issue of the RECORDER? A “high level” truism from this work is that the value of seismic can be very significant when it has a high reliability and when the variation in outcome economics is large. Sensitivity analysis of the seismic reliability and cost can be performed with this method, which could determine the maximum amount that should be spent on seismic, or the minimum change in the probability of success that the seismic can afford while still enjoying an advantage in expected value. Therefore, there is an economic penalty to deferring drilling a well within a pad when the rest of the pad is being developed. If we have the seismic, should we accept the interpreter’s advice? Figure 3 illustrates the results graphically. ROA allows computing the correct discount rate using the replicating portfolio technique or risk-neutral probability method. Table 4 below shows the solution to Bayes’ theorem if B’ indicates an average well. There are two broad strategies that can be employed: the first being the so-called “gamble strategy” where we choose to drill the well without seismic. But which of the decisions yield the highest expected NPV? The high capital nature of resource plays often make geophysical data of economic benefit throughout the life of projects. Specifically, it shows how well known decision analysis techniques can be applied to estimate the value of seismic for resource play examples. The gamble scenario has an invariant expected value of +$1,062,500. The reliability of the seismic is subject to many factors including the quality of its processing, the effort and appropriateness of the acquisition parameters to the specific geologic target, and also to the interpretive techniques being employed. Some interpretive techniques require additional processing and investment. admin@csegrecorder.com. The seismic method expected values are shown by red circles, and change linearly with the reliability of the interpretation. Even in resource plays there is commonly a significant variation in the flow rates, ultimate recoveries, and values of individual wells. The expected value of the seismic is the chance the seismic would interpret channel multiplied by the expected value of the subsequent branch plus the chance the seismic would interpret no channel multiplied by the expected value of the subsequent branch.