Expectation
Expectation in petroleum engineering and geoscience refers to the expected value (statistical mean) of a probability-weighted outcome in reservoir characterization, reserve estimation, and investment decision analysis, where uncertain quantities (recoverable reserves, production rates, exploration success probability, commodity prices) are represented not as single deterministic values but as probability distributions whose expectation provides the probability-weighted average outcome used for economic valuation and portfolio optimization; the expectation of a probability distribution is calculated as the integral (for continuous distributions) or sum (for discrete distributions) of each possible outcome multiplied by its probability of occurring, so a prospect with a 30% probability of success (Pg) and an expected reserves of 100 million barrels if successful has an risked expectation of 30 million barrels (0.30 x 100 MMBBL) that accounts for the probability of the well being dry; in the Society of Petroleum Engineers PRMS (Petroleum Resources Management System) framework, the expected case is distinct from the low case (P90, or 1C/2C in the PRMS notation) and the high case (P10, or 3C in the PRMS notation), with the expected value sometimes corresponding to the mean (probability-weighted average of the distribution) rather than the P50 median in asymmetric distributions; the concept of expected monetary value (EMV) extends the statistical expectation to the dollar outcome of investment decisions under uncertainty, combining the probability and magnitude of each scenario to calculate the probability-weighted average net present value that forms the rational basis for exploration and development investment decisions in a portfolio context.
Key Takeaways
- The distinction between expected value (mean of the distribution) and the most likely value (mode) or the median (P50) is often misunderstood in petroleum engineering contexts and leads to systematic errors in probabilistic reserve estimation: for a right-skewed lognormal distribution (which is typical of petroleum reserve estimates, where there is a hard lower bound of zero reserves but an unbounded upper tail of potentially very large accumulations), the mean is always larger than the median, which is larger than the mode; stating the "expected reserves" as the mode (the single most likely outcome) significantly understates the probability-weighted average reserves that the portfolio will deliver across many similar prospects; in a portfolio of 10 exploration prospects each with a mode of 50 MMBBL and a lognormal distribution with geometric standard deviation of 2, the mean (expected value) might be 70-100 MMBBL per prospect, and the portfolio total expected reserves would be 700-1,000 MMBBL even if the most likely outcome for any single prospect is 50 MMBBL; understanding which statistic is being reported (mode, median, or mean) when a reserve estimate is labeled as "expected" is critical to using the number correctly in portfolio analysis and capital allocation decisions.
- Expected monetary value (EMV) decision analysis provides a rigorous framework for comparing exploration and development investment alternatives under uncertainty by calculating the probability-weighted net present value (NPV) of each possible outcome (success, partial success, or dry hole) and selecting the alternative with the highest EMV; for an exploration well with a 20% probability of commercial success (Pg = 0.2), an NPV of $500 million if successful, and a dry hole cost (sunk cost) of $50 million with probability 0.8, the EMV = 0.2 x $500M + 0.8 x (-$50M) = $100M - $40M = $60M positive, meaning the EMV-positive exploration opportunity should be pursued if the company has the portfolio scale to absorb the 80% probability of losing the drilling cost; EMV analysis properly applied requires that the company has a portfolio of similar-sized opportunities (so that the law of large numbers brings the actual outcome close to the EMV expectation over many prospects) and that the outcomes of individual prospects are approximately independent (so that a string of dry holes does not threaten the company's financial survival before the portfolio statistics can work in its favor).
- The geological expectation in volumetric reserve estimation is computed from the probability distributions of the input parameters (drainage area, average net pay, porosity, hydrocarbon saturation, formation volume factor, and recovery factor) by Monte Carlo simulation, which draws random samples from each input distribution and calculates the resulting hydrocarbon volume for each draw to build a distribution of possible reserves; the mean (expected value) of the resulting reserves distribution is the input to EMV calculations and portfolio analysis, while the P10 (high case, 10% probability that reserves exceed this value) and P90 (low case, 10% probability that reserves fall below this value) bracket the uncertainty range; the width of the uncertainty range (P10/P90 ratio, which is typically 5:1 to 20:1 for early-stage exploration prospects) reflects the geological uncertainty in the input parameters, and reducing this uncertainty through additional seismic acquisition, well data, or analog study is the technical basis for appraisal drilling investment in a discovered accumulation before committing to full field development capital.
- Portfolio expectation and the aggregation problem arise when summing expected reserves or NPVs across a portfolio of prospects or fields: the expected value of the sum of multiple uncertain outcomes equals the sum of their individual expected values (linearity of expectation), so portfolio expected reserves = sum of individual prospect expected reserves regardless of the shape of the individual distributions; however, the variance of the portfolio sum depends on the correlation between individual prospect outcomes, with perfectly correlated prospects (all succeed or all fail together) having portfolio variance equal to the sum of individual variances, and independent prospects having portfolio variance equal to the sum, but the portfolio standard deviation much less than the sum of individual standard deviations; this diversification effect (reduction in portfolio-level uncertainty through combination of independent prospects) is the mathematical foundation of portfolio theory in exploration management and explains why diversified exploration portfolios (many small prospects across different geological plays and basins) provide more stable returns than concentrated portfolios (a few large prospects in a single play) even when the expected value is the same.
- Bias in expected value estimation is a persistent and well-documented problem in petroleum engineering and exploration geology: optimism bias causes geological estimates to systematically overstate prospect size and understate dry hole probability (the planning fallacy applied to subsurface uncertainty); base rate neglect causes analysts to underweight the prior probability of success (the historical success rate of similar prospects in the basin) in favor of the specific geological evidence from the individual prospect; anchoring to early estimates causes subsequent data to insufficiently update the prior expected value; and availability bias causes analysts to weight recent exploration successes or failures disproportionately in estimating prospect probability; the cumulative effect of these cognitive biases tends to produce systematic overestimates of portfolio expected value that mean actual drilling results fall below pre-drill expectations at the portfolio level; implementing structured probabilistic workflows (using formal prior distributions for play fairway parameters, explicitly tracking updating from new data, and calibrating estimates against historical drilling results) is the primary technical mitigation for these biases in exploration portfolio management.
Fast Facts
The concept of expected monetary value as a decision criterion for oil and gas investment was formalized in the petroleum engineering literature in the late 1960s and early 1970s by authors including John Newendorp (Decision Analysis for Petroleum Exploration, 1975), who introduced the decision tree and EMV frameworks that are now standard in petroleum economics curricula worldwide. Newendorp's work translated the formal expected utility theory of decision science into practical petroleum engineering terms, providing the conceptual framework that allows exploration geologists and engineers to communicate probabilistic thinking about prospects in economic terms that portfolio managers and executives can use for capital allocation decisions. The adoption of probabilistic reserve estimation and EMV-based portfolio optimization was one of the most significant improvements in petroleum industry decision quality of the late 20th century, shifting the discipline from deterministic "best estimate" thinking to explicit quantification of uncertainty.
What Is Expectation?
Expectation in petroleum decision-making is the probability-weighted average outcome — not the best case, not the most likely case, but the average of all possible cases weighted by how likely each is to occur. For an exploration prospect, the expected reserves account for the probability that the well comes up dry (zero reserves) as well as the probability that it finds more than expected (upside). For a development decision, the expected NPV accounts for the probability of lower-than-expected production, lower-than-expected commodity prices, and higher-than-expected costs — and their upside equivalents. The expected value is not what anyone expects to happen on any single project; it is what the average outcome will be across a large portfolio of similar decisions. That distinction — between the single-case outcome and the portfolio-average outcome — is the foundation of rational decision-making under uncertainty. Individual wells can only succeed or fail. Portfolios can approach their expected value. Running a portfolio of exploration prospects using expected value arithmetic is the closest thing the petroleum industry has to a scientific method for managing irreducible geological uncertainty.
Synonyms and Related Terminology
Expectation is also called expected value, mean value, or probability-weighted average in petroleum engineering contexts. Related terms include expected monetary value (EMV, the probability-weighted net present value of an investment decision under uncertainty, computed by multiplying the NPV of each possible outcome by its probability of occurrence and summing, providing the rational basis for comparing exploration and development investment alternatives), Monte Carlo simulation (the probabilistic computation method that generates the reserve or NPV distribution by drawing random samples from input parameter distributions and calculating the resulting output thousands of times, with the mean of the resulting distribution being the expected value used in portfolio analysis), probability of geological success (Pg, the probability that an exploration prospect contains a commercial hydrocarbon accumulation, which multiplies the reserves-if-successful estimate to give the risked expected reserves used in portfolio EMV calculations), P50 (the median of a probability distribution at which there is a 50% probability that the actual outcome will be above and a 50% probability it will be below, which equals the expected value only for symmetric distributions and differs from the mean expected value for the asymmetric lognormal distributions typical of petroleum reserve estimates), and decision tree (the graphical representation of sequential decisions and uncertain outcomes used in petroleum investment analysis, in which each branch is assigned a probability and a value, and the expected value of each decision node is calculated by probability-weighting the branch outcomes to identify the decision path with the highest EMV).