Expected Value Calculations

Why This Matters

Expected value is the mathematical backbone of every business decision made under uncertainty—and that's essentially every business decision. When you're calculating whether to launch a product, hire a contractor, or invest in a project, you're implicitly asking: "What's the average outcome if I played this scenario thousands of times?" This concept connects directly to probability distributions, decision analysis, risk assessment, and financial modeling—all core testable areas in your course.

Here's what the exam is really testing: Can you move beyond plugging numbers into formulas to actually interpret what expected values tell decision-makers? You'll need to recognize when to use discrete vs. continuous approaches, understand how linear transformations work, and evaluate whether gathering more information is worth the cost. Don't just memorize formulas—know what business problem each calculation solves and when to apply it.

---

The Foundation: Core Formulas and Variable Types

Before applying expected value to business problems, you need to master the fundamental calculations. The type of random variable determines which mathematical approach you'll use.

Basic Expected Value Formula

• $E(X) = \sum x \cdot P(x)$—multiply each outcome by its probability, then sum all products
• Interpretation matters more than calculation—$E(X)$ represents the long-run average outcome, not a prediction for any single trial
• Weighted average concept—probabilities serve as weights, giving more influence to likely outcomes

Discrete Random Variables

• Countable, specific outcomes—scenarios like units sold (0, 1, 2, 3...), defective items, or customer arrivals
• Apply the summation formula directly—list all possible values, multiply each by its probability, sum the results
• Business applications include inventory decisions—expected demand calculations drive stocking and ordering policies

Continuous Random Variables

• Any value within a range—time to complete a project, exact revenue amounts, or precise measurements
• Use integration instead of summation—$E(X) = \int x \cdot f(x) \, dx$ where $f(x)$ is the probability density function
• Common distributions have known formulas—for normal, exponential, and uniform distributions, expected values are often given directly

---

Mathematical Properties: Simplifying Complex Calculations

The linearity property is your best friend for efficient calculations. Understanding these shortcuts prevents errors and saves time on exams.

Linear Transformation Property

• $E(aX + b) = aE(X) + b$—constants pull out of expected value calculations cleanly
• Scaling and shifting applications—if base salary is $40,000 plus 5% commission on sales, and $E(\text{Sales}) = \$200,000$, then $E(\text{Total Pay}) = 0.05(\$200,000) + \$40,000 = \$50,000$
• Portfolio calculations rely on this—expected return of a portfolio equals the weighted sum of individual expected returns

Conditional Expected Value

• $E(X|Y)$ adjusts for known information—recalculates expected value using only probabilities relevant to the given condition
• Updated probability distributions—if you know a customer is a repeat buyer, expected purchase amount may differ from the overall average
• Bayes' theorem connection—conditional expectations often require updating probabilities based on new evidence

---

Decision Analysis: Choosing Among Alternatives

This is where expected value becomes a decision-making tool. Business strategy often requires comparing expected outcomes across multiple options.

Decision Trees

• Visual framework for sequential decisions—nodes represent choices (squares) or uncertain outcomes (circles), branches show options and probabilities
• "Fold back" calculation method—start at endpoints, calculate expected values at each chance node, then select optimal path at decision nodes
• Handles multi-stage problems—launch now vs. delay vs. abandon decisions with multiple possible market responses

Expected Monetary Value (EMV)

• $EMV = \sum (\text{Payoff}_i \times P_i)$—the expected dollar outcome of a business strategy
• Decision criterion: choose highest EMV—when comparing investment options, marketing strategies, or project bids
• Risk-neutral assumption—EMV treats a certain $100 the same as a 50% chance of $200; real decision-makers may not

---

Risk and Information: Advanced Applications

These concepts address how uncertainty affects decisions and what we'd pay to reduce it. Information has quantifiable value when it improves decision-making.

Expected Utility and Risk Preferences

• Utility functions capture risk attitudes—risk-averse individuals have concave utility (diminishing marginal value of money)
• $E[U(X)]$ replaces $E(X)$—expected utility may rank options differently than expected monetary value
• Certainty equivalent concept—the guaranteed amount a person values equally to a risky gamble; lower than EMV for risk-averse individuals

Expected Value of Perfect Information (EVPI)

• $EVPI = E[\text{with perfect info}] - E[\text{without info}]$—maximum you'd pay to eliminate all uncertainty
• Upper bound on information value—no sample or study can be worth more than EVPI
• Calculated using decision tree analysis—compare expected value when you always choose optimally (knowing outcomes) vs. choosing under uncertainty

Expected Value of Sample Information (EVSI)

• Value of imperfect information—what's a market study, pilot test, or survey actually worth?
• $EVSI = E[\text{with sample info}] - E[\text{without info}]$—always less than or equal to EVPI
• Cost-benefit decision—conduct the research only if $EVSI > \text{cost of obtaining information}$

---

Quick Reference Table

---

Self-Check Questions

1. A company can either launch Product A (60% chance of $500K profit, 40% chance of $100K loss) or Product B (certain $150K profit). Calculate the EMV of Product A and identify which option a risk-neutral manager would choose.

2. Compare and contrast EVPI and EVSI: Under what circumstances would they be equal, and why is EVSI typically lower?

3. If $E(X) = 200$ and you need to find $E(3X - 50)$, which property do you apply, and what's the result?

4. A decision tree has two chance nodes following an initial choice. Explain the "fold back" method and why you start calculations at the endpoints rather than the beginning.

5. Why might a risk-averse investor reject a project with positive EMV? Which concept—expected value or expected utility—better captures their decision-making process?