Calculating the Optimal Choice: The Expected Value Approach
To decide which alternative is mathematically superior, we use the Expected Value (EV) approach. This method calculates the weighted average of all possible outcomes for a specific alternative, using the probabilities of the states of nature as weights.
The formula for the Expected Value of a decision di is:
EV(d_i) &= \sum (\text{Probability of outcome } X \times \text{Payoff of outcome } X)
\end{aligned}$$
$$
Let's apply this to our grocery store example, assuming the probability of strong demand is $0.45$ and weak demand is $0.55$:
* **EV for Pilot Launch ($d_{1}$):**
$$\begin{aligned}
EV(d_{1}) &= (0.45 \times 120,000) + (0.55 \times -30,000) \\
&= 54,000 - 16,500 \\
&= \$37,500
\end{aligned}$$
* **EV for Full Launch ($d_{2}$):**
$$\begin{aligned}
EV(d_{2}) &= (0.45 \times 300,000) + (0.55 \times -180,000) \\
&= 135,000 - 99,000 \\
&= \$36,000
\end{aligned}$$
In this case, even though the "Full launch" has a much higher potential upside, the **Pilot launch** is the optimal decision because it has the highest expected value ($37,500).
> **Exam Tip:** When calculating Expected Value, always pay close attention to negative signs. A "loss" must be treated as a negative number in your summation, or your final EV will be incorrectly inflated.
## Visualizing the Process: Decision Trees
While payoff tables are excellent for simple problems, complex decisions involving multiple stages are often better represented by **Decision Trees**. A decision tree is a visual map of the decision process, showing the various paths, probabilities, and potential outcomes.
By tracing the branches of a tree and calculating the expected values at each "node," decision-makers can visually compare different strategic paths to see which one leads to the best long-term outcome.
![A decision tree comparing the expected monetary value of a pilot launch versus a full delivery service launch. This illustrates risk analysis and strategic decision making.](
## Measuring the Value of Certainty: EVPI
Sometimes, a decision-maker might wonder: *"How much should I pay for better market research?"* To answer this, we use the **Expected Value of Perfect Information (EVPI)**.
EVPI represents the maximum amount a decision-maker should be willing to pay to eliminate all uncertainty regarding the states of nature. It is calculated as the difference between the expected value if we had perfect knowledge and the expected value we currently have:
$$\begin{aligned}
EVPI &= EVwPI - EVwoPI
\end{aligned}$$
$$
Where:
* **EVwPI (Expected Value with Perfect Information)**: The expected value if you knew exactly which state of nature would occur and chose the best possible payoff for that state.
* **EVwoPI (Expected Value without Perfect Information)**: The highest expected value currently available from your existing decision alternatives.
### Worked Example of EVPI
Using our grocery store data:
1. **EVwoPI**: We already calculated this as the highest EV from our alternatives: **$$37,500$** (the Pilot launch).
2. **EVwPI**: If we had perfect information, we would know if demand is strong or weak.
* If we knew demand was strong (45% chance), we would choose the Full Launch ($$300,000$).
* If we knew demand was weak (55% chance), we would choose not to launch anything ($$0$).
$$\begin{aligned}
EVwPI &= (0.45 \times 300,000) + (0.55 \times 0) \\
&= \$135,000
\end{aligned}$$
3. **EVPI**:
$$\begin{aligned}
EVPI &= 135,000 - 37,500 \\
&= \$97,500
\end{aligned}$$
![A blue dart hitting the center of a red and white target with the word TARGET written above. This illustrates goal setting and decision making accuracy.](
This result tells the grocery store that they should never spend more than **$$97,500$** on market research to predict demand, as any cost higher than that would exceed the value the information provides.
## Summary of Key Tools
To master risk modeling, remember how these tools work together:
* **Payoff Tables** organize your data.
* **Expected Value** provides the mathematical "best" choice.
* **Decision Trees** visualize complex paths.
* **EVPI** sets the budget for reducing uncertainty.
