In practice, we find situations where it is not possible to make any probability assessment. What criterion can be used in decision-making situations where the probabilities of outcomes are unknown?

 Q. In practice, we find situations where it is not possible to make any probability assessment. What criterion can be used in decision-making situations where the probabilities of outcomes are unknown?

In decision-making situations where the probabilities of outcomes are unknown, several decision-making criteria can be employed to guide rational choices. These situations are common in real-world contexts, such as in business, economics, healthcare, and other complex environments, where uncertainty and incomplete information are prevalent. Traditional decision theory often relies on probability assessments to determine the optimal course of action based on expected outcomes. However, in the absence of probabilistic information, decision-makers need alternative methods to navigate uncertainty. The following discussion explores several decision-making criteria that can be effectively used when the probabilities of outcomes are unknown.

1. Maximin (or "Worst-case Scenario") Criterion

The maximin criterion is a conservative approach to decision-making under uncertainty. It involves selecting the decision alternative that maximizes the minimum payoff. Essentially, the decision-maker considers the worst possible outcome for each alternative and then chooses the alternative with the best of these worst-case outcomes. This criterion is particularly useful when the decision-maker is risk-averse and prefers to avoid the worst-case scenario at all costs.

For example, in a business context, if a company is considering launching a new product but is uncertain about the market's response, the maximin approach would involve analyzing the worst possible profit or loss for each product and choosing the one that offers the least bad outcome. This approach minimizes potential regret by focusing on securing the best possible worst-case scenario.

Application:

• A company evaluating different investment opportunities might look at the worst-case scenario for each potential investment (e.g., total loss of capital) and select the one that minimizes the loss.
• In a healthcare setting, the maximin criterion might be used by a doctor considering treatment options for a patient with an uncertain prognosis, choosing the treatment with the least harmful possible outcome if the treatment does not succeed.

2. Maximax (or "Optimistic") Criterion

The maximax criterion represents an optimistic approach to decision-making under uncertainty. In contrast to maximin, maximax focuses on the maximum possible payoff rather than minimizing the worst-case scenario. The decision-maker selects the alternative that offers the highest potential reward, assuming that the best-case scenario will occur. This approach is typically used by decision-makers who are more risk-seeking or optimistic about the outcomes of their choices.

For instance, in a startup environment, an entrepreneur might apply the maximax criterion when considering several product ideas. Even though the probability of success for each product idea is uncertain, the entrepreneur would focus on the potential for the highest return if the product succeeds. This is common in venture capital, where investors are willing to take on significant risk in exchange for the possibility of large returns.

Application:

• An entrepreneur choosing between multiple business ventures may use the maximax criterion by selecting the one that could potentially generate the greatest profit, even if the probability of that outcome is low.
• In gambling or investing in high-risk stocks, a decision-maker might focus on the best possible outcome, hoping for a high reward, despite the uncertainty.

3. Laplace (or "Equally Likely") Criterion

The Laplace criterion assumes that each possible outcome is equally likely, and therefore, it treats the uncertain situation as if the probabilities of all outcomes are identical. Under this criterion, the decision-maker calculates the expected payoff for each alternative by averaging the payoffs across all possible outcomes. This method is useful when the decision-maker has no reason to believe that some outcomes are more likely than others, and no additional information is available to differentiate between them.

For example, if a company has no information about customer preferences and is choosing between different marketing strategies, the Laplace criterion could be applied. Each strategy would be evaluated based on the average payoff, assuming each possible market response is equally probable.

Application:

• In situations where a person has no prior knowledge of the likelihood of various outcomes (e.g., launching a product in an unknown market), the Laplace criterion provides a simple, rational approach to decision-making by assuming that all outcomes are equally likely.
• A political candidate deciding between various policy platforms with no prior data on which one would be most popular might use the Laplace criterion to evaluate the average potential benefit of each policy.

4. Minimax Regret Criterion

The minimax regret criterion is designed to minimize the potential regret a decision-maker might feel after making a decision. Regret in this context refers to the difference between the outcome of the chosen alternative and the best possible outcome that could have been achieved if the decision-maker had perfect foresight. The minimax regret criterion seeks to minimize the maximum possible regret by choosing the alternative that results in the least amount of regret in the worst-case scenario.

This approach is particularly useful in situations where decision-makers are concerned with future regret, particularly when the outcome of a decision may result in feelings of missed opportunities or poor choices. The decision-maker first calculates the regret associated with each alternative for each possible outcome, and then selects the alternative with the smallest worst-case regret.

Application:

• A company considering multiple product designs may use the minimax regret criterion by evaluating how much regret they would feel if they chose a particular design and the other designs turned out to be more successful. The design that results in the least regret in the worst-case scenario would be chosen.
• In healthcare, a doctor might use the minimax regret criterion when choosing between different treatments with uncertain outcomes. The doctor would select the treatment that minimizes the regret of choosing incorrectly if the outcome turns out poorly.

5. Hurwicz Criterion (or Weighted Average Criterion)

The Hurwicz criterion is a compromise between the optimistic maximax approach and the conservative maximin approach. It involves assigning a weight to the best possible outcome (optimistic scenario) and a weight to the worst possible outcome (pessimistic scenario). The decision-maker then calculates a weighted average of the outcomes, combining the best-case and worst-case scenarios according to the assigned weights. The Hurwicz criterion is particularly useful when the decision-maker wants to consider both the potential rewards and risks of their choices.

For example, if a company is deciding whether to invest in a new technology, the Hurwicz criterion allows the company to assign more weight to the optimistic outcome (successful implementation) and less weight to the worst-case scenario (failure), reflecting their level of optimism or risk tolerance.

Application:

• An entrepreneur deciding between two business opportunities could apply the Hurwicz criterion by assigning a weight to the potential success and failure outcomes for each opportunity, then calculating the weighted average to make the best decision.
• A medical researcher evaluating different treatment options for a disease might use the Hurwicz criterion to balance the risks and rewards of each option, depending on their risk tolerance.

6. Savage Criterion

The Savage criterion, also known as the minimax regret criterion, focuses on minimizing the maximum regret, or the maximum difference between the payoff of the chosen decision and the best possible payoff. It involves determining the opportunity loss (regret) associated with each decision and choosing the decision that minimizes the worst-case regret. This is an effective criterion in situations where a decision-maker is uncertain about the probabilities of various outcomes but wants to minimize the worst-case emotional or financial regret from making a suboptimal decision.

Application:

• A business that is evaluating different strategic options may use the Savage criterion to select the option that minimizes the potential for regret by considering the opportunity losses involved with each choice.
• A policymaker choosing between various public policies under uncertainty may apply the Savage criterion to avoid the worst possible regret by minimizing the opportunity loss associated with any decision.

7. The Principle of Insufficient Reason (or Principle of Indifference)

The principle of insufficient reason, also known as the principle of indifference, is based on the assumption that when no information is available to suggest that one outcome is more likely than another, all outcomes should be considered equally likely. This is a simple and pragmatic approach that is often used when a decision-maker has no prior knowledge about the probabilities of different outcomes and no reliable way of estimating them.

Application:

• A business considering whether to enter a new market without any market data might use the principle of insufficient reason to treat each potential market outcome as equally likely and evaluate each option accordingly.
• An individual choosing between different vacation destinations without any knowledge about the weather or costs at each location might apply the principle of insufficient reason to make an unbiased decision.

8. Decision Trees and Payoff Matrices

While decision trees and payoff matrices are often used in situations where probabilities are known, they can also be useful in situations where probabilities are unknown. In such cases, decision trees help visualize the various alternatives, possible outcomes, and the sequence of decisions. The decision-maker can use qualitative assessments or subjective estimates to evaluate the potential payoffs, even in the absence of concrete probabilities.

For example, a company considering multiple product designs might use a decision tree to map out the potential outcomes of each design, along with subjective estimates of success or failure, and then assess which design is most likely to achieve the desired results.

Conclusion

When probabilities of outcomes are unknown, decision-makers can apply various decision-making criteria, each of which has its advantages and limitations. The maximin and maximax criteria are appropriate for risk-averse or risk-seeking decision-makers, respectively, while the Laplace criterion is useful when all outcomes are assumed to be equally likely. The minimax regret criterion helps minimize potential regret by selecting alternatives that avoid the worst possible emotional or financial losses. The Hurwicz criterion provides a balance between optimism and pessimism, while the Savage criterion focuses on minimizing opportunity losses. The principle of insufficient reason applies when there is no reason to favor any particular outcome over another, and decision trees and payoff matrices can also be adapted to handle uncertainty without known probabilities.

Ultimately, the choice of criterion depends on the decision-maker’s attitude toward risk, the specific nature of the decision, and the available information.

• Tweet
• Share
• Share
• Share