The Four Types of Decision Theory
Decision theory is a branch of applied probability theory that is concerned with how we make decisions. The branch of probability theory is concerned with assigning probabilities to various factors and the numerical consequences of an outcome. There are four main types of decision theory: Normative, Expected value, and Savage. Let’s take a look at the characteristics of each type and see how they apply to decision-making. If you’re confused about which type to use, consider reading our introduction to decision theory.
Savage’s decision theory
In order to understand the mechanisms underlying rational choice, it is helpful to first understand the nature of the certain-thing principle, which Savage developed. This principle says that we cannot rationally fear a risk, and so, we must always choose the most likely alternative. However, this principle is not always applicable, as many situations call for a different choice. For example, you might want to avoid an unpleasant situation, even if you are sure that it will bring about a better outcome. The sure-thing principle is a fundamental assumption of Savage’s decision theory, which serves as a basis for determining what is a desirable outcome.
The idea that rationality can be modeled mathematically is central to Savage’s decision theory. His work also helped advance Bayesian statistics and canonical subjective expected utility theory. He also tried to remake all statistical methods along Bayesian lines, including decision-making under uncertainty. Savage’s work embodies the central ideas of action-guiding theory. If we consider the role of probability in rational choice, we can see how subjective expected utility maximization affects human choices.
Expected value theory
The expected utility of a particular action is a weighted average of the utilities of all possible outcomes. Utility is the degree to which a person prefers a certain outcome. The theory describes the relationship between utility and choice and identifies two kinds of arguments for expected utility. There are two types of arguments for expected utility theory: representation theorems and long-run statistics. The theory also addresses objections to its assumptions.
The expected utility theory defines an “object’s utility as the weighted average of all possible levels of its utility at any given point in time.” It explains why individuals would choose an action with the highest expected utility in the presence of uncertainty. If a person does not have much money, the utility of the money may be low. In this case, they may take out insurance. In some cases, the utility of money is low and the utility of insurance is high.
Probability theory is a fundamental mathematical concept that is used by statisticians and business analysts to understand the relationships between two simultaneous events. For example, a share price may indicate the value of a dollar. There are numerous applications of probability theory, from the study of the probability of a person winning the lottery to how the stock market works. These examples highlight some of the benefits and drawbacks of probabilistic models in decision-making.
In business, many managers and economists face uncertainty when making decisions. For example, launching a new product may require weighing market conditions, customer perceptions, and the usefulness of the product in the targeted market. Using probability theory to guide decision-making will help managers choose the most appropriate markets for their new products, and determine the most appropriate time to launch them. While businesses often make subjective decisions, a growing amount of data is enabling managers to make more data-driven decisions.
Normative decision theory
Normative decision theory is concerned with optimal decisions, where optimality is determined by the ideal, rational decision maker. In contrast, prescriptive decision theory focuses on explaining observed behaviors with conceptual models, while descriptive theory aims to analyze how individuals make decisions. In addition, it is an interdisciplinary field, encompassing economists, philosophers, computer scientists, and biologists.
Differences between descriptive and normative decision theories
Descriptive decision theory focuses on the choice behavior of agents in hypothetical, idealized environments. These idealized environments are not typically found in reality, which makes it impossible for social scientists to establish a direct connection between idealized states and real ones. As such, normative decision theory has evolved a constraining role. This theory was the first to study decision-making behavior, but has since been overshadowed by descriptive theory.
Functional decision theory
Essentially, it’s a decision theory, but it has some important distinctions from other decision, such as causal and evidential decision theory. In particular, FDT focuses on choosing the output of a decision function rather than merely the action itself. Its goal is to guide decision makers to the ‘best’ action, regardless of how bad the results might turn out.
According to Yudkowsky and Soares, standard decision theories are based on general arguments and brute intuitions. Moreover, they say, representation theorems reveal that the maximising of expected utility is derived from the basic constraints on rational preference and belief. Therefore, they reject the preference norms of the CDT and offer a representation theorem for FDT. This argument is relevant to the question of which decision theory is best.