Terms to model decisions

What if the decisions that define your life—from choosing a career to launching a satellite—could be entirely dismantled into the same mathematical skeleton as a game of tic-tac-toe? In his foundational work *Theory of Games and Economic Behavior*, [John von Neumann and Oskar Morgenstern](https://en.wikipedia.org/wiki/Theory_of_Games_and_Economic_Behavior) revolutionized how we understand human choice. They demonstrated that any decision, no matter how complex or emotionally charged, can be modeled using a precise, universal lexicon. By translating vague human desires into rigorous mathematical terms, decision theory allows us to map the invisible forces guiding our lives. ## The Core Lexicon of Decision Modeling To model a decision is to strip away psychological noise and isolate its structural components. Practitioners rely on five core terms: 1. **Acts (Alternatives):** The set of mutually exclusive actions available to the decision-maker. In formal modeling, this is represented as a set where the agent has complete agency over which element to select. 2. **States of Nature:** External, uncontrollable scenarios that determine the consequences of an act. As [Leonard J. Savage](https://en.wikipedia.org/wiki/Leonard_J._Savage) articulated in his seminal 1954 book *The Foundations of Statistics*, these states must be mutually exclusive and exhaustive, representing every possible way the world might look. 3. **Outcomes (Consequences):** The specific result of a chosen Act colliding with a realized State of Nature. 4. **Utility:** A numerical value assigned to an outcome, representing its subjective desirability. Under the Von Neumann-Morgenstern utility theorem, if an agent's preferences obey specific axioms (such as transitivity and continuity), their choices can be modeled as maximizing expected utility. 5. **Probability Distribution:** The credences or objective likelihoods assigned to each State of Nature. Savage famously illustrated this structural anatomy with the simple act of preparing an omelet: > "Your good friend has just broken five good eggs into a bowl when you come upon him... A sixth egg, which is unbroken, also has to be used... The acts are: 1. Break it into the bowl... 2. Break it into a saucer... 3. Throw it away. The states are: Good, and Rotten. The consequences are: a six-egg omelet, a ruined five-egg omelet, or a five-egg omelet and a wasted egg." By mapping these variables, we convert intuitive guesswork into a systematic calculation of **Expected Value**—the sum of all possible utilities weighted by their probability of occurrence. ## Deepening the Inquiry To push beyond the basics of classical decision theory, consider these profound systemic challenges: - **The Problem of Bounded Rationality:** How do decision models adapt when agents face cognitive limitations, as pioneered by Herbert Simon's concept of ["satisficing"](https://plato.stanford.edu/entries/bounded-rationality/)? - **Ambiguity vs. Risk:** In a true crisis, we face "Knightian uncertainty"—where the probabilities of the States of Nature are themselves entirely unknown. How do we model decisions when we do not even know the odds? - **The Newcomb's Paradox Challenge:** What happens to decision models when our choices seemingly influence the probability of the States of Nature, forcing a conflict between Evidential and Causal Decision Theory?