Skip to content
Tutorials Center
My learning

Topics

Lesson 35 of 49 ยท Decision Analysis and Game Theory

Chapter 7 Overview: Decision Analysis and Game Theory

A map of chapter 7: choosing under uncertainty with payoff tables and decision trees, valuing information, and finding optimal strategies against a rival.

Optimization models in earlier chapters assumed that you know everything that matters. Real decisions are made before the facts arrive: you build a factory before you know the demand, sign a contract before you know the price of fuel, or launch a product while a competitor plans its own. This chapter gives you two groups of tools for those situations. Decision analysis handles uncertainty that is nobody's strategy (weather, demand, test results). Game theory handles uncertainty that comes from another decision-maker who reacts to you.

Three cards: against chance, against a rival, buying information
Caption: Three kinds of uncertain decisions, each with its own toolkit.

What you will learn

  • The difference between a decision against chance and a decision against a rival.
  • The vocabulary of payoff tables, states of nature and strategies.
  • How the chapter's lessons fit together.
  • Why information has a value that you can calculate.

Key terms

Alternative (action): one of the options you can choose.

State of nature: an uncontrollable future condition, such as high demand.

Payoff: the result (profit, cost, utility) of an alternative in a state of nature.

Strategy: a complete plan of what a player will do in a game.

Expected monetary value (EMV): the probability-weighted average payoff.

Two kinds of uncertainty

When the uncertainty comes from nature, you choose an alternative, then a state of nature occurs by chance, and you receive a payoff. A shop owner who must choose between a small, a medium and a large kiosk faces this: low, medium or high demand is nobody's strategy. The method is to list payoffs, attach probabilities or a risk attitude, and calculate.

When the uncertainty comes from a rival, the other side picks its action knowing you are thinking too. Two companies setting prices, two armies choosing where to attack, and a goalkeeper facing a penalty taker are games. The right question changes from "what is likely?" to "what is a strategy that does well whatever they do?" In some games the best strategy is to randomize, which is what makes the penalty taker unpredictable.

The big picture

Decisions against chance

Payoff table Options by states
Criteria Maximin, EMV, regret
Trees Roll back the choices
1 / 4

A first calculation

A small business is deciding on the size of a kiosk. The table shows profit in thousands of dollars. Without any theory, you can already compute how an option does on average.

Option Low demand (0.3) Medium demand (0.5) High demand (0.2)
Small 40 50 50
Medium 20 80 90
Large โˆ’30 60 150
Python Average payoff of each option

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The Large kiosk has the best best case, the Small kiosk the best worst case, and the Medium kiosk the best average. That conflict is the reason decision theory exists: "best" depends on how you weigh risk, and the next lessons make the choices explicit.

Where the chapter goes

Lesson What you will learn
Decision trees, criteria and the value of information Maximin, regret, EMV, tree rollback, EVPI, Bayes updating and sensitivity
Zero-sum games Saddle points, dominance, mixed strategies, the minimax theorem and a linear programming solution
Chapter 7 review A recap picture and a ten-question quiz

For more on strategic reasoning, see the full game theory track.

Common misconceptions

"The highest expected value is always the right choice." EMV suits repeated decisions with small stakes. For a one-time bet that could ruin you, risk attitude matters.

"Game theory only applies to competitions." It applies to any interaction where outcomes depend on several decision-makers: auctions, negotiations, traffic, pricing.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Which of these is a state of nature?

    A state of nature is an uncontrollable condition that occurs by chance and is not a decision-maker's strategy.
  2. 2. A payoff table has probabilities 0.5, 0.5 and payoffs 10 and 30 for an option. What is its EMV?

    EMV = 0.5 ร— 10 + 0.5 ร— 30 = 20.
  3. 3. Which situation is best modeled as a game?

    In a game, the other side's choices react to yours.
  4. 4. Why might a manager not pick the highest-EMV option?

    EMV ignores attitude to risk, which matters for large one-off decisions.

Summary

  • Decision analysis handles uncertainty from nature; game theory handles uncertainty from rivals.
  • A payoff table lists options, states and results; criteria and probabilities turn it into a choice.
  • Information has a calculable value, and sensitivity analysis shows how robust a choice is.
  • Optimal play against a rival can require randomizing.

Next: Decision trees, criteria and the value of information