Toolkit

Decision trees

A diagram of a decision, its options and their uncertain outcomes, used to calculate the expected value of each option.

SL and HL Verification: Verified In the IB toolkit Updated 29 October 2026

Section 01: What it is for

To compare options whose results are uncertain, using estimated probabilities and payoffs.

Class
Decision-making (IB teacher support material, p. 26)
Suggested units
Unit 1, Unit 5 (guide, p. 14)

When to use it

  • When the outcomes of a decision are uncertain but can be estimated.
  • When options differ in cost and in the spread of their possible results.

How to build it

  1. Draw a square decision node, with a branch for each option labelled with its cost.
  2. At the end of each option, draw a circular chance node with a branch for each possible outcome.
  3. Write the probability and the payoff on each outcome branch, and check that the probabilities at each chance node add up to 1.
  4. Calculate each option's expected value, the sum of probability times payoff, then subtract the option's cost.
  5. Compare the expected values, then weigh the risks and the factors that the numbers leave out.

The evidence it needs

  • Estimates of the probability of each outcome.
  • Estimates of each outcome's payoff and of each option's cost.

Section 02: What it shows, and what it hides

It reveals
  • Which option has the highest expected value on the estimates.
  • How sensitive the ranking is to a change in a probability or a payoff.
It does not reveal
  • The spread of outcomes. Two options with the same expected value can have very different worst cases.
  • Factors that are not in the numbers, such as control, quality or ethics.
  • Whether the estimates are right.

Limitations

  • Probabilities and payoffs are estimates, and are often little more than informed guesses.
  • The expected value is an average over many repetitions, but most strategic decisions are made once.
  • It ignores the timing of payoffs unless they are discounted.

Section 03: Applied: Three of Marigold's routes, from the estimates in Exhibit 8

From the case Marigold Kitchens.

Source: Simulated data

Organic growth: (0.6 × ₹150m) + (0.4 × ₹60m) − ₹72m = ₹42m. Franchising: (0.6 × ₹95m) + (0.4 × ₹30m) − ₹10m = ₹59m. Staying focused: (0.6 × ₹40m) + (0.4 × ₹25m) − ₹3m = ₹31m.

What it showed

On these estimates franchising has the highest expected value. The case evidence on finding franchisees and protecting quality is the reason the decision is still open.

Read the case file: Marigold Kitchens

Interaction

Expected value: organic growth, franchising or staying focused?

Source: Simulated data
1. Predict first

Which route has the highest expected value, and should Meera choose it? Before you calculate, which route do you expect to have the highest expected value?

2. The tree

Decision treeOrganic growth: expected value ₹42m. Franchising: expected value ₹59m. Stay focused: expected value ₹31mOrganic growthCost ₹72mEV ₹42mStrong demand (p = 0.6)₹150mWeak demand (p = 0.4)₹60mFranchisingCost ₹10mEV ₹59mStrong demand (p = 0.6)₹95mWeak demand (p = 0.4)₹30mStay focusedCost ₹3mEV ₹31mStrong demand (p = 0.6)₹40mWeak demand (p = 0.4)₹25m
Tree The square is the decision, each circle a chance node. Payoffs are before the cost of each route. Simulated teaching data.
Expected value of each route
Route Expected payoff Cost Expected value Worst case Best case
Organic growth₹114m₹72m₹42m−₹12m₹78m
Franchising Highest₹69m₹10m₹59m₹20m₹85m
Stay focused₹34m₹3m₹31m₹22m₹37m

Reading. On these estimates franchising has the highest expected value, ₹59m, ₹17m more than organic growth. Organic growth has the highest expected payoff (₹114m) but also costs ₹72m, which is why the cost must be subtracted. Stay focused has the best worst case (₹22m).

Sensitivity. Stay focused leads when the probability of strong demand is below 0.04; franchising leads above 0.04.

Change the estimates
The probability of the other outcome is set to 1 minus this, so they always add up to 1.
Organic growth

Strong demand

Weak demand

Franchising

Strong demand

Weak demand

Stay focused

Strong demand

Weak demand

The workings

Organic growth: expected value ₹42m
Given
  • Strong demand: probability and payoff: 0.6 and ₹150m
  • Weak demand: probability and payoff: 0.4 and ₹60m
  • Cost of organic growth: ₹72m
Formula
Expected value = Σ (probability × payoff) − cost of the option
Substitution
= (0.6 × ₹150m) + (0.4 × ₹60m) − ₹72m
Working
0.6 × 150 = 90
0.4 × 60 = 24
90 + 24 = 114
114 − 72 = 42
Answer
₹42m
Unit
₹ millions, expected value of the option
Interpretation
The average result over many repetitions of this decision, after the cost of the route.
Business implication
Compare it with the other routes, then with their risks and with the evidence the numbers leave out.
Franchising: expected value ₹59m
Given
  • Strong demand: probability and payoff: 0.6 and ₹95m
  • Weak demand: probability and payoff: 0.4 and ₹30m
  • Cost of franchising: ₹10m
Formula
Expected value = Σ (probability × payoff) − cost of the option
Substitution
= (0.6 × ₹95m) + (0.4 × ₹30m) − ₹10m
Working
0.6 × 95 = 57
0.4 × 30 = 12
57 + 12 = 69
69 − 10 = 59
Answer
₹59m
Unit
₹ millions, expected value of the option
Interpretation
The average result over many repetitions of this decision, after the cost of the route.
Business implication
Compare it with the other routes, then with their risks and with the evidence the numbers leave out.
Stay focused: expected value ₹31m
Given
  • Strong demand: probability and payoff: 0.6 and ₹40m
  • Weak demand: probability and payoff: 0.4 and ₹25m
  • Cost of stay focused: ₹3m
Formula
Expected value = Σ (probability × payoff) − cost of the option
Substitution
= (0.6 × ₹40m) + (0.4 × ₹25m) − ₹3m
Working
0.6 × 40 = 24
0.4 × 25 = 10
24 + 10 = 34
34 − 3 = 31
Answer
₹31m
Unit
₹ millions, expected value of the option
Interpretation
The average result over many repetitions of this decision, after the cost of the route.
Business implication
Compare it with the other routes, then with their risks and with the evidence the numbers leave out.

Section 04: The calculation

Decisions

Expected value (decision trees)

SL and HL Verification: Verified IB 1.5

Formula

Expected value = Σ (probability of outcome × payoff of outcome) − cost of the option

p
Probability of an outcome (between 0 and 1)
x
Payoff of the outcome (currency)

Unit. currency

The average result over many repetitions of the same decision. It does not show the spread of outcomes or the risk of the worst one.

Worked example: from the case Marigold Kitchens

Marigold Kitchens: the expected value of franchising

Given
  • Strong demand: probability and payoff: 0.6 and ₹95m
  • Weak demand: probability and payoff: 0.4 and ₹30m
  • Cost of franchising: ₹10m
Formula
Expected value = Σ (probability × payoff) − cost of the option
Substitution
= (0.6 × ₹95m) + (0.4 × ₹30m) − ₹10m
Working
0.6 × 95 = 57
0.4 × 30 = 12
57 + 12 = 69
69 − 10 = 59
Answer
₹59m
Unit
₹ millions, expected value of the option
Interpretation

On these estimates franchising is worth ₹59m on average, after its ₹10m cost.

Business implication

It has the highest expected value of the three routes, but the decision also depends on control, quality and the founder's aims, which the number leaves out.

Errors that cost marks
  • Probabilities at a chance node that do not add up to 1.
  • Forgetting to subtract the cost of the option.

1.5: Decision trees are a toolkit tool, used here on a growth decision

Section 05: In the exam

Decision tree calculations reward clear working: each expected value set out, the cost subtracted, and a comment on what the result does and does not show.

Common errors

  • Forgetting to subtract the cost of an option.
  • Probabilities at a chance node that do not add up to 1.
  • Recommending the option with the highest expected value without discussing risk or non-financial factors.