Understanding Decision Trees

In some projects and assets, risk is not only discrete but also sequential. For the asset to have value, it has to pass through a series of tests, with failure at any point potentially translating into a complete loss of value. Decision trees allow us not only to consider the risk in stages but also to devise the right response to outcomes at each stage.

Steps in Decision Tree Analysis

1. Distinguish among root nodes, decision nodes, event nodes, and end nodes

Root node represents the start of the decision tree, where a decision maker can be faced with a decision choice or an uncertain outcome. The objective of the exercise is to evaluate what a risky investment is worth at this node

Event nodes represent the possible outcomes on a risky gamble; We have to figure out the possible outcomes and the probabilities of the outcomes occurring, based on the information we have available today.

Decision nodes represent choices that the decision maker can make—to expand from a test market to a national market after a test market’s outcome is known

End nodes usually represent the outcomes of earlier risky outcomes and decisions made in response

2. Divide analysis into risk phases
3. Estimate the probabilities of the outcomes in each phase
4. Define decision points
5. Compute cash flows/value at end nodes
6. Fold back the tree

Use in Decision Making

1. Dynamic response to risk
2. Value of information
3. Risk Management

Flexible and powerful approach for dealing with risk that occurs in phases, with decisions in each phase depending on outcomes in the previous one. In addition to providing us with measures of risk exposure, they force us to think through how we will react to both adverse and positive outcomes that may occur at each phase.