An Introduction to

Decision Theory
Chapter 20

McGraw-Hill/Irwin Copyright © 2012 by The McGraw-Hill Companies, Inc. All rights reserved.
Learning Objectives
LO1 Identify and apply the three components of a
decision.
LO2 Compute and interpret the expected values
for a payoff table.
LO3 Explain and interpret opportunity loss.
LO4 Describe three strategies for decision making.
LO5 Compute and describe the expected value of
information.
LO6 Organize possible outcomes into a decision
tree and interpret the result.
20-2
 LO1 Identify and apply the three
components of a decision.

Statistical Decision Theory

 Classical statistics focuses on estimating a parameter, such as the population
mean, constructing confidence intervals, or hypothesis testing.

 Statistical Decision Theory (Bayesian statistics) is concerned with determining

which decision, from a set of possible decisions, is optimal.

Three components to any decision-making situation:

1. The available choices (alternatives or acts).

2. The states of nature, which are not under the control of the decision maker -
uncontrollable future events.
3. The payoffs - needed for each combination of decision alternative and state of
nature.

 A Payoff Table is a listing of all possible combinations of decision alternatives and

states of nature.

 The Expected Payoff or the Expected Monetary Value (EMV) is the expected value
for each decision.

20-3
 LO1

Decision Making

20-4
 LO2 Compute and interpret the
expected values for a payoff table.

Calculating the EMV

EMV ( Ai )  [ P( S j ) V ( Ai , S j )

 Let Ai be the ith decision alternative.

 Let P(Sj) be the probability of the jth state of nature.
 Let V(Ai, Sj) be the value of the payoff for the
combination of decision alternative Ai and state of
nature Sj.
 Let EMV (Ai) be the expected monetary value for
the decision alternative Ai.

20-5
 LO2
Decision Making Under Conditions of
Uncertainty - Example
EXAMPLE
Bob Hill, a small investor, has $1,100 to
invest. He has studied several common
stocks and narrowed his choices to (.60) (.40)
three, namely, Kayser Chemicals, Rim
Homes, and Texas Electronics. He
estimated that, if his $1,100 were
invested in Kayser Chemicals and a
strong bull market developed by the end
of the year (that is, stock prices
increased drastically), the value of his
Kayser stock would more than double, to
$2,400. However, if there were a bear
market (i.e., stock prices declined), the
value of his Kayser stock could
conceivably drop to $1,000 by the end of Calculation
the year. His predictions regarding the
value of his $1,100 investment for the (A1)=(.6)($2,400)+(.4)($1,000) =$1,840
three stocks for a bull market and for a
bear market are shown below. A study of
historical records revealed that during (A2)=(.6)($2,400)+(.4)($1,000) =$1,760
the past 10 years stock market prices
increased six times and declined only
four times. According to this information, (A3)=(.6)($2,400)+(.4)($1,000) =$1,600
the probability of a market rise is .60 and
the probability of a market decline is .40.

20-6
 LO3 Explain and interpret opportunity loss table.

Opportunity Loss

Opportunity Loss or Regret is the loss because the exact state of nature is not
known at the time a decision is made.

 The opportunity loss is computed by taking the difference between the optimal
decision for each state of nature and the other decision alternatives.

EOL( Ai )  [ P( S j )  R( Ai , S j )

20-7
 LO3

Opportunity Loss EOL( Ai )   [ P( S j )  R( Ai , S j )

Opportunity Loss when Market Rises Opportunity Loss when Market Declines

Kayser: Kayser:
$2,400 - $2,400= $0 $1,150 - $1,000= $150

Rim Homes: Rim Homes:

$2,400 - $2,200 = $200 $1,150 - $1,100 = $50

Texas Electronics: Texas Electronics:

$2,400 - $1,900 = $500 $1,150 - $1,150 = $0
20-8
 LO3

Expected Opportunity Loss

EOL( Ai )   [ P( S j )  R( Ai , S j )

0.60 0.40

(A1)=(.6)($0)+(.4)($150) =$60

(A2)=(.6)($200)+(.4)($50) =$140

(A3)=(.6)($500)+(.4)($0) =$300
20-9
 LO4 Describe three strategies for decision making.

Maximin, Maximax, and Minimax Regret Strategies

Maximin strategy maximizes the minimum gain. It is a pessimistic strategy.

Maximax strategy maximizes the maximum gain. Opposite of a maximin approach, it is an optimistic strategy

Minimax regret strategy minimizes the maximum regret (opportunity loss). This is another pessimistic strategy

Payoff Table

Maximin Maximax
1,000 2,400
1,100 2,200
1,150 1,900

Opportunity Loss Table

Minimax
Regret
150
200
500
20-10
 LO5 Compute and describe the expected value of
information.

Value of Perfect Information

What is the worth of information known in advance before a strategy
is employed?

Expected Value of Perfect Information (EVPI) is the difference

between the expected payoff if the state of nature were known and
the optimal decision under the conditions of uncertainty.

20-11
 LO5
Value of Perfect
Information
Step 1: Compute the Expected Value Under Certainty

Expected Value Under Certainty

Step 2: Compute the Expected Value Under Uncertainty Expected Value Under Uncertainty

(.60) (.40)

Step 3: Subtract the Expected Value Under Uncertainty from

the Expected Value Under Certainty

20-12
 LO6 Organize possible outcomes into a
decision tree and interpret the result.

Sensitivity Analysis and Decision Trees

Sensitivity Analysis examines the effects of various probabilities for

the states of nature on the expected values for the decision
alternatives.

Decision Trees are useful for structuring the various alternatives. A

decision tree is a picture of all the possible courses of action and the
consequent possible outcomes.
 A box is used to indicate the point at which a decision must be
made,
 The branches going out from the box indicate the alternatives
under consideration

20-13
 LO6

Decision Trees

$2,400

20-14