Chapter 3

Decisions under risk (part I)

Amor Messaud
Associate Professor of Statistics

10/03/2016 Decision theory (ISB) 1

Content
1. Introduction
2. Maximum likelihood criterion
3. Expected monetary value (EMV)
4. Expected opportunity loss
5. Decision trees

10/03/2016 Decision theory (ISB) 2

1. Introduction
2. Maximum likelihood criterion
3. Expected monetary value (EMV)
4. Expected opportunity loss
5. Decision trees

10/03/2016 Decision theory (ISB) 3

Decision making under risk
• Probabilistic decision situation
• States of nature have probabilities of occurrence
• Maximum Likelihood Criterion
• Maximize Expected Monitary Value (Bayes
Decision Rule)

10/03/2016 Decision theory (ISB) 4

1. Introduction
2. Maximum likelihood criterion
3. Expected monetary value (EMV)
4. Expected opportunity loss
5. Decision trees

10/03/2016 Decision theory (ISB) 5

Example
• For a special CNC machine, consider three
machine movements: Gears & Levers, Spring
Action, and Weights & Pulleys
• There are three possible states of nature describing
demand level: light demand, moderate demand
and heavy demand.
• The probability of occurrences and corresponding
payoffs are given in the table below

10/03/2016 Decision theory (ISB) 6

Maximum likelihood criterion
• Maximum Likelihood: Identify most likely event,
ignore others, and pick act with greatest payoff.
 Personal decisions are often made that way.
 Collectively, other events may be more likely.
 Ignores lots of information.

10/03/2016 Decision theory (ISB) 7

Example #2: Maximum likelihood
criterion
• A grocer solves a problem of how much pastry to order
every day. His profit depends on a demand that can be
low, moderate, or high.
• Values of the profit per day (in $) for these situations and
for small, medium or large order are shown in Table. The
body of this table represents a decision matrix.
• The maximum likelihood decision criterion ?
P(Low) = 0.3 P(Moderate) = 0.5 P(High) = 0.2
Low Moderate High
Small 50 50 50
Medium 42 52 52
Large 34 44 54

10/03/2016 Decision theory (ISB) 8

Criticisms of the maximum likelihood
criterion
• Since the maximum-likelihood criterion takes only
one uncertain state of nature into account it may
lead to bad decisions
• This criterion chooses an alternative without
considering its payoffs for states of nature other
than the most likely one
• For alternatives that are not chosen, this criterion
ignores their payoffs for states of nature other than
the most likely one

10/03/2016 Decision theory (ISB) 9

Criticisms of the maximum likelihood
criterion
• If the differences in the payoffs for the most likely
state of nature are much less than for another
somewhat likely state of nature, then it might make
sense to focus on this latter state of nature instead
• If there are many states of nature and they are
nearly equally likely, then the probability that the
most likely state of nature will be the true one is
fairly low.

10/03/2016 Decision theory (ISB) 10

1. Introduction
2. Maximum likelihood criterion
3. Expected monetary value (EMV)
4. Expected opportunity loss
5. Decision trees

10/03/2016 Decision theory (ISB) 11

 Bayes Decision Rule

 It is not a perfect criterion because it can lead to the

less preferred choice.
 Consider the Far-Fetched Lottery decision:

ACTS
EVENTS Probability Gamble Don’t Gamble
Head .5 +$10,000 $0
Tail .5 -5,000 0
Would you gamble?

10/03/2016 Decision theory (ISB) 12

 The Far-Fetched Lottery Decision
ACTS
Gamble Don’t Gamble
Proba-
Payoff × Prob. Payoff × Prob
EVENTS bility
Head .5 +$5,000 $0
Tail .5 -2,500 0
Expected Payoff: $2,500 $0
Most people prefer not to gamble!
 That violates the Bayes decision rule.
 But the rule often indicates preferred choices even though
it is not perfect.

10/03/2016 Decision theory (ISB) 13

 Expected Monetary Value (EMV)

N: Number of states of nature

k: Number of alternative decisions
Xij: Value of Payoff for alternative i in state of
nature j, i=1,2,...,k and j=1,2,...,N.
Pj: Probability of state of nature j

EMV ( Ai )   j  1 X ij P j
N

10/03/2016 Decision theory (ISB) 14

Example 1: Marketing a new product

• Assume that a feasibility study at Getz company of

a new product led to encouraging the introduction
of this product to the market
• The management of Getz is not sure whether a
large plant or a small one should be built to
manufacture the product
• The relevant data is presented in the table below

10/03/2016 Decision theory (ISB) 15

 Example 1:

States of Nature
Alternatives Favorable Unfavorable
Market Market P(0.5)
P(0.5)
Construct $200,000 -$180,000
large plant
Construct $100,000 -$20,000
small plant
Do nothing $0 $0

10/03/2016 Decision theory (ISB) 16

 Example 1:

States of Nature
Alternatives Favorable Unfavorable Expected
Market Market P(0.5) value
P(0.5)
Construct $200,000 -$180,000 $10,000
large plant
Construct $100,000 -$20,000 $40,000 Best choice
small plant
Do nothing $0 $0 $0

10/03/2016 Decision theory (ISB) 17

Example 2

Consider the payoff table below with 3 alternatives and 3

states of nature.

Payoff Table
S1 S2 S3
A1 3 3.5 4
A2 -1 6 7
A3 -10 -2 12
Probabilities 0.30 0.60 0.10

10/03/2016 Decision theory (ISB) 18

Example 2

Consider the payoff table below with 3 alternatives and 3

states of nature.

Payoff Table
S1 S2 S3 EV
A1 3 3.5 4 3.40
A2 -1 6 7 4.00
A3 -10 -2 12 -3.00
Probabilities 0.30 0.60 0.10

The best alternative under the EV criterion is A2, with a maximum

EV of 4.00, which is better than the other two EVs.

10/03/2016 Decision theory (ISB) 19

 Example 3:
Burger Prince Restaurant is considering opening a new
restaurant on Main Street. It has three
different models, each with a different
seating capacity. Burger Prince
estimates that the average number of
customers per hour will be 80, 100, or
120. The payoff table for the three
models is on the next slide.

10/03/2016 Decision theory (ISB) 20

Payoff Table

Average Number of Customers Per Hour

s1 = 80 s2 = 100 s3 = 120

Model A $10,000 $15,000 $14,000

Model B $ 8,000 $18,000 $12,000
Model C $ 6,000 $16,000 $21,000

10/03/2016 Decision theory (ISB) 21

Expected Value Approach

Calculate the expected value for each

decision. The decision tree on the next slide can
assist in this calculation. Here d1, d2, d3 represent
the decision alternatives of models A, B, C, and s1,
s2, s3 represent the states of nature of 80, 100, and
120.

10/03/2016 Decision theory (ISB) 22

Expected Value Approach
• Solution Spreadsheet

A B C D E F
1 PAYOFF TABLE
2
3 Decision State of Nature Expected Recommended
4 Alternative s1 = 80 s2 = 100 s3 = 120 Value Decision
5 d1 = Model A 10,000 15,000 14,000 12600
6 d2 = Model B 8,000 18,000 12,000 11600
7 d3 = Model C 6,000 16,000 21,000 14000 d3 = Model C
8 Probability 0.4 0.2 0.4
9 Maximum Expected Value 14000

10/03/2016 Decision theory (ISB) 23

1. Introduction
2. Maximum likelihood criterion
3. Expected monetary value (EMV)
4. Expected opportunity loss
5. Decision trees

10/03/2016 Decision theory (ISB) 24

Expected Opportunity Loss (EOL)

Compute the weighted average of the opportunity losses

for each alternative to yield the EOL.

10/03/2016 Decision theory (ISB) 25

Expected Opportunity Loss (EOL)

Compute the weighted average of the opportunity losses

for each alternative to yield the EOL.
Opportunity Loss (Regret) Table
S1 S2 S3 EOL
A1 0 2.5 8 2.30
A2 4 0 5 1.70
A3 13 8 0 8.70
Probabilities 0.30 0.60 0.10

The best alternative under this criterion is A2, with a minimum

EOL of 1.70, which is better than the other two EOLs.

Note that EV + EOL is constant for each alternative! Why?

10/03/2016 Decision theory (ISB) 26
1. Introduction
2. Maximum likelihood criterion
3. Expected monetary value (EMV)
4. Expected opportunity loss
5. Decision trees

10/03/2016 Decision theory (ISB) 27

Decision tree
• A decision tree is used for visualising a formal
representation of a decision problem graphically. It
is an alternative (Graphical) way to represent and
solve decision problems under risk
• Particularly useful for sequential decisions, i. e.,
decisions that are divided into several separate
steps
• All decision matrices can be converted into
decision trees, but some decision trees (e.g. trees
with more than one choice node) cannot be
converted into decision matrices

10/03/2016 Decision theory (ISB) 28

Decision tree
• Two basic symbols
 (square) = decision or choice node
 (circle) = chance or event node
• At the choice node the decision maker decides
whether to go up or down in the tree. If there are
more than two acts to choose from, one simply
adds more lines.
• At the chance nodes nature decides which line to
follow, and
• the rightmost boxes represent the possible
outcomes

10/03/2016 Decision theory (ISB) 29

Decision tree
• General Procedure:
 Start with a decision node followed by several
branches representing decision alternatives
 Each alternative branch leads to a decision
node or a chance node which is followed by
several branches representing possible states of
nature
 Repeat above steps as necessary until all
scenarios have been considered
 Put all probabilities and payoffs on the tree

10/03/2016 Decision theory (ISB) 30

 Decision Trees
• Graphical display of decision process, i.e.,
alternatives, states of nature, probabilities,
payoffs.
• Decision tables are convenient for problems
with
one set of alternatives and states of nature.
• With several sets of alternatives and states of
nature (sequential decisions), decision trees
are used!
• EMV criterion is the most commonly used
criterion in decision tree analysis.
10/03/2016 Decision theory (ISB) 31
 Softwares for Decision Tree Analysis

• DPL
• Tree Plan
• Supertree

Analysis with less effort.

Full color presentations for managers

10/03/2016 Decision theory (ISB) 32

 Steps of Decision Tree Analysis

• Define the problem

• Structure or draw the decision tree
• Assign probabilities to the states of nature
• Estimate payoffs for each possible
combination of alternatives and states of
nature
• Solve the problem by computing
expected monetary values for each
state-of-nature node

10/03/2016 Decision theory (ISB) 33

 Decision Tree

State 1
Outcome 1
1 State 2
Outcome 2

State 1
Outcome 3
2 State 2
Outcome 4
Decision
Node
State of Nature Node

10/03/2016 Decision theory (ISB) 34

 Ex1:Getz Products Decision Tree
Payoffs
EMV for node 1 = $10,000
Favorable market (0.5)
$200,000
1
Unfavorable market (0.5) -$180,000
Favorable market (0.5) $100,000
Construct
small plant 2
Unfavorable market (0.5) -20,000
EMV for node 2 = $40,000

10/03/2016 Decision theory (ISB) 35

Getz products decision tree

A state of nature node Favorable market

1
Unfavorable market
A decision node Favorable market
Construct
small plant 2
Unfavorable market

10/03/2016 Decision theory (ISB) 36

Decision Tree
Payoffs
s1 .4
10,000
s2 .2
2 s3 15,000
.4
d1 14,000
s1 .4
d2 8,000
1 s2 .2
3 18,000
d3 s3 .4
12,000
s1 .4
6,000
s2 .2
4 16,000
s3
.4
21,000
10/03/2016 Decision theory (ISB) 37
Expected Value for Each Decision

EMV = .4(10,000) + .2(15,000) + .4(14,000)

= $12,600
d1 2
Model A
EMV = .4(8,000) + .2(18,000) + .4(12,000)
Model B d2 = $11,600
1 3

d3 EMV = .4(6,000) + .2(16,000) + .4(21,000)

Model C = $14,000
4

Choose the model with largest EV, Model C.

10/03/2016 Decision theory (ISB) 38

10/03/2016 Decision theory (ISB) 39