Block 7: Decision-making under uncertainty using

decision trees
There is an introduction video on the VLE, you can access it here:
https://emfssvideo.s3.amazonaws.com/MT%26ST/ST2187/ST2187_Block7_Intro.mp4
This block provides a formal framework for analysing decision problems which involve
uncertainty. Specifically, the following topics are included:
 criteria for choosing among alternative decisions
 how probabilities are used in the decision-making process
 how early decisions affect decisions made at a later stage
 how a decision-maker can quantify the value of information
 how attitudes toward risk can affect the analysis.
A powerful graphical tool which guides that analysis is a decision tree. A decision tree
enables a decision-maker to view all important aspects of the problem at once - the decision
alternatives, the uncertain outcomes and their probabilities, the economic consequences, and
the chronological order of events.
After completing this block, you should be able to:
 analyse decision-making under uncertainty using decision trees.

Readings
Albright, S and Winston, W.L, Business Analytics Data Analysis & Decision Making,
(Cengage Learning, 2017) 6th edition [ISBN 9781305947542] Chapter 6.

Decision making under uncertainty using decision trees

Elements of decision analysis

Although decision-making under uncertainty occurs in a wide variety of contexts, all
problems have three common elements:
 the set of decisions (or strategies) available to the decision-maker
 the set of possible outcomes and the probabilities of these outcomes
 a value model which prescribes monetary values for the various decision-outcome
combinations.
Once these elements are known, the decision-maker can find an optimal decision, depending
on the optimality criterion chosen.
A payoff table lists the payoff for each decision-outcome pair. Positive values correspond to
rewards (or gains), and negative values correspond to costs (or losses).
The maximin criterion finds the worst payoff in each row of the payoff table and chooses
the decision corresponding to the best of these.
The maximax criterion finds the best payoff in each row of the payoff table and chooses the
decision corresponding to the best of these.
The expected monetary value, or EMV, for any decision is a weighted average of the
possible payoffs for this decision, weighted by the probabilities of the outcomes.
Some of the quantities in a decision analysis, particularly the probabilities, are often
intelligent guesses at best. Therefore, it is important, especially in real-world business
problems, to accompany any decision analysis with a sensitivity analysis. In a sensitivity
analysis, we systematically vary inputs to the problem to see how (or if) the outputs - the
EMVs and the best decision - change.

Decision trees
Decision trees are composed of nodes (circles, squares, and triangles) and branches (lines).
The nodes represent points in time. A decision node (a square) represents a time when the
decision-maker makes a decision. A probability node (a circle) represents a time when the
result of an uncertain outcome becomes known. An end node (a triangle) indicates that the
problem is completed - all decisions have been made, all uncertainty has been resolved, and
all payoffs and costs have been incurred.
Time proceeds from left to right. This means that any branches leading into a node (from the
left) have already occurred. Branches leading out of a decision node represent the possible
decisions; the decision-maker can choose the preferred branch.
Probabilities are listed on probability branches. These probabilities are conditional on the
events which have already been observed (those to the left).
Monetary values are shown to the right of the end nodes. EMVs are calculated through a
‘folding back’ process. Starting from the right of the decision tree and working back to the
left:
 at each probability node, calculate an EMV - a sum of products of monetary values
and probabilities
 at each decision node, take a maximum of EMVs to identify the optimal decision.
The risk profile for a decision is a ‘spike’ chart which represents the probability distribution
of monetary outcomes for this decision. By looking at the risk profile for a particular
decision, you can see the risks and rewards involved. By comparing risk profiles for different
decisions, you can gain more insight into their relative strengths and weaknesses.

Bayes’ rule
In a multistage decision tree, all probability branches at the right of the tree are conditional on
outcomes which have occurred earlier, to their left. Therefore, the probabilities on these
branches are of the form 𝑃(𝐴 | 𝐵), read as ‘A given B, where A is an event corresponding to a
current probability branch, and B is an event which occurs before event A in time.
When gathering data for the problem, it is sometimes more natural to assess conditional
probabilities in the opposite order, that is, 𝑃(𝐵 | 𝐴). Whenever this is the case, Bayes’
rule must be used to obtain the probabilities needed on the tree.
To develop Bayes’ rule, let 𝐴1 , … , 𝐴𝑛 An be any outcomes. Without any further information,
we believe the probabilities of the 𝐴𝑖 s, for 𝑖 = 1, … , 𝑛, are 𝑃(𝐴1 ), … , 𝑃(𝐴𝑛 ). These are
called prior probabilities. Because an information outcome might influence our thinking
about the probabilities of the 𝐴𝑖 s, we need to find the conditional probability 𝑃(𝐴𝑖 | 𝐵), for
each outcome 𝐴𝑖 . This is called the posterior probability of 𝐴𝑖 .
Bayes’ theorem is:
𝑃(𝐵 | 𝐴𝑖 ) 𝑃(𝐴𝑖 )
𝑃(𝐴𝑖 | 𝐵) = .
𝑃(𝐵 | 𝐴1 ) 𝑃(𝐴1 ) + ⋯ + 𝑃(𝐵 | 𝐴𝑛 ) 𝑃(𝐴𝑛 )
In words, Bayes’ rule says that the posterior is the likelihood times the prior, divided by a
sum of likelihoods times priors.
As a side benefit, the denominator in Bayes’ rule is also useful in multistage decision trees. It
is the probability 𝑃(𝐵) of the information outcome. This formula is important in its own
right. For B to occur, it must occur along with one of the 𝐴𝑖 s. The denominator of the formula
simply decomposes the probability of B into all of these possibilities. It is sometimes called
the law of probability.

The value of information

In a multistage decision problem, the first-stage decision is whether to purchase information
which will help make a better second stage decision. In this case the information, if obtained,
typically changes the probabilities of later outcomes. To revise the probabilities once the
information is obtained, you often need to apply Bayes’ rule. In addition, you typically want
to learn how much the information is worth. After all, information usually comes at a
price(!), so you want to know whether the information is worth its price. This leads to an
investigation of the value of information.
How much is the information worth, and if it costs a given amount, should you purchase it?
Presumably, information which will help you make your ultimate decision should be worth
something, but it might not be clear how much the information is worth. In addition, even if
the information is worth something, it might not be worth as much as its actual price.
Sample information is the information from the experiment itself. Perfect information is
information from a perfect test - that is, a test which will indicate with certainty which
ultimate outcome will occur.
The expected value of sample information, or EVSI, is your maximum willingness to
pay for the sample information. It is calculated as:
EVSI = EMV with (free) sample information – EMV without information.
The expected value of perfect information, or EVPI, is your maximum willingness to
pay for perfect information. The amount you should be willing to spend for information is the
expected increase in EMV you can obtain from having the information. If the actual price of
the information is less than or equal to this amount, you should purchase it; otherwise, the
information is not worth its price.
Incorporating attitudes toward risk
Rational decision-makers are sometimes willing to violate the EMV maximisation criterion
when large amounts of money are at stake. These decision-makers are willing to sacrifice
some EMV to reduce risk. The reason for this action has been studied extensively by many
researchers, both mathematically and behaviourally.
Although there is still not perfect agreement, most researchers believe that if certain basic
behavioural assumptions hold, people are expected utility maximisers - that is, they choose
the alternative with the largest expected utility.
A utility function is a mathematical function which transforms monetary values - payoffs
and costs - into utility values. Essentially, an individual’s utility function specifies the
individual’s preferences for various monetary payoffs and costs and, in doing so, it
automatically encodes the individual’s attitudes toward risk.
Most individuals are risk-averse, which means intuitively that they are willing to sacrifice
some EMV to avoid risky gambles.
Utility assessment is tedious and is not always straightforward. Classes of ready-made utility
functions have been developed to help assess the utilities of decision-makers. An exponential
utility function has only one adjustable numerical parameter, called the risk tolerance, and
there are straightforward ways to discover an appropriate value of this parameter for a
particular individual or company. The advantage of using an exponential utility function is
that it is relatively easy to assess. The risk tolerance for an exponential utility function is a
single number which specifies an individual’s aversion to risk.

Conclusion
The most important skill to gain from this block is the ability to approach decision problems
with uncertainty in a systematic manner.
This systematic approach requires you to:
 list all possible decisions or strategies
 list all possible uncertain outcomes
 assess the probabilities of these outcomes (possibly with the aid of Bayes’ rule)
 calculate all necessary monetary values
 do the necessary calculations to obtain the best decision.

Excel self-study examples

SciTools bidding decision

The file SciTools_bidding_decision.xlsx develops a decision model that finds the EMV for
various bidding strategies and indicates the best bidding strategy.
It is necessary to examine the three elements of the SciTools problem:
 submit a bid or not
  uncertain outcomes and their probabilities
 value model which transforms decisions and outcomes into monetary values for
SciTools.

Baynes’ rule
The file Bayes’_rule.xlsx uses Bayes’ rule to revise the probability of being a drug user, given
the positive or negative results of the test.
Let D and ND denote that a randomly chosen athlete is or is not a drug user, and let 𝑇 + and
𝑇 − indicate a positive or negative test result. Using Bayes’ rule, you can calculate probability
of the events 𝑃(𝐷 | 𝑇 + ) and 𝑃(𝑁𝐷 | 𝑇 − ).

Drug testing decision

The file Drug_testing_decision.xlsx uses a multistage decision framework to see whether
mandatory drug testing can be justified, given a somewhat unreliable test and a set
of ‘reasonable’ monetary values.
Assume that there are only two alternatives: perform drug testing on all athletes or do not
perform any drug testing. The decision model is straightforward.

Acme marketing decisions

The file Acme_marketing_decisions.xlsx develops a decision tree to find the best strategy for
Acme, to perform a sensitivity analysis on the results, and to find EVSI and EVPI.
Identify the three basic elements of this decision problem:
 the possible strategies
 the possible outcomes and their probabilities
 the value model.
In a contingency plan, later decisions can depend on earlier decisions and information
received. Bayes’ rule is implemented in this example exactly the same way as in the drug
example.
The role of the test market is to provide information in the form of more accurate
probabilities of national market results. Information usually costs something, as it does in
Acme’s problem. You might ask how much the test market is really worth. Intuitively,
running the test market is worth something because it changes the optimal decision. With no
test market information, the best decision is to market nationally.
This is what makes information worth something - its outcome affects the optimal decision.

Exponential utility
The file Exponential_utility.xlsx looks at how a company’s risk averseness, determined by its
risk tolerance in an exponential utility function, affects its decision. Assume that Venture
Limited has an exponential utility function.
Block 7 learning activities

Conceptual questions on the block’s topics

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&forceview=1
1. Your company needs to make an important decision which involves large monetary
consequences. You have listed all of the possible outcomes and the monetary payoffs
and costs from all outcomes and all potential decisions. You want to use the EMV
criterion, but you realise that this requires probabilities and you see no way to find the
required probabilities. What can you do?
2. If your company makes a particular decision in the face of uncertainty, you estimate
that it will either gain £10,000, gain £1,000, or lose £5,000, with probabilities 0.4, 0.3
and 0.3, respectively. You (correctly) calculate the EMV as £2,800. However, you
distrust the use of this EMV for decision-making purposes. After all, you reason that
you will never receive £2,800; you will receive £10,000, £1,000, or lose £5,000.
Discuss this reasoning.
3. In the previous question, suppose you have the option of receiving £2,700 in cash
instead of making the risky decision described. Would you take the risky decision,
where you could lose £5,000, or would you take the sure £2,700? What would
influence your decision?
4. In a classic oil-drilling example, you are trying to decide whether to drill for oil in a
field which might or might not contain any oil. Before making this decision, you have
the option of hiring a geologist to perform some seismic tests and then predict
whether there is any oil or not. You assess that if there is actually oil, the geologist
will predict there is oil with probability 0.85. You also assess that if there is no oil, the
geologist will predict there is no oil with probability 0.90. Why will these two
probabilities not appear on the decision tree? Which probabilities will be on the
decision tree?
5. Your company has signed a contract with a good customer to ship the customer an
order no later than 20 days from now. The contract indicates that the customer will
accept the order even if it is late, but instead of paying the full price of £10,000, it will
be allowed to pay 10% less, i.e. £9,000, due to lateness. You estimate that it will take
anywhere from 17 to 22 days to ship the order, and each of these is equally likely.
You believe you are in good shape, reasoning that the expected days to ship is the
average of 17 through 22, or 19.5 days. Because this is less than 20, you will get your
full £10,000. What is wrong with your reasoning?
6. You must make one of two decisions, each with possible gains and possible losses.
One of these decisions is much riskier than the other, having much larger possible
gains but also much larger possible losses, and it has a larger EMV than the safer
decision. Because you are risk-averse and the monetary values are large relative to
your wealth, you base your decision on expected utility, and it indicates that the
certainty equivalent for the risky decision is £210,000, whereas its EMV is £540,000.
What do these two numbers mean? What do you know about the certainty equivalent
of the safer decision?
 7. A potentially large hurricane is forming offshore, and there is some chance that it
might make landfall in a city where you are in charge of emergency preparedness.
You have made plans for evacuating everyone from the city, but such an evacuation is
obviously costly and upsetting for all involved, so the decision to evacuate should not
be made lightly. Discuss how you would make such a decision. Is EMV a relevant
concept in this situation? How would you evaluate the consequences of uncertain
outcomes?
8. It seems obvious that if you can purchase information before making an ultimate
decision, this information should generally be worth something, but explain exactly
why (and when) it is sometimes worth nothing.
9. Insurance companies would not exist unless customers were willing to pay the price
of the insurance and the insurance companies were making a profit. So explain how
insurance is a win-win proposition for customers and the company.
10. You often hear about the trade-off between risk and reward. Is this trade-off part of
decision-making under uncertainty when the decision-maker uses the EMV criterion?
For example, how does this work in investment decisions?

Spreadsheet solutions to the end-of-chapter cases

Case 6.1 Jogger Shoe Company - Excel file: Case 6.1 - solutions.xlsx
Case 6.2 Westhouser Paper Company - Excel file: Case 6.2 - solutions.xlsx
Case 6.3 Electronic Timing System for Olympics - Excel file: Case 6.3 - solutions.xlsx
Case 6.4 Developing a helicopter component for the Army - Excel file: Case 6.4 -
solutions.xlsx

Practice problem – Production decision

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7669
Tyson Manufacturing (a maker of industrial products) is interested in marketing a new
product. The company must decide whether to manufacture this product essentially on its
own or employ a subcontractor to manufacture it. Below are two tables which represent the
information related to the estimated probability distribution of the cost of one unit of this
product under each alternative.

Cost if Tyson makes the new product itself

Cost per unit Probability

$40 0.20

$45 0.25
 Cost if Tyson makes the new product itself

Cost per unit Probability

$50 0.35

$55 0.20

Cost if Tyson makes the new product itself

Cost per unit Probability

$40 0.15

$45 0.30

$50 0.40

$55 0.15

Cost if Tyson buys the products from a subcontractor

Assuming that Tyson seeks to minimise the expected unit cost of manufacturing of buying
the new product, should the company make the new product or buy it from a subcontractor?
Show your work.

Practice problem – Investment decision

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7680
An investor has $25,000 in assets and faces a difficult choice between two investments. If she
invests in the first opportunity there is a 70% chance that she will increase her assets by
$75,000 and a 30% chance that she will increase her assets by $20,000. If she invests in the
second option there is a 40% chance that she will increase her assets by $150,000 and a 60%
chance that she will increase her assets by $5,000.
a. Construct a decision tree to help the investor make her decision. Make sure to label all
decision and chance nodes and include appropriate costs, payoffs and probabilities.
b. What is the best choice for the investor? Why?

Practice problem – Leasing decision

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7670
A construction company has obtained a contract for a highway project and will need to lease
an additional road grader for a month to fill out its equipment fleet. The company is trying to
decide between two different lease options for the grader: (i) lease an older grader for $8,500,
or (ii) lease a newer grader for $10,000. The newer grader is still under warranty, so the lease
cost covers all repair expenses. However, the company would be responsible for any repair
expenses if it leases the older grader. The construction company’s maintenance foreman
believes there is a 30% chance that there will be no need for repairs with the older grader, but
also thinks there is a 45% chance that some repairs ($2,000) could be needed, and a 25%
chance that significant repairs ($5,000) might be required.
a. Construct a decision tree to help the company make its decision. Make sure to label
all decision and chance nodes and include appropriate costs, payoffs and probabilities.
b. Which is the better lease option? Why?

Practice problem – Expansion decision

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7671
The Waco Tire Company (WTC) is considering expanding production to meet possible
increases in demand. WTC’s alternatives are to construct a new plant, expand the existing
plant, or do nothing in the short run. It will cost them $1 million to build a new facility and
$600,000 to expand their existing facility. The market for this particular product may expand,
remain stable, or contract. WTC’s marketing department estimates the probabilities of these
market outcomes as 0.30, 0.45 and 0.25, respectively. The expected revenue for each
alternative is presented in the table below.

Market expands Market stable Market contracts

Build new plant $1,650,000 $1,000,000 $450,000

Expand plant $1,000,000 $850,000 $450,000

Do nothing $0 $0 $0

a. What is WTC’s payoff if they build a new plant and the market expands?
b. What is WTC’s payoff if they build a new plant and the market is stable?
c. What is WTC’s payoff if they build a new plant and the market contracts?
d. What is WTC’s payoff if they expand the plant and the market expands?
e. What is WTC’s payoff if they expand the plant and the market is stable?
f. What is WTC’s payoff if they expand the plant and the market contracts?
g. What is WTC’s payoff if they do nothing and the market expands?
 h. What is WTC’s payoff if they do nothing and the market is stable?
i. What is WTC’s payoff if they do nothing and the market contracts?
j. Construct a decision tree to identify the course of action which maximises WTC’s
expected profit. Make sure to label all decision and chance nodes and include
appropriate costs, payoffs and probabilities.
k. What course of action is optimal for WTC? What is the expected profit in that case?
l. Generate a risk profile for each of WTC’s possible decisions in this problem.
Characterise the differences in risk for the different options.

Practice problem – Oil exploration

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7673
A landowner is offered $200,000 for the exploration rights to oil on her land, along with a
25% royalty on the future profits if oil is discovered. The landowner is also tempted to
develop the field herself, believing that the interest in her land is a good indication that oil is
present. In that case, she will have to contract a local drilling company to drill an exploratory
well on her own. The cost for such a well is $750,000, which is lost forever if no oil is found.
If oil is discovered, however, the landowner expects to earn future profits of $7,500,000.
Finally, the landowner estimates (with the help of her geologist friend) the probability of
finding oil on this site to be 60%.
a. Construct a decision tree to help the landowner make her decision. Make sure to label
all decision and chance nodes and include appropriate costs, payoffs and probabilities.
b. What should the landowner do? Why?
c. Suppose the landowner is uncertain about the reliability of her geologist friend’s
estimate of the probability that oil will be found on her land. If she thinks the
probability could be anywhere between 40% and 80%, would that change her
decision?
d. Suppose that, in addition to the uncertainty about the probability of finding oil, the
landowner is also uncertain about the cost of the exploratory well (could vary +/-
25%) and the future profits (could vary +/- 50%). To which of these variables is the
expected value most sensitive?
e. What does the risk profile show about the relative risk levels for the landowner’s two
options?
f. Suppose the landowner suspects that she may be a somewhat risk-averse decision-
maker, because she does not feel there is much of a difference between the two
options as their expected values would indicate. She consults with a decision analysis
expert who asks her to decide between two hypothetical alternatives: (i) a gamble
with equal probabilities of winning an amount $X and losing an amount -$X/2, and (ii)
doing nothing, with a payoff of $0. The point at which she cannot decide between (i)
and (ii) is when X = $1,500,000. What is her risk tolerance if she uses an exponential
utility function to model her preferences?
 g. Apply the risk tolerance given in your answer to part f to the landowner’s decision
tree in part a. What is the optimal decision in this case? What is the resulting certainty
equivalent?

Practice problem – Football ordering

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7674
A buyer for a large sporting goods store chain must place orders for professional footballs
with the football manufacturer six months prior to the time the footballs will be sold in the
stores. The buyer must decide in November how many footballs to order for sale during the
upcoming late summer and autumn months. Assume that each football costs the chain $45.
Furthermore, assume that each football can be sold for a retail price of $90. If the footballs
are still on the shelves after next Christmas, they can be discounted and sold for $35 each.
The probability distribution of consumer demand for these footballs during the upcoming
season has been assessed by market research specialists and is presented below. Finally,
assume that the sporting goods store chain must purchase the footballs in lots of 100 units.

Demand Probability

400 0.30

500 0.50

600 0.20

a. Derive the payoff table if the order quantity is 400, 500 and 600, respectively.
b. Construct a decision tree to identify the buyer’s course of action which maximises the
expected profit earned by the chain from the purchase and subsequent sale of footballs
in the coming year.
c. What is the optimal strategy for order quantity, and what is the expected profit in that
case?
d. Generate a risk profile for each possible decision in this problem. Would this have
any impact on your decision?

Practice problem – Credit union

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7672
A customer has approached a local credit union for a $20,000 1-year loan at a 10% interest
rate. If the credit union does not approve the loan application, the $20,000 will be invested in
bonds that earn a 6% annual return. Without additional information, the credit union believes
that there is a 5% chance that this customer will default on the loan, assuming that the loan is
approved. If the customer defaults on the loan, the credit union will lose the $20,000.
a. Construct a decision tree to help the credit union decide whether or not to make the
loan. Make sure to label all decision and chance nodes and include appropriate costs,
payoffs and probabilities.
b. What should the credit union do? What is their expected profit?
c. Suppose that an actual (not perfectly reliable) credit report has the following
characteristics based on historical data: in cases where the customer did not default on
the approved loan, the probability of receiving a favourable recommendation on the
basis of the credit investigation was 80%, while in cases where the customer defaulted
on the approved loan, the probability of receiving a favourable recommendation on
the basis of the credit investigation was 25%. Given this information, what are the
posterior probabilities that a customer will default and will not default, given the
credit report predictions?

Practice problem – Television network

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7678
A television network earns an average of $1.6 million each season from a hit program and
loses an average of $400,000 each season on a program which turns out to be a flop, and of
all programs picked up by this network in recent years, 25% turn out to be hits and 75% turn
out to be flops.
a. Construct a decision tree to help the television network identify the strategy which
maximises its expected profit in responding to a newly proposed television program.
Make sure to label all decision and chance nodes and include appropriate costs,
payoffs and probabilities.
b. What should the network do? What is their expected profit?
c. The network can conduct market research to determine whether a program will be a
hit or a flop. If the market research report is perfectly reliable, what is the most the
network should be willing to pay for it?
d. Suppose that an actual (not perfectly reliable) market research report has the
following characteristics based on historical data: if the program is actually going to
be a hit, there is a 90% chance that the market researchers will predict the program to
be a hit, and if the program is actually going to be a flop, there is a 20% chance that
the market researchers will predict the program to be a hit. Given this information,
what are the posterior probabilities that a show will be a hit or a flop, given the
market research report?
e. Should the network purchase the report if it costs $160,000?

Practice problem – Radio station

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7677
The owner of a radio station in a rapidly growing community is about to begin operations and
must decide what type of programme format to offer. She is considering three formats; rock,
country and rap. The number of listeners for a particular format will depend on the type of
potential audience which is available. Income from advertising depends on the number of
listeners the station has. Three broad categories of audience type can be described as A1, A2
and A3. The rock music format draws mainly from the A1 listener, the country music format
draws mainly from the A2 listener and the rap music format draws mainly from the A3
listener.

The station owner does not know which type of audience will dominate the community once
its growth has stabilised. Probabilities have been assigned to the potential dominant audience,
based on the community growth which has already occurred in the area. Since she wants to
begin building a brand image now, the decision as to which format to adopt must be made in
an environment of uncertainty. The station owner has been able to construct the following
payoff table, in which the entries are average monthly revenue in thousands of dollars.

Format A1 A2 A3

Rock $110 $80 $70

Country $90 $120 $50

Rap $70 $60 $140

Probability 0.3 0.5 0.2

a. Construct a decision tree to help the station identify its optimal format. Make sure to
label all decision and chance nodes and include appropriate costs, payoffs and
probabilities.
b. What format is optimal? What is the expected profit in that case?
c. The station is most uncertain about the average monthly revenue associated with the
rock format and an A1 audience. Construct a strategy region chart for this input
variable with a possible range from $85,000 to $200,000. Does the optimal decision
to select the country format change at any point in this range?
d. As the average monthly revenue associated with the rock format and an A1 audience
varies between $85,000 to about $140,000, what happens to the maximum expected
revenue? Briefly explain why.
e. As the average monthly revenue associated with the rock format and an A1 audience
varies between about $140,000 and $200,000, what happens to the maximum
expected revenue? Briefly explain why.

Practice problem – Nuclear power

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7679
A nuclear power company is deciding whether to build a nuclear plant at Chico Canyon or at
Pleasantville. The cost of building the power plant is $14 million at Chico and $20 million at
Pleasantville. If the company builds at Chico, however, and an earthquake occurs at Chico
during the next 5 years, construction will be terminated and the company will lose $14
million (and will still have to build a power plant at Pleasantville). Without further
information, the company believes there is a 20% chance that an earthquake will occur at
Chico during the next 5 years.
a. Construct a decision tree to help the nuclear power company decide what to do. Make
sure to label all decision and chance nodes and include appropriate costs, payoffs and
probabilities.
b. Where should the nuclear power company build the plant? What is the expected cost?
c. Suppose that a (not perfectly reliable) geologist can be hired to analyse the earthquake
risk. The geologist’s past record indicates that he will predict an earthquake on 90%
of the occasions for which an earthquake will occur and no earthquake on 85% of the
occasions for which an earthquake will not occur. Given this information, what are
the posterior probabilities that an earthquake will and will not occur, given the
geologist’s predictions?
d. Should the nuclear power company hire the geologist if his fee is $1.5m?

Practice problem – Car insurance

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7675
Ms. Rich has just bought a new $30,000 car. As a reasonably safe driver, she believes that
there is only a 5% chance of being in an accident in the forthcoming year. If she is involved
in an accident, the damage to her new car depends on the severity of the accident. The
probability distribution for the range of possible accidents and the corresponding damage
amounts (in dollars) are shown in the table below. Ms. Rich is trying to decide whether she is
willing to pay $170 each year for collision insurance with a $300 deductible. Note that with
this type of insurance, she pays the first $300 in damages if she causes an accident, and the
insurance company pays the remainder.

Conditional
Type of accident Damage to car
probability

Minor 0.60 $200

Moderate 0.20 $1,000

Serious 0.10 $4,000

Catastrophic 0.10 $30,000

Distribution of accident types and corresponding damage amounts
a. Determine the payoffs associated with each possible decision and type of accident.
b. Construct a decision tree to help Ms. Rich decide whether or not to purchase
insurance. Note that the tree should minimise Ms. Rich’s annual expected total cost,
including the possible insurance premium, deductible payment, and damage payment.
In your tree, make sure to label all decision and chance nodes and include appropriate
costs, payoffs and probabilities.
c. What should Ms. Rich do? What is her expected cost in that case?
d. Generate a statistical summary and risk profile for each of Ms. Rich’s possible
decisions. Does this information impact her decision?
e. Perform a sensitivity analysis on the optimal decision and summarise your findings.
Vary the probability of being in an accident from 0% to 10%, the insurance premium
from $50 to $300, and the deductible amount from $0 to $600. In response to which
model inputs is the expected total cost value most sensitive?
f. Using a strategy region graph, determine what impact, if any, the probability of being
in an accident has on her decision. Briefly explain your answer.
g. Using a strategy region graph, determine what impact, if any, the insurance premium
cost has on her decision. Briefly explain your answer.
h. Using a strategy region graph, determine what impact, if any, the insurance deductible
amount has on her decision. Briefly explain your answer.
i. Why is there a kink in the line for the ‘Buy Insurance’ line in the above strategy
region chart?

Practice problem – Utility

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7681
Suppose that a decision-maker’s risk attitude toward monetary gains or losses, x, is given by
the utility function:
𝑈(𝑥) = (𝑋 + 10,000)0.5
a. Show that this decision-maker is indifferent between gaining nothing and entering a
risky situation with a gain of $80,000 (probability 1/3) and a loss of $10,000
(probability 2/3).
b. If there is a 10% chance that one of the decision-maker’s family heirlooms, valued at
$5,000, will be stolen during the next year, what is the most that she would be willing
to pay each year for an insurance policy which completely covers the potential loss of
her cherished items?

Practice problem – Utility

Solutions can be viewed on the VLE through the link below:
https://emfss.elearning.london.ac.uk/mod/book/view.php?id=28425&chapterid=7682
Suppose that a decision-maker’s utility as a function of her wealth, x, is given by 𝑈(𝑥) = ln 𝑥
(the natural logarithm of x).
a. Is this decision-maker risk-averse? Explain why or why not.
b. The decision-maker now has $10,000 and two possible alternatives. For Alternative 1,
she loses $0 (x = $10,000) with probability 0.9 and loses $5,000 (x = $5,000) with
probability 0.1. For Alternative 2, she loses $500 for certain (x = $9,500). Which
alternative maximises the expected utility of her net wealth?
c. The decision-maker now has $15,000 and two possible alternatives. For Alternative 1,
she loses $0 with probability 0.9 and loses $4,000 with probability 0.1. For
Alternative 2, she loses $1,000 for certain. Which decision maximises the expected
utility of her net wealth?

Textbook problem – Elements of decision analysis

Some decision-makers prefer decisions with low risk, but this depends on how risk is
measured. Variance is one measure of risk, but it includes both upside and downside risk.
That is, an outcome with a large positive payoff contributes to variance, but this type of ‘risk’
is good. Consider a decision with some possible payoffs and some possible costs, with given
probabilities. How might you develop a measure of downside risk for such a decision?

A solution file is available here: Chapter_6_Problem_3_solutions.xlsx

Textbook problem – One-stage decision problems

The fixed cost of $6 million in the Acme problem is evidently not large enough to make
Acme abandon the product at the current time. How large would the fixed cost need to be to
make the abandon option the best option?

A solution file is available here: Chapter_6_Problem_4_solutions.xlsx

Textbook problem – One-stage decision problems

Perform a sensitivity analysis on the probability of a great market. To do this,
enter formulae in cells B9 and B10 to ensure that the probabilities of fair and awful remain on
the same ratio, 35 to 20, and that all three probabilities continue to sum to 1. Then let the
probability of great vary from 0.25 to 0.50 in increments of 0.05. Is it ever best to abandon
the product in this range?

A solution file is available here: Chapter_6_Problem_5_solutions.xlsx

Textbook problem – One-stage decision problems

Sometimes a ‘single-stage’ decision can be broken down into a sequence of decisions, with
no uncertainty resolved between these decisions. Similarly, uncertainty can sometimes be
broken down into a sequence of uncertain outcomes. Here is a typical example. A company
has a chance to bid on a government project. The company first decides whether to place a
bid, and then if it decides to place a bid, it decides how much to bid. Once these decisions
have been made, the uncertainty is resolved. First, the company observes whether there
are any competing bids. Second, if there is at least one competing bid, the company observes
the lowest competing bid. The lowest of all bids wins the contract.

Draw a decision tree that reflects this sequence. There should be two ‘stages’ of decision
nodes. Then label the tree with some reasonable monetary values and probabilities, and
perform the folding back process to find the company’s best strategy. Note that if the
company wins the contract, its payoff is its bid amount minus its cost of completing the
project minus its cost of preparing the bid, where these costs are assumed to be known.

A solution file is available here: Chapter_6_Problem_7_solutions.xlsx

Textbook problem – Multistage decision problems

In Example 6.2, Acme’s probability of technological success, 0.8, is evidently large enough
to make ‘continue development’ the best decision. How low would this probability have to be
to make the opposite decision best?

A solution file is available here: Chapter_6_Problem_11_solutions.xlsx

Textbook problem – Multistage decision problems

In Example 6.2, the fixed costs are split $4 million for development and $2 million for
marketing. Perform a sensitivity analysis where the sum of these two fixed costs remains at
$6 million but the split changes. Specifically, let the fixed cost of development vary from $1
million to $5 million in increments of $0.5 million. Does Acme’s best strategy change in this
range?

A solution file is available here: Chapter_6_Problem_12_solutions.xlsx

Textbook problem – Multistage decision problems

In Example 6.2, use a two-way PrecisionTree sensitivity analysis to examine the changes in
both of the two previous problems simultaneously. Let the probability of technological
success vary from 0.6 to 0.9 in increments of 0.05, and let the fixed cost of development vary
as indicated in the previous problem. Explain in a short memo exactly what the results of the
sensitivity analysis imply about Acme’s best strategy.

A solution file is available here: Chapter_6_Problem_13_solutions.xlsx

Textbook problem – Multistage decision problems

In Example 6.2, a technological failure implies that the game is over - the product must be
abandoned. Change the problem so that there are two levels of technological failure, each
with probability 0.1. In the first level, Acme can pay a further development cost, D, to fix the
product and make it a technological success. Then it can decide whether to market the
product. In the second level, the product must be abandoned. Modify the decision tree as
necessary. Then answer the following questions.
 a. For which values of D should Acme fix the product and then market it, given that the
first level of technological failure occurs?
b. For which values of D is Acme’s best first decision still to ‘continue development’?
c. Explain why these two questions are asking different things and can have different
answers.
A solution file is available here: Chapter_6_Problem_21_solutions.xlsx

Textbook problem – The role of risk aversion

You saw in Example 6.4 how Acme prefers to abandon the product when the risk tolerance in
cell B12 is $5000 (really $5 million). This is despite the fact that the EMV from continuing
with the product is well above 0. Using this same risk tolerance, experiment with the sales
volume from a great market in cell C8 to see approximately how large it has to be for Acme
to prefer the ‘continue with product’ decision. (One way to do this is with a data table, using
the TRUE/FALSE value in cell B14 as the single output.) With this sales volume, what is the
certainty equivalent of the ‘continue with product’ decision? How does it compare to the
decision’s EMV? Explain the difference between the two.

A solution file is available here: Chapter_6_Problem_27_solutions.xlsx

Textbook problem – Consolidation exercise

A local energy provider offers a landowner $180,000 for the exploration rights to natural gas
on a certain site and the option for future development. This option, if exercised, is worth an
additional $1,800,000 to the landowner, but this will occur only if natural gas is discovered
during the exploration phase. The landowner, believing that the energy company’s interest in
the site is a good indication that gas is present, is tempted to develop the field herself. To do
so, she must contract with local experts in natural gas exploration and development. The
initial cost for such a contract is $300,000, which is lost forever if no gas is found on the site.
If gas is discovered, however, the landowner expects to earn a net profit of $6,000,000. The
landowner estimates the probability of finding gas on this site to be 60%.
a. Use PrecisionTree to identify the strategy that maximises the landowner’s expected
net earnings from this opportunity.
b. Perform a sensitivity analysis on the optimal decision, letting each of the inputs vary
one at a time plus or minus 25% from its base value, and summarise your findings.
Which of the inputs appears to have the largest effect on the best solution?
A solution file is available here: Chapter_6_Problem_32_solutions.xlsx

Textbook problem – Consolidation exercise

A customer has approached a bank for a $100,000 one-year loan at an 8% interest rate. If the
bank does not approve this loan application, the $100,000 will be invested in bonds that earn
a 6% annual return. Without additional information, the bank believes that there is a 4%
chance that this customer will default on the loan, assuming that the loan is approved. If the
customer defaults on the loan, the bank will lose $100,000. At a cost of $1000, the bank can
thoroughly investigate the customer’s credit record and supply a favourable or unfavourable
recommendation. Past experience indicates that the probability of a favourable
recommendation for a customer who will eventually not default is 0.80, and the chance of a
favourable recommendation for a customer who will eventually default is 0.15.
a. Use a decision tree to find the strategy the bank should follow to maximise its
expected profit.
b. Calculate and interpret the expected value of information (EVI) for this decision
problem.
c. Calculate and interpret the expected value of perfect information (EVPI) for this
decision problem.
d. How sensitive are the results to the accuracy of the credit record recommendations?
Are there any ‘reasonable’ values of the error probabilities that change the optimal
strategy?
A solution file is available here: Chapter_6_Problem_41_solutions.xlsx

Textbook problem – Consolidation exercise

The senior executives of an oil company are trying to decide whether to drill for oil in a
particular field in the Gulf of Mexico. It costs the company $600,000 to drill in the selected
field. Company executives believe that if oil is found in this field its estimated value will be
$3,400,000. At present, this oil company believes that there is a 45% chance that the selected
field actually contains oil. Before drilling, the company can hire a geologist at a cost of
$55,000 to perform seismographic tests. Based on similar tests on other fields, the tests have
a 25% false negative rate (no oil predicted when oil is present) and a 15% false positive rate
(oil predicted when no oil is present).
a. Assuming that this oil company wants to maximise its expected net earnings, use a
decision tree to determine its optimal strategy.
b. Calculate and interpret EVI for this decision problem. Experiment with the accuracy
probabilities of the geologist to see how the EVI changes as they change.
c. Calculate and interpret EVPI for this decision problem.
A solution file is available here: Chapter_6_Problem_44_solutions.xlsx

Block 7: Test your understanding

This activity is a multiple-choice quiz, please complete it on the VLE using the link below:
https://emfss.elearning.london.ac.uk/mod/quiz/view.php?id=28450&forceview=1

Learning Outcomes Checklist

Use this to assess your own understanding of the chapter. You can always go back and
amend the checklist when it comes to revision!
 Analyse decision-making under uncertainty using decision trees.