COMPLETE

BUSINESS
STATISTICS
by
AMIR D. ACZEL
&
JAYAVEL SOUNDERPANDIAN
7th edition.
Prepared by Lloyd Jaisingh, Morehead State
University

Chapter 15

Bayesian Statistics and Decision Analysis

McGraw-Hill/Irwin

Copyright 2009 by The McGraw-Hill Companies, Inc. All

15-2

15 Bayesian Statistics and Decision

Analysis

Using Statistics
Bayes Theorem and Discrete Probability Models
Bayes Theorem and Continuous Probability Distributions
The Evaluation of Subjective Probabilities
Decision Analysis: An Overview
Decision Trees
Handling Additional Information Using Bayes Theorem
Utility
The Value of Information
Using the Computer

15-3

15 LEARNING OBJECTIVES
After studying this chapter you should be able to:

Apply Bayes theorem to revise population parameters

Solve sequential decision problems using the decision tree technique
Conduct decision analysis for cases without probability data
Conduct decision analysis for cases with probability data

15-4

15 LEARNING OBJECTIVES (2)

After studying this chapter you should be able to:
Evaluate the expected value of perfect information
Evaluate the expected value of sample information
Use utility functions to model the risk attitudes of decision
makers
Solve decision analysis problems using spreadsheet
templates

15-5

Bayesian and Classical Statistics

Data

Classical
Inference

Statistical
Conclusion

Data

Bayesian
Inference

Statistical
Conclusion

Prior
Information
Bayesian
Bayesianstatistical
statisticalanalysis
analysisincorporates
incorporatesaaprior
priorprobability
probability
distribution
distributionand
andlikelihoods
likelihoodsofofobserved
observeddata
datatotodetermine
determineaa
posterior
posteriorprobability
probabilitydistribution
distributionofofevents.
events.

15-6

Bayes Theorem: Example 2-10

medical test
test for
for aa rare
rare disease
disease (affecting
(affecting 0.1%
0.1% of
ofthe
the
AAmedical
population [[ P ( I ) 0.001 ])]) isis imperfect:
imperfect:
population

When
Whenadministered
administeredto
toan
anill
illperson,
person,the
thetest
testwill
willindicate
indicateso
so

withprobability
probability0.92
0.92 [[ P ( Z I ) .92 P ( Z I ) .08 ]]
with
Theevent
event ( Z I ) isisaafalse
falsenegative
negative
The
When
Whenadministered
administeredto
toaaperson
personwho
whoisisnot
notill,
ill,the
thetest
testwill
will

erroneouslygive
giveaapositive
positiveresult
result(false
(falsepositive)
positive)with
with
erroneously
probability0.04
0.04[[P ( Z I ) 0.04 P ( Z I ) 0.96 ]]
probability
Theevent
event ( Z I ) isisaafalse
falsepositive.
positive.
..
The

15-7

15-2 Bayes Theorem and Discrete Probability

Models - Example 2-10 (Continued)
Applying Bayes Theorem
P ( I ) 0.001
P ( I ) 0.999
P ( Z I ) 0.92
P ( Z I ) 0.04

P ( I Z )
P( Z )
P ( I Z )

P ( I Z ) P ( I Z )
P( Z I ) P( I )

P( Z I ) P( I ) P( Z I ) P( I )

P( I Z )

(.92)( 0.001)
(.92)( 0.001) ( 0.04)(.999)
0.00092
0.00092

0.00092 0.03996
.04088
.0225

15-8

Example 2-10: Decision Tree

Prior
Probabilities

Conditional
Probabilities
P ( Z I ) 0.92

P( Z I ) 0.08

P( I ) 0.001

P( I ) 0.999

P( Z I ) 0.04

P( Z I ) 0.96

Joint
Probabilities
P ( Z I ) (0.001)( 0.92) .00092

P( Z I ) (0.001)(0.08) .00008

P( Z I ) (0.999)(0.04) .03996

P ( Z I ) (0.999)(0.96) .95904

15-2 Bayes Theorem and Discrete

Probability Models
The likelihood function is the set of conditional probabilities
P(x|) for given data x, considering a function of an unknown
population parameter, .
Bayes theorem for a discrete random variable:
P(x ) P( )
P( x)
P(x i )P( i )
i

where is an unknown population parameter to be estimated

from the data. The summation in the denominator is over all
possible values of the parameter of interest, i, and x stands for
the observed data set.

15-9

Example 15-1: Prior Distribution and

Likelihoods of 4 Successes in 20 Trials
Prior
Prior
Distribution
Distribution
P(S)
SS
P(S)
0.1
0.05
0.1
0.05
0.2
0.15
0.2
0.15
0.3
0.20
0.3
0.20
0.4
0.30
0.4
0.30
0.5
0.20
0.5
0.20
0.6
0.10
0.6
0.10
1.00
1.00

Likelihood
Likelihood

Binomialwith
withnn==20
20and
andpp==0.100000
0.100000
Binomial
xx P(P(XX==x)x)
4.00
0.0898
4.00
0.0898
Binomialwith
withnn==20
20and
andpp==0.200000
0.200000
Binomial
xx P(P(XX==x)x)
4.00
0.2182
4.00
0.2182
Binomialwith
withnn==20
20and
andpp==0.300000
0.300000
Binomial
xx P(P(XX==x)x)
4.00
0.1304
4.00
0.1304
Binomialwith
withnn==20
20and
andpp==0.400000
0.400000
Binomial
xx P(P(XX==x)x)
4.00
0.0350
4.00
0.0350
Binomialwith
withnn==20
20and
andpp==0.500000
0.500000
Binomial
xx P(P(XX==x)x)
4.00
0.0046
4.00
0.0046
Binomialwith
withnn==20
20and
andpp==0.600000
0.600000
Binomial
xx P(P(XX==x)x)
4.00
0.0003
4.00
0.0003

15-10

Example 15-1: Prior Probabilities,

Likelihoods, and Posterior Probabilities
Prior
Prior
Distribution Likelihood
Likelihood
Distribution
P(S) P(x|S)
P(x|S)
SS
P(S)
0.1
0.05 0.0898
0.0898
0.1
0.05
0.2
0.15 0.2182
0.2182
0.2
0.15
0.3
0.20 0.1304
0.1304
0.3
0.20
0.4
0.30 0.0350
0.0350
0.4
0.30
0.5
0.20 0.0046
0.0046
0.5
0.20
0.6
0.10 0.0003
0.0003
0.6
0.10
1.00
1.00

Posterior
Posterior
Distribution
Distribution
P(S)P(x|S) P(S|x)
P(S|x)
P(S)P(x|S)
0.00449
0.06007
0.00449
0.06007
0.03273
0.43786
0.03273
0.43786
0.02608
0.34890
0.02608
0.34890
0.01050
0.14047
0.01050
0.14047
0.00092
0.01230
0.00092
0.01230
0.00003
0.00040
0.00003
0.00040
0.07475
1.00000
0.07475
1.00000

15-11

93%
Credible
Set

15-12

Example 15-1: Prior and Posterior

Distributions
P rio r D is trib utio n o f Marke t S hare

0.5

0.5

0.4

0.4

0.3

0.3

P (S)

P (S)

P o s te rio r D is trib utio n o f M arke t S hare

0.2

0.2

0.1

0.1

0.0

0.0
0.1

0.2

0.3

0.4

0 .5

0.6

0.1

0.2

0.3

0.4

0.5

0.6

Example 15-1: A Second Sampling

with 3 Successes in 16 Trials
PriorDistribution
Distribution
Prior
P(S)
SS
P(S)
0.1
0.06007
0.1
0.06007
0.2
0.43786
0.2
0.43786
0.3
0.34890
0.3
0.34890
0.4
0.14047
0.4
0.14047
0.5
0.01230
0.5
0.01230
0.6
0.00040
0.6
0.00040
1.00000
1.00000

Likelihood
Likelihood

Binomialwith
withnn==16
16and
andpp==0.100000
0.100000
Binomial
xx P(P(XX==x)x)
3.00
0.1423
3.00
0.1423
Binomialwith
withnn==16
16and
andpp==0.200000
0.200000
Binomial
xx P(P(XX==x)x)
3.00
0.2463
3.00
0.2463
Binomialwith
withnn==16
16and
andpp==0.300000
0.300000
Binomial
xx P(P(XX==x)x)
3.00
0.1465
3.00
0.1465
Binomialwith
withnn==16
16and
andpp==0.400000
0.400000
Binomial
xx P(P(XX==x)x)
3.00
0.0468
3.00
0.0468
Binomialwith
withnn==16
16and
andpp==0.500000
0.500000
Binomial
xx P(P(XX==x)x)
3.00
0.0085
3.00
0.0085
Binomialwith
withnn==16
16and
andpp==0.600000
0.600000
Binomial
xx P(P(XX==x)x)
3.00
0.0008
3.00
0.0008

15-13

Example 15-1: Incorporating a

Second Sample
Prior
Prior
Distribution Likelihood
Likelihood
Distribution
P(S)
P(x|S)
SS
P(S)
P(x|S)
0.1 0.06007
0.06007 0.1423
0.1423
0.1
0.2 0.43786
0.43786 0.2463
0.2463
0.2
0.3 0.34890
0.34890 0.1465
0.1465
0.3
0.4 0.14047
0.14047 0.0468
0.0468
0.4
0.5 0.01230
0.01230 0.0085
0.0085
0.5
0.6 0.00040
0.00040 0.0008
0.0008
0.6
1.00000
1.00000

P(S)P(x|S)
P(S)P(x|S)
0.0085480
0.0085480
0.1078449
0.1078449
0.0511138
0.0511138
0.0065740
0.0065740
0.0001046
0.0001046
0.0000003
0.0000003
0.1741856
0.1741856

15-14

Posterior
Posterior
Distribution
Distribution
P(S|x)
P(S|x)
0.049074
0.049074
0.619138
0.619138
91%
0.293444
Credible Set
0.293444
0.037741
0.037741
0.000601
0.000601
0.000002
0.000002
1.000000
1.000000

15-15

Example 15-1: Using the Template

Applicationof
ofBayes
BayesTheorem
Theoremusing
usingthe
theTemplate.
Template. The
The
Application
posteriorprobabilities
probabilitiesare
arecalculated
calculatedusing
usingaaformula
formulabased
based
posterior
onBayes
BayesTheorem
Theoremfor
fordiscrete
discreterandom
randomvariables.
variables.
on

Example 15-1: Using the Template

(Continued)

Display of the Prior and Posterior probabilities.

15-16

Example 15-1: Using the Template with

Second sample; n = 16, x = 3.
Applicationof
ofBayes
BayesTheorem
Theoremusing
usingthe
theTemplate.
Template. The
The
Application
posteriorprobabilities
probabilitiesare
arecalculated
calculatedusing
usingaaformula
formulabased
based
posterior
onBayes
BayesTheorem
Theoremfor
fordiscrete
discreterandom
randomvariables.
variables.
on

15-17

Example 15-1: Using the Template

with Second sample; n = 16, x = 3.

Display of the Prior and Posterior probabilities.

15-18

15-19

15-3 Bayes Theorem and Continuous

Probability Distributions
Wedefine
definef()
f()as
asthe
theprior
priorprobability
probabilitydensity
densityof
ofthe
theparameter
parameter
We
. We
Wedefine
definef(x|)
f(x|)as
asthe
theconditional
conditionaldensity
densityof
ofthe
thedata
datax,x,given
given
.
thevalue
valueof
of.. This
Thisisisthe
thelikelihood
likelihoodfunction.
function.
the
Bayes' theorem
theoremfor
forcontinuous
continuousdistributions:
distributions:
Bayes'
f ((xx))ff (())
f ((xx))ff (())
f
f
ff ((xx))
Total area under f (x )
ff ((xx))ff (())dd Total area under f (x )

15-20

The Normal Probability Model

Normalpopulation
populationwith
withunknown
unknownmean
meanand
andknown
knownstandard
standarddeviation
deviation
Normal
Populationmean
meanisisaarandom
randomvariable
variablewith
withnormal
normal(prior)
(prior)distribution
distributionand
andmean
meanM
M
Population
andstandard
standarddeviation
deviation.
.
and
Drawsample
sampleofofsize
sizen:n:
Draw

The posterior mean and variance of the normal population of

the population mean, :
1
n
M 2 M

2


1
2
M =

1 n
1 n

2 2
2 2

The Normal Probability Model:

Example 15-2
M 15

n 10

M 1154
.

s 684
.

1
n
2 M 2 M


1
2
M =

1 n
1 n
2 2
2 2


1
10
.
2 15
1154
2
8
684

1
.
2
M =

1 10
1 10
2

2
2
2
8 684

.
8
684
.
2
M = 11.77
2.077
95% Credible Set: M 196
. 1177
. (196
. ) 2.077 [ 7.699 ,15841
. ]

15-21

15-22

Example 15-2
Density

Posterior
Distribution

Likelihood
function

Prior
Distribution

11.54
11.77

15

15-23

Example 15-2 Using the Template

15-24

Example 15-2 Using the Template

(Continued)

15-4 The Evaluation of Subjective

Probabilities

Basedon
onnormal
normaldistribution
distribution
Based
95% of normal distribution is within 2 standard

95% of normal distribution is within 2 standard

deviations of
of the
the mean
mean
deviations
P(-1 << xx << 31)
31) == .95=
.95= 15,
15, =
= 88
P(-1
68% of normal distribution is within 1 standard

68% of normal distribution is within 1 standard

deviation of
of the
the mean
mean
deviation
P(7 << xx << 23)
23) == .68
.68 =
=15,
15, =
= 88
P(7

15-25

15-26

15-5 Decision Analysis

Elementsof
ofaadecision
decisionanalysis
analysis
Elements
Actions
Actions

Anythingthe
thedecision-maker
decision-makercan
cando
doatatany
anytime
time
Anything

Chanceoccurrences
occurrences
Chance

Possibleoutcomes
outcomes(sample
(samplespace)
space)
Possible

Probabilitiesassociated
associatedwith
withchance
chanceoccurrences
occurrences
Probabilities
Final outcomes
Final outcomes

Payoff,reward,
reward,ororloss
lossassociated
associatedwith
withaction
action
Payoff,

Additionalinformation
information
Additional

Allowsdecision-maker
decision-makertotoreevaluate
reevaluateprobabilities
probabilitiesand
andpossible
possiblerewards
rewardsand
andlosses
losses
Allows

Decision
Decision

Courseofofaction
actiontototake
takeinineach
eachpossible
possiblesituation
situation
Course

15-6: Decision Tree: New-Product

Introduction
Decision
Decision

Chance
Chance
Occurrence
Occurrence
Product
successful
(P = 0.75)

Final
Final
Outcome
Outcome
$100,000

Market

Do not
market

Product
unsuccessful
(P = 0.25)

-$20,000

$0

15-27

15-28

15-6: Payoff Table and Expected Values of

Decisions: New-Product Introduction

Action
Action
Marketthe
theproduct
product
Market
Donot
notmarket
marketthe
theproduct
product
Do

Productisis
Product
Successful Not
NotSuccessful
Successful
Successful
$100,000
-$20,000
$100,000
-$20,000
$0
$0
$0
$0

Theexpected
expectedvalue
valueof
ofXX,, denoted
denotedEE((XX):):
The
xP((xx))

EE((XX))
xP
allallxx
Outcome)) (100,
(100,000
000)()(00.75
.75))((20
20,,000
000)()(00.25
.25))
EE((Outcome
750000--5000
5000==70,000
70,000
== 750000

15-29

Solution to the New-Product

Introduction Decision Tree
Clipping the Nonoptimal Decision Branches
Expected
Expected
Payoff
Payoff
$70,000
$70,000

Product
successful
(P = 0.75)

$100,000

Market
Product
unsuccessful
(P = 0.25)
Nonoptimal
Nonoptimal
decisionbranch
branch
decision
clipped
isisclipped

Do not
market

Expected
Expected
Payoff
Payoff
$0
$0

-$20,000

$0

15-30

New-Product Introduction:
Extended-Possibilities
Outcome
Outcome
Extremelysuccessful
successful
Extremely
Verysuccessful
successful
Very
Successful
Successful
Somewhatsuccessful
successful
Somewhat
Barelysuccessful
successful
Barely
Breakeven
even
Break
Unsuccessful
Unsuccessful
Disastrous
Disastrous

Payoff Probability
Probability xP(x)
xP(x)
Payoff
$150,000
0.1
15,000
$150,000
0.1
15,000
120.000
0.2
24,000
120.000
0.2
24,000
100,000
0.3
30,000
100,000
0.3
30,000
80,000
0.1
8,000
80,000
0.1
8,000
40,000
0.1
4,000
40,000
0.1
4,000
0.1
00
0.1
00
-20,000
0.05
-1000
-20,000
0.05
-1000
-50,000
0.05
-2,500
-50,000
0.05
-2,500
ExpectedPayoff:
Payoff:
Expected

$77,500
$77,500

New-Product Introduction:
Extended-Possibilities Decision Tree
Chance
Occurrence

Decision

Expected
Expected
Payoff
Payoff
$77,500
$77,500
Market

Payoff
0.1

$150,000
0.2 $120,000
0.3
$100,000
0.1

$80,000
0.1 $40,000
0.1
$0
0.05
0.05

-$20,000
-$50,000

Nonoptimal
Nonoptimal
decisionbranch
branch
decision
clipped
isisclipped

Do not
market

$0

15-31

15-32

Example 15-3: Decision Tree

Not Promote

$700,000
P r = 0.4

Pr = 0.5

Promote

$680,000

Pr = 0.6
$740,000

Lease

Pr = 0.3
Pr = 0.15

Not Lease

Pr = 0.05
Pr = 0.9
Pr = 0.1

$800,000
$900,000
$1,000,000
$750,000
$780,000

15-33

Example 15-3: Solution

Expected payoff:
Not Promote $700,000
Expected payoff:
0.5*425000
+0.5*716000=
$783,000

Lease

Pr=0.5

Pr = 0.4
Promote

Expected payoff:
$425,000

Expected payoff:
$753,000

$680,000

Pr = 0.6
Expected payoff:
$716,000
Pr = 0.3

Pr = 0.15
Pr = 0.05

Not Lease

$700,000

Pr = 0.9
Pr = 0.1

$740,000
$800,000
$900,000
$1,000,000
$750,000
$780,000

15-7 Handling Additional Information

Using Bayes Theorem
Payoff

Successful
Market
Test indicates
success

Do not market

$95,000
-$25,000

Failure

-$5,000
Market

Test
Test indicates
failure

Successful

$95,000

Failure

-$25,000
-$5,000

Do not market
Not test

New-ProductDecision
Decision
New-Product
Treewith
withTesting
Testing
Tree

Market

Successful Pr=0.75
Failure

Do not market

$100,000
Pr=0.25
-$20,000
0

15-34

15-35

Applying Bayes Theorem

P(S)=0.75
P(IS|S)=0.9
P(IF|S)=0.1
P(F)=0.75
P(IS|F)=0.15
P(IF|S)=0.85
P(IS)=P(IS|S)P(S)+P(IS|F)P(F)=(0.9)(0.75)+(0.15)(0.25)=0.7125
P(IF)=P(IF|S)P(S)+P(IF|F)P(F)=(0.1)(0.75)+(0.85)(0.25)=0.2875
P(IS|S)P(S)
P(S| IS) =
P(IS|S)P(S) P(IS| F)P(F)
( 0.9 )( 0.75)

0.9474
( 0.9 )( 0.75) ( 0.15)( 0.25)
P(F| IS) 1 P(S| IS) 1 0.9474 0.0526
P(IF|S)P(S)
P(S| IF) =
P(IF|S)P(S) P(IF| F)P(F)
( 0.1)( 0.75)

0.2609
( 0.1)( 0.75) ( 0.85)( 0.25)
P(F| IF) 1 P(S| IF) 1 0.2609 0.7391

15-36

Expected Payoffs and Solution

$86,866
Market

$86,866 P(S|IS)=0.9474

P(IS)=0.7125
$66.003
Test

Do not market
$6,308

Market

$6,308

P(IF)=0.2875
$70,000
Not test

P(F|IS)=0.0526

$95,000
-$25,000
-$5,000

P(S|IF)=0.2609
P(F|IF)=0.7391

Do not market
$70,000
$70,000
P(S)=0.75
Market
P(F)=0.25
Do not market

Payoff

$95,000
-$25,000
-$5,000
$100,000
-$20,000
0

Example 15-4: Payoffs and

Probabilities
PriorInformation
Information
Prior
Levelofof
Level
Economic
Economic
Profit
Activity Probability
Probability
Profit
Activity
$3million
million Low
Low
0.20
$3
0.20
$6million
million Medium
Medium 0.50
0.50
$6
$12million
million High
High
0.30
$12
0.30
Consultantssay
sayLow
Low
Consultants
Event Prior
Prior Conditional
Conditional
Event
Low
0.20
0.90
Low
0.20
0.90
Medium0.50
0.50
0.05
Medium
0.05
High 0.30
0.30
0.05
High
0.05
P(Consultantssay
sayLow)
Low)
P(Consultants

ReliabilityofofConsulting
ConsultingFirm
Firm
Reliability
Future
Future
Stateofof Consultants
ConsultantsConclusion
Conclusion
State
Economy High
High Medium
Medium Low
Low
Economy
Low
0.05 0.05
0.05
0.90
Low
0.05
0.90
Medium 0.15
0.15 0.80
0.80
0.05
Medium
0.05
High
0.85 0.10
0.10
0.05
High
0.85
0.05

Joint Posterior
Posterior
Joint
0.180
0.818
0.180
0.818
0.025
0.114
0.025
0.114
0.015
0.068
0.015
0.068
0.220
1.000
0.220
1.000

15-37

Example 15-4: Joint and Conditional

Probabilities
Consultantssay
sayMedium
Medium
Consultants
Event Prior
Prior Conditional
Conditional
Event
Low
0.20
0.05
Low
0.20
0.05
Medium 0.50
0.50
0.80
Medium
0.80
High 0.30
0.30
0.10
High
0.10
P(Consultantssay
sayMedium)
Medium)
P(Consultants

Joint Posterior
Posterior
Joint
0.010 0.023
0.023
0.010
0.400 0.909
0.909
0.400
0.030 0.068
0.068
0.030
0.440 1.000
1.000
0.440

Consultantssay
sayHigh
High
Consultants
Event Prior
Prior Conditional
Conditional
Event
Low
0.20
0.05
Low
0.20
0.05
Medium0.50
0.50
0.15
Medium
0.15
High 0.30
0.30
0.85
High
0.85
P(Consultantssay
sayHigh)
High)
P(Consultants

Joint Posterior
Posterior
Joint
0.010 0.029
0.029
0.010
0.075 0.221
0.221
0.075
0.255 0.750
0.750
0.255
0.340 1.000
1.000
0.340

15-38

AlternativeInvestment
Investment
Alternative
Profit
Probability
Profit
Probability
$4million
million
0.50
$4
0.50
$7million
million
0.50
$7
0.50
Consultingfee:
fee:$1
$1million
million
Consulting

15-39

Example 15-4: Decision Tree

Hire consultants

Do not hire consultants

6.54
L

0.22

0.34

0.44
7.2

9.413

Alternative

Invest

5.5
0.5

Alternative

7.2 L
M

0.5

5.339

4.5
0.5

Alternative
4.5

H 9.413 L
M

0.5

0.5

Invest

Alternative
4.5

H 5.339 L
M

0.5

0.5

Invest
H 2.954 L
M

0.5

0.068 0.909 0.023

0.068 0.114 0.818

$2 million

$5 million

$11million

$3 million

$6 million

$2 million

$5 million

$11 million

$3 million

$6 million

$2 million

$5 million

$3 million

$11 million

$6 million

0.750 0.221 0.029

$3 million

0.5

$6 million

0.2

$12 million

$4 million

$7 million

0.3

Invest

4.5

15-40

15-8 Utility and Marginal Utility

Utilityisisaameasure
measureof
ofthe
thetotal
totalworth
worthof
ofaaparticular
particularoutcome.
outcome.
Utility
reflectsthe
thedecision
decisionmakers
makersattitude
attitudetoward
towardaacollection
collectionof
of
ItItreflects
factorssuch
suchasasprofit,
profit,loss,
loss,and
andrisk.
risk.
factors

Utility
Additional
Utility

Additional
Utility

{
}

}
Additional $1000

Additional $1000

Dollars

15-41

Utility and Attitudes toward Risk

Utility

Utility

Risk Averse

Risk Taker

Dollars
Utility

Dollars
Utility

Risk Neutral

Dollars

Mixed

Dollars

15-42

Example 15-5: Assessing Utility

Possible
Possible
Returns
Returns
$1,500
$1,500
4,300
4,300
22,000
22,000
31,000
31,000
56,000
56,000

Initial
Indifference
Initial
Indifference
Utility
Probabilities
Utility
Utility
Probabilities
Utility
00
00
(1500)(0.8)+(56000)(0.2) 0.2
0.2
(1500)(0.8)+(56000)(0.2)
(1500)(0.3)+(56000)(0.7) 0.7
0.7
(1500)(0.3)+(56000)(0.7)
(1500)(0.2)+(56000)(0.8) 0.8
0.8
(1500)(0.2)+(56000)(0.8)
11
11

Utility
1.0

0.5

Dollars

0.0
0

10000

20000

30000

40000

50000

60000

15-43

15-9 The Value of Information

Theexpected
expectedvalue
valueof
ofperfect
perfectinformation
information(EVPI):
(EVPI):
The
EVPI == The
Theexpected
expectedmonetary
monetaryvalue
valueofofthe
thedecision
decisionsituation
situationwhen
when
EVPI
perfectinformation
informationisisavailable
availableminus
minusthe
theexpected
expectedvalue
valueof
ofthe
the
perfect
decisionsituation
situationwhen
whenno
noadditional
additionalinformation
informationisisavailable.
available.
decision
ExpectedNet
NetGain
Gainfrom
fromSampling
Sampling
Expected
Expected
Net Gain
Max

Sample Size
nmax

15-44

Example 15-6: The Decision Tree

Competitors
Fare

Airline
Fare
$200
Fare

$300
Fare

Payoff
$8 million

Competitor:$200
Pr = 0.6

8.4

6.4

Competitor:$300
Pr = 0.4
Competitor:$200
Pr = 0.6

$9 million
$4 million

Competitor:$300
Pr = 0.4

$10 million

Example 15-6: Value of Additional

Information

noadditional
additionalinformation
informationisisavailable,
available,the
thebest
beststrategy
strategyisis
IfIfno
toset
setthe
thefare
fareatat$200
$200
to
E(Payoff|200)==(.6)(8)+(.4)(9)
(.6)(8)+(.4)(9)==$8.4
$8.4million
million
E(Payoff|200)
E(Payoff|300)==(.6)(4)+(.4)(10)
(.6)(4)+(.4)(10)==$6.4
$6.4million
million
E(Payoff|300)

Withfurther
furtherinformation,
information,the
theexpected
expectedpayoff
payoffcould
couldbe:
be:
With
E(Payoff|Information)==(.6)(8)+(.4)(10)=$8.8
(.6)(8)+(.4)(10)=$8.8million
million
E(Payoff|Information)

EVPI=8.8-8.4==$.4
$.4million
million
EVPI=8.8-8.4

15-45