Chapter 4

Making Choices

Making Hard Decisions

Slide 1 of 58
 Texaco Versus Pennzoil
In early 1984, Pennzoil and Getty Oil agreed to the
terms of a merger. But before any formal documents
could be signed, Texaco offered Getty a substantially
better price, and Gordon Getty, who controlled mos of
the Getty Stock, reneged on the Pennzoil deal and sold to
Texaco. Naturally, Pennzoil felt as if it had been dealt
with unfairly and immediately files a lawsuit against
Texaco alleging that Texaco had interfered illegally in the
Pennzoil-Getty negotiations. Pennzoil won the case: in
late 1985, it was awarded $11.1 billion, the largest
judgment ever in the United States. A Texas appeal court
reduced the judgement to $2 billion, but interest and
penalties drove the total back up to $10.3 billion. James
Kinnear, Texaco’s Chief executive officer, had said that
Texaco would file for bankruptcy if Pennzoil obtained
court permission to secure the judgment by filing liens
against Texaco’s assets.

Slide 2 of 58
 Texaco Versus Pennzoil - Continued
Furthermore, Kinnear had promised to fight the case all
the way to the U.S. Supreme Court if necessary, arguing
in part that Pennzoil had not followed Security and
Exchange Commission regulations in its negotiations with
Getty. In April 1987, just before Pennzoil began to file
liens, Texaco offered to Penzoil $2 billion dollars to
settle the entire case. Hugh Liedtke, chairman of
Pennzoil, indicated that his advisors were telling him that
a settlement between $3 billion and $5 billion would be
fair.

What should Hugh Liedtke do?

1. Accept $2 Billion
2. Refuse $2 Billion and counter
offer $5 Billion

Slide 3 of 58
Texaco Versus Pennzoil – Decision Tree
Max Settlement
Amount
($ Billion )
Accept $2 Billion
2

Texaco Accepts $5 Billion

5

High
10.3
Counteroffer Texaco Refuses Medium
Final Court 5
Decision
$5 Billion Counteroffer
Low
0

High
10.3

Refuse Final Court Medium 5

Decision
Texaco Low
0
Counter -
offers $3 Billion
Accept $3 Billion
3

Slide 4 of 58
 Texaco Versus Pennzoil - Continued

• Given tough negotiation positions of the two executives,

their could be an even chance (50%) that Texaco will
refuse to negotiate further.
• Liedtke and advisor figure that it is twice as likely that
Texaco would counter offer $3 billion than accepting the
$5 billion. Hence, because there is a 50% of refusal,
there must be a 33% chance of a Texaco counter offer
and a 17% chance of Texaco accepting $5 billion.
• What are the probabilities of the final court decision?
 Liedtke admitted that Pennzoil could lose the case. Thus there is
a significant possibility the outcome would be zero. It’s probability
is assessed at 30%.
 Given the strength of the Pennzoil case it is also possible that the
court will upheld the judgment as it stands. It’s probability is
assessed at 20%.
 Finally, the possibility exists that the judgment could be reduced
somewhat to $5 billion. Thus there must be a chance of 50% of
this happening.

Slide 5 of 58
 Texaco Versus Pennzoil - Continued

• Given tough negotiation positions of the two executives, it

could be an even chance (50%) that Texaco will refuse to
negotiate further.
• Liedtke and advisor figures that it is twice as likely that
Texaco would counter offer $3 billion than accepting the
$5 billion. Hence, because there is a 50% of refusal,
there must be a 33% chance of a Texaco counter offer
and a 17% chance of Texaco accepting $5 billion.
• What are the probabilities of the final court decision?
 Liedtke admitted that Pennzoil could lose the case. Thus there is
a significant possibility the outcome would be zero. It’s probability
is assessed at 30%.
 Given the strength of the Pennzoil case it is also possible that the
court will upheld the judgment as it stands. It’s probability is
assessed at 20%.
 Finally, the possibility exists that the judgment could be reduced
somewhat to $5 billion. Thus there must be a chance of 50% of
this happening.

Slide 6 of 58
Texaco Versus Pennzoil – Decision Tree
Max Settlement
Amount
($ Billion )
Accept $2 Billion
2

Texaco Accepts $5 Billion (0.17)

5

High (0.20)
10.3
Counteroffer Texaco Refuses (0.50) Medium (0.50)
Final Court 5
Decision
$5 Billion Counteroffer
Low (0.30)
0

High (0.20)
10.3

Refuse Final Court Medium (0.50) 5

Decision
Texaco Low (0.30)
(0.33) 0
Counter -
offers $3 Billion
Accept $3 Billion
3

Slide 7 of 58
Decision Tree and Expected Monetary Value (EMV)

When objective is measured in dollars

First Suggestion:
Solve decision problem by choosing
that alternative that maximizes the EMV

Expected value of discrete random variable Y:

n n

EY [Y ]   yi  Pr(Y  yi )   yi  pi
i1 i1

Slide 8 of 58
 A double-risk dillema
y Pr(Y=y) y*Pr(Y=y)
Max Profit $24.00 0.2 $4.80
Win (0.20) -$1.00 0.8 -$0.80
$24 $4.00 = EMV
EMV= $4 $25
Trade Ticket

EMV= -$1
$4.5 Lose (0.80) -$1
$0
Win (0.45)
EMV= $4.5 $10
$10
Keep Ticket
$0 y Pr(Y=y) y*Pr(Y=y)
Lose (0.55) $0 $10.00 0.45 $4.50
$0 $0.00 0.55 $0.00
$4.50 =EMV

Interpretation EMV: Playing the same lottery a lot of times

will result over time in an average pay-off equal to the EMV

Slide 9 of 58
Texaco Versus Pennzoil – Decision Tree
Max Settlement
Amount
($ Billion )
Accept $2 Billion
2

Texaco Accepts $5 Billion (0.17)

5

High (0.20)
10.3
Counteroffer Texaco Refuses (0.50) Final Court Medium (0.50) 5
Decision
$5 Billion Counteroffer
Low (0.30)
0

High (0.20)
10.3

Refuse Final Court Medium (0.50) 5

Decision
Texaco Low (0.30)
(0.33) 0
Counter -
offers $3 Billion
Accept $3 Billion
3

Solve tree using EMV by folding back the tree

Slide 10 of 58
 Decision Tree and Expected Monetary Value (EMV)

Step 1: Calculate EMV of court decision uncertainty node

High (0.20)
EMV= $4.56 10.3

Final Court Medium (0.50) 5

Decision
Low (0.30)
0

Step 1

y Pr(Y=y) y*Pr(Y=y)
10.300 0.2 $2.06
5.000 0.5 $2.50
0.000 0.3 $0.00
$4.56 =EMV

Slide 11 of 58
 Decision Tree and Expected Monetary Value (EMV)

Step 2: Evaluate decision regarding Texaco’s counter offer

EMV= High (0.20)

4.56 10.3

Refuse Final Court Medium (0.50) 5

EMV=
4.56 Decision
Low (0.30)
0

Accept $3 Billion
3

Slide 12 of 58
 Decision Tree and Expected Monetary Value (EMV)

Step 3: Calculate EMV Texaco’s reaction uncertainty node

y Pr(Y=y) y*Pr(Y=y)
5.000 0.17 $0.85
4.560 0.5 $2.28
Accept $2 Billion 2 4.560 0.33 $1.50
$4.63 = EMV

Texaco Accepts $5 Billion (0.17) 5

EMV= EMV=
4.63 4.56
Counteroffer Texaco Refuses (0.50)

$5 Billion Counteroffer

EMV=
4.56
Texaco (0.33)
Counter -
offers $3 Billion

Slide 13 of 58
 Decision Tree and Expected Monetary Value (EMV)

Step 4: Evaluate the immediate decision

Max Result
Accept $2 Billion
EMV= 2
4.63

EMV=
4.63
Counteroffer

$5 Billion

Optimal decision: Counteroffer $5 Billion

Optimal decision strategy: Counteroffer $5 Billion
and if Texaco counteroffers $3 Billion, then
refuse this counteroffer.
Slide 14 of 58
Folding back the Decision Tree from right to left using EMV
Max Result
Accept $2 Billion
2
EMV=
4.63
Texaco Accepts $5 Billion (0.17)
5

EMV= High (0.20)

EMV=
4.63 10.3
4.56
Counteroffer Texaco Refuses (0.50) Final Court Medium (0.50) 5
Decision
$5 Billion Counteroffer
Low (0.30)
0

EMV= High (0.20)

4.56 10.3

Refuse Final Court Medium (0.50) 5

EMV=
4.56 Decision
Texaco Low (0.30)
(0.33) 0
Counter -
offers $3 Billion
Accept $3 Billion
3
Slide 15 of 58
 Definitions Decision Path and Strategy

Definition decision path:

A path starting at the left most node up to the values at the end of a
branch by selecting one alternative from decision nodes or by following
one outcome from uncertainty nodes. Represents a possible future
scenario.

Definition decision strategy:

The collection of decision paths connected to one branch of the
immediate decision by selecting one alternative from each decision
node along these paths. Represents specifying at every decision in the
decision problem what we would do, if we get to that decision (we
may not get there due to outcome of previous uncertainty nodes).

Optimal decision strategy:

That decision strategy which results in the highest EMV if we
maximize profit and the lowest EMV if we minimize cost.

Slide 16 of 58
 Counting Strategies
How many decision strategies in Example 1?

Example 1

How many decision strategies in Example 2?

Example 2

Slide 17 of 58
 Counting Strategies

How many decision strategies in Example 3?

Example 3

Slide 18 of 58
 Counting Strategies
How many decision strategies in Example 1?

Example 1
Strategy 1
Strategy 2
Strategy 3

How many decision strategies in Example 2?

Strategy 1
Example 2

1 Strategy 2 (11)
0 Strategy 3 (00)
1 Strategy 4 (10)
0 Strategy 5 (01)

Slide 19 of 58
 Counting Strategies

How many decision strategies in Example 3?

Strategy 1
Strategy 2 (111)
1
Strategy 3 (001)
0
Strategy 4 (101)
1
Example 3 Strategy 5 (011)
0
1 Strategy 6 (110)
0 Strategy 7 (000)
Strategy 8 (100)
Strategy 9 (010)

Slide 20 of 58
Decision Strategies Texaco-Pennzoil Case

How many decision strategies do we have in

the Texaco – Penzoil decision tree?

First strategy: “Accept $2 billion”

Accept $2 Billion
2

Slide 21 of 58
 Decision Strategies Texaco-Pennzoil Case

Second strategy: “Counter $5 billion and if Texaco counter

offers $3 billion refuse this counteroffer of $3 Billion”

Texaco Accepts $5 Billion (0.17)

5

High (0.20)
10.3
Counteroffer Texaco Refuses (0.50) Final Court Medium (0.50) 5
Decision
$5 Billion Counteroffer
Low (0.30)
0

High (0.20)
10.3

Refuse Final Court Medium (0.50) 5

Decision
Texaco Low (0.30)
(0.33) 0
Counter -
offers $3 Billion

Slide 22 of 58
 Decision Strategies Texaco-Pennzoil Case

Third strategy: “Counter $5 billion and if Texaco counter

offers $3 billion accept this counteroffer of $3 Billion”

Texaco Accepts $5 Billion (0.17)

5

High (0.20)
10.3
Counteroffer Texaco Refuses (0.50) Medium (0.50)
Final Court 5
Decision
$5 Billion Counteroffer
Low (0.30)
0
(0.33)
Texaco
Counter -
offers $3 Billion Accept $3 Billion
3

Slide 23 of 58
 Risk Profiles and Cumulative Risk Profiles

RISK PROFILES = Graph that shows probabilities for each of

the possible outcomes given a particular decision strategy.

Note: Risk Profile is a probability mass function for the discrete random
variable Y representing the outcomes for the given decision strategy.

CUMMULATIVE RISK PROFILES = Graphs that shows cumulative

probabilities associated with a risk profile

Note: Cumulative risk profile is a cumulative

distribution function for the discrete random variable Y representing
the outcomes for the given decision strategy.

Slide 24 of 58
 Risk Profiles

First strategy: “Accept $2 billion”

Accept $2 Billion
2

Outcome x ($Billion) Pr(Outcome|D) Risk Profile

D="Accept $2 Billion"
2 1
1 1

Pr(Outcome|D)
0.8
0.6
0.4
0.2
0
-1 2 5 8 11
Outcome ($Billion)

Slide 25 of 58
 Risk Profiles

Second strategy: “Counter $5 billion and if Texaco counter

offers $3 billion refuse this counteroffer of $3 Billion”

Calculation Prob
Texaco Accepts $5 Billion (0.17)
5 0.17 0.170

High (0.20)
10.3 0.50*0.20 0.100
Counteroffer Texaco Refuses (0.50) Final
Medium (0.50) 5 0.50*0.50 0.250
Court
$5 Billion Counteroffer Decision
Low (0.30) 0.50*0.30 0.150
0

High (0.20)
10.3 0.33*0.20 0.066
Final
Refuse Medium (0.50) 5 0.33*0.50 0.165
Court
Decision
Texaco Low (0.30) 0.33*0.30 0.099
0
Counter - (0.33)
offers $3 Billion Total 1.000

Slide 26 of 58
 Risk Profiles

Second strategy: “Counter $5 billion and if Texaco counter

offers $3 billion refuse this counteroffer of $3 Billion”
Outcome x ($Billion) Calculation Pr(Outcome| D)
0 0.150+0.099 0.249
5 0.170+0.250+0.165 0.585
10.3 0.100+0.066 0.166
1.000
Risk Profile D="Counter $5 Billion, refuse counter
offer of $3 Billion if given"

1
Pr(Outcome|D)

0.8
0.585
0.6
0.4
0.249
0.2 0.166
0
-1 2 5 8 11
Outcome ($Billion)

Slide 27 of 58
 Risk Profiles

Third strategy: “Counter $5 billion and if Texaco counter

offers $3 billion accept this counteroffer of $3 Billion”

Calculation Prob

Texaco Accepts $5 Billion (0.17) 0.17 0.170

5

High (0.20)
10.3 0.50*0.20 0.100
Counteroffer Texaco Refuses (0.50) Final
Court Medium (0.50) 5 0.50*0.50 0.250
$5 Billion Counteroffer
Decision
Low (0.30) 0.50*0.30 0.150
Texaco (0.33) 0
Counter -
offers $3 Billion
Accept $3 Billion
3 0.33 0.330

Total 1.000

Slide 28 of 58
 Risk Profiles
Third strategy: “Counter $5 billion and if Texaco counter
offers $3 billion accept this counteroffer of $3 Billion”
Outcome x ($Billion) Calculation Pr(Outcome| D)
0 0.15 0.15
3 0.33 0.33
5 0.170+0.250 0.42
10.3 0.1 0.1
1.000
Risk Profile D="Counter $5 Billion, Accept Counter
Offer of $3 Billion if given"

1
Pr(Outcome|D)

0.8
0.6
0.33 0.42
0.4
0.2 0.15 0.10
0
-1 2 5 8 11
Outcome ($Billion)

Slide 29 of 58
 Cumulative Risk Profiles

First strategy: “Accept $2 billion” Risk Profile

D="Accept $2 Billion"

1 1
Outcome x ($Billion) Pr(Outcome|D)

Pr(Outcome|D)
0.8
2 1
0.6
0.4
0.2
0
-1 2 5 8 11
Outcome ($Billion)

Outcome x ($Billion) Pr(Outcome  x|D) Cumulative Risk Profile

D="Accept $2 Billion"
2 1
1 1

Pr(Outcome  x|D)
0.8

0.6

0.4

0.2
0
0
-1 1 3 5 7 9 11
Outcome ($Billion)

Slide 30 of 58
 Cumulative Risk Profiles
Second strategy: “Counter $5 billion Risk Profile D="Counter $5 Billion, refuse counter
offer of $3 Billion if given"
and if Texaco counter offers $3 billion
1
refuse this counteroffer of $3 Billion”

Pr(Outcome|D)
0.8
0.585
0.6
0.4
Outcome x ($Billion) Pr(Outcome|D) 0.2
0.249
0 0.249 0.166
0
5 0.585 -1 2 5 8 11
10.3 0.166 Outcome ($Billion)

Cumulative Risk Profile D="Counter $5 Billion,

refuse counter offer of $3 Billion if given"

1 1
Outcome x ($Billion) Pr(Outcome  x|D) 0.834

Pr(Outcome  x|D)
0.8
0 0.249 0.834
0.6
5 0.249 + 0.585 = 0.834
0.4
10.3 0.834 + 0.166 = 1 0.249
0.2 0.249
0
0
-1 2 5 8 11
Outcome ($Billion)

Slide 31 of 58
 Cumulative Risk Profiles
Third strategy: “Counter $5 billion Risk Profile D="Counter $5 Billion, Accept Counter
Offer of $3 Billion if given"
and if Texaco counter offers $3 billion
1
accept this counteroffer of $3 Billion”

Pr(Outcome|D)
0.8
0.6
Outcome x ($Billion) Pr(Outcome|D) 0.42
0.4
0 0.15 0.15
0.33
3 0.33 0.2 0.10
5 0.42 0
-1 2 5 8 11
10.3 0.1
Outcome ($Billion)

Cumulative Risk Profile D="Counter $5 Billion,

accept counter offer of $3 Billion if given"
Outcome x ($Billion) Pr(Outcome  x|D) 1
0.90
1
0 0.15 0.90

Pr(Outcome  x|D)
0.8
3 0.15 + 0.33 = 0.48
0.6
5 0.48 + 0.42 = 0.90 0.48
0.48
0.4
10.3 0.90 + 0.10 = 1 0.15
0.2
0.15
0
-1 2 5 8 11
Outcome ($Billion)

Slide 32 of 58
 Deterministic Dominance

Original Tree
FALSE 0.000
Accept $2 Billion
2.000 2.000
Decision
Penzoill-Texaco
4.635

Texaco Accept $5 Biilion 17% 0.169

5.0 5.000
Counteroffer $5 Billion TRUE Chance
0.000 4.635

High Award 20% 0.100

10.3 10.300
Texaco Refuses Counteroffer 50% Chance
0.000 4.6

Medium Award 50% 0.249

5.0 5.000
Low Award 30% 0.150
0.0 0.000
High Award 20% 0.066
10.3 10.300
TRUE Chance
branch
0.000 4.6

Medium Award 50% 0.166

5.0 5.000
Low Award 30% 0.100
0.0 0.000
Texaco Counteroffers $3 Billion 33% Decision
0.000 4.560
FALSE 0.000
branch
3.0 3.000

Slide 33 of 58
 Deterministic Dominance

Modified Tree
FALSE 0.000
Accept $2 Billion
2.000 2.000
Decision
Penzoill-Texaco
5.257

Texaco Accept $5 Biilion 17% 0.169

5.0 5.000
Counteroffer $5 Billion TRUE Chance
0.000 5.257

High Award 20% 0.100

10.3 10.300
Texaco Refuses Counteroffer 50% Chance
0.000 5.3

Medium Award 50% 0.249

5.0 5.000
Low Award 30% 0.150
2.5 2.500
High Award 20% 0.066
10.3 10.300
TRUE Chance
branch
0.000 5.3

Medium Award 50% 0.166

5.0 5.000
Low Award 30% 0.100
2.5 2.500
Texaco Counteroffers $3 Billion 33% Decision
0.000 5.310
FALSE 0.000
branch
3.0 3.000

Slide 34 of 58
 Deterministic Dominance

Based on EMV analysis we still choose

the alternative “Counteroffer $5 Billion”

Max Result
Accept $2 Billion
EMV= 2
5.26

EMV=
5.26
Counteroffer

$5 Billion

Could we have made a decision

here without an EMV analysis ?

Slide 35 of 58
 Deterministic Dominance

Formal Definition: Deterministic Dominance:

If the worst outcome of Alternative B is at least as good
as that of the best outcome of Alternative A, then
Alternative B deterministically dominates Alternative A.

• Deterministic dominance may also be concluded by

drawing cumulative risk profiles and using the
definition:

Definition: Range of a Cumulative Risk Profile = [L,U],

where L= Smallest 0% point in Cumulative Risk Profile
and U= Largest 100% point in Cumulative Risk Profile

Slide 36 of 58
 Deterministic Dominance

• Deterministic dominance via cumulative risk

profiles:
 Step 1: Draw cumulative risk profiles in one graph
 Step 2: Determine range for each risk profile
 Step 3: If ranges are disjoint or their intersections contain a single point
then deterministic dominance
Cumulative is present
Risk Profiles
Revised Texaco-Penzoil Case Range 1: {2}
1 Range 2: [2.5,10.3]
Pr(Outcome  x)

0.8
0.6
Ranges 1 and 2
are disjoint. The
0.4
2.5 10.3 Objective is
0.2
Max Result,
0
hence Green CRP
-1 2 5 8 11
Outcome ($Billion)
deterministically
Accept $2 Billion dominates the Red
Counteroffer $5 Billion and Refuse $3 Billion one.
Slide 37 of 58
 Stochastic Dominance: Example 1

Firm A: Original Tree

FALSE 0.000
Accept $2 Billion
2.000 2.000
Decision
Penzoill-Texaco
4.635

Texaco Accept $5 Biilion 17% 0.169

5.0 5.000
Counteroffer $5 Billion TRUE Chance
0.000 4.635

High Award 20% 0.100

10.3 10.300
Texaco Refuses Counteroffer 50% Chance
0.000 4.6

Medium Award 50% 0.249

5.0 5.000
Low Award 30% 0.150
0.0 0.000
High Award 20% 0.066
10.3 10.300
TRUE Chance
branch
0.000 4.6

Medium Award 50% 0.166

5.0 5.000
Low Award 30% 0.100
0.0 0.000
Texaco Counteroffers $3 Billion 33% Decision
0.000 4.560
FALSE 0.000
branch
3.0 3.000

Slide 38 of 58
 Stochastic Dominance: Example 1

Firm B: Modified Tree

FALSE 0.000
Accept $2 Billion
2.000 2.000
Decision
Penzoill-Texaco
4.718

Texaco Accept $5 Biilion 17% 0.169

5.0 5.000
Counteroffer $5 Billion TRUE Chance
0.000 4.718

High Award 20% 0.100

10.3 10.300
Texaco Refuses Counteroffer 50% Chance
0.000 4.7

Medium Award 50% 0.249

5.2 5.200
Low Award 30% 0.150
0.0 0.000
High Award 20% 0.066
10.3 10.300
TRUE Chance
branch
0.000 4.7

Medium Award 50% 0.166

5.2 5.200
Low Award 30% 0.100
0.0 0.000
Texaco Counteroffers $3 Billion 33% Decision
0.000 4.660
FALSE 0.000
branch
3.0 3.000

Slide 39 of 58
Stochastic Dominance: Example 1

Based on EMV analysis we still choose

the alternative “Firm B”

Max Result
EMV=
4.63
Firm A
EMV=
4.72

EMV=
4.72
Firm B

Could we have made a decision

here without an EMV analysis ?
Slide 40 of 58
Stochastic Dominance: Example 1

Optimal Cumulative risk profiles in

“Firm A” Tree and “Firm B” Tree

Cumulative Risk Profiles: Firm A and Firm B

1
Pr(Outcome  x)

0.8
0.6
0.4

0.2
0
-1 2 5 8 11
Outcome ($Billion)
Firm A Firm B

Slide 41 of 58
 Stochastic Dominance: Example 1

Cumulative Risk Profiles: Firm A and Firm B

Note that for all
possible values of x: 1

Pr(Outcome  x)
Pr(Outcome  x| Firm B)  0.8

Pr(Outcome  x| Firm A) 0.6

0.4

0.2
or equivalently:
0
-1 2 5 8 11
Pr(Outcome  x| Firm B)  Outcome ($Billion)
Pr(Outcome  x| Firm A) Firm A Firm B

Hence the chances of winning with Firm B

are always better than that of Firm A.

Conclusion: Firm B stochastically dominates Firm A

Slide 42 of 58
 Stochastic Dominance: Example 2

Firm A: Original Tree

FALSE 0.000
Accept $2 Billion
2.000 2.000
Decision
Penzoill-Texaco
4.635

Texaco Accept $5 Biilion 17% 0.169

5.0 5.000
Counteroffer $5 Billion TRUE Chance
0.000 4.635

High Award 20% 0.100

10.3 10.300
Texaco Refuses Counteroffer 50% Chance
0.000 4.6

Medium Award 50% 0.249

5.0 5.000
Low Award 30% 0.150
0.0 0.000
High Award 20% 0.066
10.3 10.300
TRUE Chance
branch
0.000 4.6

Medium Award 50% 0.166

5.0 5.000
Low Award 30% 0.100
0.0 0.000
Texaco Counteroffers $3 Billion 33% Decision
0.000 4.560
FALSE 0.000
branch
3.0 3.000

Slide 43 of 58
 Stochastic Dominance: Example 2

Firm C: Modified Tree

FALSE 0.000
Accept $2 Billion
2.000 2.000
Decision
Penzoill-Texaco
6.005

Texaco Accept $5 Biilion 17% 0.169

5.0 5.000
Counteroffer $5 Billion TRUE Chance
0.000 6.005

High Award 30% 0.150

10.3 10.300
Texaco Refuses Counteroffer 50% Chance
0.000 6.2

Medium Award 60% 0.299

5.2 5.200
Low Award 10% 0.050
0.0 0.000
High Award 30% 0.100
10.3 10.300
TRUE Chance
branch
0.000 6.2

Medium Award 60% 0.199

5.2 5.200
Low Award 10% 0.033
0.0 0.000
Texaco Counteroffers $3 Billion 33% Decision
0.000 6.210
FALSE 0.000
branch
3.0 3.000

Slide 44 of 58
Stochastic Dominance: Example 2

Based on EMV analysis we still choose

the alternative “Firm C”

Max Result
EMV=
4.63
Firm A
EMV=
6.00

EMV=
6.00
Firm C

Could we have made a decision

here without an EMV analysis ?
Slide 45 of 58
Stochastic Dominance: Example 2

Optimal Cumulative risk profiles in

“Firm A” Tree and “Firm C” Tree

Cumulative Risk Profiles: Firm A and Firm C

1
Pr(Outcome  x)

0.8
0.6
0.4

0.2
0
-1 2 5 8 11
Outcome ($Billion)
Firm A Firm C

Slide 46 of 58
 Stochastic Dominance: Example 2

Cumulative Risk Profiles: Firm A and Firm C

Note that for all
possible values of x: 1

Pr(Outcome  x)
Pr(Outcome  x| Firm C)  0.8

Pr(Outcome  x| Firm A)
0.6
0.4

0.2
or equivalently:
0
-1 2 5 8 11
Pr(Outcome  x| Firm C)  Outcome ($Billion)
Pr(Outcome  x| Firm A) Firm A Firm C

Hence the chances of winning with Firm C

are always better than that of Firm A.

Conclusion: Firm C stochastically dominates Firm A

Slide 47 of 58
 Stochastic Dominance: Examples 1 & 2
Cumulative Risk Profiles: Firm A and Firm B
Commonality CRP plots:
1

Pr(Outcome  x)
0.8

• Cumulative risk profiles in 0.6

0.4
both plots do not cross 0.2
0
-1 2 5 8 11
• The CRP that is toward the Outcome ($Billion)

“right and below” Firm A Firm B

stochastically dominates Cumulative Risk Profiles: Firm A and Firm C

• The objective in both plots

Pr(Outcome  x)
0.8
0.6

is to Maximize the Result 0.4

0.2
0

• What if the objective is -1 2 5 8 11

Outcome ($Billion)
Minimize the Result? Firm A Firm C

Slide 48 of 58
 Making Decisions with Multiple Objectives

•Two Objectives:
Amount of
Fun
Making Money
(Measured in $)
Fun

Having Fun Job Overall

Decision Satisfaction
(Measured on
Constructed attribute Salary

scale, see page 138):

Best(5), Good(4),
Amount
Middle(3), Bad(2), of Work

Worst (1)

Slide 49 of 58
Making Decisions with Multiple Objectives
Consequences

Salary Fun Level

5 (0.10) $2600.00 5
Fun 4 (0.25) $2600.00 4
level
Forest Job 3 (0.40)
$2600.00 3

2 (0.20) $2600.00 2
1 (0.05)
$2600.00 1
# hours
per week 40 hours (0.35) $2730.00 3
In-Town Job 34 hours (0.50)
$2320.50 3

30 hours (0.15) $2047.50 3

Slide 50 of 58
 Analysis Salary Objective
Forest Job In-Town Job
Salary Prob Salary*Prob Prob Salary*Prob
$2,047.50 0.15 $307.13
$2,320.50 0.50 $1,160.25
$2,600.00 1.00 $2,600.00
$2,730.00 0.35 $955.50
E[Salary]= $2,600.00 E[Salary]= $2,422.88

Conclusion:
1
• Forest Job preferred

Pr(Outcome  x)
0.8
Over In-Town job 0.6
0.4

• CRP’s cross. Hence, 0.2

No Stochastic 0
$2,000 $2,200 $2,400 $2,600 $2,800 $3,000
Dominance Salary

Forest Job In-Town Job

Slide 51 of 58
 Fun Level Objective
Forest Job In-Town Job
Outcome Fun Level Prob Fun Level*Prob Prob Fun Level*Prob
5 -BEST 100.00% 0.10 10.0%
4 -GOOD 90.00% 0.25 22.5%
3 - MIDDLE 60.00% 0.40 24.0% 1.00 60.00%
2 - BAD 25.00% 0.20 5.0%
1 - WORST 0.00% 0.05 0.0%
E[Fun Level]= 61.5% E[Fun Level]= 60.00%

Conclusion:
• Forest Job preferred 1

Pr(Outcome  x)
0.8
Over In-Town job 0.6
0.4
• CRP’s cross. Hence, 0.2
No Stochastic 0
Dominance 0% 20% 40% 60% 80% 100% 120%
Fun Level

Forest Job In-Town Job

Slide 52 of 58
 Multiple Objective Analysis

• It is clear from both objective analyses that the Forest-Job

is the strongly preferred, although neither Stochastic
nor Deterministic Dominance can be observed in them.

• Careful as you are in your decisions you decide to

trade-off the salary objective and having fun objective
in a multiple objective analysis.

• Before trade-off analysis can be conducted both objectives

have to be measured on a “comparable” scale.

Slide 53 of 58
Multiple Objective Analysis: Construct 0-1 Scale

• Having Fun Objective already has a 0-1 scale:

Transformed to 0-1 scale or 0%-100% scale

Set Best=100%, Worst=0%, Determine intermediate values

Having Fun objective:

Best(100%), Good(90%), Middle(60%), Bad(25%), Worst (0%)
• Construct 0-1 scale for Salary Objective:
$2730.00=100%, $2047.50=0%

Intermediate dollar amount X:

X  $2047.50
100%
$2730  $2047.50

Slide 54 of 58
 Multiple Objective Analysis: Assess Trade-Off

ks = weight for salary kf = weight for fun

ks  k f  1
Using Expert Judgment:
Going from worst to best in salary objective is 1.5 times more important
than going from worst to best in having fun objective. Hence: ks  1.5 k f

k 1

2
 ks  k f  1 1.5 k f  k f  1  f  2.5 5
k  1.5 k   k  1.5 k   3 3 3
 s f  s f ks   
 2 5 5

Slide 55 of 58
Multiple Objective Analysis: Convert Scales
Consequences

Salary (0.6) Fun Level (0.4)

5 (0.1 0) 81% 100%

Fun 4 (0.2 5) 81% 90%
level
Forest Job 3 (0.4 0)
81% 60%

2 (0.2 0) 81% 25%

1 (0.0 5)
81% 0%

# hours
per week 40 hours (0.35) 100% 60%
In-Town Job 34 hours (0.50)
40% 60%

30 hours (0.15) 0% 60%

Slide 56 of 58
Multiple Objective Analysis: Combine Objectives

Total Score

5 (0.10) 88.6%
Fun 4 (0.25) 84.6%
level
Forest Job 3 (0.40)
72.6.%

2 (0.20) 58.6%
1 (0.05)
48.6%

# hours
per week 40 hours (0.35) 84.0%
In-Town Job 34 hours (0.50)
48.0%

30 hours (0.15) 24.0%

Slide 57 of 58
 Analysis Overall Satisfaction
Forest Job In-Town Job
Overall Satisfaction Prob OS*Prob Overall Satisfaction Prob OS*Prob
88.57% 0.10 8.9% 84.00% 0.35 29.40%
84.57% 0.25 21.1% 48.00% 0.50 24.00%
72.57% 0.40 29.0% 24.00% 0.15 3.60%
58.57% 0.20 11.7% E[OS]= 57.00%
48.57% 0.05 2.4%
E[OS]= 73.2%

Conclusion:
1
• Forest Job preferred

Pr(Outcome  x)
0.8
Over In-Town job 0.6
0.4
• CRP’s do not cross. 0.2
Hence, Stochastic 0
0% 20% 40% 60% 80% 100%
Dominance present.
Overall Satisfaction

Forest Job In-Town Job

Slide 58 of 58