Value of Perfect Information (cont.

EPPI = Σ CP* (State i) P (State i)

i

EVPI Expected Value of Perfect

Information

EVPI = EPPI - EMV*

Value of Perfect Information (cont.)

EVPI Expected Value of Perfect
Information
EVPI = $ 18,600 - 16,500= $ 2,100
Do you recognize this number???
It is the same as EOL*
EVPI = EOL* (always)
WHY???

Decision Trees and Sequential

Decision Making
Decision Tree:
• Organizes all elements of complex decision
into a meaningful arrangement
• Useful when set of possible states is
different for each possible decision
• Useful when survey or experiment is
possible for decision maker to use

1
 Decision Trees and Sequential
Decision Making
Components:
Branches- represent event or decision
Decision node- point at which decision is
made
Event (state) node- point at which state
occurs
Outcome (payoff)- result of following path
of decisions and states
CP

Decision Trees
State 1 (prob) CP1
Dec. 1 State 2 (prob) CP2
State 3 (prob) CP3
State 1 (prob) CP1
Dec. 2 State 2 (prob) CP2

State 1 (prob)
Dec. 3 State 2 (prob)
State 3 (prob)

Decision Tree

EMV = 16,500 Strong (.2) 60,000

Manuf. Avg. (.5) 12,000
Weak (.3) -5,000

Strong (.2) 40,000

Sell Avg. (.5) 7,000
EMV = 12,100 Weak (.3) 2,000

2
 Decision Trees and Sequential
Decision Making
Sequential decision: There is more than one
decision to be made, and earlier decisions
have an effect on later decisions.

State 1 (prob) Dec. 1 CP

Dec. 3 State 2 (prob) CP
State 3 (prob) CP
CP
Dec. 2
CP
CP

Risk Profiles
Show individual and cumulative probs.
associated with all payoffs for a given
decision
1.0
.75
For Manufacture decision
.50
.25

-5000 12,000 60,000

Payoff Table

Act Act 1 Act 2

Event Prob. Manuf. Sell
Strong Sales .2 $60,000 $40,000
Average Sales .5 12,000 7,000
Weak Sales .3 -5,000 2,000

3
 Risk Profiles
Cumulative risk profile

1.0
.75
For Manufacture decision
.50
.25

-5000 12,000 60,000

Payoff Table

Act Act 1 Act 2

Event Prob. Manuf. Sell
Strong Sales .2 $60,000 $40,000
Average Sales .5 12,000 7,000
Weak Sales .3 -5,000 2,000

Risk Profiles
Individual risk profile

1.0
For Sell decision
.75
.50
.25

2000 7000 40,000

4
 Risk Profiles
Cumulative risk profile

1.0
.75
For Sell decision
.50
.25

2000 7,000 40,000

Risk Profiles

For Manufacture decision

1.0
.75
For Sell decision
.50
.25

-5000 2000 7,000 12,000 40,000 60,000

Risk Profiles
Neither decision is dominant
For Manufacture decision
1.0
.75
For Sell decision
.50
.25

-5000 2000 7,000 12,000 40,000 60,000

5
 Payoff Table

Act Act 1 Act 2

Event Prob. Manuf. Sell
Strong Sales .2 $60,000 $40,000
Average Sales .5 12,000 7,000
Weak Sales .3 5,000 2,000

Manufacture is dominant because for every

state, payoff is higher than for Sell

Risk Profiles
Now Manufacture is dominant decision
(all of line is under and/or to right)
For Manufacture decision
1.0
.75
For Sell decision
.50
.25

-5000 2000 7,000 12,000 40,000 60,000

Sequential Decision Making

Decision trees are very useful when there are
multiple decisions to be made and they follow a
sequence in time. There are also usually multiple
sets of states. ion 1 CP
Decis
State 1 Decision 2 State 1 CP
Decision 1 State 2 CP State 2
e 1 State 1 CP
St at CP ecision 1 CP
Decision 2 State 2 D Stat
Decision 2 e2
e 1 CP CP
Stat CP ecision 1 CP
Decision 3 State 2 D
Decision 2
CP

6
 Sequential Decision Tree Problem
Suppose that you are trying to decide which of
three companies to invest in, Company A, B, or
C. If you choose A, there is a 50/50 chance of
going broke or earning $50,000. If you go
broke with A, you then have three choices:
accept a debt of $2,000; embezzle $35,000 of
company money [disclaimer: this problem is
for illustration purposes only!] and leave the
country; or file for personal bankruptcy at the
hands of a court-appointed trustee.

Sequential Decision Tree

Invest in A

Invest in B

Invest in C

Sequential Decision Tree Problem

Suppose that you are trying to decide which of
three companies to invest in, Company A, B, or
C. If you choose A, there is a 50/50 chance of
going broke or earning $50,000. If you go
broke with A, you then have three choices:
accept a debt of $2,000; embezzle $35,000 of
company money [disclaimer: this problem is
for illustration purposes only!] and leave the
country; or file for personal bankruptcy at the
hands of a court-appointed trustee.

7
 Sequential Decision Tree Debt

ke(.5) Embezzle
Go Bro
Invest in A
Ea Bankrupt
rn
$$
(.5
)
$50,000
Invest in B

Invest in C

Sequential Decision Tree Problem

If you embezzle money and leave the country,

there is a 95% chance of being extradited and
fined $10,000. If you file for personal
bankruptcy, there is a 95% chance that your
debts will be wiped out and a 5% chance that
you will have to pay back $4,000.

Sequential Decision Tree

Debt -$2K 5)
t r ad(.9 -$10K
x
Go Broke(.5) Embezzle E N
A ot(
.05
Ea )
rn $35K
$$ P ay b ac
(.5 Bankrupt k(.05)
) W
$50,000 ipe -$4K
B d ou
t(.9
5)
$0

8
 Sequential Decision Tree Problem

If you choose Company B, there is an 80

percent chance of earning $25,000. If
Business B fails, you still have the option
of either settling for $500 or taking a stock
option in the company that will be worth
$50,000 with probability 0.1 or zero with
probability 0.9.

Sequential Decision Tree

Debt -$2000 )
(.95
x t rad -$10000
Go Broke(.5) Embezzle E N
A ot(
.05
Ear )
n$ $35000
$(.5 P ay b ac
) Bankrupt k(.05)
$50000 W
Earn $$(.8) $25000 ipe-$4000
B ou d
t(.9
B fails(.2) Settle $500 5)
$0
Stock Earn(.1)$50000
C option No
t(.9
)$ 0

Sequential Decision Tree Problem

Finally if you choose Company C, you will

either earn $10,000 with probability 0.6, or be
in debt for $1,000 with probability 0.4.

9
 Sequential Decision Tree
Debt )
-$2000
(.95
x t rad -$10000
E
Go Broke(.5) Embezzle No
A t(.0
Ear 5)
n$ $35000
$(.5 P
) Bankrupt a y back(.05
$50000 W )
Earn $$(.8) $25000 ipe -$4000
B d ou
t(.9
B fails(.2) Settle $500 5)
$0
Earn $$(.6) Stock Earn(.1)$50000
$10000 No
C option t(.9
Debt(.4) -$1000 )
$0

Sequential Decision Tree Problem

- Solving by Folding Back the tree
For each set of state branches, find the
EMV for the connected decision.
For each set of decisions, select the one
with the highest EMV and carry the EMV*
forward (to the left)
Trees are drawn from left to right; they are
folded back from right to left.

Sequential Decision Tree

Debt -$2000
EMV*= )
(.95
$-200 EMV=-$7750 xtrad -$10000
e(.5) Embezzle E
Go Brok No
A t(.0
Ear 5)
n$ EMV=-$200 Pa $35000
$(.5
) Bankrupt y back (.05)
$50000 Wi
ped -$4000
B Earn $$(.8) $25000 o ut (
.95
Settle )
B fails(.2) $500
$0
Earn $$(.6) Stock Earn(.1)
$10000 $50000
C option No
t(.9
)
Debt(.4) -$1000 $0

10
 Sequential Decision Tree
Debt -$2000
EMV*= )
.95
EMV= $-200 EMV=-$7750 xtrad( -$10000
$24,900 e(.5) Embezzle E
Go Brok No
A t(.0
Ear 5)
n$ EMV=-$200 Pa $35000
$(.5
) Bankrupt y back (.05)
$50000 Wi
ped -$4000
B Earn $$(.8) $25000 o ut (
.95
Settle )
B fails(.2) $500
$0
Earn $$(.6) Stock Earn(.1)
$10000 $50000
C option No
t(.9
)
Debt(.4) -$1000 $0

Sequential Decision Tree

Debt -$2000 )
(.95
x t rad -$10000
Go Broke(.5) Embezzle E N
A ot(
.05
Ear )
n$ $35000
$(.5 P ay b ac
) Bankrupt k(.05)
$50000 W
Earn $$(.8) $25000 ipe -$4000
B do
ut(
EMV*=$5000 Settle .95
B fails(.2) $500 )
$0
Earn $$(.6) Stock Earn(.1)
$10000 No $50000
C option
EMV=$5000 t(.9)
Debt(.4) -$1000 $0

Sequential Decision Tree

Debt -$2000 5)
t r ad(.9-$10000
x
Go Broke(.5) Embezzle E N
A ot(
.05
Ear )
n$ $35000
$(.5 P ay b ac
EMV= ) Bankrupt k(.05)
$21,000 $50000
ipe -$4000 W
B Earn $$(.8) $25000 do
EMV*=$5000 Settle ut(
.95
B fails(.2) $500 )
$0
Earn $$(.6) Stock Earn(.1)
$10000 No $50000
C option
EMV=$5000 t(.9)
Debt(.4) -$1000 $0

11
 Sequential Decision Tree
Debt -$2000 )
(.95
x t rad -$10000
Go Broke(.5) Embezzle E N
A ot(
.05
Ear )
n$ $35000
$(.5 P ay b ac
) Bankrupt k(.05)
$50000 W
Earn $$(.8) $25000 -$4000
ipe
B ou d
t(.9
B fails(.2) Settle $500 5)
EMV= $0
$5600 Earn $$(.6) Stock Earn(.1)
$10000 N $50000
C option ot(
.9 )
Debt(.4) -$1000 $0

Sequential Decision Tree

Debt -$2000
EMV*= )
EMV= (.95
$-200 EMV=-$7750 xtrad -$10000
$24,900 e(.5) Embezzle E
Go Brok No
A t(.0
Ear 5)
n$ EMV=-$200 Pa $35000
$(.5
) Bankrupt y back (.05)
$50000 Wi
ped -$4000
B EMV= Earn $$(.8) $25000 o ut (
$21,000 EMV*=$5000 Settle .95
)
B fails(.2) $500
$0
EMV= Earn $$(.6) Earn(.1)
Stock
$10000 $50000
No
option
$5600 t(.9
C EMV=$5000 )
Debt(.4) -$1000 $0

Decision Tree Solution

After you fold back the tree and determine
the best initial decision, then state the
complete optimal sequence of decisions:

Invest in Company A. If you go broke, then

file for bankruptcy. Otherwise enjoy the
$50,000!!

12
 Sequential Decision Tree
Debt -$2000 )
(.95
x t rad -$10000
E
Go Broke(.5) Embezzle No
A t(.0
Ear 5)
n$ $35000
$(.5 P
) Bankrupt a y back(.05
$50000 W )
Earn $$(.8) $25000 ipe -$4000
B d ou
t(.9
B fails(.2) Settle $500 5)
$0
Earn $$(.6) Stock Earn(.1)$50000
$10000 No
C option t(.9
Debt(.4) -$1000 )
$0

Risk Profiles for ABC Investment problem

Invest in A; Debt
1.00

.75

.50

.25

-$2000 $50,000

Sequential Decision Tree

Debt -$2000 )
.95
d(-$10000
x t r a
Go Broke(.5) Embezzle E N
A ot(
.05
Ear )
n$ $35000
$(.5 P ay b ac
) Bankrupt k(.05)
$50000 W
Earn $$(.8) $25000 ipe -$4000
B do
ut(
Settle .95
B fails(.2) $500 )
$0
Earn $$(.6) Stock Earn(.1)$50000
$10000 No
C option t(.9
Debt(.4) -$1000 )
$0

13
Risk Profiles for ABC Investment problem

Invest in A; Embezzle
1.0

.75

.50

.25

-$10,000 $35,000 $50,000

Sequential Decision Tree

Debt -$2000 )
(.95
x t rad -$10000
E
Go Broke(.5) Embezzle No
A t(.0
Ear 5)
n$ $35000
$(.5 P
) Bankrupt a y back(.
$50000 05)
W
Earn $$(.8) $25000 ipe -$4000
B d ou
t(.9
B fails(.2) Settle $500 5)
$0
Earn $$(.6) Stock Earn(.1)$50000
$10000 No
C option t(.9
Debt(.4) -$1000 )
$0

Risk Profiles for ABC Investment problem

Invest in A; Bankrupt
1.00

.75

.50

.25

-$4000 $ 0 $50,000

14