MODULE 8

THE DECISION
THEORY
THE DECISION THEORY
• Decision means to make a choice or judgment or come to a conclusion.
• Decision theory refers to a logical and systematic approach to decision
making.

Department of Accountancy - Management Science

STEPS IN DECISION MAKING
1. Define the problem clearly.
2. List all the viable alternatives that must be considered in making the
decision.
3. List the future events or possible outcomes that may occur.( These
future events or possible outcomes are beyond the control of the
decision makers).
4. List the profits or any other measures of benefits resulting from each
possible combination of decision alternative and possible outcome.
5. Apply any one of the decision criterion and make the decision.

Department of Accountancy - Management Science

8.1 BASIC CONCEPT OF PROBABILITY
• Probability is a measure of certainty. Its value ranges from zero to one
(1.00). A probability of zero means the event will never or impossible to
occur, whereas a probability of one means that the event is certain to
occur.
• The probability of success (Ps) plus the probability of failure (Pf) is equal
to one(1.00) or 100%. In symbol,
P(s) + P(f)= 1
And P(s)= 1- P(f)
Or P(f) = 1- P(s)
the probability can be expressed as a decimal, as fraction or as a percent

Department of Accountancy - Management Science

8.2 MATHEMATICAL EXPECTATION (ME) OR EXPECTED VALUE (EV)
•   Mathematical expectation (ME) or Expected Value (EV) is the amount one
is expected to receive (gain) or to pay (loss) for the occurrence of an
event.
• ME is dependent on two factors; the probability that an event will occur,
and the amount expected to be received.
• If X represents the amount to be received and P the probability that an
event will occur , then, ME = PX and if several events are expected to
occur , then,
ME==+…
• Sometimes , the expected value is negative (-) if instead of receiving an
amount for the occurrence of an event, one may have to pay or lose.
• In a fair and equitable game, ME=0.
Department of Accountancy - Management Science
8.2 MATHEMATICAL EXPECTATION (ME) OR EXPECTED VALUE (EV)
Example 8.2.1.
• In a game involving the roll of a single die , the following rules were
suggested. I pay P2 for each roll of the die. If a 5 comes out, I receive
P2.50, if 6 comes out, I receive P4, and if a 4 comes out, I receive P1.50. I
receive nothing if 1, 2, or 3 comes out.

Department of Accountancy - Management Science

8.2 MATHEMATICAL EXPECTATION (ME) OR EXPECTED VALUE (EV)
•Example
  8.2.1 – Solutions
ME=PX
=2.50-2=0.50 =1/6
=4.00-2=2.00 =1/6
=1.50-2=-0.50 =1/6
=0-2=-2.00 =1/6+1/6+1/6=1/2

ME=1/6(0.50)+1/6(2)+1/6(-0.50)+1/2(-2)=-P0.67
This means that in the long run I expect to lose P0.67 per game.

Department of Accountancy - Management Science

8.2 MATHEMATICAL EXPECTATION (ME) OR EXPECTED VALUE (EV)
•Example
  8.2.2
• A sidewalk vendor usually earns P500 (net) on a fair day. When it rains
hard, he stands to lose P300 for the day. What would be his expectation if
the probability of rain is 1/3?

Example 8.2.2 – Solution

=2/3 =P500
=1/3 =-P300 (negative sign for loss)
ME=2/3(500)+1/3(-300)= P233.33

Department of Accountancy - Management Science

8.2 MATHEMATICAL EXPECTATION (ME) OR EXPECTED VALUE (EV)
•Example
  8.2.3
• If the concert is to be held outdoors, the promoter stands to lose P5,000 if it
rains, or gain P35,000 if it does not rain. If the concert is held in the
gymnasium, he stands to gain P32,000 if it will not rain and lose P4,000 if it
rains. Which choice has a higher expectation if the probability of rain is 43%?

Example 8.2.3 – Solution

Outdoor concert: Gymnasium:
=0.43 =-P5,000 (loss) =0.43 =-P4,000
=0.57 =P35,000(gain) =0.57 =P32,000
ME=(0.43)(-5,000)+0.57(35,000) ME=0.43(-4,000)+0.57(32,000)
= P 17,800 =P16,520
Department of Accountancy - Management Science
8.3 THE DECISION TREE ANALYSIS
• Graphical representation of a decision situation
• It is a graphic tool that facilitates the decision process by giving a clear
picture of the decision problem, thereby helping decision makers analyze
possible outcomes
• A decision tree has two nodes:
1. Square Node – representing the decision point or the point where a
choice must be made
2. Circle Node – representing the chance event or expected value

Department of Accountancy - Management Science

8.3 THE DECISION TREE ANALYSIS
Example 8.3.1
• Because the lead actor in a new TV program of a local TV network is involved
in a drug related controversy and is generating bad publicity, the management
of the local TV network has decided whether to show the new program on TV
or not. The cost of producing the program is P1M. If the program is shown,
the network may earn of P3M (from advertisement) if it is accepted by the
viewers regardless of the bad publicity. If it fails to be accepted by the viewing
public it has to be shelved, and the network stands to lose P1.3M. The
network estimated that there is a 59-41 chance of the program being
accepted. Should the network go ahead and show the new program?
• The branches of the tree emanating from a square node represent
alternatives and branches emanating from circle nodes represent the possible
outcome that are beyond the control of the decision maker.
Department of Accountancy - Management Science
8.3 THE DECISION TREE ANALYSIS
Example 8.3.1 – Solution

To show

No t
to s
h ow
Decision point

Department of Accountancy - Management Science

8.3 THE DECISION TREE ANALYSIS
Example 8.3.1 – Solution
• Now let us indicate the financial values and probabilities
Program
accepted
B (Success) P3M - P1M
A P=0.59 = P2m

w
To sho
-P1.3M – P1M
Program = P2.3M
No Not
t to
sho Accepted
w P1M
Decision point (failure)
P=0.41

Department of Accountancy - Management Science

8.3 THE DECISION TREE ANALYSIS
Example 8.3.1 – Solution

Success (P3M – P1M)(0.59) = P1.18M

To show
no
t
Failure (-P1M-P1.3M)(0.41) = -P0.943M
EV = P0.237M
-1M
The final alternatives are: P0.237M and –P1M
Decision: The local TV networks should decide to show the new program
since they still gain profit.

Department of Accountancy - Management Science

8.3 THE DECISION TREE ANALYSIS
Example 8.3.2
• A canteen concessionaire has to decide whether to prepare a bid or not
for the canteen concession of PCU. The cost of preparing the bid is P200.
If the bid is submitted it has a 60-40 chance that the concession will be
awarded to her. If she gets the contract, she may earn a monthly income
of P52,000 if it succeeds, (that is, if the students patronize the canteen) or
incurs a monthly loss of P13,000 if it fails. The Concessionaire are
estimated the probability of success to be 75%. Should she prepare the
bid or not?

Department of Accountancy - Management Science

8.3 THE DECISION TREE ANALYSIS
Example 8.3.2 – Solution

P52,000 – P200
C ss = P51,800
su cce
e d
B a rd P35,550 P=0.75
A aw
prepare fai
P21,250 lur
Decision No e
no t -P13,000 – P200
t aw
ar = - P13,200
de
d

Department of Accountancy - Management Science

8.3 THE DECISION TREE ANALYSIS
Example 8.3.2 – Solution
For the success branch: P52,000 – P200 = P51,800
For the failure branch: - P13,000 – P200 = -P13,200
• Then solve for the EV of nodes C and B using the backward pass
technique.
Node C: EV = 0.75 (51,800) + ).25 (-13,200) = P35,550
Node B: EV = 0.60 (35,550) + 0.40 (-200) = P13,200
• The final alternatives are: to prepare the bid with an income of P21,250
and not to prepare with no income (0) at all.
• Decision: The Canteen concessionaire should prepare the bid.

Department of Accountancy - Management Science

8.3 THE DECISION TREE ANALYSIS
Example 8.3.3
• An investor who would like to invest in the Philippines has to decide
whether to build a large plant or a small plant which is expected to have a
market life of 10 years to manufacture their own products. The cost of
building the large plant and putting it into operation is P280M. While that
of the small plant is P140M. The investor estimates that the distribution
of sales over the 10-year period is as follows:
High Demand Probability = 0.49
Moderate DemandProbability = 0.31
Low Demand Probability = 0.20

Department of Accountancy - Management Science

8.3 THE DECISION TREE ANALYSIS
Example 8.3.3
• The investor made surveys and investigations and come up with the
following:
• A large plant with high demand would yield an annual profit of P100M while
that with moderate demand would yield P60M. A large plant with low
demand would lose P20M annually due to production inefficiencies.
• A small plant with high demand would yield only P25M annual profits due to
lost sales because of insufficient supply, while that with moderate would yield
an annual profit of P45M because the lost sales would be somewhat lower. A
small plant with low demand would yield P55M annually since the plant size
matched the market size.
• Armed with this information, should the investor build a large plant or a small
plant if he is to base his decision on expected profit?
Department of Accountancy - Management Science
8.3 THE DECISION TREE ANALYSIS
Example 8.3.3 – Solution
• The immediate problem is whether to build a large plant or a small
plant.

build large
plant
A

Decision

build small
plant

Department of Accountancy - Management Science

8.3 THE DECISION TREE ANALYSIS
Example 8.3.3 – Solution
• The starting point is square A (decision point A). The investor has to
decide whether to build a large plant or small plant
high demand
B
build large moderate demand
plant
A
low demand

Decision
high demand
C
build small
plant moderate demand

low demand
Department of Accountancy - Management Science
8.3 THE DECISION TREE ANALYSIS
Example 8.3.3 – Solution
• All this points, add the monetary and probability values and compute
for the expected value or profit (EV) of the event nodes.
B (0.49) (100) (10)
HD
= P490 M
MD (0.31) (60) (10)
356M = 186 M
build large plant LD
(0.20) (-20) (10)
A = -40 M

Total EV = P 636 M
Decision C - Plant Cost = 280 M
net EV = 356 M

(0.49) (25) (10)

HD = P122.5 M
build small plant
MD (0.31) (45) (10)
232M = 139.5 M
LD
(0.20) (55) (10)
= 110 M
Total EV = P 372 M
- Plant Cost = 140 M
Department of Accountancy - Management Science net EV = P 232 M
8.3 THE DECISION TREE ANALYSIS
Example 8.3.3 n d (0.49)(100)(10) = P 490M
m a
All financial data are in B de
h
millions of pesos (M) Hig (0.31)(60)(10) = 186M
a nt Mod. demand
uild pla P356M
Lo
B rge de w (0.20)(-20)(10) = - 40 M
A la m
an
d Total EV = P 636 M
Decision Plant cost = - 280 M
Net EV = P 356 M
Bu C High and
pla ild m
de (0.49)(25)(10)
nt a sm = P122.5M
all P232M
Mod. demand
Lo (0.31)(45)(10) = 139.5 M
w
de
m = 110 M
an (0.20)(55)(10)
d Total EV = P 372M
Plant cost = - 140M
Department of Accountancy - Management Science Net EV = P 232M
8.3 THE DECISION TREE ANALYSIS
Example 8.3.3 – Solution
• The final alternatives are: P356M for the large plant and P232M for a
small plant.
• Decision: The investor should build the large plant since the profit is
greater.

Department of Accountancy - Management Science

8.4 VALUE OF PERFECT INFORMATION
EVPI
• Expected profit with perfect information – EMV (of final decision)
• Example 8.4.1 Art-Sion Co. is debating whether or not to obtain a patent for a new
product they are developing. It costs P30,000 to obtain the patent. If the Company
obtains the patent, it should develop the product. There are two methods of
developing the product – Method I and method II. Cost for Method I is P40,000
and for method II is P35,000. Probability of success for Method I is 55% and for
Method II is 60%. If the product is successfully developed, the company may earn
P200,000. Should the company purchase the patent?
Assume that before deciding the company has the option to buy perfect
information from a research firm which has made research on these two methods
and has perfect information on whether only Method I, only Method II, either
method or neither can be successful. Should the company buy the information.
What is the EVPI
Department of Accountancy - Management Science
8.4 VALUE OF PERFECT INFORMATION
Solution
• Let financial data be in 000’s pesos if the company does not decide to buy
s
the perfect information. C ce
s 200 – 30 – 40 =
c
Su 0.55 130
=
PP
P 40 Fa =0
A Purchase
B ilu
re
.45
-30 – 40 = -70
Patent
Decision P55

D
NO

s
Pu Pat

s
ce
rc en

c 200 – 30 – 35 =
ha t

u
S .60 135
se

0
0 P 55 PP
=
Fa =0
ilu .40
re
-30 – 35 = -65
Department of Accountancy - Management Science
8.4 VALUE OF PERFECT INFORMATION
The final alternatives are P55,000 and 0.
Decision
• The President should obtain the patent and use Method II in developing the
product.

If, however, the company decides to buy the perfect information the company
will have to determine the expected value ( or profit ) of the additional
information and compare it with the cost of this additional information.
Solution:
Let all financial data be in 000’s pesos if the company does not decide to buy
the perfect information.

Department of Accountancy - Management Science

8.4 VALUE OF PERFECT INFORMATION
• The probability of success and failure with perfect information of each of the four
branches of our decision tree was determined in the following manner:

Method I only: P = P of Method I times Pf of Method II

= 0.55 (0.40) = 0.22
Method II only: P = P Method II times Pf of Method I
= 0.60 (0.45) = 0.27
Either method: P = P of Method I times P of Method II
= 0.55 (0.60) = 0.33
Neither method: P = P of method I times P of Method II
= 0.45 (.40) = 0.18

Department of Accountancy - Management Science

8.4 VALUE OF PERFECT INFORMATION
• Cost of perfect information = EV with perfect information – EV without
perfect information.

EVPI of cost of PI = P104.2 thousand – 55 = P49.20 thousand

• Decision: The company should buy the perfect information and use
Method II in developing the product, since it will earn a higher expected
value (or profit).

Department of Accountancy - Management Science

8.4 VALUE OF PERFECT INFORMATION
d I = 0.22 (130) = 28.6
ho
et o d II
Solution: M
Met
h
Eith = 0.27 (135) = 36.45
P104.20 er

N
= 0.33 (135) = 44.55

ei
th
er
nfo
= 0.18 (-30) = - 5.40
Total = 104.20

tI
Buy Perfec
ess 200 – 30 – 40 =
c c 5
Su = 0.5 130
A B P 40
P

Fa
Purchase -30 – 40 = -70

ilu
P
Patent

=0

re
P104.20 Not

.45
Decision
Bu to c e ss
c 200 – 30 – 35 =
y Su 60
in
NO ase

.
Pu Pat

0 135
fo P55 P 55 P=
rc en

T
h t

Fa
i
P

lu
-30 – 35 = -65
0

re
0.
40
Department of Accountancy - Management Science
8.4 VALUE OF PERFECT INFORMATION
• Since the EV with perfect information is already computed (EV = P104.2 thousand)
then compute for the cost of perfect information (EVPI).
EV = 0.22 (130) + 0.27 (135) + 0.33 (135) + 0.18 (-30)
EV = P 104.20 thousand

Cost of Perfect Information =

EV with Perfect Information – EV without Perfect Information

EVPI or Cost of PI = 104.4 – 55

= P 49.20 thousand
Decision
The company should buy the perfect information and use Method II in developing
the product since it will earn a higher expected value (or profit).
Department of Accountancy - Management Science
8.5 DECISION ANALYSIS USING CONDITIONAL PROFIT AND
EXPECTED VALUE
Following procedure in solving the problem:
1. Calculate the conditional profit table
2. Compute for the expected profit resulting from each stock level
3. Choose the stock level with the highest profit

Department of Accountancy - Management Science

8.5 DECISION ANALYSIS USING CONDITIONAL PROFIT AND
EXPECTED VALUE
1. Calculate the conditional profit table
For demand = 20 baskets For demand = 22 baskets
P 45 (20) – P 30 (20) = P 300 P 45 (20) – P 30 (20) = P 300
P 45 (20) – P 30 (21) = P 270 P 45 (21) – P 30 (21) = P 315
P 45 (20) – P 30 (22) = P 240 P 45 (22) – P 30 (22) = P 330
P 45 (20) – P 30 (23) = P 210 P 45 (22) – P 30 (23) = P 300

For demand = 21 baskets For demand = 23 baskets

P 45 (20) – P 30 (20) = P 300 P 45 (20) – P 30 (20) = P 300
P 45 (21) – P 30 (21) = P 315 P 45 (21) – P 30 (21) = P 315
P 45 (21) – P 30 (22) = P 285 P 45 (22) – P 30 (22) = P 330
P 45 (21) – P 30 (23) = P 255 P 45 (23) – P 30 (23) = P 345

Department of Accountancy - Management Science

8.5 DECISION ANALYSIS USING CONDITIONAL PROFIT AND
EXPECTED VALUE
Table 8.1 Conditional Profit Table of Ruby

Total = 86 Total = 1.00

Department of Accountancy - Management Science
8.5 DECISION ANALYSIS USING CONDITIONAL PROFIT AND
EXPECTED VALUE
2. Compute for the expected profit resulting from each stock level
Expected Monetary Value = Conditional Profit × Probability

Department of Accountancy - Management Science

8.5 DECISION ANALYSIS USING CONDITIONAL PROFIT AND
EXPECTED VALUE
Table 8.2 Expected Monetary Value of Ruby

Highest Expected Value

Department of Accountancy - Management Science
8.5 DECISION ANALYSIS USING CONDITIONAL PROFIT AND
EXPECTED VALUE
3. Choose the stock level with the highest profit
Decision: Ruby should buy 21 baskets daily with a maximum expected
value of P306.

Conditional Profit with Salvage Value =

No. of left over × Selling Price + Conditional Profit Without Salvage Value

Department of Accountancy - Management Science

8.5 DECISION ANALYSIS USING CONDITIONAL PROFIT AND
EXPECTED VALUE
Calculating Profit With Salvage Value For demand = 22 baskets
For demand = 20 baskets P 300 + P 5 (0) = P 300
P 300 + P 5 (0) = P 300 315 + P 5 (0) = P 315
270 + P 5 (1) = P 275 330 + P 5 (0) = P 330
240 + P 5 (2) = P 250 300 + P 5 (1) = P 305
210 + P 5 (3) = P 225
For demand = 23 baskets
For demand = 21 baskets P 300 + P 5 (0) = P 300
P 300 + P 5 (0) = P 300 315 + P 5 (0) = P 315
315 + P 5 (0) = P 315 330 + P 5 (0) = P 330
285 + P 5 (1) = P 290 345 + P 5 (0) = P 345
255 + P 5 (2) = P 265

Department of Accountancy - Management Science

8.5 DECISION ANALYSIS USING CONDITIONAL PROFIT AND
EXPECTED VALUE
Table 8.3 Conditional Profit Table with Salvage Value of Ruby

Department of Accountancy - Management Science

8.5 DECISION ANALYSIS USING CONDITIONAL PROFIT AND
EXPECTED VALUE
Table 8.4 Expected Monetary Value with Salvage

Highest EMV of Ruby

Department of Accountancy - Management Science
8.5 DECISION ANALYSIS USING CONDITIONAL PROFIT AND
EXPECTED VALUE
Decision: Ruby should stock 21 baskets daily for a maximum profit of P307.

Department of Accountancy - Management Science

8.6 EXPECTED PROFIT WITH PERFECT INFORMATION
• Perfect Information is the knowledge that a future state of nature will
occur with certainty, that is, being sure of what will occur in the future.

Department of Accountancy - Management Science

8.6 EXPECTED PROFIT WITH PERFECT INFORMATION
Example
• If the demand is known for 20 baskets of strawberry, she will buy exactly
20 baskets for the day for a conditional profit of P300.If the known
demand is for 21 baskets, she will buy exactly 21 baskets and sell 21
baskets for a conditional profit of P315

Department of Accountancy - Management Science

8.6 EXPECTED PROFIT WITH PERFECT INFORMATION
Table 8.5

Department of Accountancy - Management Science

8.6 EXPECTED PROFIT WITH PERFECT INFORMATION
Table 8.6
Demand Conditional profit, with certainty Probability Expected Profit with perfect information

20 P300 x 0.20 P 60
21 315 x 0.40 126
22 330 x 0.30 99
23 345 x 0.10 34.50
1.00 319.50

Department of Accountancy - Management Science

8.7 MINIMIZING EXPECTED LOSSES OR REGRET
• There are two kind of losses:
1. Real loss: Supply > Demand
2. Opportunity loss: Supply < Demand
 Zero Regret: Supply = Demand

Department of Accountancy - Management Science

8.7 MINIMIZING EXPECTED LOSSES OR REGRET
Table 8.7

Department of Accountancy - Management Science

8.7 MINIMIZING EXPECTED LOSSES OR REGRET
Table 8.8 Expected Monetary Regret Table of Ruby

Department of Accountancy - Management Science

8.7 MINIMIZING EXPECTED LOSSES OR REGRET
Table 8.9

Department of Accountancy - Management Science

REFERENCE:
Quantitative Techniques in Business Management/
Management Science by Asuncion C. Mercado del
Rosario

Department of Accountancy - Accounting 10 & 11 Intermediate

Accounting Part 2
 THANK YOU
STAY SAFE

Department of Accountancy - Management Science