Decision Analysis:

Decision Tree

Prof. U.K. Bhattacharya

1
 Introduction to Decision Analysis

• The field of decision analysis provides a framework for

making important decisions.
• Decision analysis allows us to select a decision from a set of
possible decision alternatives when uncertainties regarding
the future exist.
 The goal is to optimize the resulting payoff in terms of a
decision criterion.
 Example
2
 Criteria
• Maximizing expected profit is a common criterion when
probabilities can be assessed.

• Maximizing the decision maker’s utility function is the mechanism

used when risk is factored into the decision making process.

3
 Payoff Table Analysis
• Payoff Tables
Application: Tom Brown Investment Decision
Tom Brown has inherited $1000 from a distant relative. Since he still has
another year of studies before graduation from Iowa State University, Tom
has decided to invest the $1000 for a year. Literally tens of thousands of
different investment possibilities are available to him, including growth
stocks, income stocks, corporate bonds, municipal bonds, government
bonds, futures, limited partnerships, annuities, and bank accounts.
Given the limited amount of money he has to invest, Tom has decided that it is
not worthwhile to spend the countless hours required to fully understand
those various investments. Therefore he has turned to a broker for
investment guidance.

4
The broker has selected five potential investments she believes would be
appropriate for Tom: gold, a junk bond, a growth stock, a certificate of deposit
and a stock option hedge. Tom would like to set up a payoff table to help him
choose the appropriate investment.

5
 The states of nature are mutually
exclusive and collectively exhaustive.

The Payoff Table

DJA is up more DJA is up DJA moves DJA is down DJA is down more
than1000 points [+300,+1000] within [-300, -800] than 800 points
[-300,+300]

Decision Define the states

States of nature.
of Nature
Alternatives Large Rise Small Rise No Change Small Fall Large Fall
Gold -100 100 200 300 0
Bond 250 200 150 -100 -150
Stock 500 250 100 -200 -600
C/D account 60 60 60 60 60
Stock option 200 150 150 -200 -150
6
 Alternatives Large Rise Small Rise No Change Small Fall Large Fall
Gold -100 100 200 300 0
Bond 250 200 150 -100 -150
Stock 500 The Payoff Table
250 100 -200 -600
C/D account 60 60 60 60 60
Stock option 200 150 150 -200 -150

Determine the
set of possible
decision
alternatives.

7
 The Payoff Table

Decision States of Nature

Alternatives Large Rise Small Rise No Change Small Fall Large Fall
Gold -100 100 200 300 0
Bond 250 200 150 -100 -150
Stock 500 250 100 -200 -600
C/D account 60 60 60 60 60
Stock option 200 150 150 -200 -150

The stock option alternative is dominated by the

bond alternative 8
Classifying Decision Making Criteria

– Decision making under certainty.

• The future state-of-nature is assumed known.
• Example
– Decision making under risk.
• There is some knowledge of the probability of the states of
nature occurring.
-Decision making under uncertainty.
• There is no knowledge about the probability of the states of
nature occurring.
9
 Decision Making Under Uncertainty

• The decision criteria are based on the decision maker’s

attitude toward life.
• The criteria include the
– Maximin Criterion - pessimistic or conservative approach.
– Minimax Regret Criterion - pessimistic or conservative approach.
– Maximax Criterion - optimistic or aggressive approach.
– Principle of Insufficient Reasoning – no information about the
likelihood of the various states of nature.
10
 Decision Making Under Uncertainty -
The Maximin Criterion
• This criterion is based on the worst-case scenario.
– It fits both a pessimistic and a conservative decision
maker’s styles.

11
 TOM BROWN - The Maximin Criterion
Th
e op
The Maximin Criterion
The Maximin Criterion tim Minimum
Minimum
Decisions Large
LargeRise
Rise Small
Smallrise
rise NoNoChange
Change Small
al Large
SmallFall
Decisions Fall
e d LargeFall
Fall Payoff
Payoff
Gold
Gold -100
-100 100
100 200
200 300 cisio 00
300 -100
-100
Bond
Bond 250
250 200
200 150
150 -100
-100 n
-150
-150 -150
-150
Stock
Stock 500
500 250
250 100
100 -200
-200 -600
-600 -600
-600
C/D
C/Daccount
account 60
60 60
60 60
60 60
60 6060 60
60

12
Decision Making Under Uncertainty -
The Minimax Regret Criterion

13
 Decision Making Under Uncertainty -
The Minimax Regret Criterion

• The Minimax Regret Criterion

– This criterion fits both a pessimistic and a
conservative decision maker approach.
– The payoff table is based on “lost opportunity,” or
“regret.”
– The decision maker incurs regret by failing to choose
the “best” decision.

14
 Decision Making Under Uncertainty -
The Minimax Regret Criterion
• The Minimax Regret Criterion
– To find an optimal decision, for each state of nature:
• Determine the best payoff over all decisions.
• Calculate the regret for each decision alternative as the
difference between its payoff value and this best payoff
value.
– For each decision find the maximum regret over all
states of nature.
– Select the decision alternative that has the minimum of
these “maximum regrets.”
15
Investing in Stock generates no
regret when the market exhibits
TOM BROWN – Regret Table
a large rise

The
ThePayoff
PayoffTable
Table
Decision
Decision Large
Largerise
rise Small
Smallrise
rise No
Nochange
changeSmall
Smallfall
fall Large
Largefall
fall
Gold
Gold -100
-100 100
100 200
200 300
300 00
Bond
Bond 250
250 200
200 150
150 -100
-100 -150
-150
Stock
Stock 500
500 250
250 100
100 -200
-200 -600
-600
C/D
C/D 60
60 60
60 60
60 60
60 60
60

The
TheRegret
RegretTable
Table
Decision
Decision Large
Largerise
riseSmall
Smallrise
riseNo
Nochange
changeSmall
Smallfall
fall Large
Largefall
fall
Gold
Gold 600
600 150
150 00 00 60
60
Bond 250 Let us
50 build the
50 Regret Table
400 210
Bond 250 50 50 400 210
Stock
Stock 00 00 100
100 500
500 660
660
C/D
C/D 440
440 190
190 140
140 240
240 0 16
0
 TOM BROWN – Regret Table
The
ThePayoff
PayoffTable
Table
Decision
Decision Large
Largerise
rise Small
Smallrise
rise No
Nochange
changeSmall
Smallfall
fall Large
Largefall
fall
Gold
Gold -100
-100 100
Investing 200
100 in200 300
300 a regret
gold generates 00
Bond
Bond 250
250 200
200 150
of 600 when -100
150 the market -150
-100 exhibits-150
Stock
Stock 500
500 250
250 100
100
a large -200
Th rise
-200 -600
-600
C/D 60 60 60 e op 60 60
C/D 60 60 60 60
tim 60
500 – (-100) = 600 al
de
The
The Regret
Regret Table
Table
cis
Maximum
Maximum
ion
Decision
Decision Large
Large rise
rise Small
Smallrise
riseNo
No change
change Small
Smallfall
fall Large
Large fall
fall Regret
Regret
Gold
Gold 600
600 150
150 00 00 60
60 600
600
Bond
Bond 250
250 50
50 50
50 400
400 210
210 400
400
Stock
Stock 00 00 100
100 500
500 660
660 660
660
C/D
C/D 440
440 190
190 140
140 240
240 00 17
440
440
 Decision Making Under Uncertainty -
The Maximax Criterion
• This criterion is based on the best possible scenario.
It fits both an optimistic and an aggressive decision maker.

• An optimistic decision maker believes that the best possible

outcome will always take place regardless of the decision
made.

• An aggressive decision maker looks for the decision with the

highest payoff (when payoff is profit).
18
 Decision Making Under Uncertainty -
The Maximax Criterion

• To find an optimal decision.

– Find the maximum payoff for each decision
alternative.
– Select the decision alternative that has the maximum
of the “maximum” payoff.

19
 TOM BROWN - The Maximax Criterion

Th
eo
pti
al m
The Maximax Criterion Maximum
de
Decision Large rise Small rise No change Small fall Largecisfall Payoff
Gold -100 100 200 300 0 ion 300
Bond 250 200 150 -100 -150 200
Stock 500 250 100 -200 -600 500
C/D 60 60 60 60 60 60

20
 Decision Making Under Uncertainty - The
Principle of Insufficient Reason

• This criterion might appeal to a decision maker who

is neither pessimistic nor optimistic.
– It assumes all the states of nature are equally likely to
occur.
– The procedure to find an optimal decision.
• For each decision add all the payoffs.
• Select the decision with the largest sum (for profits).

21
 TOM BROWN - Insufficient Reason

• Sum of Payoffs
– Gold 600 Dollars
– Bond 350 Dollars
– Stock 50 Dollars
– C/D 300 Dollars
• Based on this criterion the optimal decision
alternative is to invest in gold.
22
 Decision Making Under Risk

• The probability estimate for the occurrence of

each state of nature (if available) can be
incorporated in the search for the optimal
decision.
• For each decision calculate its expected payoff.

23
 Decision Making Under Risk –
The Expected Value Criterion
• For each decision calculate the expected payoff
as follows:

Expected Payoff == (Probability)(Payoff)

Expected Payoff (Probability)(Payoff)

(The summation is calculated across all the states of nature)

• Select the decision with the best expected payoff

24
 TOM BROWN - The Expected Value Criterion
The
opt
im a
The Expected Value Criterion l deci Expected
sion
Decision Large rise Small rise No change Small fall Large fall Value
Gold -100 100 200 300 0 100
Bond 250 200 150 -100 -150 130
Stock 500 250 100 -200 -600 125
C/D 60 60 60 60 60 60
Prior Prob. 0.2 0.3 0.3 0.1 0.1

EV = (0.2)(250) + (0.3)(200) + (0.3)(150) + (0.1)(-100) + (0.1)(-150) = 130

25
When to use the expected value approach

• The expected value criterion is useful generally

in two cases:
– Long run planning is appropriate, and decision
situations repeat themselves.
– The decision maker is risk neutral.

26
 Expected Regret Criterion
Expected Regret Approach
1. Determine the best value (maximum payoff or minimum cost) for each
states of nature
2. For each state of nature, the regret corresponding to a decision
alternative is the difference between its payoff value and this best value.
3. Find the expected regret for each decision alternative
4. Select the decision alternative that has the maximum expected regret.

27
 Applications-6.3
National foods has developed a new sports beverage it would like to advertise on Super
Bowl Sunday. National’s advertising agency can purchase either one, two.or three 30 second
commercials advertising the drink. It estimates that the return will be based on super Bowl
viewer ship, which in turn based on fans’ perception on whether the game is dull, average,
above average, or exciting.
National foods ad agency has constructed the following payoff table giving its estimate of the
expected profit (in $100,000’s ) returning from purchasing one, two or three advertising
spots. (Another possible decision is for National foods not to advertise at all during the
Super Bowl). The States of nature of the game being “dull”, “average”, “ above average” or
“exciting”.

28
_____________________________________________________________________
Number of 30 Perceived Game Excitement
Sec. Commercials ______________________________________
Purchased Dull Average Above exciting
average
______________________________________________________________________
One -2 3 7 13
Two -5 6 12 18
Three -9 5 13 22
_______________________________________________________________________
a. What is the optimal decision if the national foods ad manager is optimistic
b. What is the optimal decision if the national Foods advertising manager is pessimistic
c. What is the optimal decision if the National Foods ad manager wishes the minimize the
firm’s maximum regret?
29
Consider the data given in problem 3 for national Foods. Based on passed super bowl
games, suppose the Decision maker believes that the following probabilities hold for the
states of nature.
P(Dull Game)=0.20
P(Average Game)=.40
P(Above Average game)=.30
P(exciting)=0.10
a. Using the expected value criterion, determine how many commercials National Foods
should purchase?
b. Based on the probabilities given here, determine the expected value of perfect
information.

30
Expected Value of Perfect Information

• The gain in expected return obtained from knowing

with certainty the future state of nature is called:

Expected Value of Perfect Information

(EVPI)

31
Expected Value of Perfect Information

 Expected return with 

   Expected return without 
 perfect informatio n as   
 Expected    additional information as 
  to which state of nature 
 value of Perfect      to which state of nature 
 Information   will occur prior to   
  
making decision   will occur prior to making 
   decision 

 

32
 TOM BROWN - EVPI
If it were known with certainty that there will be a “Large Rise” in the market

-100
The Expected Value of Perfect Information
Decision Large
Largerise
rise Small rise No change Small fall Large fall
Gold 250-100 100 200 300 0
Bond 250 200 150 -100 -150
Stock
Stock 500 500 250 100 -200 -600
C/D
Probab.
60 60
0.2
60
0.3
60
0.3
60
0.1
60
0.1

... the optimal decision would be to invest in...

Similarly,…

33
 TOM BROWN - EVPI

The-100
Expected Value of Perfect Information
Decision Large rise Small rise No change Small fall Large fall
Gold 250
-100 100 200 300 0
Bond 250 200 150 -100 -150
Stock 500
500 250 100 -200 -600
C/D 60 60 60 60 60
Probab. 600.2 0.3 0.3 0.1 0.1

Expected Return with Perfect information =

ERPI = 0.2(500)+0.3(250)+0.3(200)+0.1(300)+0.1(60) = $271
Expected Return without additional information =
Expected Return of the EV criterion = $130
EVPI = ERPI - EREV = $271 - $130 = $141 34
Bayesian Analysis - Decision Making with
Imperfect Information
• Bayesian Statistics play a role in assessing
additional information obtained from various
sources.

• This additional information may assist in refining

original probability estimates, and help improve
decision making.
35
Tom Brown Investment Decision (Continued)
Tom has learned that, for only $50, he can receive the results of noted
economist Milton Samuelman’s multimillion dollar economic forecast,
which predicts either “positive” or “negative” economic growth for the
upcoming year. Samuelman has offered the following variable statistics
regarding the results of his model:
1. When the stock market showed the large rise, the forecast predicted
“positive” 80% of the time and “negative” 20% of the time.
2. When the stock market showed a small rise, the forecast predicted
“positive” 70% of the time and “negative” 30% of the time.
3. When the stock market showed no change , the forecast was equally
likely to predict “positive” or “negative”.
4. When the stock market showed a small fall, the forecast predicted
“positive” 40% of the time and negative 60% of the time.
36
5. When the stock market showed a large fall, the forecast always
Predicted “negative”
Tom would like to know whether it is worthwhile to pay $50 for the result of
the Samuelman forecast.

37
 TOM BROWN – Solution
Using Sample Information
• If the expected gain resulting from the decisions made
with the forecast exceeds $50, Tom should purchase
the forecast.
The expected gain =
Expected payoff with forecast – EREV
• To find Expected payoff with forecast Tom should
determine what to do when:
– The forecast is “positive growth”,
– The forecast is “negative growth”.
38
 Conditional Probabilities
P(forecast predicts”positive”/large rise in the market)=0.80
P(forecast predicts “ negative”/large rise in the market)=0.20
P(forecast predicts ‘positive”/ small rise in the market)=0.70
P(forecast predicts “negative”/small rise in the market)=0.30
P(forecast predicts “positive”/ no change in the market)=0.50
P(forecast predicts “negative”/ no change in the market) =0.50
P (forecast predicts “positive”/ small fall in the market)=0.40
P(forecast predicts “ negative” / small fall in the market)=0.60
P(forecast predicts ‘positive”/ large fall in the market)=0 39
 TOM BROWN – Solution
Using Sample Information
• Tom needs to know the following probabilities
– P(Large rise | The forecast predicted “Positive”)
– P(Small rise | The forecast predicted “Positive”)
– P(No change | The forecast predicted “Positive ”)
– P(Small fall | The forecast predicted “Positive”)
– P(Large Fall | The forecast predicted “Positive”)
– P(Large rise | The forecast predicted “Negative ”)
– P(Small rise | The forecast predicted “Negative”)
– P(No change | The forecast predicted “Negative”)
– P(Small fall | The forecast predicted “Negative”)
– P(Large Fall) | The forecast predicted “Negative”) 40
 TOM BROWN – Solution
Bayes’ Theorem
• Bayes’ Theorem provides a procedure to calculate
these probabilities
P(B|Ai)P(Ai)
P(Ai|B) =
P(B|A1)P(A1)+ P(B|A2)P(A2)+…+ P(B|An)P(An)

Posterior Probabilities Prior probabilities

Probabilities determined Probability estimates
after the additional info determined based on
becomes available. current info, before the
new info becomes available.
41
 Posterior Probability “ Positive” Forecast
for Tom Brown
States of nature Si Prior Conditional Joint Probability P(Si/positive)
Probability Probability P(positive  Si)
P(Si) P(positive/Si)
LR 0.20 0.80 0.16 0.286

SR 0.30 0.70 0.21 0.375

NC 0.30 0.50 0.15 0.268

SF 0.10 0.40 0.04 0.071

LF 0.10 0 0 0

Total 0.56

P(positive)=0.56 42
 Posterior Probability: “negative” forecast
for Tom Brown
States of nature Si Prior probability Condl Probability Joint Probability P(si/ negative)
P(si) P(negative/Si) P(negative 
Si)
LS 0.20 0.20 0.04 0.091

SR 0.30 0.30 0.09 0.205

NC 0.30 0.50 0.15 0.341

SF 0.10 0.60 0.06 0.136

LF 0.10 1.00 0.10 0.227

Total 0.44

43
 Expected value of sample informtion
If the Samuelman’s Forecast is positive economic growth, the revised expected values for the decision
alternatives (rounded to the nearest dollar) are

EV(gold/”positive”)=$84
EV(Bond/’Positive)=$180
EV(Stock/”Positive”)=$249
EV(C/D/”Positive”)=$60

Decision: So if the samuelman’s forecast is positive then buying the stock would be optimal.

44
 Samuelman’s forecast is negative
The expected returns are
EV(Gold/ negative)=$120
EV( Bond/ negative )=$67
EV (Stock/ “negative”)=-$33
EV(C/D/ “ Negative”)=$60

45
Expected value of sample Information
 Expected return 
 
 Expected Value   Expected return   
     
     without sample 
     
 of Sample Informatio n   with sample informatio n   
     
 
 informatio n 

EVSI  ERSI - EREV

46
 Decision
ERSI=.56(249.11)+.44(120.45)=$192.50

EREV=$130

EVSI=$192.50-130=$62.50

Decision: Tom should acquire it

47
 Extension of Application 6.3

Consider the data given in problems (3 and 4 ) for National Foods. The firm can hire the noted
sports pundit Jim Worden to give his opinion as to whether or not the Super Bowl game will be
interesting. Suppose the following probabilities holds for Jim’s predictions.
P(Jim Predicts game will be interesting /game is dull)=.15,
P(Jim predicts game will be interesting / game is average).25
P(Jim predicts game will be interesting / game is above average)=.50,
P(Jim predicts game will be interesting / game is exciting)=.80,
P(Jim predicts game will not be interesting / game is Dull)= .85
P(Jim predicts game will not be interesting / game is average)= .75
P(Jim predicts game will not be interesting / game is above average)=.50,
48
P(Jim predicts game will not be interesting / game is exciting)=.20,
 Questions
a. If Jim predicts the game will be interesting what is the probability that the
game will be dull.
b. What is the National’s optimal strategy if Jim predicts the game will be (I)
interesting (ii) Not interesting
c. What is the expected value of Jim’s information.

49
TOM BROWN – Conditional Expected Values

• The revised expected values for each decision:

Positive forecast Negative forecast
EV(Gold|Positive) = 84 EV(Gold|Negative) = 120
EV(Bond|Positive) = 180 EV(Bond|Negative) = 65
EV(Stock|Positive) = 250 EV(Stock|Negative) = -37
EV(C/D|Positive) = 60 EV(C/D|Negative) = 60

If the forecast is “Positive” If the forecast is “Negative”

Invest in Stock. Invest in Gold.
50
TOM BROWN – Conditional Expected Values

• Since the forecast is unknown before it is

purchased, Tom can only calculate the expected
return from purchasing it.
• Expected return when buying the forecast = ERSI =
P(Forecast is positive)(EV(Stock|Forecast is positive)) +
P(Forecast is negative”)(EV(Gold|Forecast is negative))
= (.56)(250) + (.44)(120) = $192.5

51
Expected Value of Sampling Information
(EVSI)
• The expected gain from buying the forecast is:
EVSI = ERSI – EREV = 192.5 – 130 = $62.5

• Tom should purchase the forecast. His expected

gain is greater than the forecast cost.

• Efficiency = EVSI / EVPI = 63 / 141 = 0.45

52
 DECISION TREE APPLICATION
• You have the chance to invest in three mutual
funds utility, aggressive growth and global. The
value of your investment will change depending on
the market conditions. There is a 10% chance the
market will go down, 50% chance it will remain
moderate , and 40% chance it will per form well.
The following table provides the percentage change
in the investment value under the three conditions

53
 CONTINUED

PERSENT RETURN ON INVESTMENT

_______________________________
ALTERNATIVE DOWN MARKET (%) MODERATE MARKET (%) UP MARKET (%)
________________________________________________________________________________________
UTILITY +5 +7 +8
AGGREGATE GROWTH -10 +5 +30
GLOBAL +2 +7 +20
________________________________________________________________________________________
_
(A)REPRESENT THE PROBLEM AS A DECISION TREE.
(B) WHICH MUTUAL FUND SHOULD YOU SELECT?
54
 APPLICATION DECISION TREE
• Farmer McCoy can plant either corn or soybean. The
probabilities that the next harvest prices of these commodities
will go up, stay the same or go down are .25, .30, and .45
respectively. If the prices go up corn crop will net $30,000 and
the soybean will net $10000. If the prices remain unchanged ,
McCoy will (barely) breakeven. But if the prices go down, the
corn and soybean crop will sustain losses of $35,000, & $5000
respectively.
• (a) Represent McCoy problem as a decision tree.
• (b) Which crop McCoy should plant

55