Petro-Co Project Minimax Regret
The first question is about mini-max
regret analysis of the oil discovery project carried out by the Petro-Co. According
to the information presented in the question statement Petro-company wants to
take a decision about oil discovery project. The company has two options: drill
an exploratory for oil well (spending cost on drilling or other activities) or not
to drill a well because chances are very limited for oil discovery (Kochenderfer). According to
estimation, in case of oil discovery Petro-company will get a net profit of
800000 and in case of failure 800000 will be considered a loss. Considering this
information following calculations are made for minimax regret criterion
analysis.
Min-Max
regret
|
STATES OF NATURE
|
DECISION
|
Oil
|
No oil
|
Drill
|
800000
|
-800000
|
Not to drill
|
200000
|
200000
|
|
Drill
|
0
|
1000000
|
Not to drill
|
600000
|
0
|
|
Drill
|
1000000
|
|
Not
to drill
|
600000
|
Minimum
|
Decision: Not to drill
|
Above mentioned min-max regret
creation explain that drilling an exploratory well is not favorable for
Petro-Co. Min-max regret criterion explains that the minimum amount is 600000
for this project that the company may suffer. In Min-max regret creation at the
first information about possible payoffs for both options are discussed. In the
second box maximum value is selected from each column to subtract all values of
this column (Preuschoff, Mohr and Hsu). After that maximum
values are selected from the third box. The final decision is taken by checking
the minimum value of the third box. According to this criterion, the minimum
value is 600000 for an oil discovery project.
Question
2. What is the best choice Petro-Co should go for using the expected value of
this major project?
Best Choice using Expected value
Expected value calculations for the
major project is the major source to select the appropriate decision for a
project. In this Petro-Co project best choice is selected by using the expected
value of this major project (Conejo, Carrión and Morales). Expected value is
calculated through multiplying the oil discovery payoff and no oil discovery
with the probabilities. The probability of oil discovery is 45% while the
remaining 55% represents chances of failure in oil discovery. According to the
expected value calculation formula (see presented below formula) expected value
for drilling is -800000 that represents chances of loss rather than profit.
Drill expected value calculation
Expected
value for not-drilling the well
Expected
value
|
STATES OF NATURE
|
DECISION
|
Oil
|
No oil
|
Expected Value
|
Drill
|
800000
|
-800000
|
-80000
|
Not to drill
|
200000
|
200000
|
200000
|
|
|
Probability
|
0.45
|
0.55
|
|
Maximum
|
200000
|
Decision
|
Not to drill
|
Expected value calculation concludes
that the company should not invest in this project. Because of the high
probability of loss (as there is 55% probability that company will fail to
discover oil after drilling an exploratory well) expected value is negative (Conejo,
Carrión and Morales). If the company will drill oil well
then they will face the loss of 800000. Somehow, the maximum expected value is
200000 that support the option of “not to drill”.
Question:
3 What is the best decision based on the EOL?
Best
Decision based on EOL
EOL refers to the expected
opportunity loss for a project. In this Petro-Co project, EOL will explain how
much loss the company can face in case they avail of this opportunity. Expected
opportunity loss is calculated through the use of a simple EOL formula (Preuschoff,
Mohr and Hsu).
In EOL calculation maximum value is used to subtract the other loss related
options (for instance in this case “not to drill” option was representing the
loss). After the subtraction of these values, calculated “drill” value “0” and
drilling loss “1000000” are multiplied by the probabilities 45% and 55%
(converted to decimal as 0.45 and 0.55). After that multiplication results are added up
to get the calculated value of EOL.
EOL
|
Drill
|
0
|
Drilling loss (No oil)
|
1000000
|
Probability
|
|
Oil
|
0.45
|
No oil
|
0.55
|
EOL
|
550000
|
Thus in the light of EOL
calculation and above-mentioned table, expected opportunity loss for Petro-Co
is 550000. Calculated EOL value loss is almost double as compared to the
initial cost of the project (Conejo, Carrión and Morales). Considering this
information it is clear that the best decision for Petro-Co is to avoid
investing in this project. The project is even not capable to meet breakeven
point. Therefore, instead of investing
in this project they should search for a better opportunity (Kochenderfer).
Question:
4 What is the expected value of perfect information?
The
expected value of Perfect Information
The expected value of perfect
information for Petro-Co project elaborate increase in the overall possible or
expected net profit that company or managers of the company (who is going to
take decision for investment) will ensure through getting information about
uncertainty and certainty of an event occurrence. The expected value of perfect
information deals with the upper bound of the expected project value. The
simple formula to calculate the expected value of perfect information is to
subtract “A” value from “B” value.
If EVwPI (Expected
value without perfect information) is 470000 then
Expected value with perfect information
|
STATES
OF NATURE
|
DECISION
|
Oil
|
No oil
|
Expected Value
|
Drill
|
800000
|
-800000
|
-80000
|
Not to
drill
|
200000
|
200000
|
200000
|
|
|
Probability
|
0.45
|
0.55
|
|
Evwpi
|
470000
|
Max.
Expected value
|
200000
|
EVPI
|
270000
|
|
The above-mentioned table provides an
overview of the calculations and results of the expected value without and with
perfect information. The total calculated value of EVPI is 270000 which means
that Petro-Co has to spend this amount to get perfect information about the
project. The expected value of perfect information and Expected value without
perfect information cannot be all the same for a project. The margin between
these values presents the requirements for market research, cost estimation,
and other important information.
Question: 5 Petro Co is
considering conducting a seismic survey which costs $10,350, to see what the
underground conditions are like. The survey history shows that
? There is a 96% chance of a
favorable wave transmission, given that there is an oil discovery.
? There is a 84% chance of an
unfavorable wave transmission, given that there is a dry well.
According to the given data,
find the posterior probabilities.
Posterior
Probabilities
Posterior probabilities have
significant importance in the decision-making process. Posterior Probabilities
put light on the possible outcomes of an event. Bayes Theorem is the basis of
Posterior Probabilities analysis (Conejo, Carrión and Morales). In statistics and
business decision analysis related courses, posterior probabilities are also
known as branch probabilities for decision trees (Preuschoff,
Mohr and Hsu).
With the help of these Posterior Probabilities, managers can understand
possible outcomes of their decisions. In this case, information related to the
Petro-Co oil discovering project is used to search out Posterior Probabilities
of events.
In the analysis chances of favorable
wave transmission (oil discovery) is 96% while 84% chances stand for unfavorable
wave transmission. According to the unfavorable wave transmission well are dry
therefore the company can not extract oil from these well. Posterior
Probabilities calculation is used as a technique to evaluate available
alternative options to select the best one for Petro-Co. In the presented below
tables information about Posterior Probabilities are projected.
Posterior
Probabilities |
STATES OF NATURE |
DECISION | Oil | No oil | Expected Value |
Drill | 800000 | -800000 | 96000 |
Not to drill | 10350 | 10350 | 18630 |
| |
Probability | 0.96 | 0.84 | |
Posterior Probabilities
of favorable wave transmission
|
State
|
Prior
|
Likelihoods
|
Joint
|
Posterior
|
Oil
|
0.96
|
0.960
|
0.9216
|
0.964824121
|
No oil
|
0.84
|
0.04
|
0.0336
|
0.035175879
|
|
|
|
0.9552
|
1
|
|
|
|
|
|
Posterior Probabilities
of Un-favorable wave transmission
|
State
|
Prior
|
Likelihoods
|
Joint
|
Posterior
|
Oil
|
0.96
|
0.840
|
0.8064
|
0.857142857
|
No oil
|
0.84
|
0.16
|
0.1344
|
0.142857143
|
|
|
|
0.9408
|
1
|
In the above-mentioned table, two
states are used for events. Prior probabilities are simple probability given in
the question statement (Preuschoff, Mohr and Hsu). While likelihoods,
joint, and posterior are calculated. Joint is calculated by multiply the prior
values with the likelihoods of the states. Posterior value is calculated by
dividing the joint value from total joint values (Conejo, Carrión and Morales). In the above-presented
tables of two events, favorable wave transmission and un-favorable wave
transmission total posterior is equal to 1. Somehow, total posterior value 1
represents the accuracy of Posterior
Probabilities calculation.
Following image represents the
posterior probabilities for both options (drill the well or not to drill a well
for oil). Red color values represent not to drill options while purpose color
is for Drill option.
Figure
1 Posterior Probabilities
Question:
6 What is the value of sample information?
Value
of Sample Information
The expected value of sample
information and the expected value of perfect information both are relatively
the same. Both values relate to the additional expected profit caused by the
detailed information about the project. Somehow, both cannot be interpreted as
the same value because of the difference. EVSI concerns with the sample while
EVPI relates to perfect information. Survey results and knowledge gained from
the samples is not perfect information, therefore, values of EVSI and EVPI
cannot be the same. In this case, EVIS is calculated through the use of the
presented below formula.
18630
According to the calculations expected
value for sample information is -8280.
Question:
7 Examine how your decision might change with different oil discovery
probabilities. Let p denote the probability of a dry well and 1-p denote the
probability of well discovery. What are the ranges of p that affect your
decision? Solve this part as a risk neutral decision maker without any perfect
or sample information.
Risk
Neutral
Decision making is not as easy as
usual, we take it. Organizational success and business profitability high
depend on effective and appropriate decision making. Managers taking the wrong
decision because of limited available information contributes to the failure of
the organization in the market. As competition and need for profitability are
growing it's becoming essential for the project manager and business managers
to take the right decision at right time after evaluating all available
alternative options. In the case of Petro-Co risk factor is really high for the
company.
The company is going to invest a
huge amount of money with less 45% probability of getting net profit ($800000).
Changes in the probability of oil discovery can change decision also. Even then
the low probability is not the indicator of loss (Mittelhammer). It does not mean
that the company should not invest in their project there are many other
factors also that need to be taken into consideration while making the
investment. For instance, in this case, we cannot ignore the expected return on
this investment and possible loss. If the loss is not really high then the
company can take the risk.
There are a number of ways and
methods supported by academic theories that can be used for risk analysis. Ranges
of p that can influence or change my
decision are between the ranges 65-100%. Risk factor shows how much company
will suffer in case they invest in drilling project and get dried well in
discovery (Kochenderfer). Of course, dry well cannot provide oil,
therefore, dry well discovery is a direct loss for the company. The risk-neutral
person always guesses 1 as the sum of whole probability should be equal to one.
Less than one indicate that outcomes are not 100 %, therefore, there is a
possibility of failure.
Conejo, Antonio J., Miguel Carrión and Juan M.
Morales. Decision Making Under Uncertainty in Electricity Markets.
Springer Science & Business Media, 2010. 03 03 2019.
Kochenderfer, Mykel J. Decision Making Under
Uncertainty: Theory and Application. MIT Press, 2015. 03 03 2019.
Mittelhammer, Ron C. Mathematical Statistics for
Economics and Business. Springer Science & Business Media, 2012. 03
03 2019.
Preuschoff, Kerstin, Peter N. C. Mohr and Ming Hsu. Decision
Making under Uncertainty. Frontiers Media SA, 2015.