Time: 2.5 hour Name:___________________ Max Points: 60 Roll No.:_________________
This is an open-book and open-note test. However, sharing of material is NOT permitted.
Instructions: Do not seek any clarifications. Provide appropriate arguments and show calculation in support of your final answer. Answer all questions in the space provided. Do not attach any additional sheets, use the back pages, if necessary.
This part is for the use of the GRADER ONLY
Question # A B C TOTAL Part # i ii iii iv v i ii iii iv v vi vii i ii iii iv bonus
Full Marks 6 2 2 4 2 4 3 4 2 4 6 2 8 4 5 2 2 60 Obtained Marks
Page 2 of 10 Question A:
International Institute of Management, Billekahali has an in-take of 250 students for its MBA program in 2002. In the IS (Introduction to Statistics) course, the students are given an assignment of drawing a sample of 30 students from a database (which has the records of prior work experience, in months, of all 250 students) using simple random sampling with replacement. Some of the relevant characteristics that are of interest are: A = average work experience of all the 250 students B = standard deviation of work experiences of all students C = proportion of students who have no work experience Since each student can only select thirty observations from this database, none of them knows the exact value of A, B, or C; however, once a student selects the sample, (s)he is required to compute the mean work experience of students in the sample, variance of work experiences of the students in the sample, number of students in the sample having no work experience. Each student reports not only these values, but also a 90% confidence interval for C (based on his/her sample) to their CR(class representative). CR then enters all the data in an Excel sheet (1 row for each students results) and also calculates the average and standard deviation of each column entries, before submitting the file to the Professor. The sheet that Professor received looks like this:
Roll No. Mean of work exp Variance of work exp No. with no exp Confidence interval for C lower limit upper limit 1 9.300 65.045 10 e F 2 7.033 56.033 14 0.317 0.616 . . . . . . . . . . . . 250 9.633 64.861 d 0.162 0.438
column mean x = 8.61 z v = 9.60 r t column s.d. y = 1.3364 u w m k
i) Calculate as many of the following missing/unknown values as possible: (if you cannot find out two or more of the missing values, but you know their inter-relations, then mention them as well) [6 points] A = z =
B = w =
C = d =
Page 3 of 10 ii) What would have been the value of y, had each student adopted simple random sampling without replacement? [2 points]
iii) How many students are expected to have their confidence interval containing the value of C? [2 points]
iv) What must be the confidence interval for the student with Roll No. 1? [4 points]
v) Which of the following statements do you agree with (and why)? [2 points]
a) It is certain that C lies in the interval (r,t). b) There is a 90% chance that C lies in the interval (r,t). c) It is not possible to comment on the likelihood of C lying in the interval (r,t)
Page 4 of 10 Question B
Hari is studying the electricity spot price behaviour in certain unregulated competitive markets. Following a special request to the relevant authority in Norway (which has one such market), he is able to get the price data for 1998 and 1999. These prices are recorded in intervals of 5 minutes, so that for all practical purpose, it may be considered to be available on a continuous time domain. When Hari analysed the data, one particular feature of the price series immediately caught his attention. From time to time, price really shot up, resulting in spike-like features in the price vs. time plot. Hari decided to study two aspects of these spikes more closely. The first was the timing of the spikes (or price hikes) --- careful investigation suggested that the timings seemed to have the memory-less property, indeed the timings of the spike may be modelled as a Poisson Process. The second aspect was the magnitudes of these spikes. Hari decided to define magnitude as the percentage of increase in price from just before the occurrence of the spike. The magnitude is clearly random; to start with, Hari thought that the magnitude must be having a Normal distribution since the 2 years of data had as many as 438 spikes. However, following suggestion from his friend Rajesh, Hari decided to look at the magnitudes more closely. The following frequency table reflects the magnitudes of the spikes in the two-years of price data:
i) Calculate the mean and the standard deviation of the magnitudes of the spikes in the last two years. [2+2=4 points]
mean =
standard deviation =
lower limit upper limit No of spikes 200% 300% 20 300% 400% 100 400% 500% 175 500% 750% 93 750% 1000% 50 Page 5 of 10 ii) Find a 99% confidence interval for the true mean magnitude of the price spikes. [3 points]
iii) Hari wants that the above confidence interval to have a length of 10% or less without sacrificing his level of confidence. How many years of data would he need to achieve that? [4 points]
iv) What is the average number of price spikes per day? [2 points]
Page 6 of 10 v) What is the chance that the gap between any two consecutive price spikes is less than a day? [4 points]
vi) Given that there were 6 price spikes on a given week, what is the probability that more than 1 of them occurred during the weekend (Saturday + Sunday) of that week? [6 points]
vii) How critical are Haris assumptions (spike timings constituting a Poisson Process and spike magnitudes following a Normal distribution) in your answers to parts i) through vi)? [2 points]
Page 7 of 10 Question C Mavuru Islands (popularly known as Mavi) attained independence 3 years ago. At the time of independence, the islands were divided into two separate countries namely Northern Mavi and Southern Mavi, divided by the 37 th parallel. As a result of the division, the oil refineries are in the Southern Mavi where as the oil wells are in the Northern Mavi. There is an irrevocable treaty signed between the two countries with regard to the crude and the refined products. Southern Mavi has to buy the required crude only from Northern Mavi and Northern Mavi has to buy the refined oil products only from Southern Mavi. Southern Mavi has three refineries namely Amity Oil, Fortune Oil and Peoples Oil. The first two refineries, which account for 25% each of the refining capacity, are in the private sector where as Peoples Oil, accounting for 50% of the refining capacity is a public sector company. Recently, the Government of Southern Mavi decided to privatize the People Oil and accordingly invited bids from the two private companies. The company with the higher bid will acquire Peoples Oil at the bid amount. Amity Oil is considering two possible bids, one at $2 billion and the other at $3 billion. They came to know that Fortune Oil is also considering bids of similar amounts; but, of course, the chance that the two bids will be exactly identical is zero. If Fortune Oils bid happens to be higher (even very marginally) than that of Amitys bid, Amity will fail in the acquisition race. On the other hand, if Amitys bid is higher (even very marginally) than Fortunes, then Amity will succeed in the acquisition race. Helen Reddy, the Managing Director of Amity estimated that there is a 70% chance that Fortunes bid will be higher than that of Amity, irrespective of Amitys bid. Amity has to borrow heavily at a high interest rate, from the banks abroad to finance the acquisition, if they decide to bid $ 3 billion. Acquiring such a large additional capacity has its own problems. The crude prices in the Northern Mavi have been fluctuating widely in the recent months. The treaty between the two countries requires that the higher cost of the crude, if the price goes up, has to be borne by the refining company. On the other hand, Southern Mavi has an administered pricing mechanism for the refined products. As a result, the refineries have to sell their products pre-determined prices. If Amity successfully bids $ 3 billion and acquires Peoples oil, their interest commitment will be very high and be a drain on their resources. Consequently, if the crude prices go up, they will not be able to operate at full capacity and end up making heavy losses. On the other hand, if they bid only $2 billion and are successful in the acquisition, there are no interest costs and they will make decent profits, irrespective of the crude prices. Helen Reddy estimated the profits under each possible scenario. This data is given in the table below: Profits under each scenario Bid Acquisition Crude Prices Increase Decrease $3 Billion Success -70 80 Fail 40 -40 $2 Billion Success 30 40 Fail -20 -50
She has also estimated that the probability of the crude prices going up in the next year is 0.60.
Page 8 of 10 i) Draw the decision tree and advise Helen Reddy on the appropriate action. [8 points]
ii) What is the maximum possible value that Helen would be willing to pay for any kind of information? [Hint: EVPI] [4 points]
Page 9 of 10 Question C (cont.) John Denver, a freelance consultant approached Helen Reddy with very valuable information. He told Ms. Reddy that he has a mole inside Fortune Oil, who can get the exact bid value of Fortune, well in advance. He also collected data on the crude prices and the terrorist activity in Northern Mavi in the past 50 weeks. He presented that following frequency table to Helen Reddy.
Terrorist Activity Crude Prices Total Increased Decreased Intensified 20 5 25 Weakened 10 15 25 Total 30 20 50
He has also informed Helen that he will be able to predict the terrorist activity with absolute certainty because he has direct contacts with the terrorist leaders. Helen considered Mr. Denvers offer and decided that she will put in a bid such a way that it will be successful only if it is predicted that the terrorist activity will be weakened.
iii) Should Helen Reddy take the information from Mr. Denver? What is the maximum amount that she can afford to pay Mr. Denver? [5 points]
Page 10 of 10 iv) What is the efficiency of the information provided by Mr. Denver? [2 points]
[Bonus question: worth 2 points]
How do you justify the efficiency of Mr. Denvers information being over 100%?