Tutorial 2

PNC
August 26, 2017

1 Instructions
1. The tutorial will be discussed in the class next week. The exact date will be mailed to you.

2. In case of any doubt or clarification wrt any question, please open a thread on the google
group [email protected] and initiate a discussion.

2 Questions
1. Obtain the closed form for the following expression
1 + 21 + 13 + ..... + n1 (Rating: ***)
2. During the world war II, the western allies wanted to estimate the number of tanks possessed
by the German army. The tanks produced by the Germans were marked with consecutive
numbers starting from 1 to n. The alias captured some of these German tanks and noted these
numbers. The problem was to estimate the number n of actual tanks Germans possessed by
looking at the sample of the passing by tanks.
Simply stated in another form: Martha has n numbers, she gives you k of these numbers
uniformly at random, without repetitions. By looking at these k numbers, can you find the
expected value of n?
We discussed a simple solution to this problem in the class. If the maximum of these n
k
numbers happens to be ak , we have proved that E[ak ] = k+1 (n + 1).
What is the standard deviation of ak ? (Rating: ***)
3. Assuming you have a function which gives you random numbers from 1 to n, devise an
efficient algorithm to permute an array containing k elements. (Rating: **)

4. There are two coins with us. First is the normal unbiased coin having tail on one side and
head on the other, the probability of both- head or tail being equal to appear on flipping.
The other is a coin having head on both the sides, so always head appears on flipping it. We
call this coin the faulty coin. We choose one of these coins uniformly at random and flip it.
What is the probability that the tossed coin is the faulty one, provided :

• We witness a tail on flipping this coin ?

• We witness a head on flipping this coin ?
(Rating: **)
5. A professor comes up with an intelligent software to detect whether a student has cheated
in an exam or not, based on the student’s handwriting in the answer script. There is a small
amount of false positives and false negatives associated with the software. If a student cheats,
the software might declare the student honest with a probability of 0.8. If a student has not
cheated, the software might hold the student guilty with a probability of 0.0001. 10% of the
students in the class actually cheated in the mid semester exams. The professor picks an
answer script uniformly at random and his software says that the student has cheated. What
is the probability that the student has actually cheated in the exam? (Rating: **)
6. We know the meaning of mutually independent and dependent events. Give examples of
three events A, B, and C where:

1
 • A, B and C, all are mutually dependent.
• A, B and C, all are independent.
• Pick any two of them, they are independent. But as soon as you pick the third one,
they become dependent. i.e. A pair is independent but all three of them together are
dependent.
(Rating: **)

7. A new faculty joins a school and the school administration decides to shift some of the
students from the existing class to the class of the newly joined teacher.
• Let the class be represented by a set of students, say S = s1 , s2 , s3 , ..., sn . Some students
have to be selected out of this set to be taught by the newly joined teacher in the school.
We know that there are 2n ways of selecting some students out of all students, i.e.
number of subsets of the set S. The following is what the school administration does for
it- A coin is flipped for each student in the class, if head turns up, the student is shifted
to the new class, else not. Prove that this technique used for shifting the students is
same as choosing one of the 2n possible subsets of the set S uniformly at random. We
call it activity 1.
• Assume we perform activity 1 two times and obtain two sets of students X and Y .
What is the probability that (i) X is a subset of Y (ii) Y is a subset of X. (iii) X and
Y are equal. (iv) X ∪ Y = S
(Rating: **)
8. Consider the tic-tac-toe game.

• Mark the places uniformly at random with O and X. What is the probability that X
has won? What is the probability of a tie? Note: The number of O and X need not be
equal. We are marking each place uniformly at random.
• Consider the full fledged tic-tac-toe game, where O and X make moves alternately but
uniformly at random in the empty spaces. Now, what is the probability that X wins?

(Rating: **)
9. If you are going to Milan, you better smile. It is illegal not to smile in Milan and can cost you
a fine of $100. Assume a hypothetical situation where the citizens of Milan were allowed to
vote for this rule, to smile, or not to smile. We assume all the 100,000 citizens participated in
the voting. As it happens, the people in Milan love smiling and 70,000 people are in support
of the smiling law and 30,000 are in opposition. On the voting day, there seemed to be a
slight confusion among the citizens because of the layout of the ballot. Hence, each voter
with a probability of 1/1000 voted for the wrong choice (and ended up voting for what she
did not want). What is the expected number of votes in the favor of the “SMILE PLEASE”
rule? (Rating: **)

10. Assume a gambler playing the following game. At the beginning, her net profit is zero. She
played for n rounds. In every round, the net profit increases by 1 with a probability of 1/3
and decreases by 1 with a probability of 2/3. Show that the expected number of steps in
which her net profit is positive can be upper bounded by an absolute constant, independent
of the value of n. (Rating: **)