Assignment 1 - Probability and statistics

Mohieddine Farid Imane Fjer

February 21, 2023

1 Borel - Cantelli Theorem

Consider an infinite sequence of trials. The probability of success at the i-th trial
is some positive number pi . Let N be the event that there is no success, and I
the event that there isPan infinite number of successes. Assume that the trials are
∞
independent and that i=1 pi = +∞.
Show that P (N ) = 0

Solution. Let’s assume for the sake of contradiction, that there are only finitely
many indices i, with i ∈ N∗ such that pi > 0. Because the set of natural numbers
is infinite, the other probabilities are thus all equal to 0. Consider i1 , i2 , · · · , ir as
the indices of the non-null probabilities. We then have,
∞
X r
X
+∞ = pi = pik
i=1 k=1

Which is impossible, because of the right-hand-side sum is finite.

The last equality, follows from the fact that the other probabilities are all equal
to 0. Hence, there are infinitely many indices i, with i ∈ N∗ such that pi > 0.
Now let’s try to compute P (N ), the event has no success if every trial is un-
successful, the probability of success being pi for a trial, we thus have:
∞
Y
P (N ) = (1 − pi )
i=1

Because of our assumption, there are infinitely many pi ’s greater than 0, if follows
that the quantity 1 − pi is strictly less than 1 for infinitely many indices i ∈ N∗ .
Hence, P (N ) is the product of quantities, less than 1, within which infinitely many
are < 0. It follows that the product converges to 0, i.e P (N ) = 0.

1
Pr. Hajar El Hammouti Mohieddine Farid, Imane Fjer

2 Newton Binomial
n+1 n n




2.1 Show that k
= k
+ k−1
for all (k, n) ∈ N × N, and
1≤k≤n
Solution. Consider n + 1 people, we will count the number of ways to choose k
people

from these n + 1 in two ways. The first way is can be obviously done in
n+1
k ways.
Now let’s consider a specific person from the group, call him A. The groups we
are going to form either contains A or not. If the group does not contain A, then
we need to choose k people to form the group from the remaining n people (we
excluded A), which is done in nk ways. If the group contains A, then we need to

the remaining k − 1 people from n to complete the group, which gives us

choose
n
k−1 ways to do it.
Thus we finally get the desired Pascal’s property:
     
n+1 n n
= +
k k k−1
Pn n

k n−k
2.2 Show that (a + b)n = k=0 k a b , for all (a, b) ∈ R × R
and n ∈ N
Solution. We use induction on n, the case n = 0 is trivial. Assume we have the
property for n, let’s prove it for n + 1. We have:

(a + b)n+1 = (a + b)n (a + b)
n  
!
X n k n−k
= a b (a + b)
k
k=0
n   n  
X n k+1 n−k X n k n−k+1
= a b + a b
k k
k=0 k=0
n+1
X n  n  
k n−k+1
X n k n−k+1
= a b + a b
k−1 k
k=1 k=0

Where the second and fourth equalities follow respectively from the induction
hypothesis andPa change of

kthe index in the firstPoperand.

n+1 n n n
Note that k=1 k−1 a bn−k+1 = an+1 + k=1 k−1 ak bn−k+1 .
Pn n k n−k+1

n+1
P n n k n−k+1


Also k=0 k a b =b + k=1 k a b
Pr. Hajar El Hammouti Mohieddine Farid, Imane Fjer

Furthermore, we simply the expression,

n    
X n n
(a + b)n+1 = an+1 + bn+1 + + ak bn+1−k
k−1 k
k=1
n  
n+1 n+1
X n + 1 k n+1−k
=a +b + a b
k
k=1
n+1
X n + 1 
= ak bn+1−k
k
k=0

Where the second line follows from the first question: n+1 = nk + k−1
n




k .
Hence the result follows by induction.

3 Communication through a noisy channel

A binary (0 or 1) message transmitted through a noisy communication channel
is received incorrectly with probability ϵ0 and ϵ1 , respectively. Errors in different
symbol transmissions are independent. The source transmits a 0 with probability p
and transmits a 1 with probability 1 − p.

3.1 What is the probability that a randomly chosen symbol is

received correctly?
Solution. Consider A the event of receiving the symbol correctly.
We will use the total probability theorem.
P (A) = P (A|0)P (0) + P (A|1)P (1)
= (1 − ϵ0 )p + (1 − ϵ1 )(1 − p)

3.2 Suppose that the string of symbols 1011 is transmitted.

What is the probability that all the symbols in the string
are received correctly?
We know that the errors in different symbol transmissions are independent, meaning
that receiving a symbol is an independent event.
P (1011) = P (1)P (0)P (1)P (1)
= P (0)P (1)3
= (1 − ϵ1 )3 (1 − ϵ0 )

3.3 What is the probability that a transmitted 0 is correctly

decoded?
A transmitted 0 is correctly decoded if the 3-string message is decoded with two or
three zeros. Let B be the event of correctly decoding zero.
Pr. Hajar El Hammouti Mohieddine Farid, Imane Fjer

P (B) = P (000) + P (001) + P (100) + P (010)

= P (0)3 + 3P (001)
= (1 − ϵ0 )3 + 3ϵ0 (1 − ϵ0 )2

with

P (001) = P (100) = P (010) = ϵ0 (1 − ϵ0 )2

The above formula derives from the fact that we transmitted zero twice in the
correct way while one try was unsuccessful.

3.4 For what values of ϵ0 is this an improvement over sending

a single 0?
To have an improvement, the value of P (B) should be greater than the probability
of successfully decoding a 0 from one string, i.e. P (B) < 1 − ϵ0

(1 − ϵ0 )3 + 3(1 − ϵ0 )2 ϵ0 > (1 − ϵ0 )
(1 − ϵ0 )2 + 3(1 − ϵ0 )ϵ0 > 1
1 − 2ϵ0 + ϵ20 + 3ϵ0 − 3ϵ20 > 1
−2ϵ20 + ϵ0 + 1 > 1
(1 − 2ϵ0 )ϵ0 > 0
1 − 2ϵ0 > 0
1
> ϵ0
2

3.5 Suppose that the scheme of part (c) is used. What is

the probability that a 0 was transmitted given that the
received string is 101?
Consider E the event of receiving string is 101.
Where P (0) (P (1) respectively) refers to the probability that the source transmitted
a 0 (1 respectively). From question c, we have
P (E|0) = P (101|0) = ϵ20 (1 − ϵ0 )
and P (E|1) = P (101|1) = ϵ1 (1 − ϵ1 )2

Hence, using Bayes rule:

P (0)P (E|0)
P (0|E) =
P (0)P (E|0) + P (1)P (E|1)
pϵ20 (1 − ϵ0 )
= 2
pϵ0 (1 − ϵ0 ) + (1 − p)ϵ1 (1 − ϵ1 )2
Pr. Hajar El Hammouti Mohieddine Farid, Imane Fjer

4 Problem 4
Let A and B be independent events. Use the definition of independence to prove
the following:

4.1 The events A and B ∁ are independent

We are given that P (A ∩ B) = P (A)P (B), we need to prove P (A ∩ B ∁ ) =
P (A)P (B ∁ ).
We will use the following lemma:
If B ⊆ A, then P (A − B) = P (A) − P (B).

Solution. It is trivial that B ∁ is a subset of A ∪ B ∁ . Hence we can apply the

above formula:

P (A ∪ B ∁ ) − P (B ∁ ) = P (A ∪ B ∁ − B ∁ ) = P (A ∩ B) (1)
The independency of A and B ∁ follows from:

P (A ∩ B ∁ ) = P (A) + P (B ∁ ) − P (A ∪ B ∁ )
= P (A) + P (B ∁ ) − (P (B ∁ ) + P (A ∩ B))
= P (A) − P (A ∩ B)
= P (A) − P (A)P (B)
= P (A)(1 − P (B))
= P (A)P (B ∁ )

Where the second line follows from (1).

4.2 The events A∁ and B ∁ are independent.

Solution. We have:

P (A∁ ∩ B ∁ ) = 1 − P (A ∪ B)
= 1 − (P (A) + P (B) − P (A ∩ B))
= 1 − P (A) − P (B) + P (A ∩ B)
= 1 − P (A) − P (B) + P (A)P (B)
= (1 − P (A))(1 − P (B))
= P (A∁ )P (B ∁ )

It follows that A∁ and B ∁ are independent.