Lecturenotes7 8 Probability

Introduction to Probability Theory
K. Suresh Kumar
Department of Mathematics
Indian Institute of Technology Bombay
August 14, 2017

2
LECTURES 7 - 8
Example 0.1 (Ballot problem) In an election between two candidates, can-

didate I get n votes and II get 10 votes, where n > 11. Find the probability
that the winner was leading throughout the ballot counting.
For k, l non negative integers, let Akl denote the event that I is leading
throughout the counting and Fl+kk denote the event that the last ballot was
casted for I when I has k votes and II has l votes. Also denote P (Akl ) =
P (k, l). Observe that
P (k, l) = 0, l ≥ k, P (k, 0) = 1 for all k, l.
Then using Law of total probability, we get

c c
P (Akl ) = P (Akl |Fl+kk )P (Fl+kk ) + P (Akl |Fl+kk )P (Fl+kk ).
Now
(l + k − 1)! k
P (Fl+k,k ) = k × = .
(l + k)! l+k
Hence
c l
P (Fl+k,k ) = .
l+k
Now P (Akl |Fl+k,k ) can be viewed as the probability that I is leading from 1
to l + k − 1 ballot counts when I has l + k − 1 votes and II has k votes.
Hence
P (Akl |Fl+k,k ) = P (l − 1, k).
c
Similarly P (Akl |Fl+kk ) can be viewed as the probability that I leading from
1 to l + k − 1 ballot counts where I has k votes and II has l − 1 votes. Hence
c
P (Akl |Fl+kk ) = P (k, l − 1).
Hence we get the following recursive relation

k l
P (k, l) = P (k − 1, l) + P (k, l − 1)
l+k l+k
with the boundary conditions P (k, l) = 0, l ≥ k, P (k, 0) = 1 for all k =
0, 1, · · · , n, l = 0, 1, · · · , 10. Now by induction one can see that
k−l
P (k, l) = , k ≥ l.
l+k
n−10
This implies the required probability is n+10 .
3
Foundations
In this chapter, we quickly go through some basic mathematics used in

this course.
Real number system is a set R (called the set of real numbers) together
with two binary operations addition + and multiplication · and an order
relation <, i.e. (R, +, ·, <) satisfying the following.
I (Algebraic properties) Let a, b, c ∈ R.
1 (associative property)
(a + b) + c = a + (b + c), (a · b) · c = a · (b · c).
2 (commutative property)
a + b = b + a, a · b = b · a.
3 (additive and multiplicative identity ) There exists 0, 1 ∈ R such that
a + 0 = a = 0 + a, ∀ a ∈ R, a · 1 = a = 1 · a.
4 (additive and multiplicative inverses) For a ∈ R, there exists −a ∈ R

and 0 6= a ∈ R, there exists a−1 ∈ R such that
a + (−a) = 0, a · a−1 = 1.
5 (distributive property)
a · (b + c) = a · b + a · c.
II (Order properties) R contains a subset R+ (called the set of positive reals)

such that
1 For each a ∈ R, exactly one of the statements hold true.
a ∈ R+ , a = 0, −a ∈ R+ .
2 If a, b ∈ R+ , then a + b, a · b ∈ R+ .
4
Order properties defines an order < on R given by, a < b if b − a ∈ R+

and a ≤ b if a < b or a = b.
III (Least upperbound property) If A ⊆ R has an upper bound1 , then it

has a least upper bound.
On R, we introduce a metric (i.e., a distance) using the ’modulus’ func-
tion 
 x if x > 0
|x| = 0 if x = 0
−x if x < 0,

i.e., distance between x, y ∈ R is given by |x − y|.

Sets and Operations on sets: We will briefly describe naive theory of
sets. Let U be a universal set, i.e. a set which contains all ’well defined”
objects we may come across. For us it is going to the ’Sample space’. A is a
subset of U , if every element of A is an element of U and P (U ) denote the
collection of all subsets of U .
Given a family of subsets {Ai |i ∈ I} (indexed by I), ∪i∈I Ai and ∩i∈I Ai
are defined as
{x ∈ U |x ∈ Ai for some i}, {x ∈ U |x ∈ Ai for all i}
respectively. For a subset A of U , compliment of A (in U ) denoted by Ac is

defined as
Ac = {x ∈ U |x ∈ / A}.
We will now list, some of the useful properties of sets below.
1. For A, B, C subsets of U ,
(commutative property) A ∪ B = B ∪ A, A ∩ B = B ∩ A,
(associative property) (A∪B)∪C = A∪(B∪C), (A∩B)∩C = A∩(B∩C).
(Ac )c = A.
2. For subset of U , given by A, Bi ; i ∈ I, I an index set,
A ∩ (∪i∈I Bi ) = ∪i∈I (A ∩ Bi ), A ∪ (∩i∈I Bi ) = ∩i∈I (A ∪ Bi ) .

1
M ∈ R is said to be an upper bound for the set A if a ≤ M for all a ∈ A
5
3. For subsets of U given by Bi , i ∈ I,

c c
∪i∈I Bi = ∩i∈I Bic , ∩i∈I Bi = ∪i∈I Bic .
Functions and its properties: Let A and B be two non empty sets, by
f : A → B, we mean a function from A to B. For C ⊆ A, f (C) defined by
f (C) = {f (a)|a ∈ C}
is called the image set of C under f and for D ⊆ B, f −1 (D) defined by
f −1 (D) = {a ∈ A|f (a) ∈ D}
is called the pre image of D under f .

A function f : A → B is said to be 1 − 1 if a1 , a2 ∈ A, with f (a1 ) =
f (a2 ) ⇒ a1 = a2 . A function f : A → B is said to be onto if for each b ∈ B,
there exists a a ∈ A such that f (a) = b. A function f : A → B is said to be
a bijection if it is both 1 − 1 and onto.
Now we list some properties which are left as exercises.
Let f : A → B and {Ai , i ∈ I} be a family of subsets of A and {Bi , i ∈ I}
be a family of subsets of B, then
1.
f (∪i∈I Ai ) = ∪i∈I f (Ai ).
2.
f (∩i∈I Ai ) ⊆ ∩i∈I f (Ai ),
the inclusion is in general strict.
3.
f −1 (∪i∈I Bi ) = ∪i∈I f −1 (Bi ).
4.
f −1 (∩i∈I Bi ) = ∩i∈I f −1 (Bi ).
5.
f −1 (B1c ) = (f −1 (B1 ))c .
6
Sequences and series: We consider real valued sequences and series. A

sequence {an } is said to converge to a if for each ε > 0, there exists a
natural number N such that |an − a| < ε for all n ≥ N . In this case, we
write lim an = a.
n→∞
Now we list some useful results.
1. If {an } is an increasing sequence (i.e., an ≤ an+1 for all n), which

is bounded from above, then lim an = sup an . An analogous result
n→∞ n≥1
holds true for decreasing sequences.
2. (i) Let {an } and {bn } be two convergent sequences such that an ≤ bn
for all n. Then
lim an ≤ lim bn .
n→∞ n→∞
(ii) Let {an }, {bn } and {cn } be three sequences such that an ≤ bn ≤ cn
for all n and limn→∞ an = limn→∞ cn = a, then {bn } is convergent and
limn→∞ bn = a.
A series ∞
P
an is said to converge to a number a, if the sequence
n=1 P of
partial sums sn = nk=1 ak converges to a. In this case, we write ∞
P
n=1 n =
a
a.
A series ∞
P P∞
n=1 an is said to converge absolutely if the series n=1 |an |
converges.
In general
P∞convergence doesn’t imply absolute convergence. For example,
(−1)n
the series n=1 n converges but not absolutely. This is left it as an
exercise.
Real valued functions Real valued functions f : I → R, where I is an
interval is very often used. Given an interval I, x0 is said to be an interior
point, if there exists ε > 0 such that (x0 − ε, x0 + ε) ⊆ I. An interval I is
said to be open if all x ∈ I are interior points.
We say that f : I → R, I an open interval, is continuous at x0 ∈ I, if

for each ε > 0, there exists a δ > 0 such that
|x − x0 | < δ ⇒ |f (x) − f (x0 )| < ε.
f : I → R is said to be continuous if it is continuous at all x0 ∈ R.
We will be using the following result. f : R → R is continuous iff f −1 (O)
is open for all O open in R.
7
Now I will do a couple of exercises in detail to illustrate how to solve

problems. This is intended as an illustration of ’how to approach a problem’
Problem 1: Let F be a field with finitely many elements. Show that F is

a σ-field.
Solution-explained: By comparing the definitions of field and σ-field, it
is evident that one need to show the following ”If A1 , A2 , · · · ∈ F, then
∪∞n=1 An ∈ F.”
One may immediately notice that ”since there are only finitely many
elements in F, any countable union is going to be a finite union”. Intuition
is correct but how do one establish this? One may say, it is ’trivial’ because
An are taken from a finite collection, so An ’s should repeat after some n
onwards. Convincing but some thing is missing in the ’proof’, in fact it is
too vague.
I will give a proof, using the set theory we know. From given, there
exists n such that #F = n. Hence F = {B1 , B2 , · · · , Bn } for some Bi ∈
P (Ω), i = 1, 2, · · · , n. Now since An ∈ F for all n, we have {A1 , A2 , · · · } ⊆
{B1 , B2 , · · · , Bn } (definition of subset). Hence there exists {i1 , i1 , · · · , ik } ⊆
{1, 2, · · · , n} such that {A1 , A2 , · · · } = {Bi1 , Bi1 , · · · , Bik }. This is what
one meant by An ’s repeat after some time (no ambiguity here). Now we
claim that
∪∞ k
n=1 An = ∪l=1 Bil .
The above follows by verification as follows.
x ∈ ∪∞
n=1 An ⇔ x ∈ An for some n
⇔ x ∈ Bil for some il (since, An ∈ {B1 , B2 , · · · , Bn })
⇔ x ∈ ∪kl=1 Bil .
i.e. the proof of countable union is indeed a finite union is complete. Now
∪∞n=1 An ∈ F, since F is a field.
Problem 2: Let {A1 , A2 , · · · , An } be a partition of Ω. Show that σ({A1 , A2 , · · · , An })

is the set of all finite union of Ai ’s.
Solution-explained First set
D = set of all finite union of A0i s.
Our task to show that σ({A1 , A2 , · · · , An }) = D.
By looking at the definition of σ({A1 , A2 , · · · , An }), we need to show
that
8
1. D is a σ-field,
2. A1 , A2 , · · · , An ∈ D,
3. D is the smallest σ-field satisfying 1. and 2.
Once roughly the tasks are identified, it is about doing them. Note first two
tasks, i.e. 1 and 2 are just verifications.
For checking 1., we need to verify the definition of σ-field for D. Let us
do this.
Note Ω = A1 ∪ · · · ∪ An , a finite union of Ai ’s, and hence in D.
For A ∈ D, since A is a finite union of Ai ’s, A = Ai1 ∪ · · · ∪ Aik for some
i1 , · · · , ik .
Now,
Ac = ∪i∈{1,2,··· ,n}\{i1 ,··· ,ik } Ai ∈ D.
Now for A = Ai1 ∪ · · · ∪ Aik , B = Aj1 ∪ · · · ∪ Ajl ∈ D,
A ∪ B = Ai1 ∪ · · · ∪ Aik ∪ Aj1 ∪ · · · ∪ Ajl ∈ D.
Hence D is a field and hence a σ-field by using Problem 1.
For checking 2., note that each Ai is a ’one’ union of Ai and hence are
in D. This prove 2.
Now for 3., one need to think how to prove this. For example, one if can
show the statement
’if G is a σ-field containing {A1 , · · · , An }, then D ⊆ G’
then from the definition of σ-field generated and 1. and 2. it follows that
σ({A1 , A2 , · · · , An }) = D.
Hence proving 3. is same as proving the statement ’if G is a σ-field
containing {A1 , · · · , An }, then D ⊆ G’.
Let G is a σ-field containing {A1 , · · · , An }, then from the definition of
σ-field, it contains all the finite union of Ai ’s. Hence D ⊆ G. That’s all.

Lecturenotes7 8 Probability

Uploaded by

Copyright:

Available Formats

Lecturenotes7 8 Probability

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecturenotes7 8 Probability

Uploaded by

Copyright:

Available Formats

Introduction to Probability Theory

August 14, 2017

Example 0.1 (Ballot problem) In an election between two candidates, can-

P (k, l) = 0, l ≥ k, P (k, 0) = 1 for all k, l.

Then using Law of total probability, we get

Hence we get the following recursive relation

In this chapter, we quickly go through some basic mathematics used in

3 (additive and multiplicative identity ) There exists 0, 1 ∈ R such that

4 (additive and multiplicative inverses) For a ∈ R, there exists −a ∈ R

II (Order properties) R contains a subset R+ (called the set of positive reals)

1 For each a ∈ R, exactly one of the statements hold true.

Order properties defines an order < on R given by, a < b if b − a ∈ R+

III (Least upperbound property) If A ⊆ R has an upper bound1 , then it

i.e., distance between x, y ∈ R is given by |x − y|.

{x ∈ U |x ∈ Ai for some i}, {x ∈ U |x ∈ Ai for all i}

respectively. For a subset A of U , compliment of A (in U ) denoted by Ac is

(associative property) (A∪B)∪C = A∪(B∪C), (A∩B)∩C = A∩(B∩C).

2. For subset of U , given by A, Bi ; i ∈ I, I an index set,

A ∩ (∪i∈I Bi ) = ∪i∈I (A ∩ Bi ), A ∪ (∩i∈I Bi ) = ∩i∈I (A ∪ Bi ) .

3. For subsets of U given by Bi , i ∈ I,

is called the image set of C under f and for D ⊆ B, f −1 (D) defined by

f −1 (D) = {a ∈ A|f (a) ∈ D}

is called the pre image of D under f .

the inclusion is in general strict.

Sequences and series: We consider real valued sequences and series. A

Now we list some useful results.

1. If {an } is an increasing sequence (i.e., an ≤ an+1 for all n), which

We say that f : I → R, I an open interval, is continuous at x0 ∈ I, if

Now I will do a couple of exercises in detail to illustrate how to solve

Problem 1: Let F be a field with finitely many elements. Show that F is

Problem 2: Let {A1 , A2 , · · · , An } be a partition of Ω. Show that σ({A1 , A2 , · · · , An })

3. D is the smallest σ-field satisfying 1. and 2.

A ∪ B = Ai1 ∪ · · · ∪ Aik ∪ Aj1 ∪ · · · ∪ Ajl ∈ D.

Hence D is a field and hence a σ-field by using Problem 1.

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.