Section 2.

2 Sample Space and Events

We consider an experiment whose outcome is not predictable with certainty. However, we
suppose that the set of all possible outcomes is known.
DEFINITION: The set of all possible outcomes of an experiment is the sample space of an
experiment, and is denoted by S. Subsets of S are called events. Elements of S, x S, are
called outcomes.
EXAMPLES:
1. A coin is tossed twice and the outcome of each is recorded. Then,
S = {(H, H), (H, T ), (T, H), (T, T )}
The event that the second toss was a Head is the subset
E = {(H, H), (T, H)}
2. Consider 3 light-bulbs. Our experiment consists of finding out which light-bulb burns out
first, and how long (in hours) it takes for this to happen.
S = {(i, t) : i {1, 2, 3}, t 0}
i tells you which one burns out, and t gives how long it lasted, in hours. The event that the
2nd bulb burns out first, and it lasts less than 3 hours is the set
E = {(2, t) : t < 3}
3. You roll a four sided die until a 4 comes up. The event you are interested in is getting a
three with the first two rolls.
S = { a1 a2 . . . an | n 1, an = 4, ai {1, 2, 3} for i 6= n}
E = { 33a3 a4 . . . an | n 3, an = 4, ai {1, 2, 3} for i 6= n}
DEFINITION: The union of E and F , denoted E F , consists of those outcomes that are in
either E or F (or both). (we write x E F if x E or x F.)
EXAMPLE: If E = {(H, H), (H, T )} (Head on first toss) and F = {(T, T ), (H, T )} (tail
on the second toss). Then,
E F = {(H, H), (H, T ), (T, T )}
only missing (T, H).
DEFINITION: The intersection of E and F , denoted EF or E F , consists of those events
that are in both E and F. (we write x EF if x E and x F.)
1

EXAMPLE: If E = {(H, H), (H, T )} (Head on first toss) and F = {(T, T ), (H, T )} (tail
on the second toss). Then
EF = {(H, T )}
DEFINITION: The empty set {} = is the set consisting of nothing.
DEFINITION: Two sets are mutually exclusive if EF = . A set of sets, {E1 , E2 , . . . , } are
mutually exclusive if Ei Ej = for all i 6= j.
EXAMPLE: Let S = {(i, j) | i, j {1, . . . , 6}} be the outcome from two rolls of a die. Let
E be those such that i + j = 6 (i.e. E = {(1, 5), (2, 4), . . .}) and F be those such that i + j = 7
(i.e. F = {(1, 6), (2, 5), . . .}). Then EF = and they are mutually exclusive.
The definition for the union and intersection of a sequence of events E1 , E2 , . . . are similar:

[
1.
En is the event consisting of those outcomes that are in at least one En , for n = 1, 2, . . ..
n=1

2.

En is the event consisting of those outcomes that are in each En for n = 1, 2, . . ..

n=1

DEFINITION: The complement of E, denoted E c , consists of those outcomes that are not in
E. Note that S c = and that (E c )c = E.
DEFINITION: E is a subset of F if x E implies x F . Notation: E F .
DEFINITION: E and F are equal, denoted E = F , if E F and F E.

Important Set Relations

Commutative laws:
E F = F E,

EF = F E

Associative laws:
E (F G) = (E F ) G,

E(F G) = (EF )G

Distributive laws:
(EF ) H = (E H)(F H),
De Morgans laws:

i=1

Ei

!c
c

(E F )H = (EH) (F H)

Eic

i=1
c

Note that if Ei = for i 3, we have (E F ) = E F c .

!c

\
[
Ei =
Eic
i=1

i=1

Note that if Ei = for i 3, we have (E F )c = E c F c .

2

Proof: Let x
that Ei and so

i=1

Ei

!c

. Then, x 6

Ei . Thus, x 6 Ei for at least one Ei . So x Eic for

i=1

Eic

i=1

Let x

Eic .

Then x

Eic

i=1

Ei

!c

Eic

i=1

for at least one Ei and so x 6 Ei and, hence, x 6

i=1

Ei . Therefore

i=1

Ei

i=1

!c

and

Eic

i=1

Ei

i=1

!c

Section 2.3 Axioms of Probability

Basic Question: What do we mean by the probability of an event? Good idea: Relative
frequency interpretation. Let n(E) be the number of times that E occurs during n performances
of the experiment. Finally, define
n(E)
n
n

P (E) = lim
EXAMPLES:

1. Rolling a fair die. Will find P (6 is rolled) 1/6.

2. Flipping a fair coin twice. Will find P (first heads, then tails) 1/4.
REMARKS:
n(E)
actually converges, this has to be an assumption.
n
n

1. No reason to believe that lim

2. Notions that do not have repeatability do not have meaning: for example, probability of
rain tomorrow or guilt/innocence in criminal cases.
We will take a different route. An axiomatic approach. We will eventually show that after
assuming three basic axioms, the limit above does exist (law of large numbers, chapter 8).

Axioms of Probability
Let S be a sample space for some experiment. For each event E S, we assume that a number
P (E) is defined and satisfies the following three axioms:
Axiom 1: 0 P (E) 1.
Axiom 2: P (S) = 1.
Axiom 3: If {Ei } are mutually exclusive ( Ei Ej = if i 6= j), then P


S

i=1

We call P (A) the probability of E.

3

Ei

i=1

P (Ei ).

THEOREM: P () = 0.
Proof: We have S = S

. Note that {S, , , . . .} are mutually exclusive. Thus,

i=2

P (S) = P

i=2

= P (S) +

P ()

i=2

Thus, P () = 0. 
THEOREM: Let A1 , . . . , An be mutually exclusive (but finite). Then,
!
n
n
X
[
P
Ai =
P (Ai )
i=1

i=1

Proof: Use previous result and take sum to infinity with the rest being the empty set. 
EXAMPLES:
1. Consider flipping a coin and recording the outcome. Then S = {H, T }. Need a probability.
Assume fair coin. Then
P ({H}) = 1/2
P ({T }) = 1/2
P ({H, T }) = 1
Note that
P ({H} {T }) = 1 = 1/2 + 1/2 = P ({H}) + P ({T })
This is not the only probability possible for this experiment/sample space. Could have an
unfair coin.
P ({H}) = p, P ({T }) = 1 p, P ({H, T }) = 1
2. Rolling a die. Assuming that the die is fair, P ({1}) = P ({2}) = . . . = P ({6}) = 1/6.
Probability of rolling an even number is
P ({2, 4, 6}) = P ({2}) + P ({4}) + P ({6}) = 1/6 + 1/6 + 1/6 = 1/2

Section 2.4 Some Simple Propositions

THEOREM: For any event E we have
P (E c ) = 1 P (E)
Proof: For any set, E and E c are mutually exclusive. Therefore,
1 = P (S) = P (E E c ) = P (E) + P (E c ) 
4

EXAMPLE: Roll two fair dice. Sample space

S = {(i, j) : 1 i, j 6}
Assume fair dice, so probability of each event is 1/36. Let A be the event that the sum of the
rolls equals 1, 2, 3, 5, 6, . . . , or 12. Hard to calculate. But
P (Ac ) = P (sum is 4) = P ({(1, 3), (2, 2), (3, 1)}) = 3/36
So, P (A) = 1 3/36 = 33/36 = 11/12.
THEOREM: If A B, then P (A) P (B).
Proof: Because A B, we can write B as
B = A (Ac B)
These are mutually exclusive and so
P (B) = P (A) + P (Ac B)
P (Ac B) 0 and so the result is shown. 
THEOREM: P (E F ) = P (E) + P (F ) P (EF ).
Proof: Note that E F = E E c F , and these are mutually exclusive. Thus,
P (E F ) = P (E) + P (E c F )
Also, F = EF E c F (which are mutually exclusive) and so
P (F ) = P (EF ) + P (E c F )
Combining the two equations shows
P (E F ) = P (E) + P (E c F ) = P (E) + P (F ) P (EF ) 
THEOREM: For any two sets E and F we have
P (F ) = P (EF ) + P (E c F )
EXAMPLE: You believe the following: with probability 0.5 it will snow today. With probability
0.3 it will snow tomorrow. With probability 0.2 it will snow both days. What is the probability
that it will not snow either day?

EXAMPLE: You believe the following: with probability 0.5 it will snow today. With probability
0.3 it will snow tomorrow. With probability 0.2 it will snow both days. What is the probability
that it will not snow either day?
Solution: Let E, F be the events that it snows today/tomorrow, resp. Then,
P (E) = 0.5, P (F ) = 0.3,

and P (EF ) = 0.2

We need P (E c F c ). Not snowing, is the complement of snowing at least once, E F , and so,
P (E c F c ) = 1 P (E F ) = 1 (P (E) + P (F ) P (EF )) = 1 (0.5 + 0.3 0.2) = 1 0.6 = 0.4
We can generalize the previous theorem:
THEOREM: For any three sets A, B, C, we have
P (A B C) = P (A) + P (B) + P (C) P (AB) P (BC) P (AC) + P (ABC)
Proof: Proof by Venn diagram. 
THEOREM (inclusion-exclusion identity): Let E1 , E2 , . . . , En be a sequence of sets. Then
P (E1 E2 . . . En ) =

n
X

P (Ei )

i=1

i1 <i2

+(1)r+1

P (Ei1 Ei2 ) + . . .
P (Ei1 Ei2 . . . Eir )

i1 <i2 <...<ir

the summation

i1 <i2 <...<ir

set {1, 2, . . . , n}.

+ . . . + (1)n+1 P (E1 E2 . . . En )
 
n
possible sets of size r of the
P (Ei1 Ei2 . . . Eir ) is taken over the
r

Proof: Can be done by induction. 

Section 2.5 Sample Spaces Having Equally Likely

Outcomes
We have already seen that in many experiments it is natural to assume all outcomes are equally
likely. That is, if S = {1, 2, . . . , N} then,
P ({1}) = P ({2}) = . . . = P ({N}) =
It then follows by Axiom 3 that for any event E
P (E) =

# elements of E
# elements in S
6

1
N

EXAMPLES:
1. A number is chosen at random from S = {1, . . . , 520}. What is the probability that it is
divisible by 5 or 7?

2. If 3 balls are randomly drawn from a bowl containing 6 white and 5 black balls, what is the
probability that one of the balls is white and the other two are black?

3. A committee of 6 is to be chosen from 15 people. There are 5 Americans, 3 Canadians, and 7

Europeans in the group. Assuming the committee is randomly chosen, what is the probability
that the committee will consist of 2 Americans, 1 Canadian, and 3 Europeans?

EXAMPLES:
1. A number is chosen at random from S = {1, . . . , 520}. What is the probability that it is
divisible by 5 or 7?
Solution: Assume each number has a probability of 1/520 of being chosen. Let A be the event
that it is divisible by 5. Let B be the event that it is divisible by 7. We have 520/5 = 104, so
P (A) = 104/520. Similarly, 520/7 = 74 + 2/7, so P (B) = 74/520. Want P (A B). So, by the
theorem, we need P (A B). x A B if and only if divisible by 35. 520/35 = 14 + 30/35. So
P (A B) = 14/520. Thus,
P (A B) = P (A) + P (B) P (A B) = 104/520 + 74/520 14/520 = 164/520
2. If 3 balls are randomly drawn from a bowl containing 6 white and 5 black balls, what is the
probability that one of the balls is white and the other two are black?
Solution 1: Think of order being relevant. There are 11 10 9 = 990 total possible choices of a
first, second, third ball. There are 6 (5 4) ways to first choose a white, then two black, there
are 5 6 4 ways to choose black, white, black, and 5 4 6 ways to choose two black, then white.
So there are a total of 3 6 5 4 = 360 ways to choose one white and two black. Thus, the
answer is
360
4
=
990
11
 
11
ways to choose three
Solution 2: Now suppose order is not important. There are now
3
 
 
5
6
ways to choose one white and two black balls, respectively.
and
balls. There are
2
1
Therefore, the probability is
  
6 5
5!
5!
6
54332
4
1 2
3!2! =
2!
  =
=
=
11 10 9
11!
11
11 10 9
11
3!
8!3!
3
3. A committee of 6 is to be chosen from 15 people. There are 5 Americans, 3 Canadians, and 7
Europeans in the group. Assuming the committee is randomly chosen, what is the probability
that the committee will consist of 2 Americans, 1 Canadian, and 3 Europeans?
 
   
 
7
3
5
15
, and
,
possible committees. There are
Solution: There are a total of
3
1
2
6
ways to choose 2 Americans, 1 Canadian, and 3 Europeans, respectively. Therefore, by the
generalized counting process, the probability is
   
5 3 7
30
2 1 3
 
0.2097
=
15
143
6
8