1.

5 Special Counting Cases:

1.5.1. Arrangements around a circle.

Key: Make one seated object in the circle fixed, and permute the
(n-1) objects around this fixed reference point.

Formula: number of ways to arrange n objects in a circle =(n-

1)!

Ex: 5 people sat around a circular table. In how many ways can they
be seated?

Set one person on top, count the number of ways that

the other 4 people can be seated in the remaining 4
seats.

= 4x3x2x1
= 24 ways.

Contrasting Example:
If the Last Supper happened to be on a round table, and John sat to
Jesus’ right and Judas Iscariot to his left. In how many ways could the
12 Apostles and Jesus have seated around the table?

Jo Je Ju This time, the formula (n-1)! does

not apply anymore. This is a
simple multiplication rule
application. Place Jesus on top of
the circular table, put John to his
right and Judas to his left. You now
have 10 other Apostles who can
freely seat around the remaining
10 seats. So 10P10 = 10! =3,628,800 ways.

Probability and Counting Theorypage 12

1.5.2. Permutation of n indistinguishable objects, of which
n1 is of type 1, n2 is of type 2 and so on for k different types.

n !
Formula:
n1! n2 !....nk !

where n=total number of indistinguishable objects to be ordered.

n1= number of similar objects of type 1.
n2= number of similar objects of type 2.
:
:
nk= number of similar objects of type k.
and sum of (n1+ n1+ ……+ nk) = n

Example: In how many ways can you arrarrange the letters in the word
“ENGINEER”?

Ans: n= total letters in the word = 8

n1= number of similar letters “E” = 3.
n2= number of similar letters “N” = 2 :
n3= 1 “G”.
n4= 1 ”I”
n5= 1 “R”

8 !
3! 2! 1! 1! 1!
= 3360 ways.

Example: You have three 100five 10and seven 1 resistors that

you would connect together in series. How many ways can
you arrange the 15 resistors in series?

Ans: 15! / (3! 5! 7!) = 360,360 ways.

Example: In how many ways can you arrange 5 boys, 5 girls and 4
dogs in a line with no restrictions on order?

Ans: 14! / (5! 5! 4!) = 252,252 ways.

Probability and Counting Theorypage 13

1.5.3. Partitioning n indistinguishable objects into cells of
size n1,n2, ….nk.

n !
Formula:
n1! n2 !....nk !

where n=total number of indistinguishable objects to be divided

up.
n1= number of items in cell 1.
n2= number of items in cell 2.
:
:
nk= number of items in cell 3.

Example: 15 students in a class need to be divided up into groups of

sizes 3,5, and 7. In how many ways can the groupings be
made?

Ans: 15! / (3! 5! 7!) = 360,360 ways.

Example: 15 distinct gifts need to be divided up evenly among 4

nephews. Any remainder would be kept. In how many ways
can the 15 candies be divvied up?

Ans: There should be 3 gifts per nephew and the remainder of 3

candies kept. So we would be partitioning 15 gifts into four
cells of size 3 and another cell of size 3 remains.

15!/ (3! 3! 3! 3! 3!) = 168,168,000 ways.

Example: In how many ways can 5 recruits be assigned into three

disctinct combat units if each unit is to have at least one
recruit assigned?

Label each combat unit as A, B and C, and enumerate the assignment

of recruits as possible.
Assignme A B C Number of Muliplication rule
nts ways to assign 5
recruits
1 1 1 3 5!/(1!1!3!)
2 1 3 1 5!/(1!3!1!) 5!/(1!1!3!)x3

Probability and Counting Theorypage 14

 cases
3 3 1 1 5!/(3!1!1!)
4 1 2 2 5!/(1!2!2!)
5 2 1 2 5!/(2!1!2!) 5!/(1!2!2!) x 3
cases
6 2 2 1 5!/(2!2!1!)
Total: 150 ways.
Practice Exercises: (Special Permutation cases: Circular arrangement,
ordering similar types, Partitioning)
1. Five Americans, three Italians and six Filipinos sit randomly in a
round table such that each sits together with their countrymen.
How many seating arrangements are possible? Ans: 1,036,800
arrangements.
2. Eight people are to be seated in a circular table, but 2 people do not
want to sit beside each other. How many possible seating
arrangements are possible? Ans: 3600 ways.
3. Judy is a librarian. She has 4 copies of Jane Eyre, 4 copies of A Tale
of Two Cities, and 4 copies of Of Mice and Men. If she placed all the
classics on a shelf, in how many ways can she shelf them? Ans:
34,650 ways
4. Judy has three sets of classics in literature, each set having four
volumes. In how many ways can she put them in a bookshelf so
that books of each set are not separtated? Ans:82,944 ways.
5. A seven-bit signal was be made from a sequence of 1’s and four
0’s (not necessarily in that order). However, the computer that
received the signal did not receive the signals in the correct
sequence, but is certain it received three 1’s and four zeroes. In
how many ways can the signal be incorrect? Hint: One sequence is
correct, deduct one from the total number of possible signals. Ans:
[7!/(3!4!)-1] = 34 incorrect sequences.
6. Imagine a city grid that is like a rectangular coordinate system.
You are currently at (0,0) and you want to get to corner (3,4). If a
path consists of a series of eastward (go left) and northward (go
upward) moves, how many shortest paths are possible assuming
that all streets and avenues are of the same length? Ans: 7!/ (3!4!)
= 35 paths
7. Ten people are going on a roadtrip and four of them decided to
bring their cars. In how many ways can they be transported if we
strictly abide by all the following conditions: (a) the owner will
drive his own car (b) each car can hold a maximum of five
passengers including its driver and (c) each car must transport at
least two passengers.. Hint: There should be only 6 passengers to
allocate. Ans: 1560 ways.

Probability and Counting Theorypage 15

8. A committee of 4 members is to be formed from 15 members of an
organization. One of these elected will be chairman. In how many
ways can the election results go? Ans:5460 ways.
9. A tenants’ organization with 25 members must elect a policy board
with 5 members, and a negotiating team composed of 3 of the 5
policy board members. How many different outcomes can the
election have? Ans: 531,300 ways.
10. Seven X-men want to go out for lunch. But
there are actually 3 places where they can go eat. In how many
ways can the mutants go out when they can go out as a team (or
separate teams) of at least three mutants each? Ans: 213 ways.
11. 10 students are to be assigned inside a
dormitory that has 6 rooms left. Each room can accommodate 2
students. Assume that all rooms are initially unoccupied.
 In how many ways can the students be assigned ?
Ans:4,082,000
 In how many ways can there be no unoccupied room? Ans:3,402,000

2.0 SOME RELATIONS FROM SET THEORY

An event can be thought of as a set. As a set, we may use

relationships and results from elementary set theory to study events.
The following operations will be used to construct new events from
given sets.

Definitions :

Union = The union of two events A and B (denoted by AB and read
“A or B”) is the event consisting of all outcomes that are either
in A or in B or in both events.

Intersection = The intersection of two events A and B (denoted by

AB and read “A and B”) is the event consisting of all outcomes
that are in both A and B.

Complement = the complement of an event A, denoted by A’ (read

“A”-prime or “not A”), is the set of all outcomes in the universal
set S that are not contained in A.

Probability and Counting Theorypage 16

Sets can be represented as a Venn Diagram:= a rectangle that includes
circles depicting the subsets. Named after the English logician John
Venn (1834-1923).

Probability and Counting Theorypage 17

(Ex) Experiment: Single throw of two dice.
Let Event A = getting a total of 7
Event B = first die is a 3.
Event C= getting a sum of 10 or more.

Let us depict the following events:

a. AB = “either (A=a total of 7) or (B=the first die is 3)”
b. AB = “(a total of 7) and (the first die is 3)”
c. A’ = “not A”= a total of 7 does not occur.

1st die 1 2 3 4 5 6
2nd die

1 (1, (2, (3, (4, (5, (6,

1) 1) 1) 1) 1) 1)

2 (1, (2, (3, (4, (5, (6,2

2) 2) 2) 2) 2) )

3 (1, (2, (3, (4, (5, (6,3

3) 3) 3) 3) 3) )

4 (1, (2, (3, (4, (5, (6,4

4) 4) 4) 4) 4) )

5 (1, (2, (3, (4, (5, (6,5

5) 5) 5) 5) 5) )

6 (1, (2, (3, (4, (5, (6,6

6) 6) 6) 6) 6) )

Event A

Event B

We can show the events as regions corresponding to the positions on

the table above:

1 2 3 4 5 6 1 2 3 4 5 6
1 A 1 B
2 A 2 B
3 A 3 B
4 A 4 B
5 A 5 B

Probability and Counting Theorypage 18

 6 A 6 B
Event A= Total of 7 Event B= First die
is a 3.

Probability of A = Probability of B =

=P(A) = 6/36 =P(B)=

6/36

1 2 3 4 5 6 1 2 3 4 5 6
1 AB AB 1
2 AB AB 2
3 AB AB 3
4 AB 4 AB
5 AB AB 5
6 AB AB 6

Event (AB)= Either a total of 7 occurs Event AB =A total of

7 occurs
Or the first die is a 3. and the first die
is 3.

Probability of (A or B) Probability of (A
and B)
=P(AB) = 11/36 =P(AB)= 1/36

1 2 3 4 5 6
1 A’ A’ A’ A’ A’
2 A’ A’ A’ A’ A’
3 A’ A’ A’ A’ A’
4 A’ A’ A’ A’ A’
5 A’ A’ A’ A’ A’
6 A’ A’ A’ A’ A’
Event A’= not A
= A total of 7 does not occur.
= 30/36

PROBABILITY LAWS RELATED TO SET THEORY:

Probability and Counting Theorypage 19

 1. P(A’) = 1-P(A) or P(A)=1-P(A’)
When it is easier to count either P(A) or its complement P(A’)
then this law would be useful.

Ex: If a card is drawn from a deck of playing cards, what is the

probability that it is not a court card (J,Q, or K)?

= 1-P(court card)
= 1-(3 court cards per suit)x(4 suits) /52
= 1-12/52
= 40/52
= 0.7692

2. If two events A and B are mutually exclusive, then P(AB)=0

Recall the definition of mutually exclusive events: two events are
mutually exclusive if the occurrence of one precludes the occurence
of the other. When two events are mutually exclusive, they cannot
occur simultaneously: either A or B but not both occurs. If two
events are mutually exclusive, then there are

no common occurrences between the two.

(Ex) In a single throw of 2 die, let event C= a total of 5 occurs, and

let event D=a total of 10 or more occurs. Events C and D
cannot occur simultaneously. Therefore events C and event D
are mutually exclusive, and there exists no intersection of
events C and D. P(CD)=0

1 2 3 4 5 6 1 2 3 4 5 6
1 C 1
2 C 2
3 C 3
4 C 4
5 5 D D
6 6 D D
Event C= Total of 5 Event D= total of
10 or more.

(Ex) In a single throw of 2 die, let event O=sum is an odd number,

and let event F=both die show the same number.

Probability and Counting Theorypage 20

 1 2 3 4 5 6 1 2 3 4 5 6
1 O O O 1 F
2 O O O 2 F
3 O O O 3 F
4 O O O 4 F
5 O O O 5 F
6 O O O 6 F

One can see that for all slots taken by “O” in the left-hand table do
not share any slots taken by “F” in the right-hand table. Since no
element is shared by both events, they are mutually exclusive.

Alternative reason: For event F, the sum of two identical numbers

n is always even. (2n =even). All odd numbers (like event O)
cannot be even numbers. Therefore, the two events O and F are
mutually exclusive. P(OF)=0.

3. P(AB)= P(A) + P(B) – P(AB)

If a Venn diagram of (AB) is drawn, where there exists a common
portion (AB), We could perform the set union operation (A+B) and
see that in the process of adding P(A) with P(B), their intersection
was inadvertently added twice. We must deduct the intersection
AB from the sum P(A)+P(B) to find the net sum of their areas.

(Ex) Recall the experiment on two die on pages 13-14. Let A=a
sum of 7 occurs, and B=the first die is a 3. Find P(AB) using the
formula.

P(A) = 6/36 or 1/6

P(B) = 6/36 or 1/6 P(AB)= 1/6 + 1/6 –1/36

P(AB)=1/36 = 11/36

(Ex) From a survey, 450 people said they use Brand X, while 300
people use Brand Y. Further inspection shows that 120 people use
both X and Y. It is also known that a total of 780 people were
surveyed.

Probability and Counting Theorypage 21

 a. How many people use either X or Y? Ans: 450+300-
120 = 630
b. How many people do not use X nor Y? Ans: 780 – 630 =
150
c. How many people use X exclusively? Ans: 450 –120 =
330
d. How many people use Y exclusively? Ans: 300 –120 =
180

Eqns: a. 450+300-120 c. X+120=450

b. 780-(450+300-120) d. Y+120=300

We may easily transform the questions in the previous example and

turn them into probability questions:

If a person was randomly chosen from the survey sample, what is

the probability that:
a. The person uses either X or Y? Ans: 630/780 =
0.8077
b. The person does not use X nor Y? Ans: 150/780 =
0.1923
c. The person uses X exclusively? Ans: 330/780 =
0.4231
d. The person uses Y exclusively? Ans: 180/780 =
0.2307

4. The probability of two independent events A and B occurring

together P(AB) = P(A)xP(B).

Probability and Counting Theorypage 22

 Definition: Two events A and B are independent if the probability
of one event is unaffected by the occurrence of the other.

(Ex) Take the experiment of tossing a coin twice. Let A and B be

the events of getting heads on the first and second tosses,
respectively. The two events A and B are independent because
the second toss result will not be affected by the result of the
first toss. Note that P(A)=1/2, and P(B)=1/2. and since the
result of this experiment consists of four equally probable
results: (HH, HT, TH, TT), we can see that P(AB)=P(HH) =
¼1/4.

But since P(A)xP(B) is also 0.25¼ we can say that events A and B
are independent.

(Ex) 65% of the Philippine population is female. 85% of the

Philippine population live near or with their families. If it can be
assumed that being male or female does not change the
probability that a Filipino would live near or with their families,
what is the probability that a randomly selected Filipino is
female and lives with their family? Ans: 0.65 x
0.85 = 0.5525

This situation may be shown as a tree diagram: Notice that if 2

events A and B are independent, then all branches of a certain
event B does not change with respect to varying conditions of
event A, and vice-versa.

85% Live with family

Female
15%
Not Live with family
65%

35% 85% Live with family

Male
15%
Not Live with family
This same tree diagram of independent events [(Gender) and
(Family)] may be redrawn such that the first branch is (Live with
family)-(Not live with family) and the subbranches are Male-

Probability and Counting Theorypage 23

 Female, Male-Female. The probabilities for each event branch
would be correspondingly moved.

(Ex) Let’s say an NBA player has a 85% freethrow percentage. If

you bet that he will miss his first freethrow for two different
randomly selected games . What is the probability that you
would lose your bet?

You will win your bet if the player misses both his freethrows. If
he makes at least one, then you lose. We can simply compute
for the P(miss both freethrows), which is the same as P(win bet)
and find its complement to get P(lose bet). I.e. P(lose bet) = 1-
P(win bet)

Since the freethrows were made at two different games, the

player would be presumably unaffected by the knowledge of
either freethrow being made or not. We may consider each
freethrow as independent, with each having a probability of
85% of being made.

P(win bet) = P(misses both freethrows).

= P(miss 1st) x P(miss 2nd)

P(1st is missed)= 1- P(first is made) P(2nd is

missed)=1-P(2nd is made)
= 1-0.85 = 1-0.85
= 0.15 = 0.15

P(win bet)= 0.15 x 0.15

= 0.0225 Therefore, P(lose bet) = 1-0.0225
= 0.9775

(Ex) A man has two kidneys. He only needs one functioning kidney
to live. It has been estimated that during his lifetime, his left
kidney can be reliable with a 95% probability, his right kidney
is slightly stronger with a 98% probability. What is the
probability that the man dies of kidney failure?

Ans: P(die of kidney failure)= P(both kidneys fail)

= (1-0.95) x (1-0.98)
= 0.001 or 1 in 1000 chance.

(Ex) A man and a woman decide to meet at a bar after work

anywhere between 5 to 6 pm, agreeing to wait no longer than

Probability and Counting Theorypage 24

 10 minutes. If they will randomly arrive anywhere between 5
to 5:50 pm (to clear 10 minutes til 6pm), what is the probability
that they will meet?

P(man in bar during 5-6 pm) = 10/60

P(woman in bar during 5-6 pm) = 10/60
P(man and woman coincide) = 10/60 x 10/60 = 1/36

P(they meet) = P(union of man time in bar and woman time in bar)
= 1/6 + 1/6 –1/36 = 11/36.

(Ex) Consider a bridge whose components have a probability of

failure.

A bridge is constructed such that there are 2 cantilever bars and a

middle part connected by two joints (see figure) The cantilever
beams have been found to fail with a probability of 2% if a load above
its design strength is applied. The middle portion is known to fail 2%
of the time under overloads. The two joints fail 5% of the time when
overloaded. Suppose that overload occurs, what is the reliability of
the bridge ?

A x B y C

2% 5% 2% 5% 2%

The components are arranged in a series such that if any member fails,
the whole system fails.

Ans: Reliability = (Reliability of A) (Reliability of B) (Reliability of C)

(Reliability of x) (Reliability of y)
where:
Reliability of any member= 1- P(failure of member)

Reliabilty = 0.98 x 0.98 x 0.98 x 0.95 x 0.95

= 0.8494

Contingency Table: A useful tool to present probabilities

for two events

Events A A’ Probability

Probability and Counting Theorypage 25

 B P(AB) P(A’B) P(B)
B’ P(AB’) P(A’B’) P(B’)
Probability P(A) P(A’) 1.00

Row totals and column totals are the probabilities of each event.
All probabilities must sum to unity.

(Ex)
40% of DLSU Engineering students are ECE majors. 70% of all
Engineering majors are Male. Male ECE majors comprise 25% of all
engineering students. Set up the contingency table for the events
(ECE/Non-ECE), and (male/female).

Given :
Events ECE Non-ECE Probability
Male 0.25 Not given 0.70
Female Not given Not given Not given
Probability 0.40 Not given 1.00

One can use relationships between the unknown variables and

their total probabilities to derive the other figures. Take the
column totals for ECE/non-ECE events (the last row).

Events ECE Non-ECE Probability

Probability 0.40 Not given 1.00

Using the definition of Complimentary events: P(Non-ECE) = 1-

P(ECE)
= 1.00-0.40
=0.60
We could now update the last row:
Events ECE Non-ECE Probability
Male 0.25 Not given 0.70
Female Not given Not given Not given
Probability 0.40 0.60 1.00

Probability and Counting Theorypage 26

 We could do the same for the last row: P(female)=1-P(male)
= 0.30
Events ECE Non-ECE Probability
Male 0.25 Not given 0.70
Female Not given Not given 0.30
Probability 0.40 0.60 1.00

Take the first row:

Events ECE Non-ECE Probability

Male 0.25 Not given 0.70

If the total probability of men is 0.70, of which 0.25 are ECE, then
the difference of 0.70-0.25=0.45 must be the male non-ECE
proportion.

We could now fill up the table as:

Events ECE Non-ECE Probability
Male 0.25 0.45 0.70
Female Not given Not given 0.30
Probability 0.40 0.60 1.00

The rest of the table should be quite easily filled out.

Events ECE Non-ECE Probability
Male 0.25 0.45 0.70
Female 0.15 0.15 0.30
Probability 0.40 0.60 1.00

Conditional Probabilities :

If events A and B are not independent, then we say they are

dependent. This happens when one event affects the probability of the
other. The probability of a later event B is affected by the knowledge
of a prior event A, and hence we call the later event having a
conditional probability—its probability value is conditional/contingent
on the prior event. We use the symbol P(B|A) with a standing slash
symbol (“|”) to represent the conditional probability of B given that A
has occurred. Let’s take a simple example.

Probability and Counting Theorypage 27

Illustrative Example:
A box contains slips of papers where each slip has been labelled 1,
2, 3 ,4 and 5. An experiment is done such that a first slip is picked
from the box, its label number is recorded and then a second slip is
picked and recorded. Presumably, the first slip was not returned to
the box before the second slip was chosen. Let us define event A
that an odd number was chosen in the first pick, and define event B
that an odd number was chosen on the second pick. We can create
the following event tree and its corresponding probabilities.

B
Odd
P(AB)=P(A)xP(B|A)
=3/10
P(B|A)=2/4
A
Odd

P(A)=3/5 P(B'|A)=2/4
B' P(AB')=P(A)xP(B'|A)
Even
=3/10
B
Odd P(A'B)=P(A)xP(B'|A)
P(A')=2/5
P(B|A')=3/4 =3/10
A'
Even
P(B'|A')=1/4
B'
Even
P(A'B')=P(A)xP(B'|A)
=1/10

How the probabilities (on branch stems) were arrived:

P(A) = P(odd for the first try) = Initially there are 5 possible slips to
choose from, of which 3 of them are odd: #1, #3 and #5. = 3/5
P(A’)= P(not odd) = P(Even) = From the 5 initial slips, there are 2
which are even: #2 and #4. = 2/5

P(B|A) = read as “the probability of event B given A” = This is the

conditional probability of event B (having the second slip turn out
odd) if the first event A is known to have occurred. This means
that event A (first odd) has occurred, and that the sample space
for event B has now become smaller: that is, there are now only
(5-1)=4 slips left of which (3-1)=2 of them are odd. There are
only 2 odd slips left because one odd number had already been
chosen in the prior event A. = 2/4 = 0.5

P(B’|A) = read as “the probability of not B given A” = this is the

probability of having an even (not odd) number on the second

Probability and Counting Theorypage 28

 slip (event B’) given that the first event had been an odd number
(event A). P(B’|A) is the complement of P(B|A), therefore, P(B’|
A)=1-P(B|A) = 1-0.5=0.5

P(B|A’) = read as “the probability of B given (not A)”= When the first
event is A’ (even), then there will still be 3 even numbers for the
2nd slip. There will be a total of only 4 slips left. Therefore
getting an even number on the second try should now be 3/4.

P(B’|A’) = should be the complement of P(B|A’)= 1 - 3/4 = 1/4.

The probability of the compound event (AB) = “both numbers are

odd”
= P(AB) = P(A)P(B|A)
= (3/5)x(2/4)
= 3/10

We may relate the event tree for this illustrative example and
show how it could be represented by a contingency table:

Let A=first slip is odd,

B= second slip is odd.

Events A A’ Probability
B P(AB) P(A’B) P(B)
B’ P(AB’) P(A’B’) P(B’)
Probability P(A) P(A’) 1.00

Probability and Counting Theorypage 29

 B
Odd
P(AB)=P(A)xP(B|A)
=3/10
P(B|A)=2/4
A
Odd

P(A)=3/5 P(B'|A)=2/4
B' P(AB')=P(A)xP(B'|A)
Even
=3/10
B
Odd P(A'B)=P(A)xP(B'|A)
P(A')=2/5
P(B|A')=3/4 =3/10
A'
Even
P(B'|A')=1/4
B'
Even
P(A'B')=P(A)xP(B'|A)
=1/10
Fill out the entries at the branch ends at the right
Events A A’ Probability
B 3/10 3/10 ??
B’ 3/10 1/10 ??
Probability ?? ?? 1.00

And then sum up the probabilities to get the individual probabilities of

A and B.

Events A A’ Probability
B 3/10 3/10 6/10
B’ 3/10 1/10 4/10
Probability 6/10 4/10 1.00

Note that P(AB) cannot be P(A)xP(B) when P(A)=6/10 and P(B)=6/10.

We know that P(AB)=3/10 (from the event tree) and is not the
same as the product 6/10 x 6/10 = 36/100  3/10. This means that
events A and B are not independent.

This example has also shown that a 2-stage binary Event tree
can be related to one and exactly one Contingency table.

Working with equations of Conditional Probability:

Probability and Counting Theorypage 30

It is also noteworthy to notice that the formula for conditional
probability of B given A is

P( A  B )
P ( B | A) 
P ( A)
This formula suggests that to get the conditional probability of B given
that event A is known to have occurred is simply the ratio of the
probabilities of event (AB) over the probability of event A.

On the contingency table, we can see that P(B|A) can be derived as

P(AB) / P(A) or seen as

Events A A’ Probability
B 3/10 3/10 6/10
B’ 3/10 1/10 4/10
Probability 6/10 4/10 1.00

3 / 10
P(B|A)=  0.5
6 / 10

Which should be the same probability as in the event tree (see branch
A (odd) then B (odd again) at the right-hand topmost branch of
the event tree).

Let’s use the symbols for conditional probability to solve a numeric

problem.

Given: P(A)=0.75 Find: P(AB)

P(A|B)=0.80 P(A’B’)
P(B|A) = 0.60 P(B)

P( A  B)
Since P ( B | A)  then P ( A  B )  P( A)  P( B | A)
P ( A)
= 0.75 x 0.60
=0.45

P ( A  B) P( A  B)
and since P ( A | B )  then P( B) 
P( B ) P( A | B)
0.45

0.80
=0.5625

We could now fill up the contingency table:

Probability and Counting Theorypage 31

 Events A A’ Probability
B 0.45 ?? 0.5625
B’ ?? P(A’B’)= ??
?
Probability 0.75 ?? 1.00

Since we only need 3 entries to fill up the contingency table, we could

now fill and find the last probability P(A’B’) like thus:

Events A A’ Probability
B 0.45 0.1125 0.5625
B’ 0.30 0.1375 0.4375
Probability 0.75 0.25 1.00

So the answers are:

P(AB) = 0.45
P(A’B’)= 0.1375
P(B) = 0.5625

Bayes’ Theorem:

Let us take the Illustrative Example on conditional probability again,

reproduced here: and put a twist to it.

A box contains slips of papers where each slip has been labelled 1,
2, 3 ,4 and 5. An experiment is done such that a first slip is picked
from the box, its label number is recorded and then a second slip is
picked and recorded. Let us define event A that an odd number was
P(AB)=P(A)xP(B|A)
B
Odd
chosen in
=3/10 the first
P(B|A)=2/4
pick, and
A
Odd define
P(B'|A)=2/4 event B
P(A)=3/5
B' P(AB')=P(A)xP(B'|A) that an odd
Even
=3/10 number
B
P(A'B)=P(A)xP(B'|A)
was chosen
Odd
P(A')=2/5
P(B|A')=3/4 =3/10 on the
A'
Even
Probability and Counting
P(B'|A')=1/4 Theorypage 32
B'
Even
P(A'B')=P(A)xP(B'|A)
=1/10
 second pick. We can create the following event tree and its
corresponding probabilities.

Suppose that we want to find the probability of event B’,

that the second slip chosen was even. We know from the
contingency table that P(B’) = 4/10. (page 28)

On the event tree, there are two final right-side branches where
B
Odd

A
Odd
B'
Even

B
Odd
A'
Even
B'
Even

B’ occurs. (darkened paths shows the two cases when B’ occurs)

We can also see that P(B’) = P(B’A) + P(B’A’)

= P(AB’) + P(A’B’)
= P(A)P(B’|A) + P(A’)P(B’|A’)
{from event tree branches’ probabilities}
= 3/5 x 1/2 + 2/5 x 1/4¼
= 3/10 + 1/10 Ans = 0.4
Now, what if we already know that the second slip was even, and
we want to find a prior event’s probability. A question such as the
following may be asked:

P(A|B’)=? What is the probability that the first slip was

odd
if it is known that the second slip was even?

This can be symbolized by P(A|B’). This is of course different

from P(B’|A). In compound events like in P(AB), we could interchange
the two events’ symbols and still have the same probability [i.e.
P(AB)=P(BA)] . This is because the compound event already has
two events happening together, and so it doesn’t matter in what order

Probability and Counting Theorypage 33

they are written. In contrast, P(A|B’) refers to B’ having been known
to have happened. So event B’ is known, and is different from having
A known, as in P(B’|A).

To find P(A|B’), we make use of the formula definition of

conditional probability:

P( A  B' )
P ( A | B' ) 
P( B' )

The numerator P(AB) is equal to the probability of the darkened

path on the top in the event tree where B’ occurs:

B
Odd

A
Odd
B'
Even

B
Odd
A'
Even
B'
Even

The denominator P(B’) is the sum of the two branches’

probabilities (computed earlier, page 30) = P(B’) = P(B’A) +
P(B’A’) =0.4.

We can rewrite the equation:

P ( A) P( B '| A)
P( A | B ' ) 
P ( A) P ( B '| A)  P( A' ) P ( B'| A' )
3 1
  
P( A | B' )  5  2  0.75
3  1  2 1
      
5  2 5  4

We can do away with the complement symbols on B’ and find

that the equation can also hold for P(A|B).

P( A) P ( B | A)
P( A | B)  Bayes’ Formula
P ( A) P ( B | A)  P ( A) P ( B | A)

for finding prior event A given that later event B is known.

Probability and Counting Theorypage 34

 a
This formula is easily like of this form: P ( A | B) 
ab
where a is simply one of the items in the sum (a+b) in the
denominator.

It can also be said that to get the results of Baye’s formula, a simple
cursory look at the contingency table would have yielded an easier
result: to wit:
P ( A  B ' ) 0. 3
P( A | B' )    0.75 This does not mean, however, that
P ( B' ) 0. 4
the contingency table is a tool to be used for all probability problems.
The contingency table is limited to only two binary-result events at a
time. It cannot be used if the true probability of each event (A and B)
is not known or derivable.

Baye’s Theorem is stated in probabilities, but the following example

can show that numeric representations (i.e. actual counts) can also
hold for Bayes’ formula:

Example: A pregnancy test will indicate “positive” 95% of the time if

the woman being tested truly is pregnant. It will also indicate
that “negative” 99% if she is not pregnant. These numbers
provide a margin of error of 5% for sensitivity to actual
pregnancy, and a margin of error of 1% for specificity of a true
non-pregnancy. Suppose that a woman takes the test
and got a “positive” result (meaning: she must be
pregnant), what is the probability that she actually
doesn’t have a baby? (a false alarm) For practical
purposes, suppose that at any time, 2% of the females in the
population is truly pregnant.

Illustration:
Let’s take 10,000 women as a sample. Exactly 200 out of
them should be pregnant (2% of 10,000), while 9,800 of them
should not. If the 200 pregnant women would take the
pregnancy test, 190 of the tests should come out “positive”
(95% of 200). Of the 9,800 non-pregnant women who take
the pregnancy test, about 9,702 of them would have a
“negative” result. (99% of 9800), which also means that the
other 98 non-pregnant women (1% of 9800) would have a
“positive” result.

Probability and Counting Theorypage 35

 190 "Positive"
95%
200 Pregnant
2% 5%
10 "Negative"

10,000
women
98 "Positive" (False Alarm)
98% 1%

9800 not Preg

99%
9,702 "Negative"
Using numerical values on final right-side tally.
P(False Alarm) = P(Not Pregnant|“Positive”) =
P( Not pregnant " Positive" )
P (" Positive" )
Women having False " Positive"
= Total women getting " Positive"
98
=  34.03%
190  98

Positive=
95% "Pregnant"
Pregnant
2% 5% Negative=
"Not"

98% 1% Positive
Not "Pregnant"
Pregnant
99% Negative=
"Not"

Using Bayes’ Formula =

P(False Alarm) = P(Not Pregnant|“Positive”) =
P( Not pregnant " Positive" )
P (" Positive" )
P( not Pr eg )  P (" Positive"| not Pr eg )
=
P ( preg ) P(" Positive"| Pr eg )  P ( not Pr eg ) P (" Positive"| not Pr eg )

(0.98)(0.01)
=  34.03%
(0.02)(0.95)  (0.98)(0.01)

Probability and Counting Theorypage 36

Ans: 34.03% chance that a false alarm has occurred given that the
pregnancy test concludes“Positive”.

Note that the absolute probability of having a false “positive” is only

98/10,000 = 0.0098 or about 1 in every 100 test attempts (or
alternately derived using probability figures: 0.98 x 0.01 = 0.0098
=approx. 1%) When a later (posterior) event like “positive” has
occurred, then the probability of a false alarm jumps up to 34.02%.
This number is called a relative probability. The probability is
relative to the sample of “positive” results only, and does not take
into account all absolute counts of both “positive” and “negative”
results.
When arguing about probabilities, take relative probabilities with some
reservations. Always use absolute probabilities to see clearly the
possibilities of events.
Relative probabilities (like the result of Bayes’ formula) assume
something has happened, and it is what it is--an assumption--
which cannot be proven. Only when some posterior event has
actually occurred, can we use Bayes’ formula to show the relative
possibility of a previous event. Bayes’ Formula is a “detectives’
historical digging tool”.
It can also be a way to scare the numerically illiterate.

Practice Exercises: (Probability Laws, Conditional Probabilities and Bayes’

Theorem)

1. Suppose that out of a total group of 150 athletes surveyed, it is

found that
90 take Vitamin A 53 take both A and B
88 take Vitamin B 55 take both A and C
97 take Vitamin C 57 take both B and C
and 32 take all three.
Question:
How many take none of the three vitamins?

2. In a survey of 30 consumers of detergents, 5 people said they

are loyal to brand A, while another 5 are loyal to brand B. 7
surveyed said they do not use either brands A, B nor C, but use
some other detergent not included in the survey. A total of 10
consumers said they prefer brand A, and a total of 11
consumers said they prefer brand C. There were no one who
used both brands B and C. Further, almost peculiarly, no one

Probability and Counting Theorypage 37

 said they used A, B and C simultaneously. Find the number of
consumers who:
a. Preferred brands A and B.
b. Preferred brands A and C.
c. Preferred brands A or B.
d. Preferred brands A or C.

3. There are 5 candidates for the EVP position and 6 candidates for
the VP-Academics position in a certain student organization. 5
candidates are listed in the VP-Activities, but closer inspection
shows that there are 2 individuals listed as candidates in all three
positions. Three individuals are running for both VP Activities
but not for EVP. Two candidates are vying for the EVP position
alone. What is the total number of individuals are running for
the officerships?
4. Suppose that the probability that Seven is happy is 0.53, the
probability that Seven is tense is 0.72, and the probability that
she is both happy and tense is 0.48. Calculate :
a. the conditional probability that Seven is happy given that
she is tense
b. the conditional probability that Seven is tense, given that
she is happy.
5. If P(A)=0.75 and P(B=0.8) and P(A’B’)=0.05, are events A and B
independent?
6. Given P(A)=0.25, P(B’)=0.85 and P(A|B)=2/3. Find
a. P(AB) d. P(B|A)
b. P(AB) e. P(B|A’)
c. P(A’B’) f. P(AB’)
7. Two letters, not necessarily distinct, are typed at random. What
is the probability that they are both vowels, given
a. No information at all
b. That they are different
c. That at least one is a vowel.

8. A fair coin is tossed three times. WITPT (what is the probability

that) it comes up heads at least once, given
a. No information at all
b. That all three tosses come up the same?
c. That it comes up heads at most once?
9. A lie detector test given to a suspect is known to be 90% reliable
when the person is guilty and 99% reliable when the person is

Probability and Counting Theorypage 38

 innocent. In other words, 10% of the guilty are judge innocent by
the test and 1% of the innocent are judged guilty. A suspect was
selected from a group of suspects, of which, only 5% have
committed a crime. The test indicates that he is guilty, what is
the probability that he is innocent?
10. A large industrial firm uses 3 local motels to provide
overnight accommodations for its client. From past experience, it
is known that 20% of the clients are assigned rooms at the
Ramada Inn, 50% at the Sheraton, and 30% at the Lake View
Motor Lodge. The plumbing is faulty in 5% of the rooms at the
Ramada Inn, 4% in the Sheraton, and 8% in the Lake View Motor
Lodge. What is the probability that,
a. a client will be assigned in a room with faulty plumbing?
b. a person with a room having faulty plumbing was assigned
accommodations at the Lake View Motor Lodge?

Probability and Counting Theorypage 39