MENU

Learn P 1 1.2 1.2.1 Conditional Probability

Conditional Probability

Conditional probability is the probability that an event occurs based on a condition, or

given that another event has occurred. In other words, conditional probability is the
probability of an event, assuming the conditional event is treated as the sample space.

In probability notation, the probability that event A occurs given event B has occurred
is Pr(A ∣ B) . The formula for the conditional probability of A given B is

Pr(A ∩ B)
Pr(A ∣ B) = (1.2.1)
Pr(B)

for Pr(B) ≠ 0 . Similarly, the formula for the conditional probability of A

′
given B is

′
Pr(A ∩ B)
′
Pr(A ∣ B) = (1.2.2)
Pr(B)

In addition,

′
Pr(A ∩ B) Pr(A ∩ B)
′
Pr(A ∣ B) + Pr(A ∣ B) = +
Pr(B) Pr(B)
′
Pr(A ∩ B) + Pr(A ∩ B)
= (1.2.3)
Pr(B)

From the Law of Total Probability, Pr(A ∩ B) + Pr(A ∩ B)

′
equals Pr(B). Therefore,

Pr(B)
′
Pr(A ∣ B) + Pr(A ∣ B) = = 1 (1.2.4)
Pr(B)

It is easiest and most intuitive to explain the formulas for Pr(A ∣ B) and Pr(A′ ∣ B)
visually. Consider the following Venn diagram (Figure 1.2.1) containing events A and B.
Given the fact that event B has occurred, logic dictates it is impossible for an event to
occur outside of B. Thus, given B occurs, the sample space is reduced to event B. Of
course, it is not technically a real sample space because Pr(B) may not be 1.

Figure 1.2.1

Within this new sample space, the region where event A occurs is A ∩ B , shown in
green. The conditional probability of A given B is the probability that A occurs within
the domain of event B. So, the conditional probability is the ratio of Pr(A ∩ B) to
Pr(B).

Likewise, the region where A

′
occurs within the confines of B is A ∩ B,
′
shown in blue.
Thus, the conditional probability of A given B is the ratio of Pr(A ∩ B) to Pr(B). The
′ ′

ratio for a conditional probability should be intuitive rather than memorized. If it is not
intuitive yet, review the following examples until the ratio for the conditional
probability is clear.

Rearranging Equation 1.2.1 produces

Pr(A ∩ B) = Pr(A ∣ B) ⋅ Pr(B) (1.2.5)

Reverse A and B to create a formula conditioned on event A.

 Pr(B ∩ A) = Pr(B ∣ A) ⋅ Pr(A) (1.2.6)

Equation 1.2.5 and Equation 1.2.6 are known as the Multiplicative Law of Probability.
Since Pr(A ∩ B) = Pr(B ∩ A) ,

Pr(A ∣ B) ⋅ Pr(B) = Pr(B ∣ A) ⋅ Pr(A) (1.2.7)

These formulas are crucial to know, and will be used extensively in later sections.

EXAMPLE 1.2.1

An actuary rolls two six-sided fair dice. The two numbers that appear on the top face
of the dice are observed.

Calculate the probability that the product of the two numbers is odd, given the sum of
the two numbers is between six and eight inclusive.

SOLUTION

In order to illustrate that conditioning on an event redefines the sample space, we will
work this example using two methods. First, define A as the event that the product is
odd. Then, define B as the event that the sum is between six and eight inclusive.

Begin with the sample space for rolling two dice, as shown in Table 1.2.1. There are six
numbers on each die, making 6
2
= 36 unique elements. The first number in each cell is
the outcome of the first die, while the second number is the outcome of the second
die. The product of two numbers is odd if both numbers are odd. The shaded cells
indicate where the product is odd.
 Table 1.2.1

If the question were to ask for the probability that the product is odd with no
conditions, then the final answer would simply be Pr(A) =
9

36
=
1
.
4

However, because the question asks for the probability that the product is odd, given
that the sum is between six and eight inclusive, the sample space changes. The
conditional event gives information that reduces the number of possible outcomes.
The number of possible outcomes is 36. Now, it reduces to the number of cells where
the sum of the two numbers equals 6, 7, or 8. There are 16 of these cells, highlighted in
Table 1.2.2.
 Table 1.2.2

To find Pr(A ∣ B) , divide the number of cells that satisfy event A and are in event B
by the number of cells in event B. The highlighted cells in Table 1.2.1 and Table 1.2.2
only intersect at five cells (shown in black in Table 1.2.3).

Table 1.2.3

Therefore,

5
Pr(A ∣ B) =
16

ALTERNATIVE SOLUTION

The formula for Pr(A ∣ B) is

Pr(A ∩ B)
Pr(A ∣ B) =
Pr(B)
The event that A and B occur simultaneously is the event where the product is odd
and the sum is between six and eight inclusive. The product of two numbers is odd
only if both numbers are odd. In addition, the sum of the numbers must equal 6, 7, or
8. Only (1, 5), (3, 3), (3, 5), (5, 1), and (5, 3) satisfy these conditions. Therefore,

5
Pr(A ∩ B) =
36

There are a total of 16 combinations of two numbers where the sum is 6, 7, or 8. They
are (1, 5) through (5, 1) containing five combinations, (1, 6) through (6, 1) containing six
combinations, and (2, 6) through (6, 2) containing five combinations. Therefore,

16
Pr(B) =
36

Finally,

5/36
Pr(A ∣ B) =
16/36

5
=
16

EXAMPLE 1.2.2

We draw two cards from a standard deck of 52 cards, without replacement.

Calculate the probability that both cards are kings.

SOLUTION
"Without replacement" indicates that when a card is drawn from the deck of 52 cards,
it is not placed back in the deck. So after one card is drawn, there are 51 cards
remaining. This concept contrasts with "with replacement," which indicates that when
a card is drawn from the deck of 52 cards, it is placed back in the deck.

Let K1 be the event that the first card drawn is a king, and K2 be the event that the
second card is another king. Then, the probability of drawing two kings is
Pr(K1 ∩ K2 ) .

According to Equation 1.2.5,

Pr(K1 ∩ K2 ) = Pr(K2 ∣ K1 ) ⋅ Pr(K1 )

There are a total of four kings in a standard deck of 52 cards.

Therefore,

4 1
Pr(K1 ) = =
52 13

After drawing one card without replacement, there are 51 cards remaining. If the first
draw is a king, there are three kings remaining. Given the first card is a king, the
probability that the second card is also a king is

3 1
Pr(K2 ∣ K1 ) = =
51 17

To complete this problem, calculate the probability both cards are kings as:

Pr(K1 ∩ K2 ) = Pr(K1 ) ⋅ Pr(K2 ∣ K1 )

1 1
= ⋅
13 17
1
=
221

ALTERNATIVE SOLUTION
For a visual approach, analyze the probabilities using a tree diagram.

Figure 1.2.2

There are two outcomes on the first draw: a king or not a king. The probability of a
king on the first draw is 4 out of 52. The probability of not a king on the first draw is
the complement, or 48 out of 52.

There are two outcomes on the second draw for each outcome on the first draw, for a
total of four outcomes after two draws. After a king on the first draw, 51 cards remain,
consisting of 3 kings and 48 non-kings. Therefore, the probability of a king on the
second draw after a king on the first draw is 3 out of 51. Likewise, the probability of
not a king on the second draw after a king on the first draw is 48 out of 51.

The same technique is used to assign probabilities to outcomes of the second draw
after not drawing a king on the first draw (see Figure 1.2.2). However, these
probabilities are not needed to calculate the final answer.

The probability that both draws are kings equals the probability of K1 then K2 .

Multiply the probabilities on the K1 − K2 path on the tree diagram.

4 3
Pr(K1 ∩ K2 ) = ⋅
52 51
1
=
221

■
EXAMPLE 1.2.3

An urn contains ten red balls and fifteen blue balls. Two balls are drawn at random and
without replacement.

Calculate the probability that the two balls are of different colors.

SOLUTION

Let R be the event of drawing a red ball, and B be the event of drawing a blue ball.
Drawing two balls of different colors means drawing a red ball first, then a blue ball; or
drawing a blue ball first, then a red ball. Thus,

Pr(different colors) = Pr(R first ∩ B second) + Pr(B first ∩ R second)

= Pr(R first) ⋅ Pr(B second ∣ R first)

+ Pr(B first) ⋅ Pr(R second ∣ B first)

There are 10 red balls and 15 blue balls in the urn, so

10 10
Pr (R first) = =
10 + 15 25

15 15
Pr (B first) = =
10 + 15 25

Since drawing is without replacement, the total number of balls decreases by one after
the first draw. Thus,

15 15
Pr(B second ∣ R first) = =
9 + 15 24
 10 10
Pr(R second ∣ B first) = =
10 + 14 24

Finally,

10 15 15 10
Pr(different colors) = ⋅ + ⋅
25 24 25 24
1
=
2

ALTERNATIVE SOLUTION

Figure 1.2.3

There are two outcomes on the first draw: red or blue. There are 25 balls in total. The
probability of drawing a red ball on the first draw is 10 out of 25. The probability of
drawing a blue ball on the first draw is 15 out of 25.

There are two outcomes on the second draw for each outcome on the first draw, for a
total of four outcomes after two draws. After drawing a red ball on the first draw, 24
balls remain, consisting of 9 red and 15 blue. Therefore, the probability of drawing a
red ball on the second draw after drawing a red ball on the first draw is 9 out of 24.
Likewise, the probability of drawing a blue ball on the second draw after drawing a red
ball on the first draw is 15 out of 24.

The same technique is used to assign probabilities to outcomes of the second draw
after drawing a blue ball on the first draw (see Figure 1.2.3).

The probability that the two balls are of different colors is the sum of the probability
of R then B and the probability of B then R. Add the products of the probabilities on
the R−B path and the B−R path.

10 15 15 10
Pr(different colors) = ⋅ + ⋅
25 24 25 24
1
=
2

EXAMPLE 1.2.4

An insurance company estimates their policyholders as follows:

1. 30% of policyholders who only have a life insurance policy will renew next year.

2. 50% of policyholders who only have a health insurance policy will renew next
year.

3. 70% of policyholders who have a life insurance and a health insurance policy will
renew at least one of those policies next year.

The company's data show that

a. 40% of policyholders have a life insurance policy.

b. 45% of policyholders have a health insurance policy.

c. 25% of policyholders have a life insurance and a health insurance policy.

Using the company's estimates, calculate the percentage of policyholders that will
renew a life insurance or health insurance policy next year.

SOLUTION

Let L and H denote having a life insurance policy and having a health insurance
policy, respectively. Moreover, let A be the event that a policyholder will renew a life
insurance or health insurance policy next year. From statements a, b, and c,

Pr(L) = 0.4

Pr(H ) = 0.45

Pr(L ∩ H ) = 0.25

Figure 1.2.4 is the Venn diagram showing the relationship between L and H. Use the
Law of Total Probability to calculate Pr(L ∩ H ′ ) and Pr(L′ ∩ H ).

′
Pr(L) = Pr(L ∩ H ) + Pr(L ∩ H )

′
⇒ 0.4 = 0.25 + Pr(L ∩ H )
′
⇒ Pr(L ∩ H ) = 0.15

′
Pr(H ) = Pr(H ∩ L) + Pr(H ∩ L )

′
⇒ 0.45 = 0.25 + Pr(H ∩ L )
′
⇒ Pr(H ∩ L ) = 0.20
 Figure 1.2.4

From statements 1, 2, and 3,

′
Pr(A ∣ L ∩ H ) = 0.3
′
Pr(A ∣ L ∩ H ) = 0.5

Pr(A ∣ L ∩ H ) = 0.7

The objective is to calculate Pr(A), the probability a policyholder will renew a life
insurance or health insurance policy next year. This is equivalent to the percentage of
policyholders who will renew a life insurance or health insurance policy next year.

Event Acan only occur if the policyholder has at least one of those two policies, i.e.
L ∪ H . Therefore, A has to be within L ∪ H (see Figure 1.2.5).
 Figure 1.2.5

Thus, Pr(A) can be calculated using the Law of Total Probability.

′ ′
Pr(A) = Pr[A ∩ (L ∩ H )] + Pr[A ∩ (L ∩ H )] + Pr[A ∩ (L ∩ H )]

Since the conditional probabilities of A are given, find the probabilities of the three
events:

′ ′ ′
Pr[A ∩ (L ∩ H )] = Pr(A ∣ L ∩ H ) ⋅ Pr(L ∩ H )

= 0.3 ⋅ 0.15

= 0.045

′ ′ ′
Pr[A ∩ (L ∩ H )] = Pr(A ∣ L ∩ H ) ⋅ Pr(H ∩ L )

= 0.5 ⋅ 0.2

= 0.100

Pr[A ∩ (L ∩ H )] = Pr(A ∣ L ∩ H ) ⋅ Pr(L ∩ H )

= 0.7 ⋅ 0.25

= 0.175
In the final step, sum the three parts to calculate Pr(A).

Pr(A) = 0.045 + 0.100 + 0.175

= 0.32

EXAMPLE 1.2.5

Five cards are drawn without replacement from a standard deck of 52 cards consisting
of four suits of thirteen cards each.

Calculate the probability that the five cards result in a flush (all five cards are of the
same suit).

SOLUTION

Let F be the event of a flush. There are four different suits (spades, hearts, clubs, and
diamonds). Let Si , Hi , Ci , and Di be the events that the i
th
card is a spade, heart,
club or diamond, respectively.

The probability of a flush is the sum of the probability of a flush in one of the
individual suits.

Pr(F ) = Pr(S1 ∩ S2 ∩ S3 ∩ S4 ∩ S5 ) + Pr(H 1 ∩ H 2 ∩ H 3 ∩ H 4 ∩ H 5 )

+ Pr(C1 ∩ C2 ∩ C3 ∩ C4 ∩ C5 ) + Pr(D 1 ∩ D 2 ∩ D 3 ∩ D 4 ∩ D 5 )

All four summands in the previous equation are numerically equal. For example, it is
equally likely to draw a flush of hearts as it is to draw a flush of diamonds. This is
because there are equal numbers of cards for each suit. Thus,

Pr(F ) = 4 ⋅ Pr(S1 ∩ S2 ∩ S3 ∩ S4 ∩ S5 )
The probability of a flush with spades (or any other suit) is

13 12 11 10 9
Pr(S1 ∩ S2 ∩ S3 ∩ S4 ∩ S5 ) = ⋅ ⋅ ⋅ ⋅
52 51 50 49 48

≈ 0.0004952

Because drawing is without replacement, the total number of cards and the number of
spades decreases by one after every draw.

Therefore, the desired probability is four times the probability of a flush for one of the
suits.

Pr(F ) = 4 ⋅ 0.0004952

≈ 0.001981

EXAMPLE 1.2.6

A life insurance company's policyholder cohort is defined as:

1. Each policyholder is classified as either male or female, and either a smoker or a

non-smoker.

2. There are 200 more male policyholders than female.

3. There are 50 more male non-smoker policyholders than female non-smoker

policyholders.

4. The total number of smoker policyholders is 250.

Calculate the number of policyholders who are male smokers.

SOLUTION
From statement 1, create Table 1.2.4 to classify the policyholders.

Table 1.2.4

Using statements 2, 3, and 4, insert the information given into Table 1.2.4.

Table 1.2.5

From statement 2, if X is the number of females, then X + 200 is the number of males.

From statement 3, if Y is the number of female non-smokers, then Y + 50 is the

number of male non-smokers. Therefore, the number of female smokers is X −Y .

From statement 4, the total number of smokers is 250 .

There are two ways to express the number of male smokers in terms of X and Y .
Either subtract X − Y from 250 or subtract Y + 50 from X + 200. Thus, the number of
male smokers is 250 − (X − Y ) or (X + 200) − (Y + 50).

Next, set these expressions equal to each other to solve for X −Y .

 250 − (X − Y ) = (X + 200) − (Y + 50)

⇒ 250 − (X − Y ) = (X − Y ) + 150

250 − 150
⇒ X −Y = = 50
2

Conclude that the number of male smokers is 250 − 50 = 200 .

EXAMPLE 1.2.7

A doctor is studying the relationship between blood pressure and heartbeat

abnormalities in her patients. She tests a random sample of her patients and notes
their blood pressures (high, low, or normal) and their heartbeats (regular or irregular).
She discovers that:

1. 14% have high blood pressure.

2. 22% have low blood pressure.

3. 15% have an irregular heartbeat.

4. Of those with an irregular heartbeat, one-third have high blood pressure.

5. Of those with normal blood pressure, one-eighth have an irregular heartbeat.

Determine the percentage of the patients that have

a. a regular heartbeat and high blood pressure.

b. a regular heartbeat and normal blood pressure.

c. an irregular heartbeat and low blood pressure.

SOLUTION
This example is an extension of Example 1.2.6. Begin with a table similar to Table 1.2.4.

From statements 1, 2, and 3, update the table as follows:

Table 1.2.6

Because proportions are calculated, the total must sum to 1.00. So, the percentage of
patients with normal blood pressure is 1 − 0.22 − 0.14 = 0.64 (expressed as a
decimal). In addition, the percentage of patients with a regular heartbeat is
1 − 0.15 = 0.85.

Statement 4 defines that , which means

1
Pr(High ∣ Irregular) =
3

⋅ 0.15 = 0.05 . Furthermore, statement 5 defines that

1
Pr(Irregular ∩ High) =
3

, which means ⋅ 0.64 = 0.08 .

1 1
Pr(Irregular ∣ Normal) = Pr(Irregular ∩ Normal) =
8 8

Table 1.2.7

Next, complete the table by subtracting from the totals.

 Table 1.2.8

The final answers for all three parts of the problem are:

a. Pr(Regular ∩ High) = 0.09

b. Pr(Regular ∩ Normal) = 0.56

c. Pr(Irregular ∩ Low) = 0.02

COACH'S REMARKS

Note that Venn diagrams, tree diagrams, and tables all do the same thing — organize
your solution. You can typically use any of these three methods interchangeably.

In general, Venn diagrams tend to be more useful when there are two or more
categories where each individual can fall into any number of categories. So, the
outcomes are "in the category" or "not in the category". 1.2.4 is an example of this.
There are three categories: life insurance, health insurance, and renewal.

Tables tend to be more useful when there are exactly two categories, each with two or
more outcomes. 1.2.7 is an example of this. The two categories are heartbeat and
blood pressure. Heartbeats are classified as regular or irregular, and blood pressures
are either low, medium, or high.
Tree diagrams can be used when there are two or more categories, each with two or
more outcomes. They are mainly useful when there are conditional probabilities. 1.2.3
is an example of this. Although we're not explicitly given conditional probabilities, they
are easy to infer from the information we do have.

Discussions

Search... EXPAND

ANSWERED
Question 1.2.6
2 Abdoullah Ahmed Jun 19 at 6:05 PM

ANSWERED
Q 1.2.3
2 Aaryaman Anand Jul 7 at 11:33 PM

ANSWERED
Example 1.2.1
2 Greta Meske Aug 5 at 9:27 PM

ANSWERED
Question ??
1 Mr.007 Mar 27 at 5:00 PM

Don't see what you're looking for?

Our coaches are here to answer your questions, help clarify, or further discuss the
above content. If you don't see a thread that helps you, start one of your own!

START A DISCUSSION
 Next Lesson
Watch 1.2.1 Conditional Probability