Lecture 3:
Conditional Probability
Dr. Luu Van Thanh
Vietnamese-German University
Definition
Definition Let A and B be events. The conditional probability of event A
conditional on event B is P (A ∩B)
P (A|B) =
P (B)
for P(B) > 0. This measures the probability that A occurs when it is known
that B occurs.
Some special cases are:
➢ A and B are mutually exclusive, i.e., A ∩ B = ∅. Then
P(A ∩B) 0
P(A|B) = = = 0.
P(B) P(B)
➢ B ⊂ A. Then P(A ∩B) P(B)
P(A|B) = = = 1.
P(B) P(B)
2
Consider the following
example.
We have
P(A ∩B) 0.26
P(A|B) = = = 0.464
P(B) 0.56
P(A ∩B') 0.01
P(A|B') = = = 0.023
P(B') 0.44
P(A' ∩B) 0.3
P(A'|B) = = = 0.536
P(B) 0.56
Note that P(A|B) + P(A ′ |B) = 1 (similar to P(A) + P(A ′ ) = 1).
3
Example (Defective computer chips) Consider the example of
defective chips. Recall that we define A the event that a box
has no more than 5 defective chips, which is also called the correct
event. We calculated before that P(A) = 0.71.
If the company guarantees that a box has no more than 5
defective chips, then the customers can be classified as either
satisfied or unsatisfied, depending on whether the guarantee is met.
What is the probability that a satisfied customer purchased a box
with no defective chips?
Intuitively, this conditional probability should be larger than the
unconditional probability P(0 defectives) = 0.02.
P(0 defectives ∩ A)
P(0 defectives|A) =
P(A)
= P(0 defectives)/P(A) = 0.02/0.71 = 0.028
4
Example (Power plant operation) Recall that we denote A to be the event
that plant X is idle. Note that P(A) = 0.32.
If it is known that at least 2 out of the 3 plants are generating electricity
(event B), then how does this change the probability that plant X is idle? We
have
P(A ∩B) 0.18
P(A|B) = = = 0.257.
P (B) 0.70
Therefore, whereas plant X is idle 32% of the time, it is idle only about
25.7% of the time when at least 2 plants are generating electricity.
5
Games of chance
Die rolling
If a fair die is rolled, then P( { 6 }) = 1 / 6 .
If someone rolls a die without showing you but announces that the result is
even, then the chance that a 6 is obtained should be 1/3 . This is justified
by
P( { 6 } ∩ even) P( { 6 } ) 1/6 1
P({6}\even) = = = = .
P(even) P(even) 1/6 + 1/6 + 1/6 3
Suppose a red die and a blue die are thrown with 36 equally likely outcomes.
Let A be the event that a red die scores a 6. Then P (A) = 6 / 3 6 = 1 / 6 .
Let B be the event that at least one 6 is obtained on the 2 dice.
Then P (B) = 11/36 .
Suppose someone rolls the 2 dice without showing you but announces that
at least one 6 has been scored.
Then what is the probability that the red die scored a 6?
6
Answer The question asks for P(A|B):
P(A ∩B) P(A)
P(A|B) = = = (1/6)/(11/36) = 6/11
P(B) P(B)
Let C be the event that exactly one 6 has been scored. Then
P(A ∩C)
P(A|C) = =(5/36)/(10/36) = 1/2
P(C)
7
Card playing If a card is drawn from a pack of cards, let A be the event that a card
from the heart suit is obtained. Also let B be the event that a picture card is drawn.
Recall that P (A) = 1 3 / 5 2 = 1 / 4 , and P (B) = 1 2 / 5 2 = 3 / 1 3 .
A ∩ B represents the event that a picture card is drawn from the heart suit with P (A
∩ B) = 3 / 5 2 .
Now suppose that someone draws a card and announces that it is from the heart
suit. What is the probability that it is a picture card? We have
P(B ∩ A) =(3/52)/(13/52) = 3/13
P(B|A) =
P(A)
Note that in this case, P(B) = P(B|A) as the proportion of picture cards in the heart
suit is identical to the proportion of picture cards in the whole pack. So that we say
A and B are independent event. This notion will be discussed later.
8
Next let C be the event that the A is chosen. Then P(C) = 1 / 5 2 .
If it is known that a card from the heart suit is obtained, then intuitively the
conditional probability of the card being A is 1/13, as there are 13 equally likely
cards in the heart suit.
The intuition is confirmed by
P(C ∩A) P(C) 1/52 1
P(C|A) = = = = .
P(A) P(A) 1/4 13
9
Group Assignment
1.4.1, 1.4.3, 1.4.5, 1.4.9, 1.4.10, 1.4.12
(page 38-39)