Probability
Probability
1,2,3,4
Date: August 06, 07,08,11, 2020
MA 2302: Introduction to Probability and Statistics
PROBABILITY
Instructor
Prof. Gopal Krishna Panda
Department of Mathematics
NIT Rourkela
Probability
𝑚(𝐴)
• Thus, 𝑃 𝐴 = , where 𝑚 may the counting measure or it measures length in
𝑚(𝑆)
1D, area in 2D and volume in 3D.
• Two events 𝐴 and 𝐵 are equally likely (to occur) if 𝑃 𝐴 = 𝑃(𝐵). 𝐴 and 𝐵 are mutually
exclusive if they are set theoretically disjoint leading to 𝑃 𝐴 ∩ 𝐵 = 0. However,
𝑃 𝐴 ∩ 𝐵 = 0 need not imply that 𝐴 and 𝐵 are mutually exclusive.
𝐴 𝐵
𝐴𝐵𝑐 𝐴𝐵 Ac 𝐵
• If 𝐴, 𝐵 and 𝐶 are three events, then
𝑃 𝐴∪𝐵∪𝐶 = 𝑃 𝐴∪𝐵 ∪𝐶 =𝑃 𝐴∪𝐵 +𝑃 𝐶 −𝑃 𝐴∪𝐵 ∩𝐶
= 𝑃 𝐴 ∪ 𝐵 + 𝑃 𝐶 − 𝑃 𝐴𝐶 ∪ 𝐵𝐶
= 𝑃 𝐴 + 𝑃 𝐵 − 𝑃 𝐴𝐵 + 𝑃 𝐶 − 𝑃 𝐴𝐶 − 𝑃 𝐵𝐶 + 𝑃 𝐴𝐵𝐶
= 𝑃 𝐴 + 𝑃 𝐵 + 𝑃 𝐶 − 𝑃 𝐴𝐵 − 𝑃 𝐴𝐶 − 𝑃 𝐵𝐶 + 𝑃 𝐴𝐵𝐶 .
Assignment 1: Derive the formula for 𝑃 𝐴 ∪ 𝐵 ∪ 𝐶 ∪ 𝐷 .
Assignment 2: Guess the formula for 𝑃 𝐴 ∪ 𝐵 ∪ 𝐶 ∪ 𝐷 ∪ 𝐸 .
Assignment 3: Observe that
𝑛 𝑛
𝑃 ራ 𝐴𝑖 = 𝑃 𝐴𝑖 − 𝑃 𝐴𝑖 𝐴𝑗 + 𝑃 𝐴𝑖 𝐴𝑗 𝐴𝑘 − ⋯ + −1 𝑛−1 𝑃 𝐴1 𝐴2 ⋯ 𝐴𝑛 .
𝑖=1 𝑖=1 1≤𝑖<𝑗≤𝑛 1≤𝑖<𝑗<𝑘≤𝑛
= 1 − 1 − 𝑝 𝑛.
• If 𝐴 and 𝐵 are two events and 𝐴 ⊂ 𝐵, then 𝑃 𝐴 ≤ 𝑃(𝐵).
Observe that 𝐵 = 𝐴 ∪ 𝐴𝑐 𝐵, 𝐴 and 𝐴𝑐 𝐵 are mutually exclusive.
Hence, 𝑃 𝐵 = 𝑃 𝐴 + 𝑃(𝐴𝑐 𝐵). Since 𝑃 𝐴𝑐 𝐵 ≥ 0, the 𝑩
result follows. 𝑨𝒄 𝑩
• 𝑃 𝐴𝐵 ≥ 𝑃 𝐴 + 𝑃 𝐵 − 1 follows from
𝑃 𝐴 ∪ 𝐵 = 𝑃 𝐴 + 𝑃 𝐵 − 𝑃 𝐴𝐵 ≤ 1. In general,
𝑛 𝑛
𝑃 ሩ 𝐴𝑖 ≥ 𝑃 𝐴𝑖 − 𝑛 − 1 .
𝑖=1 𝑖=1
Prove it using mathematical induction.
Conditional Probability
If we perform any experiment, then the sample space always occurs. But a event 𝐴 may or
may not occur. Let 𝐴 and 𝐵 be two events and it is known that 𝐵 has occurred. Then
depending on 𝐵, the probability of 𝐴 prior and posterior to the occurrence of 𝐵 will
certainly be affected.
40 39
𝑃 𝐴𝐵 = 𝑃 𝐴 𝑃 𝐵 ∶ 𝐴 ℎ𝑎𝑠 𝑜𝑐𝑐𝑢𝑟𝑒𝑑 = × .
90 89
𝐴 𝐵
𝐴𝐵
If 𝑃 𝐵 = 0, then 𝑃 𝐴 𝐵 = 𝑃(𝐴) since the nothing has occurred. However, if 𝑃 𝐵 = 1, then also
𝑃 𝐴 𝐵 = 𝑃 𝐴 since occurrence of 𝐵 is equivalent to the occurrence of the sample space which
always occur. In particular 𝑃 𝐴 𝜙 = 𝑃 𝐴|𝑆 = 𝑃(𝐴).
Observe the following about conditional probability: (To be proved by you)
𝑃 𝐴 𝐵 + 𝑃 𝐴𝑐 𝐵 = 1, 𝑃 𝐴 𝐵𝑐 + 𝑃 𝐴𝑐 𝐵𝑐 = 1
𝑃 𝐴∪𝐵 𝐶 =𝑃 𝐴 𝐶 +𝑃 𝐵 𝐶 −𝑃 𝐴∩𝐵 𝐶 .
𝑃(𝐴)
𝑃 𝐴 𝐴∪𝐵 = , 𝑃 𝐴 𝐴 ∩ 𝐵 = 1.
𝑃(𝐴 ∪ 𝐵)
𝑃(𝐴)
𝑃 𝐴𝐵 = .
𝑃(𝐵)
If 𝐴 ⊂ 𝐵, then
𝑃 𝐴𝐶 ≤𝑃 𝐵𝐶 .
Independent events
Observe that, in the previous example if the resistors are chosen at random with replacement, then
40 40
𝑃 𝐴𝐵 = 𝑃 𝐴 𝑃 𝐵|𝐴 = × = 𝑃 𝐴 𝑃(𝐵).
90 90
Definition: Two events 𝐴 and 𝐵 in a sample space 𝑆 are said to be independent if 𝑃 𝐴𝐵 = 𝑃 𝐴 𝑃(𝐵).
If there are more than two events, say 𝐴1 , 𝐴2 , … , 𝐴𝑛 , then they are said to be independent if
𝑃 𝐴𝑖1 𝐴𝑖2 … 𝐴𝑖𝑟 = 𝑃 𝐴𝑖1 𝑃 𝐴𝑖2 ⋯ 𝑃 𝐴𝑖𝑟 for each nonempty subset 𝑖1 , 𝑖2 , … , 𝑖𝑟 of 1,2, … , 𝑛 . If
𝑃 𝐴𝑖 𝐴𝑗 = 𝑃 𝐴𝑖 𝑃(𝐴𝑗 ) for all 𝑖 ≠ 𝑗, then 𝐴1 , 𝐴2 , … , 𝐴𝑛 are said to be pairwise independent.
𝑃 𝐴𝐵𝐶 = 𝑃 𝐴 𝑃 𝐵 P(C)
Example 1. If 𝐴 and 𝐵 are independent, show that 𝐴𝑐 and 𝐵 are also independent.
Proof: Given 𝑃 𝐴𝐵 = 𝑃 𝐴 𝑃(𝐵). To prove 𝑃 𝐴𝑐 𝐵 = 𝑃 𝐴𝑐 𝑃(𝐵).
𝐴𝑐 𝐵 ∪ 𝐴𝐵 = 𝐵. 𝐴𝑐 𝐵 and 𝐴𝐵 are mutually exclusive. Hence
𝑃 𝐵 = 𝑃 𝐴𝑐 𝐵 + 𝑃 𝐴𝐵 = 𝑃 𝐴𝑐 𝐵 + 𝑃 𝐴 𝑃 𝐵
Thus,
𝑃 𝐴𝑐 𝐵 = P 𝐵 − 𝑃 𝐴 𝑃 𝐵 = 1 − 𝑃 𝐴 𝑃 𝐵 = 𝑃 𝐴𝑐 𝑃(𝐵).
Q. If 𝐴 and 𝐵 are independent, show that 𝐴𝑐 and 𝐵𝑐 are also independent.
Proof: Given 𝑃 𝐴𝐵 = 𝑃 𝐴 𝑃(𝐵). To prove 𝑃 𝐴𝑐 𝐵𝑐 = 𝑃 𝐴𝑐 𝑃(𝐵𝑐 ).
𝑃 𝐴𝑐 𝐵𝑐 = 𝑃( 𝐴 ∪ 𝐵 𝑐 ) = 1 − 𝑃 𝐴 ∪ 𝐵 = 1 − 𝑃 𝐴 − 𝑃 𝐵 + 𝑃 𝐴𝐵
=1−𝑃 𝐴 −𝑃 𝐵 +𝑃 𝐴 𝑃 𝐵 = 1−𝑃 𝐴 1−𝑃 𝐵 = 𝑃 𝐴𝑐 𝑃(𝐵𝑐 )
Example 2. If 𝐴, 𝐵 and 𝐶 are pairwise independent and 𝐴 is independent of 𝐵 ∪ 𝐶 then prove that 𝐴, 𝐵 and
𝐶 are mutually independent.
𝑃 𝐴𝐵 +𝑃 𝐴𝐶 − 𝑃 𝐴𝐵𝐶 = 𝑃 𝐴 𝑃(𝐵)+ 𝑃 𝐴 𝑃 𝐶 − 𝑃 𝐴 𝑃 𝐵𝐶 .
Now using pairwise independence, one can get the desired results.
Example 3. The probability of hitting a target in a single shot is 𝑝. What is the probability that out of ten
shot fired, at least one shot will hit the target?
Ans. Let 𝐴𝑖 be the event that the 𝑖-th shot hits the target. Then 𝑃 𝐴𝑖 = 𝑝 for each 𝑖. We are
interested in
Example 4. A box contains three chips bearing numbers 112,121,211 and 222. A chip is chosen
at random. Let 𝐴, 𝐵 and 𝐶 be the events that the first, second and third digit of the chip
number is 1 respectively. Prove that 𝐴, 𝐵 and 𝐶 are pairwise independent but not mutually
independent.
1
𝑃 𝐴𝐵𝐶 = 0 ≠ = 𝑃 𝐴 𝑃 𝐵 P(C). Hence 𝐴, 𝐵 and 𝐶 are pairwise independent but not
8
mutually independent.
Example 5: 𝑃 𝐴𝐵𝐶 = 𝑃 𝐴 𝑃 𝐵 𝑃(𝐶), but 𝐴, 𝐵 and 𝐶 are not pairwise independent
and hence not mutually independent.
1
𝑃 𝐴2 𝑃 𝐷|𝐴2 × 0.1 2
𝑃 𝐴2 |𝐷 = 3 = 3 = .
σ𝑖=1 𝑃 𝐴𝑖 𝑃 𝐷|𝐴𝑖 1 1 1 9
× 0.2 + × 0.1 + × 0.15
3 3 3
Bayes Theorem: Have you ever thought like this?
Let 𝐴 and 𝐵 be any two arbitrary events such that 𝑃 𝐵 > 0. Then 𝐵 ⊂ (𝐴 ∪ 𝐴𝑐 ) and hence
𝐵 = 𝐵 ∩ 𝐴 ∪ 𝐴𝑐 = 𝐴𝐵 ∪ 𝐴𝑐 𝐵. Hence,
𝑃 𝐵 = 𝑃 𝐴𝐵 + 𝑃 𝐴𝑐 𝐵 = 𝑃 𝐴 𝑃 𝐵 𝐴 + 𝑃 𝐴𝑐 𝑃(𝐵|𝐴𝑐 )
which is the law of total probabilities and
𝑃 𝐴 𝑃 𝐵𝐴
𝑃 𝐴𝐵 =
𝑃 𝐴 𝑃 𝐵 𝐴 + 𝑃 𝐴𝑐 𝑃(𝐵|𝐴𝑐 )
which is the Bayes theorem. Also
𝑐 𝑃(𝐵|𝐴𝑐 )
𝑃 𝐴
𝑃 𝐴𝑐 𝐵 = 𝑐 𝑐
.
𝑃 𝐴 𝑃 𝐵 𝐴 + 𝑃 𝐴 𝑃(𝐵|𝐴 )
In addition,
𝑃 𝐴𝑐 𝑃(𝐵𝑐 |𝐴𝑐 )
𝑃 𝐴𝑐 𝐵𝑐 = .
𝑃 𝐴𝑐 𝑃 𝐵𝑐 𝐴𝑐 + 𝑃 𝐴 𝑃(𝐵𝑐 |𝐴)
Example 7. It is estimated that 2% of certain locality were infected by a virus. A certain plasma test is
used to detect infection due the virus. About 5% of the test results are false positive and 10% are false
negative. If a person was tested positive for the virus, find the probability that he was really infected. If a
person was tested negative, what was the probability that he was actually not infected?
Ans. Let 𝐴 be the event that the person was infected by the virus and 𝐵, the event that that the person
was tested positive for the virus. False positive means a person was tested positive given that he was
not infected. False negative means a person was tested negative given that he was really infected.
Thus,
𝑃 𝐴 = 0.02, 𝑃 𝐴𝑐 = 0.98, 𝑃 𝐵 𝐴𝑐 = 0.05, 𝑃 𝐵𝑐 𝐴 = 0.10
Hence, 𝑃 𝐵𝑐 𝐴𝑐 = 0.95, 𝑃 𝐵 𝐴 = 0.90. (Observe that 𝑃 𝐴 𝐵 + 𝑃 𝐴𝑐 𝐵 = 1.) Hence,
Pr 𝐴 𝑝𝑒𝑟𝑠𝑜𝑛 𝑤𝑎𝑠 𝑟𝑒𝑎𝑙𝑙𝑦 𝑖𝑛𝑓𝑒𝑐𝑡𝑒𝑑 𝑔𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡 ℎ𝑒 𝑤𝑎𝑠 𝑡𝑒𝑠𝑡𝑒𝑑 𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒
𝑃 𝐴 𝑃 𝐵𝐴 0.02 × 0.90 0.018
=𝑃 𝐴𝐵 = = =
𝑃 𝐴 𝑃 𝐵 𝐴 + 𝑃 𝐴𝑐 𝑃(𝐵|𝐴𝑐 ) 0.02 × 0.90 + 0.98 × 0.05 0.018 + 0.049
= 0.2686.
4 3 5
Hence, the required probability is 1+3× .
9 9
Ans. Let 𝐴𝑖 be the event that the number 𝑖 is in the 𝑖-th position, i = 1,2, … , 10. we need to find
𝑃 𝐴1𝑐 ∩ 𝐴𝑐2 ∩ ⋯ ∩ 𝐴10
𝑐
= 𝑃 𝐴1 ∪ 𝐴2 ∪ ⋯ ∪ 𝐴10 𝑐
= 1 − 𝑃 𝐴1 ∪ 𝐴2 ∪ ⋯ ∪ 𝐴10 ,
𝑃 𝐴1 ∪ 𝐴2 ∪ ⋯ ∪ 𝐴10 = σ10
𝑖=1 𝑃 𝐴𝑖 − σ1≤𝑖<𝑗≤10 𝑃 𝐴𝑖 𝐴𝑗 + σ1≤𝑖<𝑗<𝑘≤10 𝑃 𝐴𝑖 𝐴𝑗 𝐴𝑘 − ⋯ +
−1 10−1 𝑃 𝐴1 𝐴2 ⋯ 𝐴10
9! 8! 7!
Observe that 𝑃 𝐴𝑖 = for each 𝑖, 𝑃 𝐴𝑖 𝐴𝑗 = for 𝑖 ≠ 𝑗, 𝑃 𝐴𝑖 𝐴𝑗 𝐴𝑘 = for
10! 10! 10!
1
𝑖 ≠ 𝑗 ≠ 𝑘, …….. and finally, 𝑃 𝐴1 𝐴2 ⋯ 𝐴10 = .
10!
Thus,
10 9! 10 8! 10 7! 9
1
𝑃 𝐴1 ∪ 𝐴2 ∪ ⋯ ∪ 𝐴10 = ⋅ − ⋅ + ⋅ − ⋯ + −1 ⋅
1 10! 2 10! 3 10! 10!
and hence the required probability is
1 1 1 1 1
1− + − + −⋯+ .
1! 2! 3! 4! 10!
Example 12. Three capacitors are chosen at random and with replacement from a lot consisting of 8
capacitors of capacity 500 𝜇𝐹 and 16 capacitors of capacity 1000 𝜇𝐹. Find the probability that, out of the
three chosen capacitors, (a) there is at least one capacitors of capacity 500 𝜇𝐹 and (b) not more that two
capacitors of capacity 1000 𝜇𝐹?
Ans. Let A be the event of choosing a 500 𝜇𝐹 capacitor and B, the event of choosing a 1000 𝜇𝐹 capacitor
1 2
from the lot if one is chosen at random. The certainly, 𝑃 𝐴 = and 𝑃 𝐵 = . The three choices can result
3 3
in any one of the following:
Outcomes 𝐴𝐴𝐴 𝐴𝐴𝐵 𝐵𝐴𝐴 𝐴𝐵𝐴 𝐴𝐵𝐵 𝐵𝐴𝐵 𝐵𝐵𝐴 𝐵𝐵𝐵
Thus, the probability of choosing 0,1,2,3, 500 𝜇𝐹 capacitors are respectively 8/27, 12/27, 6/27, 1/27. The
probability of choosing 0,1,2,3, 1000 𝜇𝐹 capacitors are respectively 1/27, 6/27, 12/27, 8/27 . The
probability of choosing at least one capacitors of capacity 500 𝜇𝐹 is equal to 12/27+6/27+1/27=19/27. The
probability of choosing not more then two capacitors of capacity 1000 MFD is equal to
1/27+6/27+12/27=19/27.