Negative Binomial
Negative Binomial
Negative Binomial
Let r ∈ N. Suppose that we keep performing independent Bernoulli trials until the
r-th success is observed. Further suppose that the probability of success in each trial is
p ∈ (0, 1). Thus, the sample space is
Note: To find r-th success, we have to perform Bernoulli trials at least r-times. Thus,
(w1 , w2 , . . . , wn ) ∈ S corresponds to one of n−1
r−1
ways in which the r-th success is obtained
in the n-th Bernoulli trials wn = s and the first n−1 Bernoulli trials result in r−1 successes
and n − r failures.
MX (t) = E(etX )
X
= etx fX (x)
x∈EX
∞
X r+x−1 r x
tx
= e p q , where q = 1 − p
x=0
r − 1
∞
X r+x−1 r t x
= p (qe )
x=0
r−1
= pr (1 − qet )−r , |qet | < 1
p r
= , |t| < − ln q
1 − qet
Therefore,
(1)
MX (t) = pr {rqet (1 − qet )−r−1 }, |t| < − ln q;
(2)
MX (t) = pr {rqet (1 − qet )−r−1 + r(r + 1)(qet )2 (1 − qet )−r−2 }, |t| < − ln q;
(1) rq
E(X) = MX (0) = ;
p
(2) r(r + 1)q 2 rq
E(X 2 ) = MX (0) = + ;
p2 p
rq
and V ar(X) = E(X 2 ) − (E(X))2 = 2 , where q = 1 − p.
p
Example 2. A person repeatedly rolls a fair dice independently until an upper face with
two or three dots is observed twice. Find the probability that the person would require
eights rolls to achieve this.
Solution: In each trial, let us label the outcome of observing an upper face with two
or three dots as success and observing any other outcome as a failure. Hence success
probability in each trial is 13 . Let Z denote the number of trials required to get the second
success and X = Z − 2. Then X ∼ NB(2, 13 ). Therefore, the required probability is
2 6
7 1 2 448
P ({Z = 8}) = P ({X = 6}) = = .
1 3 3 6561
Example 3. A mathematician carries one matchbox each in his right and left pockets.
When he wants a match, he selects the left pocket with probability p and the right pocket
with probability 1 − p. Suppose that initially each box contains N matches. Consider the
moment when the mathematician for the first time discovers that one of the match boxes is
empty. Find the probability that at that moment the other box contains exactly k matches,
where k ∈ {0, 1, 2, · · · , N }.
Solution: Let us identify success with the choice of the left pocket. The left pocket box
will be empty at the moment when the right pocket box contains exactly k matches if
and only if N − k failures precede the (N + 1)-th success. A similar arguments applies to
the right pocket.
2
Now the required probability is
p = P (the left pocket is found empty, the right pocket contains k matches)
+ P (the right pocket is found empty, the left pocket contains k matches)
N + 1 + N − k − 1 N +1 N −k N +1+N −k−1
= p (1 − p) + (1 − p)N +1 pN −k
N +1−1 N +1−1
2N − k N +1 N −k 2N − k
= p (1 − p) + (1 − p)N +1 pN −k .
N N
2. Geometric Distribution