1. Suppose a series of n independent trials can end in one of three possible outcomes.
Let k1
and k2 denote the number of trials that result in outcomes 1 and 2, respectively. Let p1
and p2 denote the probabilities associated with outcomes 1 and 2. Deduce a formula for
the probability of getting k1 and k2 occurrences of outcomes 1 and 2, respectively.
Let n = k1 + k2 + k3 and k3 = n − k1 − k2 . At the same time p1 + p2 + p3 = 1 and
p3 = 1 − p1 − p2 .
Then
n!
P (k1 outcomes 1,k2 outcomes 2, k3 outcomes 3) = pk1 pk2 pk3
k1 !k2 !k3 ! 1 2 3
n!
= pk1 pk2 (1 − p1 − p2 )n−k1 −k2
k1 !k2 !(n − k1 − k2 )! 1 2
which is known as the multinomial distribution.
2. Repair calls for central air conditioners fall into three general categories: coolant leakage,
compressor failure, and electrical malfunction. Experience has shown that the probabilities
associated with the three categories are 0.5, 0.3, and 0.2, respectively. Suppose that a
dispatcher has logged in ten service requests for tomorrow morning. Use the answer to
Question 1 to calculate the probability that three of those ten will involve coolant leakage
and five will be compressor failures.
Using Q1, we have that p1 = 0.5, p2 = 0.3 and n = 10.
10!
P (3 outcomes 1,5 outcomes 2) = 0.53 0.35 (0.2)2 = 0.031.
3!5!2!
3. Find the probability mass function (pmf ) for the discrete random variable X whose cumu-
lative distribution function (cdf ) at the points x = 0, 1, . . . , 6 is given by FX (x) = x2 /36.
We have that
pX (k) = P (X = k) = FX (k) − FX (k− )
where FX (k− ) = limk0 ↑k FX (k) and the pmf pX (k) is the height of the jump of FX (k).
Then
pX (0) = P (X = 0) = 0
1
pX (1) = P (X = 1) =
36
3
pX (2) = P (X = 2) = FX (2) − pX (1) =
36
5
pX (3) = P (X = 3) = FX (3) − pX (1) − pX (2) = FX (3) − FX (2) =
36
7
pX (4) = P (X = 4) = FX (4) − FX (3) =
36
9
pX (5) = P (X = 5) = FX (5) − FX (4) =
36
11
pX (6) = P (X = 6) = FX (6) − FX (5) = .
36
1
4. Consider the following game. A fair coin is flipped until the first head appears. The win
is $2 if it appears on the first toss, $4 if it appears on the second toss, and, in general,
$2k if it first occurs on the kth toss. Let the random variable X denote the winnings.
How much must the bet be in order for this to be a fair game? [Note: A fair game is one
where the difference E[Xn+1 − Xn ] = 0.]
k
Note pX (2k ) = P (X = 2k ) = 12 . Thus
∞
X X 1
E(X) = 2k pX (2k ) = 2k = 1 + 1 + ...
2k
all k k=1
which is a divergent sum. That is X does not have a finite value, so in order for this to be
a fair game, the initial bet would have to be an infinite amount of money. This is known
as the St Petersburg paradox.
5. An urn contains four chips numbered 1 through 4. Two are drawn without replacement.
Let the random variable X denote the larger of the two. Find E(X).
There are 42 equally likely unordered outcomes
S = {(1, 2), (1, 3), (1, 4), (2, 3), (2, 4), (3, 4)}
and the cardinality of S is |S| = 6.
Then the corresponding random variable X can assume value
X = {2, 3, 4, 3, 4, 4} = {2, 3, 4}.
Thus pX (2) = 16 , pX (3) = 62 , and pX (4) = 36 .
The expected value is
4
X 2 6 12 1 10
E(X) = kpX (k) = + + = (1 + 3 + 6) = .
6 6 6 3 3
k=2
6. Suppose the experiment consists of tossing 3 fair coins. Let Y denote the number of heads
that appear, define the pmf pY (k) and show it is a valid pmf.
The random variable Y is taking values in {0, 1, 2, 3} and the pmf is
1
pX (0) = P ({T, T, T }) =
8
3
pX (1) = P ({T, T, H}, {T, H, T }, {H, T, T }) =
8
3
pX (2) = P ({T, H, H}, {H, T, H}, {H, H, T }) =
8
1
pX (3) = P ({H, H, H}) =
8
Since Y must take on one of the values in {0, 1, 2, 3}, we must have
3
X 1 3 3 1
∪3k=1 {Y
1=P = k} = P (Y = k) = + + + = 1.
8 8 8 8
k=0
2
7. Independent trials consisting of the flipping of a coin having probability p of coming up
heads are continually performed until either a head occurs or a total of n flips is made.
Let X denote the number of times the coin is flipped, find pX (k), and check that it is a
pmf.
The random variable X can take on one of values in {1, 2, 3, . . . , n}. The pmf is defined
as
P (X = 1) = P ({H}) = p
P (X = 2) = P ({T, H}) = (1 − p)p
P (X = 3) = P ({T, T, H}) = (1 − p)2 p
..
.
P (X = n − 1) = P ({T, T, . . . , T , H}) = (1 − p)n−2 p
| {z }
n-2
P (X = n) = P ({T, T, . . . , T , H}, {T, T, . . . , T , T }) = (1 − p)n−1
| {z } | {z }
n-1 n-1
As a check, note that
n
X
P (∪nk=1 {X = k}) = P (X = k)
k=1
n−1
X
= p(1 − p)k−1 + (1 − p)n−1
k=1
1 − (1 − p)n−1
=p + (1 − p)n−1
1 − (1 − p)
= 1 − (1 − p)n−1 + (1 − p)n−1 = 1
8. Find E(X) and Var(X), where X is the outcome when we roll a fair die.
1
With a fair die, we have that pX (1) = pX (2) = pX (3) = pX (4) = pX (5) = pX (6) = 6 and
1 1 1 1 1 1 7
E(X) = 1 · +2· +3· 4· +5· +6· = .
6 6 6 6 6 6 2
Also
1 1 1 1 1 1 1
E(X 2 ) = 12 · + 22 · + 32 · 42 · + 52 · + 62 · = 91
6 6 6 6 6 6 6
and 2
91 7 35
Var(X) = − = .
6 2 12
3
9. A school class of 120 students is driven in 3 buses to a symphonic performance. There
are 36 students in one of the buses, 40 in another, and 44 in the third bus. When the
buses arrive, one of the 120 students is randomly chosen. Let X denote the number of
students on the bus of that randomly chosen student, and find E(X).
Since the randomly chosen student is equally likely to be any of the 120 students, it follows
that
36
P (X = 36) =
120
40
P (X = 40) =
120
44
P (X = 44) = .
120
Hence,
3 1 11 1208
E(X) = 36 + 40 + 44 = = 40.2667.
10 3 30 30
However, the average number of students on a bus is 120/3 = 40, showing that the
expected number of students on the bus of a randomly chosen student is larger than
the average number of students on a bus. This is a general phenomenon, and it occurs
because the more students there are on a bus, the more likely it is that a randomly chosen
student would have been on that bus. As a result, buses with many students are given
more weight than those with fewer students.
4
10. Suppose there are m days in a year, and that P each person is independently born on day r
with probability pr , with r = 1, . . . , m and m
k=1 pk = 1. Let Ai,j be the event that person
i and person j are born on the same day.
(a) Find P (A1,3 ).
(b) Find P (A1,3 |A1,2 ).
(c) Show that P (A1,3 |A1,2 ) ≥ P (A1,3 ).
a. The event A1,3 =“person 1 and person 3 have the same birthday” is the union of the
mutually exclusive events that they were both born on day r, r = 1, . . . , m we have that
m
X
P (A1,3 ) = p2r .
r=1
b. Using the definition of conditional probability, we have that
P (A1,2 , A1,3 )
P (A1,3 |A1,2 ) =
P (A1,2 )
Pm 3
pr
= Pr=1
m 2
.
r=1 pr
where P (A1,2 , A1,3 ) is computing the probability of the union of the m mutually exclusive
events that 1, 2, 3 were born all on day r, r = 1, . . . , m.
c. It follows that P (A1,3 |A1,2 ) ≥ P (A1,3 ) is equivalent to
m
X m
X
p3r ≥( p2r )2 .
r=1 r=1
To prove this, let X be a random variable that is equal to pr with probability pr , i.e.
P (X = pr ) = pr for r = 1, . . . , m. Then
m
X m
X
E(X) = pr P (X = pr ) = p2r
r=1 r=1
m
X m
X
E(X 2 ) = p2r P (X = pr ) = p3r .
r=1 r=1
Since Var(X) ≥ 0, it means that E(X 2 ) ≥ [E(X)]2 , therefore
m
X m
X
p3r ≥( p2r )2 .
r=1 r=1
