MTH135/STA104: Probability

Homework # 2 Solution Due: Tuesday, Sep 13, 2005

Prof. Robert Wolpert

Reminder: On this and every homework, give (non-integer) numerical an-

swers as fractions in lowest terms or as decimals to four significant digits—
so π/1000 = 0.003142, 24/6 = 4.
1. Let U1 and U2 be independent Un[0, 1] random variables; recall that
U ∼ Un[0, 1] means that, for any real number x ∈ R,

0 −∞ < x < 0

P[U ≤ x] = x 0≤x<1

1 1 ≤ x < ∞,


and that, since U1 ⊥⊥ U2 , we may think of (U1 , U2 ) as a point drawn uni-

formly from the unit square. Evaluate the probabilities (art helps):
a) P[U1 ≤ 0.8] = 0.8∗1
1
= 0.8
√
b) P[ U1 ≤ 0.8] = P[U1 ≤ 0.64] = 0.64∗1
1
= 0.64
(0.8∗0.8)/2
c) P[U1 + U2 ≤ 0.8] = 1
= 0.32
√ h i√0.8
2
R 0.8 U13
d) P[U1 + U2 ≤ 0.8] = 0 (0.8 − U12 )dU = 0.8U1 − 3
= 0.47703
0
e) P[2 U1 − U2 ≤ 0.8] = (0.4 ∗ 1) + (.5 ∗ 1)/2 = 0.65
2. Alex, Blake, and Camden like to play Scat, a three-handed card game.
Alex wins twice as often as Blake, and Blake wins three times as often as
poor pathetic Camden.
So then, A = 2B = 6C and B = 3C
a) What is the chance that Alex will win the first game?
Make a system of equations. Since, P[B] = 3P[C] and P[A]+P[B]+P[C] = 1,

1
then P[C] + 3P[C] + 6P[C] = 1.
6
P[A]wins one game is = 10 or 35
b) If they play exactly three games, what is the probability that each
will win a game?
1

3
6
27
P[each one wins a game] = 3! 10 10 10
= 250
3. (p. 46 prob 8) A hat contains a lot of cards, with 30% white on both
sides, 50% black on one side and white on the other, and 20% black on both
sides. The cards are mixed up and then a single card is drawn at random
and placed on the table. If the top is black, what is the probability that the
other side is white?
P[W W ] = 0.3, P[BW ] = 0.5, and P[BB] = 0.2
P [one side is white and one side is black]
P[one side is white|one side is black] =
P [one side is black]
P [one side is white and one side is black]
=
P [exactly one side is black or both sides are black]
5
=
9

4. (p. 46 prob 12) Give a formula for P[F | Gc ] in terms of P[F ], P[G] and
P[F G].
P[F Gc ] P[F ] − P[F G]
P[F | Gc ] = c
=
P[G ] 1 − P[G]

5. A German Doppelkoff deck has 48 cards, 12 each of the four suits ♥,

♠, ♦, and ♣. The deck is shuffled well and five cards are dealt, without
replacement. Find the probabilities that:
a) All five cards are ♥’s.
12 11 10 9 8 1
P[All five cards are ♥’s.] = ( )( )( )( )( ) =
48 47 46 45 44 2162
b) At least one card is a ♥.
P[At least one card is a ♥’s.] = 1 − P[no card is a ♥ ]
 
36 35 34 33 32
= 1 − ( )( )( )( )( )
48 47 46 45 44
= 1 − 0.2202 = 0.7798

2
 c) Exactly three of the five cards are ♥’s.
 
5 12 11 10 36 35
P[Exactly three of the five cards are ♥’s] = ( )( )( )( )( ) = 0.008094
2 48 47 46 45 44

6. A dodecahedral (twelve-sided) die is equally likely to show any of its

twelve faces, numbered {1, 2, ..., 12}. If such a die is tossed ten times in a
row, what is the probability that two (or more) consecutive tosses will both
show 12, somewhere among the ten tosses?

• Let f (n) be the probability of at least one consecutive pair of successes

in n independent tries, each with probability p of success. Then f (0) =
0 and f (1) = 0, while for k ≥ 0,

f (k + 2) = p2 + q f (k + 1) + pq f (k) (q ≡ 1 − p). (1)

A direct solution is to use this relation to compute successively f (2),

f (3),...; this can be automated a bit with a spreadsheet or a loop in
any programming language.
Eqn. (1) can also be solved in closed form. The constant f (k) ≡ 1
satisfies Eqn. (1) (but not the boundary conditions), as does f (k) =
1 + a r k for any number r satisfying the homogeneous equation

r k+2 = q r k+1 + pq r k
r 2 = q r + pq
0 = r 2 − q r − pq
q±δ p
r = , δ ≡ q 2 + 4 pq
2

If we denote the two solutions by r+ and r− , then there are unique

numbers a+ , a− for which

f (k) ≡ 1 + a+ (r+ )k + a− (r− )k (2)

satisfies the two boundary conditions,

3
 0 = f (0) = 1 + a+ + a−
0 = f (1) = 1 + a+ r+ + a− r−
= 1 + (a+ + a− )(q/2) + (a+ − a− )(δ/2)
= 1 − q/2 + (a+ − a− )δ/2; thus
a+ + a− = −1
a+ − a− = (q − 2)/δ
a+ = (q − 2)/2δ − 1/2
a− = (2 − q)/2δ − 1/2
To solve the problem, set p = 1/12, q = 11/12, and n = 10; the answer
from Eqn. (2) with the indicated values for a± is about 0.05700001.
• Another solution: For each integer 0 ≤ k ≤ n/2, the number of ways
to pick k numbers in the range 1, 2, ..., n with no two consecutive is
the
P same as the way to choose r = k + 1 integers xj ≥ 1 with sum
xj = S = n + 2 − k (why?); it’s easy to show that this is S−1


r−1
in
n+1−k


general, or k
for us. Thus the probability in question is
5    k  10−k
X 11 − k 1 11
1 −
k=0
k 12 12
1110 119 118 117 116 115
=1 − 1 − 10 − 36 − 56 − 35 − 6
1210 1210 1210 1210 1210 1210
= 1 − 0.4189 − 0.3808 − 0.1246 − 0.0176 − 0.0010 − 0.0000156,
or (again) about 0.05700001.

7. In the following diagram, the system breaks down if either component

C3 fails or if both components C1 and C2 fail (or both).

C1

C3

C2

4
The failure probabilities the three components are P[F1 ] = 0.10, P[F2 ] = 0.70,
and P[F3 ] = 0.30, respectively; all components work or fail independently.
What is the probability that the system works?
P[system works] = (0.9)(0.3)(0.7) + (0.1)(0.3)(0.7) + (0.9)(0.7)(0.7) = 0.651

OR
P[system works] = (1 − 0.10 ∗ 0.70)(1 − .3) = 0.93 ∗ 0.7 = 0.65

8. (almost p. 54 prob 5) The fraction of persons in a population who have

a certain disease is 0.02. A diagnostic test is available to test for the disease,
but it is not perfect: a healthy person has probability 0.05 that the diagnostic
test will (incorrectly) indicate presence of the disease, while a diseased person
might have the disease go undetected by this diagnostic with probability 0.10.
For a person selected at random from the population, find
a) The probability of a positive test result (indicating disease);
P[positive test result] = (0.05)(0.98) + (0.90)(0.02) = 0.067

b) The probability the person selected at random does have the disease
but that the diagnostic does not identify it;
(0.1)(0.02) = 0.002

c) The probability that the person is correctly diagnosed and healthy;

(1 − 0.05)(1 − 0.02) = 0.931

d) The conditional probability of disease, given that the diagnostic test

result is positive (for the disease).
P[indicated disease | disease]P[disease]
P[disease | indicated disease] = =
P[indicated disease]
P[indicated disease | disease]P[disease]
= =
P[indicated disease | disease]P[disease] + P[indicated disease |no disease]P[no disease]
(0.9)(0.02) 0.18
= = = 0.2687
(0.9)(0.02) + (0.05)(0.98) 0.67

5
9. Three fair coins are tossed at once— a nickle, a dime, and a quarter.
a) What is the probability that at least one coin shows Heads?

P[at least one coin shows Heads] =

1 7
1 − P[no coin shows Heads] = 1 − =
8 8

b) What is the probability that at least $0.32 worth of coins show heads?
   
1 1 1 1
P[at least $0.32 worth of coins show heads?] = 2 =
2 2 2 4

c) Skyler wants to know the probability that all three coins show the
same face (i.e., all Heads or all Tails). She reasons that at least two of the
coins must show the same face (after all, there are only two possibilities),
and there is a 50% chance the third face will match. Is she right? Why?
1 1 1
No, the probability is + =
8 8 4

d) Find the probability that all three faces show Heads, given that at
least one of them does.
(1/8) 1
=
(1 − 1/8) 7

e) Can you find an event E with probability exactly P[E] = 1/3? Why?
No; every event in this space has probability 8j for some integer 0 ≤ j ≤ 8 ;
1
3
is not of that form.
10. Brite Lites has three lightbulb factories, one each in Athens, Brevard,
and Coallawalla. The probabilities of defective lightbulbs at these three
factories are P[A] = 0.07, P[B] = 0.05, P[C] = 0.01. A shipment contains 10
bulbs from facility A, 20 bulbs from B, and 70 bulbs from C.
a) A single light bulb is chosen at random from this shipment. If it is

6
defective, what is the probability it came from factory A?

P[Defective |A]P[A]
P[A| Defective] = =
P[Defective]
P[Defective |A]P[A]
= =
P[Defective |A]P[A] + P[Defective |B]P[B] + P[Defective |C]P[C]
0.07 ∗ 0.10 0.007
= = =
(0.07 ∗ 0.10) + (0.05 ∗ 0.20) + (0.70 ∗ 0.01) 0.007 + 0.01 + 0.007
0.007
= = 0.29167
0.024

b) What is the probability that this entire shipment contains at least

two defective bulbs?

P[at leas two defective bulbs ] = 1 − P[none or one defective bulbs]

P[none defective bulbs ] = 0.9310 ∗ 0.9520 ∗ 0.9970 = 0.08585
 
10
P[one defective bulb ] = 0.939 ∗ 0.07 ∗ 0.9520 ∗ 0.9970 +
1
 
10
+ 0.9310 ∗ 0.9519 ∗ 0.05 ∗ 0.9970 +
1
 
10
+ 0.9310 ∗ 0.9520 ∗ 0.9969 ∗ 0.01 =
1
= 0.1185
P[at least two defective bulbs ] = 1 − (0.08585 + 0.1185) = 0.79565