F.

Borgonovo -

Solutions of Problems on Probability theory

Chapter 1
P.1.1 Rolling three dices, evaluate the probability of having k equal faces, with k [0; 2; 3].
Solution.
P (0) =

(6)3
= 0, 5
63

P (2) =

635
= 0, 416
63

P (3) =

6
= 0, 027
63

P.1.2 Rolling a dice three times, evaluate the probability of having at least one 6.
Solution.
p=1

53
= 0, 42129..
63

P.1.3 Assuming women and men exist in equal number, and assuming that 5% of the men are
colour blind and that 0,25% of the women are colour blind, evaluate the probability that a
person drawn at random is colour blind. Then evaluate the probability that, having drawn
a colour-blind person, this is a male.
Solution. By the total probability theorem
P (D) =

1 0.25
1 5
+
= 0.02625
2 100 2 100

Using Bayes Theorem

P (M |D) = P (D|M )

5
1/2
P (M )
=
= 0.952
P (D)
100 0.02625

P.1.4 Drawn a card from a deck of 52 cards, verify wheter the following events are statistically
independent:
a) A = {drawing of a picture card};

B = {drawing of a hearth card}

b) What if the king of hearths is missing from the deck of cards?

c) What if a card, at random, is missing?
Solution. a)

independent

b) not independent

c) of the women

P.1.5 A dice A has four red faces and two white faces. A dice B, vice-versa, has two red faces and
four white faces. You ip a coin once, if heads the game continues with diceA, otherwise it
continues with dice B. a) On rolling the dice, what is the probability that a red face appears
on the dice? b) and at the second rolling of the same dice? c) If the rst two rollings show
a red face, what is the probability that also on the third rolling is red? d) If the rst n
rollings show a red face, what is the probability that you are using dice A?
Solution.
a)

1
2

b)

1
2

c)

3
5

d)

2n
1 + 2n

F. Borgonovo -

P.1.6 An urn contains two white balls and two black. A ball is drawn and replaced with a ball of
a dierent colour. Then a second ball is drawn. Calculate the probability p that the rst
extracted was white, when the second is white.
Solution: 1/4
P.1.7 The probabilities that three dierent archers, A, B hit the mark, independently of one another, are respectively 1/6, 1/4 and 1/3. Everyone shoots an arrow. a) Find the probability
that only one hits the mark. b) If only one hits the mark, what is the probability he is
archer A?
Solution:
a)

31
72

b)

6
31

P.1.8 A duel among three people A, B and C is carried out according to the Russian roulette. A
six round revolver is loaded with two cartridges. The duelists pass cyclically the weapon,
spinning the cylinder every time (so that each duelist has 1/3 probability of being on a
loaded chamber) and shooting themselves as long as only one remains alive. Assuming that
A is the rst, what is the probability that each duelist is the rst to die? b) and to win?
Solution:
9
6
4
pB =
pC =
19
19
19
56
69
84
pB =
pC =
2.
pA =
209
209
209
You have at least two ways to get at the result. The one is to evaluate all the sequence.
For example, if pA is the probability that A dies rst we have
1.

pA =

pA = 1/3 + (2/3)3 (1/3) + (2/3)6 (1/3) + .... = (1/3)

(8/27)i = 9/19

i=0

The second one is to observe that

pB = (2/3)pA ;

pC = (2/3)2 pA .

Since we have also

pA + pB + pC = 1
this provide an equation whose solution is pA , and the others follow.
P.1.9 From a deck of 52 cards we draw two cards. Find the probabilities of the following events
A = { rst card is a King; the second gure} ={K1 ; F2 } B = { at least one gure}
Solution: We can count favorable outcomes in the space of outcomes. Alternatively, we
can use the conditional probability denition:
p(A) = P (K1 ; F2 ) = P (K1 )P (F2 |K1 ) =

p(B) =

12 12 12 11
+

= 0, 411 . . .
52 52 52 51

4 11
= 0, 0165 . . .
52 51

F. Borgonovo -

Chapter 2
C
, determine the relationship between C e in order to
+ x2
make f (x) a pdf. (Cauchy). (3.1)

P.2.1 Given the function f (x) =

Solution. C =
P.2.2 A point P uniformly chosen in a square of Side L centered at the origin and the x-axis.
Find the pdf of RV X, coordinate of the orthogonal projection of P on the horizontal axis.
Solution.
fX (x) =

1
,
L

L/2 x L/2

P.2.3 A point P uniformly chosen in a circle of radius R centered at the origin and the x-axis. Find
the pdf of RV X, coordinate of the orthogonal projection of P on the horizontal axis.(3.2)
Solution.
fX (x) =

2
R

x
1 ( )2 ,
R

r/2 x R/2

P.2.4 Find the rst order moment of pdf f (x) = 2 xex , x 0, and 0 elsewhere.
Solution. 2/.
P.2.5 Find the rst order moment of the integer distributions
1. P (X = k) = (1 p)k1 p,
2. P (X = k) = (1

p)k p,

k 1;
k 0.

Solution. 1/p, and (1 p)/p.

P.2.6 2 points are chosen uniformly and independently in a segment of length L. Find the pdf of
RV X distance to the origin of the point closest to the origin. Find the joint pdf of (X, Y )
where Y is distance to the origin of the point farthest to the origin. Extend the result to
the case of n points.(3.6)
Solution.
As usual
P (x < X x + x) = P (one point in x; the other beyond x)
We get
fX (x) =

2 Lx
,
L L

0xL

For the joint pdf we have

P (x < X x + x; y < Y y + y) = 2
fXY (x, y) =

2
,
L2

x<y

x y
,
L L

x<y

F. Borgonovo -

P.2.7 2 points are chosen uniformly and independently in a circle of radius R. Find the pdf of RV
X distance to the center of the point closest to the center.(3.8)
Solution.


x
x
fX (x) = 2n 2 1 ( )2
R
R

n1

(0 x R)

P.2.8 Take a number X from one to six, throw three dices. You win C if X appears once, 2C if
X appears twice, 3C if it appears three times, and you lose C if X does not appear. Check
whether this is a fair game. (3.12)
Solution.
The number of possible outcomes is 63 = 216. X appears once with probability 3 25/216,
twice with probability 3 5/216, and thrice with probability 1/216. The probability of
winning is the sum, i.e, 91/216; therefore your loss is on the average (125/216)C. On the
other side, if your win is, on the average,
(75/216)C + (15/216)2C + (1/216)3C = (108/216)C
Then on the average you loose (17/216)C at each bet.
P.2.9 Assume the RV X, lifespan of a component, is uniform in [0; L]. We know that the component age is z; nd the pdf of its lifespan. Find the pdf of Y , remaining lifespan.
Solution.
We look for
fX (x|age=z) = fX (x|X > z) = lim

= lim

P (x < X x + x)
,
xP (X > z)

P (x < X x + x; X > z)
=
xP (X > z)

x>z

Therefore
fX (x|X > z) =

1
fX (x)
=
,
P (X > z)
Lz

zxL

Then
fY (y) = fX (y + z|X > z) =

1
,
Lz

0xLz

P.2.10 Repeat the previous exercise assuming that the pdf of X is negative exponential. Find the
fair amount a a customer of age z must pay to get a capital C if he dies before the year.
Solution.
As in the previous case we have
fX (x|X > z) =

ex
fX (x)
= z ,
P (X > z)
e

x>z

F. Borgonovo -

fY (y) = fX (y + z|X > z) =

e(y+z)
= ey ,
ez

x>0

Note that the remaining life span has still the same pdf as the original lifespan. This is
because the negative exponential is memoryless.
The probability of dying within a year is P (X < 1) = 1 e . The fair amount a is such
that
ae = C(1 e )
P.2.11 Check whether functions of x and y below can represent joint pdfs and if so check whether
X and Y are statistically independent. (5.1)
1. f (x, y) = 4xy

(0 x 1; 0 y 1),

2. f (x, y) = 8xy

(0 x y; 0 y 1),

3. f (x, y) = 4x2 y

(0 x 1; 0 y 1)

P.2.12 A person in phone booth makes a phone call whose duration is represented by RV X, with
negative exponential pdf with mean value 1/. A second person comes after a time Y . RV
negative exponentially with average 1/, independent of X. Find the pdf of RV W , the
time the latter has to wait to the end of the call. (5.6)
Solution.
If Y > X the second person arrives when the rst has already nished his phone call and,
therefore W = 0. On the other side, we take the condition Y = y, Y < X. RV W is then
the remaining lifespan of Problem 2.10. By this problem we have learned that, with the
negative exponential pdf, the remaining lifespan has the same pdf. Therefore
fW (w|Y = y; Y < X) = ew ,

y0

that doesnt depend on y. The, using the total probability Theorem

fW (w|Y < X) = ew ,

y0

Finally
fW (w) = (w)P (Y > X) + ew P (Y < X),

y0

See Problem 2.13 to see P (Y > X).

P.2.13 Given two independent RVs X, and Y , nd the probability of the event {Y X} when
1. fX , fY are uniform within intervals respectively [1; 3], [0; 4];
2. fX , fY with the same pdf (you do not need to know the pdf);
3. fX , fY are negative exponentials with parameters and ;

F. Borgonovo -

What about event {Y X/2}?

Solution.
The rst way nds event {Y X} in the plane x, y, and then integrates the joint pdf in
such event. Case 1) is very simple since the joint distribution is uniform and the integration
takes the volume of the regular solid where the base is the area of the event. The area of
the vent is the portion of the rectangle of base [1; 3] and height [0; 4] that lies beneath
straight line y = x. The height of the pdf is 1/16 .
The second way uses conditioning.
P (Y X) =

P (Y X|X = x)fX (x)dx

In the rst case we have

P (Y X|X = x) = x/4,
P (Y X|X = x) = 0,
P (Y X) =

0x3
1 x 0

 3
0

x/4 1/4dx = 9/32

In case 2, since both RVs obeys the same law, outcomes X > Y or Y > X are equally
probable.
In case 3 we use conditioning again.
P (Y X) =


P (Y X|Y = y)fY (y)dy =

ey ey dy =


0


0

P (X y|Y = y)fY (y)dy =

( + )e(+) dy =

P.2.14 Find the pdf of RV Z = min(X, Y ), where X and Y are two independent negative exponential RVs with parameters and respectively. (Hint: observe that min(X, Y ) > z if x > z
and Y > z. Also, we may take the condition Y = y...)
Solution.
The suggestion says
P (Z > z) = P (min(X, Y ) > z) = P (X > z; Y > z) = P (X > z)P (Y > Z) = ez ez
Therefore
P (Z > z) = e(+)
or
fZ (z) = ( + )e(+)z ,

z0

Take notice: The minimum of two negative exponential RVs is again a negative exponential
RV with parameters sum of parameters.

F. Borgonovo -

P.2.15 Take interval [0, X], where X is a RV Erlang-2. Then take a point P uniformly within the
preceding interval. Find the pdf of Y , length of 0P .
Solution.
We use the Total Probability Theorem
fY (y|X = x) =

1
,
x

yx

fY (y) =

fY (y|X = x)fX (x)dx =



1 2 x
xe
dx =
ex dx =
y

= (1 FX (y) = ex
The solution could be expected. Why?
P.2.16 n points are uniformly taken within [0; T ]. Find the probability that k out of n point lie
within an interval [0; X] where RV X is uniform in [0; T ].
Solution.
The probability can be written as:
 T

P (N (X) = k/N (T ) = n) =

P (N (X) = k/N (T ) = n, X = x)fX (x)dx =



 
 T
T x nk x k 1
n
0

dx =

1
n+1

It seems strange that the solution does not depend on k. This becomes apparent if we
solve the problem in this other way. The extreme of interval X is itself a uniform point in
[0, T ], exactly as the others n. Therefore, the sought probability is the probability that this
boundary point lies the k + 1-th position out of n + 1. But all positions are equally probable
and therefore the sought probability is 1/(n + 1).
P.2.17 Two RVs X and Y are independent and uniformly distributed in [0; 1]. Find fX (x|X > Y ),
fXY (x, y|X > Y ) and P (X > 2Y |X > Y ).
Solution.
fX (x|X > Y ) = P (X > Y |X = x)

1
fX (x)
=x
= 2x,
P (X > Y )
1/2

fXY (x, y|X > Y ) = P (X > Y |X = x, Y = y)

fXY (x, y)
=2
P (X > Y )

0x1

x>y

Since the above is uniform in 0 y x 1, P (X > 2Y |X > Y ) is simply the ratio of the
areas of events X > 2Y and X > Y , equal to 1/2.

F. Borgonovo -

Chapter 3
P.3.1 fX , fY are uniform within intervals respectively [0; 5] [3, 1]. Find the pdf of RVs (6.1)
1. Z = X + Y
2. W = X Y
Solution.

(1/10)(z + 3)

fz (z) =

3 z 1
1 z 2
2z4

1/5

(1/10)(4 z)

(1/10)(z 1)

fW (w) =

1/5

(1/10)(8 w)

1w3
3w6
6w8

P.3.2 Let X e Y be independent RVs with negative exponential pdfs and average value 1 . Find
the pdf of RVs (6.2)
1. Z = X Y
Y
2. W = X +
2
Solution.
a) fZ (z) =

1 |z|
e
(Laplace)
2

b) fW (w) = 2(ez e2z )

(z > 0)

P.3.3 Find P (Z = n) where Z = X + Y is the sum of the numbers that appear in the rolling of
two dices. (6.6) Solution.

P (Z = n) =

n1

(2 n 7)

13 n

(7 n 12)

36

36