Problem Set 3
1. Let (Ω, F, P ) be the Lebesgue measure on [0, 1] and define
(
1, if ω = 21n for some n ∈ N,
X(ω) =
0, otherwise.
Show that X is a random variable and compute E(X).
2. Let (Ω, F, P ) be the Lebesgue measure on [0, 1], and set
(
2ω, 0 ≤ ω ≤ 21 ,
X(ω) = 1
1, 2 < ω ≤ 1.
(a) Show that X is a random variable.
(b) Compute E(X).
R
(c) Compute A X dP , first for A = [0, 1], and then for A = [0, 1/2].
3. (a) Show that if X is discrete (not necessarily simple) then
X
E(X) = xi P (Ai ), where Ai = {X = xi }.
i
(b) Show that if X is non-negative integer valued, then
∞
X
E(X) = P (X ≥ k).
k=1
4. Let X be a random variable defined on the probability space (Ω, F, P ).
(a) Show that if A and B are disjoint events, then
Z Z Z
X dP = X dP + X dP.
A∪B A B
Z
(b) What is X dP for arbitrary events A and B?
A∪B
(c) Show that if A and B are independent events, then
Z Z Z
X dP = X dP dP.
A∩B A B
5. Let X1 , X2 , . . . be non-negative, independent and identically distributed ran-
dom variables with mean µ, and let N be an integer-valued random variable
with mean m, with N independent of all the Xi . Let
∞
X
S = X1 + · · · + XN = Xi I{N ≥i} .
i=1
Compute E(S) in terms of µ and m.
1
Problem 6
Suppose that (Ω, F , P ) is a probability space and X : Ω → R is a random
variable. Define the positive and negative parts of X by
X + = max{0, X}, X − = max{0, −X}.
Show that X + and X − are random variables and that:
(a) 0 ≤ X + ≤ |X|,
(b) 0 ≤ X − ≤ |X|,
(c) X + − X − = X,
(d) X + + X − = |X|.
Problem 7
Show that a constant function X is a random variable with respect to any
σ-field.
Problem 8
Let Ω = {1, 2, 3, 4} and let
F1 = {∅, Ω, {1}, {2, 3, 4}}, F2 = {∅, Ω, {1}, {2, 3, 4}, {1, 2}, {3, 4}, {2}, {1, 3, 4}}.
Verify whether the following functions are random variables with respect to the
σ-fields F1 and F2 :
(a) X ≡ 1
(b) X(ω) = 1 + ω for each ω ∈ Ω
(c) X(ω) = 1 if ω = 1, 2 and X(ω) = ω if ω = 3, 4
(d) X(ω) = 1 if ω = 1, 3 and X(ω) = 2 if ω = 2, 4
(e) X(1) = 1 and X(ω) = 2 if ω = 2, 3, 4
(f) X(ω) = 1 + ω if ω = 1, 2 and X(ω) = 2 + ω if ω = 3, 4
(g) X(ω) = 1 + ω if ω = 1, 2 and X(ω) = 1 if ω = 3, 4
Problem 9
Find the smallest σ-fields on Ω = {−2, −1, 0, 1, 2} such that the following func-
tions are random variables:
X(ω) = ω 2 ; X(ω) = 1 + ω; X(ω) = |ω|; X(ω) = 2ω.
For the above random variables X, compute E(X), (X) and σ(X) if:
(i) P ({ω}) = 0.2 for each ω ∈ Ω,
(ii) P ({−2}) = P ({−1}) = 0.2, P ({0}) = 0.4, P ({1}) = P ({2}) = 0.1.
2
Problem 10
Suppose that Ω = {−2, −1, 1, 2} and X(ω) = ω 2 , Y (ω) = 1 + ω. Find the
partitions PX and PY defined by X and Y , respectively. Compute first E(X),
E(Y ) and E(Y 2 ) and then compute
a) E(X | PY );
b) E(Y | PX );
c) E(Y 2 | PX );
if P ({−2}) = P ({2}) = 0.2, P ({−1}) = P ({1}) = 0.3.
Problem 11
Suppose
ξ(HHT) = ξ(HTH) = 10 + 20 = 30,
ξ(THT) = 10 + 50 = 60,
ξ(THH) = 20 + 50 = 70.
Show that E(ξ | D) = E(ξ | G).
Hint: The definition of E(ξ | D) involves an integral and so does the defini-
tion of E(ξ | G). How are these integrals related?
Problem 12
Show that if (
1 for ω ∈ A,
1A (ω) =
0 for ω ∈
/A
(the indicator function of A), then
E(1A | B) = P (A | B),
where
P (A ∩ B)
P (A | B) =
P (B)
is the conditional
R probability of A given B.
Hint: Write B 1A dP as P (A ∩ B).
Problem 13
Show that if η is a constant function, then E(ξ | η) is constant and equal to
E(ξ).
Hint: The event {η = c} must be ∅ or Ω for any c ∈ R.
3
Problem 14
Show that (
P (A | B) if ω ∈ B,
E(1A | 1B )(ω) =
P (A | Ω \ B) if ω ∈
/ B,
for any B such that 0 ̸= P (B) ̸= 1.
Problem 15
Assuming that η is a discrete random variable, show that
E(E(ξ | η)) = E(ξ).
R R
Hint: Observe that B E(ξ | η) dP = B ξ dP for any event B on which η
is constant. The desired equality can be obtained by covering Ω by countably
many disjoint events of this kind.
Problem 16
Show that if G = {∅, Ω}, then E(ξ | G) = E(ξ) a.s.
Hint: What random variables are G-measurable if G = {∅, Ω}?
Problem 17
Show that if ξ is G-measurable, then E(ξ | G) = ξ a.s.
Hint: The conditions of Definition 2.4 are trivially satisfied by ξ if ξ is
G-measurable.
Problem 18
Show that if B ∈ G, then
E (E(ξ | G)1B ) = E(ξ1B ).
Hint: The conditional expectation on either side of the equality satisfies the
defining property of the conditional expectation.