Intereting Posts

Logic and number theory books
Characterization of hitting time's law. (Proof check)
How to find the equation of a line tangent a circle and a given point outside of the circle
What does the math notation $\sum$ mean?
On Shanks' quartic approximation $\pi \approx \frac{6}{\sqrt{3502}}\ln(2u)$
The rank of a linear transformation/matrix
Prove that Pascals triangle contains only natural numbers, using induction.
For which values of $\alpha$ and $\beta$ does the integral $\int\limits_2^{\infty}\frac{dx}{x^{\alpha}ln^{\beta}x}$ converge?
How many ways are there to shake hands?
How to prove that $\int_0^{\infty} \log^2(x) e^{-kx}dx = \dfrac{\pi^2}{6k} + \dfrac{(\gamma+ \ln(k))^2}{k}$?
Summing digits of powers of 2 to get 1 2 4 8 7 5 pattern
Proving equivalence relations
Proving a property of an ellipse and a tangent line of the ellipse
Why is exponentiation defined as $x^y=e^{\ln(x)\cdot y}$?
Why don't we define division by zero as an arbritrary constant such as $j$?

My problem is the following:

$X$ and $Y$ are two random variables and $\mathcal{F}$ is a $\sigma$-algebra. Given that $X$ and $Y$ are independent, and that $X$ is independent of $\mathcal{F}$, can I affirm that

$$\mathbb{E}(XY \mid \mathcal{F}) = \mathbb{E}(X) \cdot \mathbb{E}(Y \mid \mathcal{F})?$$

My intuition is yes but I did not manage to find a formal proof. Any idea ?

Thanks a lot.

- What is the expected value of the number of anchors of $S$?
- Find $E(\max(X,Y))$ for $X$, $Y$ independent standard normal
- Show that $\mathbb{E}(T) = \sum\limits^\infty_{k=1}\mathbb{P}(T \geq k)$ for $T$ nonnegative integer valued
- If U1, U2, U3 are iid Uniform, then what is the probability of U1+U2>U3?
- Expectation of an increasing transformation of a random variable
- Coupon collector problem doubts

- Are X and |X| independent? where $f(x)=\exp(-|x|)/2$
- Intuitive reason why sampling without replacement doesnt change expectation?
- Show $P(A\cap B)\geq P(A)+P(B)-1$
- Probability that 2 appears at an earlier position than any other even number in a permutation of 1-20
- On “for all” in if and only if statements in probability theory and stochastic calculus
- Managing a bond fund: Simulating the maximum of correlated normal variates
- Does convergence in probability preserve the weak inequality?
- Using multivariate hypergeometric distribution to compute probability of multiple events
- Combination with repetitions.
- Should I put number combinations like 1111111 onto my lottery ticket?

No, in general the statement is not correct.

**Example** Consider sets $A_1,A_2,A_3$ such that $\mathbb{P}(A_1 \cap A_2 \cap A_3) = 0$ and $\mathbb{P}(A_i)>0$, $\mathbb{P}(A_i \cap A_j) = \mathbb{P}(A_i) \cdot \mathbb{P}(A_j)$ for any two $i,j \in \{1,2,3\}$. Set $X:=1_{A_1}$, $Y=1_{A_2}$, $\mathcal{F} := \sigma(1_{A_3})$. Then $X$ and $Y$ are independent, as well as $X$ and $\mathcal{F}$, so the assumption on the independence is satisfied. For $F:= A_3$ we have

$$\int_F X \cdot Y \, d\mathbb{P} = \int 1_{A_1} \cdot 1_{A_2} \cdot 1_{A_3} \, d\mathbb{P} = \mathbb{P}(A_1 \cap A_2 \cap A_3) = 0 \tag{1} $$

On the other hand,

$$\int_F \mathbb{E}(X) \cdot \mathbb{E}(Y \mid \mathcal{F}) = \mathbb{P}(A_1) \cdot \underbrace{\mathbb{E}(Y)}_{\mathbb{P}(A_2)} \cdot \mathbb{P}(A_3)>0 \tag{2}$$

where we used that $Y$ and $\mathcal{F}$ are independent (hence $\mathbb{E}(Y \mid \mathcal{F})=\mathbb{E}(Y)$). Since by the definition of conditional expectation

$$0 \stackrel{(1)}{=} \int_F X \cdot Y \, d\mathbb{P} \stackrel{!}{=} \int_F \mathbb{E}(X \cdot Y \mid \mathcal{F}) \, d\mathbb{P}$$

we conclude from $(2)$ that

$$\mathbb{E}(X \cdot Y \mid \mathcal{F}) \neq \mathbb{E}(X) \cdot \mathbb{E}(Y \mid \mathcal{F})$$

**But:** If $\sigma(X)$ is independent of $\sigma(Y,\mathcal{F})$ (i.e. the smallest $\sigma$-algebra containing $\mathcal{F}$ and $\sigma(Y)$), then the claim holds as the following proof shows:

Denote by $\mathcal{G} := \sigma(\mathcal{F},\sigma(Y))$ the $\sigma$-algebra generated by $\mathcal{F}$ and $Y$. Since $\sigma(X)$ and $\mathcal{G}$ are independent, we have

$$\mathbb{E}(X \mid \mathcal{G}) = \mathbb{E}(X)$$

Consequently, by the tower property

$$\mathbb{E}(X \cdot Y \mid \mathcal{F}) = \mathbb{E}(Y \cdot \underbrace{\mathbb{E}(X \mid \mathcal{G})}_{\mathbb{E}(X)} \mid \mathcal{F}) = \mathbb{E}(X) \cdot \mathbb{E}(Y \mid \mathcal{F})$$

- Mandelbrot set: periodicity of secondary and subsequent bulbs as multiples of their parent bulbs
- Small integral representation as $x^2-2y^2$ in Pell's equation
- Multiplication partitioning into k distinct elements
- Quadratic optimisation with quadratic equality constraints
- Unusual pattern in the distribution of odd primes
- How to solve $e^x=x$?
- Extending Homomorphism into Algebraically Closed Field
- Another simple rule satisfied by the Fibonacci $n$-step constants?
- Does sequences in $\ell^p$ belong to some $\ell^q$ for $q<p$?
- $F$ has undecidable positive existential theory in the language $\{+, \cdot , 0, 1, t\}$
- Borel-Cantelli lemma problem
- metric on the Euclidean Group
- Why can quotient groups only be defined for subgroups?
- Integrate $\displaystyle \int_{0}^{\pi}{\frac{x\cos{x}}{1+\sin^{2}{x}}dx}$
- The equation $3^n+4^m=5^k$ in positive integer numbers