Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

What is the conditional expectation of $\mathbb{E}(X\mid\sin(X))$ if $X$ is uniformly distributed on $[0,\pi]$?

Intuitively I expect that it is constant and equal to $\frac{\pi}{2}$, since the Borel set generated by $\sin(X)$ is symmetric around $\frac{\pi}{2}$.

Is that true? What would be the formal proof?

share|cite|improve this question
up vote 2 down vote accepted

Is that true?

Yes it is.

What would be the formal proof?

A formal proof would be to show that $\mathbb E(Xu(\sin X))=\frac\pi2\mathbb E(u(\sin X))$ for every bounded measurable function $u$.

Recall that in full generality, $\mathbb E(Y\mid Z)$ may be defined as the only $Z$-measurable random variable $T$ such that $\mathbb E(Yu(Z))=\mathbb E(Tu(Z))$ for every bounded measurable function $u$. Use this for $Y=X$, $Z=\sin X$ and $T=\frac\pi2$.

A painless way to show that $\mathbb E(Xu(\sin X))=\frac\pi2\mathbb E(u(\sin X))$ is to start with the fact that $\pi-X$ and $X$ are identically distributed hence $(\pi-X,\sin(\pi-X))=(\pi-X,\sin X)$ and $(X,\sin X)$ are identically distributed. Thus, $(*)=\mathbb E(Xu(\sin X))$ solves the identity $$ (*)=\mathbb E((\pi-X)u(\sin(X)))=\pi\mathbb E(u(\sin X))-(*), $$ and the proof is complete.

share|cite|improve this answer
I dont understand your first equation: $\mathbb{E}(Xu(\sin(x)))=\frac{\pi}{2}\mathbb{E}(u(sin(X)))$. – johanMozart Nov 18 '12 at 12:44
See Edit. $ $ $ $ – Did Nov 18 '12 at 12:47
I have never seen that definition before. Do you have a reference, where the equivalence of the definitions is explained? – johanMozart Nov 18 '12 at 13:08
That this version implies the one with events (the only fact you need in the present case) is obvious: take only functions $u$ which are indicator functions. // Reference: a perky answer would be any reference which does a good job at introducing conditional expectations... :-) Anyway you could try David Williams, Probability with martingales (small book, extremely pleasant to read, simple to follow). – Did Nov 18 '12 at 13:16

$\frac{\pi}{2}$ is correct, yes.

Proof: We want to use the following theorem: $$\mathbb{E}(X|\mathcal{F}) = Y \Leftrightarrow \forall G \in \mathcal{G}: \int_G Y \, d\mathbb{P} = \int_G X \, d\mathbb{P} \qquad (\ast)$$

where $G$ is a generator of the $\sigma$-Algebra $\mathcal{F}$. In this case we can choose $\mathcal{G} = \{\{\sin X \leq y\}; y \in \mathbb{R}\}$. Let $G = \{\sin X \leq y\} \in \mathcal{G}$, $y \in [0,1]$, then

$$\int_G X \, d\mathbb{P} = \frac{1}{\pi} \int_0^\pi x \cdot 1_{(-\infty,y]}(\sin x) \, dx = \frac{1}{\pi} \cdot \left( \int_0^{\arcsin y} x \,dx + \int_{\pi-\arcsin y}^{\pi} x \, dx \right) \\ = \ldots = \arcsin y = \ldots = \frac{1}{\pi} \cdot \int_0^\pi \underbrace{\frac{\pi}{2}}_{=:Y} 1_{(-\infty,y]}(x) \, dx = \int_G \frac{\pi}{2}$$

Clearly the equality holds also for $y \in \mathbb{R} \backslash [0,1]$. From $(\ast)$ follows $\mathbb{E}(X|\sin X)=\frac{\pi}{2}$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.