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

Give an example on $\Omega = \{a, b, c\}$ in which

$E(E(X|F_{1})|F_{2}) \neq E(E(X|F_{2})|F_{1})$


Obviously X is a random variable and $F_{1}$ and $F_{2}$ are sigma-algebras... but I'm not even sure how to get started on the actual example. Any help is appreciated.


share|cite|improve this question
Let $X$ be the random variable defined by $X: \Omega \to \mathbb{R}$ as $X(a) = 1$, $X(b) = 2$ and $X(c) = 3$. Let $F_1 = \{\{a\}, \{b,c \}, \emptyset, \{a,b,c \} \}$ and $F_2 = \{\{a,b,c\}, \emptyset \}$. Perhaps something can be gleaned from this. – PEV Mar 28 '11 at 21:08
Can you write down all possible $\sigma$-algebras? – Gerben Mar 28 '11 at 21:10
@PEV You are getting close, but the proposed counterexample cannot work with your $F_2$ (the trivial $\sigma$-field). The OP should consider "Why not?". – Byron Schmuland Mar 28 '11 at 21:13
@Byron I understand you wish the OP to construct an example by himself and concur, so will not give one. – Did Mar 28 '11 at 21:17
Or perhaps let $F_2 = \{ \{b \}, \{a,c \}, \emptyset, \{a,b,c \} \}$. – PEV Mar 28 '11 at 21:20
up vote 1 down vote accepted

Suppose, without loss of generality, that $\Omega = \{1,2,3\}$. Recall that ${\rm E}[X|Y]$ stands for ${\rm E}[X|\sigma(Y)]$, where $\sigma(Y)$ is the $\sigma$-algebra generated by $Y$. Let ${\rm P}$ be the uniform probability measure on $\Omega$, that is, ${\rm P}(\{\omega\})=1/3$, $\forall \omega \in \Omega$. Now, define random variables $X$, $Y$, and $Z$ as follows: $$ X(\omega)=\omega,\;\; \forall \omega \in \Omega , $$ $$ Y(1)=1, Y(\omega)=2.5, \omega \in \{2,3\}, $$ and $$ Z(\omega)=1.5, \omega \in \{1,2\}, Z(3)=3. $$ Then, $$ {\rm E}[X|Y] = Y $$ and $$ {\rm E}[X|Z] = Z. $$ It thus suffices to show that $$ {\rm E}[Y|Z] \ne {\rm E}[Z|Y]. $$ Indeed, ${\rm E}[Y|Z]$ takes the values $1.75$ and $2.5$ with probabilities $2/3$ and $1/3$, respectively, whereas ${\rm E}[Z|Y]$ takes the values $2.25$ and $1.5$ with probabilities $2/3$ and $1/3$, respectively. So, these random variables are never equal.

share|cite|improve this answer
(This is a friendly comment.) In relation with @Byron's comment to the OP, you might wish to meditate on Alex Bartel's advice given on meta, and also on the thread, called A suggestion for dealing with users who post homework problems. As I said, these are on the meta site, so you could have missed them. Please let me know what you think of it. – Did Mar 30 '11 at 6:57
@Didier: I agree I should have given a less detailed example. On the other hand, I left out many details (relative to a beginner), and the OP might need to work hard in order to work out a complete answer (and understand it). Anyway, I have found a general solution for this problem: give hints, and promise to elaborate in the future (say two weeks after the question was posted). In general, however, detailed answers may be valuable for the general SE community. – Shai Covo Mar 30 '11 at 10:47

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.