MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

Let (\Omega, F, P) be a probability space, which may have atoms (important), S be a set of measure-preserving transformations T:\Omega\to\Omega, that is, such that preimage T^{-1}(A) is measurable whenever A is measurable, and P(A)=P(T^{-1}(A)). Then, obviously, random variables X and Y given by Y(\omega)=X(T(\omega)), T\in S, have the same distribution. I need the fact that the converse also true: if X and Y have the same distribution, then Y(\omega)=X(T(\omega)), for some T\in S. I am sure that this fact should be known, therefore do not want to reprove it. In the worst case, it should be a corollary from some known theorems, with 2-3 line proof. The question is: is it indeed known, and if so, where to refer?

share|cite|improve this question

It is not true, at least as stated here.

For example, if $\Omega = [0,1]$ with Lebesgue measure, then $X(\omega)=\omega$ and $Y(\omega)=2\omega\text{ mod }1$ have the same distribution, and there exists a transformation that maps $X$ to $Y$, but it is not invertible, and there is no measure-preserving transformation from $Y$ to $X$.

Indeed, if $X(\omega) = Y(T(\omega))$, then $T(\omega)$ should select either $\omega/2$ or $(\omega+1)/2$ in a measure-preserving way. $P[T^{-1}([0,1/2])] = 1/2$, but note the "squeezing" map $\omega \mapsto \omega/2$ from $T^{-1}([0,1/2])$ to $[0,1/2]$ obviously does not preserve Lebesgue measure.

share|cite|improve this answer
Thabk you! I will update the question. – bogdan Aug 23 '12 at 10:43
I don't understand the question now. What is this $S$ for? And I believe that my counterexample applies here for trivial reasons: if there is no transformation at all, then surely there is no transformation in $S$, right? And by the way, if you place $'s around your LaTeX code, it would be easier to read. – Alexander Shamov Aug 23 '12 at 16:10

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.