Let $(\Omega,\mathcal{F})$ denote some measurable space. Let $P_1$ and $P_2$ denote respectively two probability measures. Now let $\mathcal{G}$ be some sub sigma-algebra of $\mathcal{F}$. Given a positive integrable random variable $X$, we can define respectively the conditional expectation

$$Y_1=E^{P_1}[X|\mathcal{G}],~ Y_2=E^{P_2}[X|\mathcal{G}]$$

Now for some $0<\alpha<1$, we can define a new probability measure $P=\alpha P_1+(1-\alpha)P_2$, then we get


Now my question is whether we can prove

$$\operatorname{esssup}{}_P(Y)\le \alpha \operatorname{esssup}_{P_1}(Y_1)+(1-\alpha)\operatorname{esssup}_{P_2}(Y_2)?$$

Here the definition of $\operatorname{esssup}_{Q}(\cdot)$ w.r.t some probability $Q$ can be found here:


Thanks a lot for the help!


1 Answer 1


Unless I am missing something, this is false, and the conditional expectation is a red herring.

Let $\Omega = \{a,b\}\newcommand{\esssup}{\operatorname{ess\,sup}}$ be a set with two points and $\mathcal{G} = \mathcal{F} = 2^\Omega$. Set $P_1 = \delta_a$ and $P_2 = \delta_b$, and $X = 1_{\{a\}}$. Since $\mathcal{G} = \mathcal{F}$ the conditional expectation doesn't do anything, and we have $Y_1 = Y_2 = Y = X$. (Technically $Y_1(b)$ and $Y_2(a)$ are undefined but it won't actually matter.) Then it is easy to see that $$\begin{align*} \esssup{}_{P_1} Y_1 &= 1 \\ \esssup{}_{P_2} Y_2 &= 0 \\ \esssup_P Y &= 1 \end{align*}$$ so your proposed inequality reads $1 \le \alpha$.

  • $\begingroup$ Thanks so much for your suggestion. But could we prove the following inequality? $\endgroup$
    – CodeGolf
    Apr 1, 2014 at 8:50
  • $\begingroup$ $$\operatorname{esssup}{}_P(Y)\le \max\Big( \operatorname{esssup}_{P_1}(Y_1), \operatorname{esssup}_{P_2}(Y_2)\Big)$$ $\endgroup$
    – CodeGolf
    Apr 1, 2014 at 8:51
  • $\begingroup$ @codegolf: This should be asked as a new question, and I think it would be better for math.stackexchange.com than for MathOverflow. $\endgroup$ Apr 1, 2014 at 13:09
  • $\begingroup$ It is better, thx! $\endgroup$
    – CodeGolf
    Apr 1, 2014 at 13:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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