Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I'm trying to see what conditions are necessary such that the following holds for a general random variable $X$: $$\mathrm{Pr}[X\ge E[X]] \ge \frac{1}{2}.$$ This seems to be intuitively true if the range of $X$ is somewhat "uniform". Is there anything more specific known for this condition to hold?

share|improve this question

1 Answer 1

up vote 4 down vote accepted

One would ask that the distribution of $X$ is symmetric with respect to its mean, that is, that ($X$ is integrable and that) there exists $m$ such that $m-X$ and $X$ are equidistributed.

Then the only possible value of $m$ (when $X$ is integrable) is $m=2\mathbb E(X)$ and the inequality $\mathbb P(X\geqslant \mathbb E(X))\geqslant\frac12$ follows from the remark that $\mathbb P(X\geqslant \mathbb E(X))=\mathbb P(A)+\mathbb P(B)$ where $A=[X\gt E(X)]$, $B=[X=E(X)]$, $C=[X\lt E(X)]$, and from the identities $\mathbb P(A)+\mathbb P(B)+\mathbb P(C)=1$ (always true) and $\mathbb P(A)=\mathbb P(C)$ (due to the symmetry).

In the absence of symmetry, such a bound cannot hold in general. For a simple counterexample, consider $X$ such that $\mathbb P(X=-1)=x/(1+x)$ and $\mathbb P(X=x)=1/(1+x)$ for some $x\gt0$. Then $\mathbb E(X)=0$ and $\mathbb P(X\geqslant0)=1/(1+x)$ can take every value in $(0,1)$.

share|improve this answer

Your Answer

 
discard

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.