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

If I have a random variable X with probability mass function, $f(u)$, would somebody give me an intuitive reason as to why the expected value of X is $\int^t_0 x f(x) dx$ ?

share|cite|improve this question
I think quoting the pdf is not so relevant to your actual question – Please Delete Account Apr 14 '11 at 7:56
Agreed, I've removed it now. Thanks. – Undercover Mathematician Apr 14 '11 at 7:58
Do you understand the discrete case, $E = \sum x p(x)$? – Alon Amit Apr 14 '11 at 8:04
up vote 1 down vote accepted

The integral there should be not on $[0,t]$ but on the set of possible values of $X$. By taking this integral you assign to each value $x$ its density (like an importance, likelihood) and then taking the average (so, the integral) such that you are summing all possible values of a random variable $X$ taking into account their likelihood (importance).

E.g. for the discrete case these arguments fall into the sum $$ \mathsf{E}[X] = \sum\limits_i x_i P(X = x_i). $$

On the other hand for the continuous case $P(X = x) = 0$ usually (if there are no singularities). That's why we use density rather than probability and integral rather than the sum.

share|cite|improve this answer

Let us suppose for simplicity that the function $f(x)$ is $0$ outside the interval $[a,b]$.

Divide the interval into $n$ subintervals, each of length $h =(b-a)/n$. Let $x_0=a$, $x_1=a+h$, $x_2=a+2h$, and so on up to $x_{n}=a+nh=b$. The smooth curve $y=f(x)$ can be approximated by the broken line curve $y=f(0)$ on the interval $[x_0,x_1)$, $y=f(x_1)$ on the interval $[x_1,x_2)$, and so on up to $y=f(x_{n-1})$ on the last interval. The probability that $X$ lies in the interval $[x_i,x_{i+1})$ is approximately $f(x_i)h$.

Our random variable behaves much like the discrete random variable which is $x_0$ with probability $f(x_0)h$, $x_1$ with probability $f(x_1)h$, and so on up to $x_{n-1}$ with probability $f(x_{n-1})h$. This discrete random variable has expectation $$\sum_{i=0}^{n-1} x_i f(x_i) h=\sum_{i=0}^{n-1} x_i f(x_i) (b-a)/n$$ Now let $n \to\infty$. The approximation gets better and better. As $n \to\infty$, the displayed sum above approaches the Riemann Integral of $xf(x)dx$ from $a$ to $b$.

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.