1
$\begingroup$

This is an example in the book (A First Course in Probability by Sheldon Ross).

A stick of length 1 is split at a point U that is uniformly distributed over $(0,1)$. Determine the expected length of the piece that contains the point $0 \leq p \leq 1 $.

The problem with this is I don't know how would I go about solving this. They have solved this in the book but I do not understand their solution.

For example I don't know how to setup the probability density formula and then to find the expected value from there.

$\endgroup$
6
  • 1
    $\begingroup$ I get $p(1-p)+\frac12$, is that the right answer? $\endgroup$
    – bof
    Apr 13 '15 at 6:16
  • $\begingroup$ Yes can you please explain how you got this? $\endgroup$ Apr 13 '15 at 6:18
  • 2
    $\begingroup$ By definition of $U$ being uniformly distributed, the answer is $$\int_0^p(1-u)du+\int_p^1udu=\tfrac12+p-p^2.$$ If $U$ has density $f_U$ on $(0,1)$, consider $$\int_0^p(1-u)f_U(u)du+\int_p^1uf_U(u)du.$$ $\endgroup$
    – Did
    Apr 13 '15 at 6:37
  • $\begingroup$ I understood your solution but how have you taken the function value to be 1 in both cases? $\endgroup$ Apr 13 '15 at 6:39
  • $\begingroup$ Which function value? $f_U$? This is the definition of being uniformly distributed on $(0,1)$, no? $\endgroup$
    – Did
    Apr 13 '15 at 6:39
3
$\begingroup$

Let $X$ be the length of the piece containing $p$.

Given that $U\gt p,$ then $X$ is uniformly distributed on $[p,1]$, so the conditional expectation is $\frac{p+1}2.$

Given that $U\lt p,$ then $X$ is uniformly distributed on $[1-p,1]$, so the conditional expectation is $\frac{(1-p)+1}2.$

So $$E(X)=P(U\gt p)\cdot\frac{p+1}2+P(U\lt p)\cdot\frac{(1-p)+1}2=(1-p)\cdot\frac{p+1}2+p\cdot\frac{(1-p)+1}2$$ which simplifies to $\frac12+p(1-p).$

I don't know if this is how you're supposed to do it, or if you're supposed to derive the probability distribution etc. Is my solution anything like the solution in the book?

$\endgroup$
3
  • $\begingroup$ In the book they've found the probability density formula and then calculated the expected value from there. Also in the method you've used may be correct but I don't know what condition expectation is. $\endgroup$ Apr 13 '15 at 6:37
  • $\begingroup$ Conditioning is not needed (and might be seen as an unnecessary complication). $\endgroup$
    – Did
    Apr 13 '15 at 6:38
  • $\begingroup$ @Did Right, your solution is much better. $\endgroup$
    – bof
    Apr 13 '15 at 6:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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