-3

Do you agree that if $(X_i)$ is a sequence of i.d.d. random variable, then for all $i$

$$\mathbb E[X_i\mid X_1,...,X_n]=\mathbb E[X_i]\ \ \ ?$$

idm
  • 11,824

1 Answers1

9

I don't agree. Given the outcomes of all $X_i$ you know exactly what the conditional expectation is, namely, $X_i$ for $i \leq n$. $$E[X_i | X_1, ..., X_n] = X_i$$ That's true regardless of if the sequence are iid or not.

Since the $X_i$ are independent, this does hold true for $i > n$ $$E[X_i | X_1, ..., X_n] = E[X_i]$$

muaddib
  • 8,267
  • 2
    This reminds me a bit of the fun task to expand the polynomial $(x-a)(x-b)(x-c)\cdots (x-z)$ :) – Hagen von Eitzen May 27 '15 at 10:34
  • @HagenvonEitzen Could you please explain it? – PSPACEhard May 27 '15 at 10:37
  • @echo What is the factor between $(x-w)$ and $(x-y)$? – Did May 27 '15 at 11:02
  • I agree with that. – muaddib May 27 '15 at 11:33
  • Sorry, my english is not very well. I'm not sure to understand everything. So idd or not, if $i>n$ $$\mathbb E[X_i\mid X_1,...,X_n]=\mathbb E[X_i]$$ right ? And if $i\leq n$ what happens ? – idm May 27 '15 at 11:41
  • 1
    No, for independent and $i > n$ that statement holds true. For $i \leq n$ the expected value is just what you observed for the variable itself $X_i$. This is just the statement that $E[X | X] = X$. – muaddib May 27 '15 at 11:44