where 
and therefore 
.
So, the mean of
can be written 
.
Now, I have
.
As t tends to infinity, this tends to the mean of r. My question is, when is this a "good" approximation? Is it when t is much larger than the mean of the PDF? Or does it depend on something else?
