When the point estimator under consideration has a pdf , the $P[T=\tau(\theta)]=0 $ , where $\tau(.)$ is some function of parameter $\theta$ and $T$ is an estimator of $\tau(\theta)$.
But I did many exercises to find point estimators of the parameters of density functions..
For example the two point estimators of mean , $\mu$ and variance $\sigma^2$ of normal density are $\bar x$ and $\frac{n}{n-1}S^2$, respectively.
But for the following statement, do those need to be zero ?
" When the point estimator under consideration has a pdf , the $P[T=\tau(\theta)]=0 .$ "