About Order Statistic
Given random variable , we can get a sample of size . At times, the cost of sort can be lower than that of calculation, we can get the so called order statistics by sorting the sample in ascending order: By this, we lose all information of the sequence of draw.
Cumulative and Distribution Function
Notice that means that at least observations of the sample is less than , we have By differentiating , we can obtain its density function: which can be simplified into Heuristically, this can be understood as selecting elements and ensure that they are less than and then select elements and ensure that they are greater than , and in the end multiply the density of the -th .
as an Estimator
An interesting property of is that its an unbiased estimator of the -th -quantile of . To deal with quantile, a lemma may come handy:
for any continuous distribution , where .
This is easy to prove: thus . We now can proceed to prove that : Now notice that , Admittedly the last step is very clumsy, will need to be fixed.