Winsorize

It is the process where a certain number of extreme values in the statistical data are replaced with smaller figures.

The method helps to build the mean that is sensitive to extreme values. If asymmetric population data is used, the winsorized mean gathered for it can get an unbiased estimate where certain values are modified to get a superior result.

The method is based on the number of counts and is used to improve the effectiveness of statistical inference.

Since the classical statistics mean, and standard deviation methods are sensitive to extreme values; the method helps to support classical statistics by lowering the impact of extremes. 

It is integrated into automated processing systems or data charts where hundreds and thousands of entries are made, but one cannot check all the extreme values entered in the fields.

The unbelievably huge values, or those considered contaminated or long-tailed, can be refined using the method.  

It can minimize the impact of outliers, but the method has certain drawbacks, as it can also create a bias.

Eliminating or modifying the powerful figures in terms of mean/average is not a good way to find accurate information.

In real-life conditions, the observations that are eliminated through the method of rejection, like in commodity market technical analysis or to find financial crisis facts and figures, discarding the observations at the time of gathering information is considered inappropriate.