The trimmed mean discards outliers, so it can measure "what is the typical value for non-outliers?". The windsorized mean reduces the weight of outliers, so it

The trimmed mean discards outliers, so it can measure "what is the typical value for non-outliers?". The windsorized mean reduces the weight of outliers, so it can measure "how many outliers are there?" instead of "how extreme are the outliers?"