Statistics do not filter outliers
@bberg
Submitted by Benjamin Berg Link to original bug (#785018)
Description
The provided summary does not filter outlier points for the power usage estimation. Unfortunately this seems to happen commonly at the start of a test with really large outliers (e.g. 100W).
An easy approach might be:
- calculate the usage for each measurement
- cut off the top/bottom 10%
- average over only these points