[Gretl-users] boxplot accuracy

Responding to Frederik's posting at http://lists.wfu.edu/pipermail/gretl-users/2017-December/012814.html "Sometimes the candlebars pretend values which do not exist. I.e., if the IQR *1.5 data does not exist (which can happen - it happens to me), the boxplot gives a false impression." Please try a current gretl snapshot. This was a known issue but we believe it's now fixed. In the current Changelog we have: Boxplots: fix for incorrect extent of "whiskers" in some cases Allin