[Bloat] [Starlink] the end of the average

[Sebastian Moeller]

Contrarian view; the average is still a pretty fine measure, assuming one does not over-interpret it, accompany it with a measure of variance and only apply it a distributions where it intuitively makes sense (e.g. for distributions like a gaussian curve, A-OK, for a bi-modal distribution with peaks symmetric around zero probably not so much).

However the same caution should be used for alternative measures, beware of the single number descriptions of complex data...

Regards
	Sebastian

P.S.: Often it is not obvious from the beginning how a distribution is going to look like, it seems dangerous to summarize to mean + variance at that stage before actually looking at the data, however an easy trap to fall into especially for sparse data.


> On Jul 10, 2022, at 18:20, Dave Taht via Starlink <starlink at lists.bufferbloat.net> wrote:
> 
> This book is my weekend's reading. Excerpt:
> 
> https://www.thestar.com/news/insight/2016/01/16/when-us-air-force-discovered-the-flaw-of-averages.html
> -- 
> FQ World Domination pending: https://blog.cerowrt.org/post/state_of_fq_codel/
> Dave Täht CEO, TekLibre, LLC
> _______________________________________________
> Starlink mailing list
> Starlink at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink