I'm aware I'm nitpicking and maybe too theoretical now, but the processing time would vary in whatever application the user is using.. I'm with negus here who b

I'm aware I'm nitpicking and maybe too theoretical now, but the processing time would vary in whatever application the user is using.. I'm with negus here who basically means the same I guess. The generic danger here (regarding benchmarking) is that you're explicitly also benchmarking curl. In case curl/"the web client" would handle 30x redirects super inefficiently, these results could lead to wrong assumptions.