As long as you go to the same test and use multiple results you should at least learn if a tweak has improved your results or not.
I guess this is the reason that Bandwidth Place had become so popular ...

Practically everyone went there and you could compare the results of your tweaks to other ppl's results without thinking abt whether diff tests showed different results for the same connection and so on ...

It had become a sort of standard.