NavList:

   

Reply

Dear Peter,

With all my respect, I have to say that several years ago
I decided not to
participate in any discussion of statistics on
this list. Because my experience showed that these
discussions tend to be infinite and fruitless.

In my previous message, in the very end of it,
I explained as clearly as I could, what really happens.
If some procedure gives a random error x,
then the same procedure repeated n times gives a random
error of the order x times sqrt(n).
For example, if a table gives a random error 0.5,
then applying this table 4 times gives an error of 1,
and applying it 16 times gives an error of 2, in the average.
Sorry for the repetition.
There is really nothing else that I can say on this subject.

And discussion on how to phrase it in ordinary words
(that "errors accumulate" or "cancel each other" or
"do both") is potentially infinite and (on my opinion)
useless, as most of the discussion about how to state
precise things in imprecise words.

Alex.

> >  They will do both things: cancel but still
> >  accumulate.
>
> Come on Alex, its one or the other, surely.


--~--~---------~--~----~------------~-------~--~----~
Navigation List archive: www.fer3.com/arc
To post, email NavList@fer3.com
To , email NavList-@fer3.com
-~----------~----~----~----~------~----~------~--~---

   

Reply