NavList:

   

Reply

Dear Peter,

> Firstly; these are not 'errors',
> which implies inaccuracy, but rather
> a limitation in precision.

"Error" is a standard term in statistics. For example,
"rounding error". Unlike in real life, in statistics this
word has no negative flavor:-)

> will NOT tend to accumulate,
> but will tend to cancel each other out,

They will do both things: cancel but still
accumulate. More precisely, suppose
that one application of some table introduces
a random error x. Then n applications of this table
can be expected to introduce an error of x times sqrt(n).
So, if in some calculation, you use a table where
precise numbers are rounded to 1', and you have to
enter this table 4 times, then each entry has expected
error 0'5, while the whole calculation 0'5 sqrt(4)=1'.
If you have to use the table 9 times then the expected
error of the whole will be 1'5 and so on.

Alex.


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

   

Reply