|
|
|
|
|||||
|
||||||
Subject: Re: On LOPs
From: Trevor J. Kenchington (Gadus@XXX.XXX)
Date: Mon Apr 15 2002 - 21:24:53 EDT
Jared,
I would agree with you that the probability contours around the MPP
would not be elliptical if there were more than two LOPs but they would
not be circular either -- though a circle might be a close approximation
in some cases. If there were four or more LOPs, those contours might get
quite complicated, with two (or more) relatively probable areas
separated by less-probable areas and irregular low-probability contours
surrounding the whole.
Also, you are surely right that a rigorous determination of the MPP
would require consideration of the various LOPs having different
precisions. Assuming that they are all equally reliable is just a
convenient simplification to ease discussion of these points.
Then again, if your imprecisions are so small that they can be
represented graphically by the thicknesses of pencil lines, I'm not sure
why you would bother in practice to do more than plot them and circle
your entire cocked hat as your fix.
Trevor Kenchington
Jared Sherman wrote:
>So far I haven't seen any justification for the "most probable" position, much less for the use of an ellipse (which has 2 focii) rather than a circle (with one focii, i.e. center) when using 3 or more LOPs (which would indicate a need for 3 or more "focii"). [snip]
>
|