NavList:

Again, this is a problem of semantics, not mathematics: You are mixing up true position (as God knows it, or a fly on the wall with a GPS in his hand) and most probable position, as you try to establish it to the best of your _knowledge_.

Your example is constructed with the assumption of non-random errors, but at the same time you make a disclaimer to the effect that these errors are unknown to you. (Without this disclaimer, I could have replied: "So, go ahead and remove them!"). But if they are unknown, they are random in the sense that they could be there or they could not be there, and therefore there is nothing you can do about them. Unpredictable, random errors affect the quality of the fix. Consequently, God (or the fly) knows that you are not where you are forced to think that you are as a result of your measurement, math and logic. But _you_ don't know that.
 

WSMurdoch@AOL.COM wrote:

In a message dated 4/15/2002 12:06:50 AM Eastern Daylight Time, hprinz@ATTGLOBAL.NET writes:
 
 

The

most probable position (MPP) is, by definition (!), inside the cocked hat.
 

I am not so sure about that.  I think it may be true if the errors are random, but not if they are not.  Letâ€�s say I have three bodies at 300, 000, and 060 degrees (the horizon is dirty to the south), and a sextant with an (unknown to me) 5 arcminute error.  I reduce my sights, get a nice small cocked hat and am actually 5 miles to the same side of all three LOPs.  It seems to me that my most probable position is outside the cocked hat.  Now these are non-random errors, but those are the pretty normal (at least for me).

Bill Murdoch