NavList:

   

Reply

Recent discussions, on rejecting observations from a lunar, are now happily
concluded with the agreement that the rule should be "USE YOUR COMMON
SENSE". Would there be any interest in a more general discussion on that
topic of rejecting rogue data?

I remember, some time ago now, a posting of what appeared to be a superb
set of lunar distances, all very close to a beautiful straight line,
measured by a member who was rather new to lunars. I won't embarass him by
naming, but he knows who he is.

In off-list correspondence, I discovered that those points were no more
than a subset of a much larger series he had taken. These were a reasonably
good set of lunars by most standards (but not comparable with Alex).
However, all the missing points were off his straight line, some way off
it. He had plotted his observations very carefully, then found that he
could draw a line accurately through some of them (less than half) and had
therefore rejected all the others.

And it's probably true, in most cases, that if you take a set of, say, 15
observations. showing reasonable scatter, there are many combinations of,
say, 5 of those points, and some of those combinations will lie beautifully
close to some arbitrary straight line, if that's your only criterion.

It wasn't unreasonable, really: He genuinely thought that was how the job
should be done, because nobody had shown him any different.

To those of us that have a background in some sort of science, if only at
long-past high-school level, it can be hard for us to understand the
difficulties others can find with some of these concepts, if they haven't
shared those same privileges.

For example, one rule might be "never reject anything just because it
disagrees with your preconceptions". I guess the seabed must be littered
with ships in which the navigator had just erased a position line, saying
"I can't possibly be THERE!", to then hear the splintering sound of keel on
rock.

We all make blunders, and somehow we need to recognise and reject them, if
we can. It would be interesting to hear from other navigators about how
they apply their judgment to that end.

George.

================================================================
contact George Huxtable by email at george@huxtable.u-net.com, by phone at
01865 820222 (from outside UK, +44 1865 820222), or by mail at 1 Sandy
Lane, Southmoor, Abingdon, Oxon OX13 5HX, UK.
================================================================

   

Reply