NavList:

   

Reply

Peter Fogg wrote:

> What really perplexes me is that it seems such a simple concept: a given
> slope and data points that can be compared with it. Why does it need to get
> more complicated than that? (leaving aside the non-linearity issue)

To study the slope and fit of isolated altitude observations is a good training
exercise. As such it has its purpose as it enables the student to check his own
work and to develop confidence. (In fact, this is the context in which Bowditch
1962 treates the subject, not under sightreduction. See Article 1507.
"Developing observational skill.", pp. 407-409)

In the real world nav station, however, the goal is to get a fix. For a fix one
needs at least two "runs" (the Power Squadron calls it so; it's a series of
observations of the same cel. body within a short time), but often one will seek
three runs to get a nice cocked hat. So one ends up with k + m + n = p
individual observations. The question is now, whether the MPP in the cocked hat
built from the three averages is the same as that obtained from all p
observations directly. In general, it is not.

It might be the same, if k = m = n AND no observation is rejected AND the
correct averaging method is used. I have never seen this investigated in the
literature. It would be only fair that those who propose averaging should also
provide this analysis. However, in my opinion this is a moot question, because
it imposes an unrealistic and unnecessary constraint on the observations in a
real life situation.

Herbert Prinz

   

Reply