NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: Averaging
From: Peter Fogg
Date: 2004 Oct 20, 08:58 +1000
From: Peter Fogg
Date: 2004 Oct 20, 08:58 +1000
Just a couple of short notes about the simple pen and paper method of comparing a number of sights of the same body over a few minutes with a line approximating the apparent rise or fall of the body, caused by the rotation of the earth: This rise or fall is not a straight line, but an arc. This is why the straight 'line of best fit' should not be extended beyond 5 minutes. The statistical approach, presumably that of simple linear regression, works from the opposite direction. A sample of observations is taken and a process employed that best fits a line to them. What I am proposing is taking the given line of slope (a function of latitude and azimuth of body observed) and best fitting the obs to it. Quite a difference. If this process has flaws in extreme situations (and I have wondered, idly, what might happen at the prime vertical) then I am indebted for that information, but it doesn't change things for all the other occasions. Properly speaking it is not a system of averaging at all. It is a system of choosing, purely by visual comparison, a line of best fit. I guess this is why it quite acceptable, indeed necessary, to ignore extreme values. To put it most simply, the given line is the fact, the observed values have to 'best fit' it. Not the other way around.