NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: Lunar distance accuracy
From: Alexandre Eremenko
Date: 2007 Oct 25, 12:23 -0400
From: Alexandre Eremenko
Date: 2007 Oct 25, 12:23 -0400
Dear Frank, > Alex often gets annoyed when I say that beginners can get The main disagreement we have on this matter is not about "novices". After all, my own very first Lunar in my life (made shortly after I purchased my first sextant) and posted on the list, was a PERFECT result. There is no doubt that "beginners CAN". In the sense that "some beginners sometimes can". The main point as I understand it is that we evaluate the same data very differently. We use different numbers derived from the same series as indicators access "accuracy". If the "average error" of some long series of observations (taken over the preiod of weeks or months, like in the case of White) is say 0'2, then what does this tell you about accuracy? Very little, indeed. In a real-life situation you don't want to risk a chance of hitting some rocks with 5% probability, do you? By the way, the measures like "quadratic mean" or "the most probable error" are also poor measures of accuracy. My recent analysis of White, Bolte and my own observations shows that a) distribution of errors is FAR from normal. Which means that the usual measures of dispersion cannot be applied. b) they are especially far from normal in their "tails". That is there are MORE LARGE ERRORS than a normal distribution would suggest. Indeed, in the 42 observations of White, we have one error of 0'.8 and 3 errors of about 0'.5 in the distance. And I say that all available data contain a certain substantial proportion of such outlayers. You can call them blunders if you wish. My 121 observations in spring 2007 are as good as White's. They contain 13 individual shots with the error 0'.5, 2 shots with the error 0'.6 and one with the error 0'.8. The error of the average is 0'.0. The average error is 0'.2. If I believe that this 0'.2 is a "reliable indicator of accuracy" and then even multiply it by 2 "for safety", I will "hit the rocks" 18 times out of 120 ! My experience and all available data show that the errors cwin the range 0.5 and more are UNAVOIDABLE, and occur perhaps in 10% of all observations. This is the main point of our disagreement. Alex. P.S. Bolte, White and my own observations of last spring are now posted on my web site under http://www.math.purdue.edu/~eremenko/accuracy.html I will soon add a digest of Bolte in English, as well, as my own statistical analysis of all these observations. --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to NavList@fer3.com To , send email to NavList-@fer3.com -~----------~----~----~----~------~----~------~--~---