To learn about the theory behind this process, see related articles online at www.starpath.com.
The goal is to effectively average a series of sights to determine which one is a proper representative of the full set. To do this, we compute the theoretical values of Hc over the time interval of the sights stored-taking into account the motion of the vessel-and then compare the slope or curve of these data with the actual sights. This process shows which sights are outside of statistical fluctuation, so they can be deleted or not considered, and lets you choose the best of the set.
To execute Sight Analyzer in StarPilot, enter a set of sights of the same body over a reasonably short time period (10 to 20 minutes or so) using the normal Sight Reduction function. If the Hs values change too much during your sight period (time is too long), then the Hs scale on the calculator display is too compressed and differences do not show up well. Once the sights are stored, set DR mode to Speed, DR time to the time of the first sight and enter the correct DR position for that time, along with your course and speed made good over the sight period.
Then from the Celestial Fix menu, select Sight Analyzer. The sample here shows its use for a series of sights at LAN, but it would more typically be used for any sequence of sights of the same body.
The Analyzer will then compute the Hc value (actually a computed Hs, since we undo all the corrections that have been applied to the sight at that time), then store this theoretical value, and then proceed on down your sight list, advancing the DR to the time of the next sight, computing Hc, and storing it, until all are done. Next the curve of theoretical sight values is drawn as a graph and your actual sights are plotted on the graph. Any sights that are far from the plot are most likely wrong.
To best compare the data, move the curve by placing the cursor at the place you want the curve to cross and press [Enter], and this way find the best fit to the data. Note the cursor cannot be right on a data point, since [Enter] then will report back the difference between line and data. Find a location off of a data point that will place the curve (line) such that there are about the same number of points above as below the line, but disregard any that are clearly far off the line. Then choose any one sight that is on the line as your best. That one sight will be as good for a fix (or better) than all the sights plotted together.
This process can also be applied to any sights. Those taken before or after LAN could be used for a "reduction to the meridian" analysis (see older versions of Bowditch for reference).
This tool is best suited for poor or sparse data. The LAN example sights have little spread, yet we can still improve the analysis as shown.
The sight analyzer provides a mechanism for selecting specific sights out of the sight array to analyze using a sight filter. By default all sights in the sight array are displayed. You can select a sub range of sights to display by entering the sight number of the first sight in the range followed by a "-" and the sight number of the last sight in the sequence into the filter. Individual sights can be enumerated and delimited with ",". Sub sequences and individual sights can be mixed and matched in the filter string. The following filter string "1-4,7,9,11-14" tells the sight analyzer to include sights 1 through 4, 7 and 9, and 11 through 14. A graphical sight picker can also be used to select subsets of the sight array for analysis. To use the graphical tool hit the [->] button and use standard Windows selection methods to pick and choose sights. The following example illustrates the use of the graphical sight picker.
LAN Analysis using Sight Analyzer
The example is a set of high LAN sights, peaking at about 87 degrees, from July, 1982 in the Tropics (note it is not common practice to take sights this high).
The first picture shows all 15 sights spanning LAN, and the analyzed results are not very interesting since the scale is too big.
But suppose we use the filter to select the few sights shown in the picture below it. Then we run the Analyzer, set the cursor where we want the line redrawn, and click the left mouse button.
The scale is now much better. Notice the difference between solid dots (exact fit) and open circles (off the curve).
With the use of the Analyzer we see that even with just these few sights, we have a good LAN, and that the next to last sight was too low and that our peak height is a reasonable one for the LAN value even though we slightly missed the actual peak time. This also means that you could use the LAN utility to get a reasonable Lon as well as Lat from the sights.