NavList:

Steve B, you wondered:
"I figured out the azimuth of the Sun at the two times from that splashdown location, and the Sun shifted from 100 degrees to 103 d, centered almost on East.by.South. That's too close for a "good" fix, but how about a BAD fix?"

You get a long skinny error ellipse from those two sights. Its short axis is aligned in the mean direction to the Sun, and yes, just as you say that's nearly EbS ("E by S" would be 101.25° exactly, while the average azimuth here is 101.5°). In that direction the position would be well determined with some small uncertainty. Let's call it +/- 0.5 miles for a pair of good sights. In the perpendicular direction, the error ellipse is long, and the length or error in that direction is approximately the width of the ellipse (about a mile by this setup) multiplied by 115/ΔZ where ΔZ is the difference in azimuth in degrees. The relationship is simple like this for "small" azimuth differences, less than 20° or so. That has the form you might expect: smaller angle between the lines of position, greater expected uncertainty in the fix. 

In this scenario, the ΔZ is 3° so the multiplier is about 38 (=115/3). The width of the error ellipse is one mile, and the length is 38 miles. So not much of a fix! But yes, there's a position in there even with a pair of sights only 15 minutes apart. Another way to describe it: if the position has been determined to +/- 0.5 miles in the direction towards the Sun, then the cross-axis uncertainty, perpendicular to the Sun, is +/- 19 miles.

What do you suppose would happen to the fix if we wait fifteen minutes more and then do another pair separated by fifteen minutes? :) Rounding the times, Josh started with Sun sights at 14:35 and  14:50. I'm suggesting we take two more at 15:05 and 15:20.

Frank Reed