NavList:

   

Reply

On 2016-07-10 10:13, Stan K wrote:
> In looking at the "Example of the Complete Reduction of a Sight" in the 
"Compact Sight Reduction Table" (Modified H.O. 211, Ageton's Table) book, I 
noticed that the value of B for the dec of 3º30'S is 81, but the table value 
is 81.1.  In fact, all table values less than 166 are shown to tenths.  Why 
bother showing the tenths if they are just going to be rounded out in the 
calculation anyway?

A few weeks ago I said, "In my opinion, extracting A and B
values to tenths of a unit from the standard Ageton table is not worth
the extra work." That's also true of the Bayless table. In my Monte
Carlo simulations (which retain tenths, if they're present in the table)
the improvement is not easy to distinguish from the statistical noise.

If altitude is restricted to 5° - 80° and observer latitude to 0° - 70°,
my Monte Carlo test with the Bayless table had a root mean square
altitude error of 2.03'. About 43% were worse than 0.5'.

Significant improvement occurs if A(R) is interpolated from B(R) when t
is within 5° of 90° (about 4.5% of sights), RMS altitude error improves
to 0.82'. There's a small but definite degradation to 0.86' if the table
gives whole numbers only.


By the way, my promised evaluation of the Sadler technique has not been
forgotten. The code is written but I still have to run the simulations
and write up the results.

   

Reply