[Roger Hunter]

(02-26-2017, 04:14 PM)Duffy Wrote: 25th Feb M 5.0 Eastern New Guinea, P.N.G. ................... 13:24:40 ut
              Sunset on bearing 75' 03' E - 14' 54' S (D3) ..... 13:24 ut

26th Feb M 5.5 Vanuatu Region ..................................... 06:44:21 ut
              Dawn on bearing 28' 49' W - 9' 59' S (B21) ...... 06:45 ut
              Sunset on bearing 173' 15' E - 11' 38' S (B17) .. 06:45 ut
              CM on bearing 78' 04' E (R4)

26th Feb M 5.1 Atacama, Chile ....................................... 08:59:16 ut
              Dusk on bearing 158' 52' E - 17' 05' S (B1) ....... 09:00 ut
              CS on bearing 48' 24' E (B5) (R22)


Roger

Thank you for being straight with me, the only thing worse than frustration is confusion. I didn't mean to portray tone, grammar has never been one of my finer qualities.  I don't know if I can be of any assistance with your program issues, but the data you posted is reminiscent of a scenario I tried last October.  I chose a bunch of numbers from the September quake list, and ran them up to new moon Oct 30th. I couldn't get it to work at first because I was testing September latitudes against October terminator zones. I had forgotten to account for the difference in sun/moon latitudes between the two months ... bad example but sometimes the most obvious things are overlooked !

This test period has now expired, so I would like to thank you both for contributing and participating in this unusual experiment. Weather there is anything to be gained from the results is not yet clear. But considering it is sunrise/sunset somewhere in the world at any time, and the quakes during this test period averaged 4 a day ... nearly all but 5 got a hit.  I have unfortunately been hindered by a less than average broadband service through out this test, but I will post an accurate evaluation when I get things back to normal. 

Duffy

Duffy;

What you need to know is total number of quakes during the time period, total number of hits, total number of days in the test period, total number of days in prediction windows.

The idea is that if 50% of the days are in predicted windows then you should catch 50% of the quakes.

It's a simple test; too simple really because quakes are not uniformly distributed, but it's a good first step.

I'm still fighting with the program. I find I can get the right answer if I shift the date a couple of days but the date calculation is from a renowned expert in the field and should be correct.

Roger