[Roger Hunter]

(02-08-2017, 02:30 AM)Skywise Wrote:

(02-07-2017, 02:16 PM)Roger Hunter Wrote: It's also a geocentric angle, if that's what you mean.

Yes. The 'cross' must be in sunrise/sunset if it is 90 degrees away from the subsolar point.

But this is an unusual definition of sunrise/sunset, as some locations, particularly those towards the poles, have very long sunrise/sunset times. For example, locations on the arctic circle around the time of winter solstice have the sun going around the sky never getting off the horizon.

But for the test here, it works.

BTW, if I've done my math right, a band +-1 degree wide 90 degrees from a point on a sphere covers about 1.75% of that sphere. In the case of a spherical earth of 6371 km radius, that's a surface area of 510,064,472 km^2.

That's equivalent to a circular area with a radius of 12,742 kilometers. Basically a bulls eye that big. From past experience this tells us that the odds of a hit are very high. Recall my Earthquake Dart Board experiment only used 1000km radii.

In fact, continuing the comparison to my EQDB, my 11 daily 1000km radius circles had a total area 34,557,519km. So Duffy's prediction area is 14.76 times larger than mine, and his windows are 30 days compared to my 7 days (4.28 times larger).

A significant difference is that my locations targeted high seismic regions by design, whereas Duffy's do not. But I think this may be offset by the much larger prediction area and time windows. Although his windows aren't truly a full 30 days since they are only valid for sunrise/sunset.

Brian

Check out my most recent posts to him. I've narrowed the band to 0.5 degrees and counted only the first hit in the 30 day window and still get around 40% hit ratio with random signal times. It's chance but I don't know why.

Roger