I've thought that there might be a chance increase during the windows but I still have a problem with every month - 1 and/or 2 8-day windows are open - therefore 16 days out of 28 days are 'windows' -

The prediction is the same for each (with little mod over time) - 3+ for la, sf, pacific nothwest and 7+ anywhere in ring of fire, which has since changed to whole world (this after turkey quake)

I don't see it by itself being significant - BUT - if taken with other types of data that may come to light to focus into locations and actual times rather than 16 days out of 28, well then maybe its something to use. It may be that added to something like measured strain accumulations and ULF signals (or the like) will more point to a specific area and a time.

Canie