Hi Hermione,
> That's (at least) sixty four (or sixty eight)
> possible circles.
>
> You then search for a fit on a subsidiary
> pyramid, on which you have nine points: corners,
> mid-point of sides and apex. All of these you're
> trying to intersect.
>
> With that many "bullets" and that many "targets",
> is it any wonder that there seem to be so many
> results?
I can understand why you might think that is true. If you look more closely, however, I think you will find that the results do not seem to fit random results. Take G1c as an example. There are 15 arcs that pass through this pyramid. Of these 15, 5 (33 percent) pass within 1 meter of a corner or the center. The total area of the five "targets" is 15.7 square meters. The total area of the base is 2,070 square meters. The targets, therefore, are .759 percent of the total area. Multiply this by 15 and you get 11.4 percent. As you can see, the actual "hits" are almost three times the expected number. The actual ratio will be a bit less due to the possibility of multiple hits from a single arc. (I think my analysis is sound.)
If this were an isolated case, one could reasonably think that it was unusual but not necessarily compelling evidence. Given that the method also establishes the location and size of G2 and G3 in only three steps each and also works for G3a, I believe the method should not be dismissed. (I need to check my work before posting G3c, then start on G3b.)
I think the evidence is there, but I don't mind if others disagree with me.
C. Wayne Taylor
Richmond, Virginia USA