I've never heard it put this way. In my head I can see this is close to the 90° string angle so often mentioned but maybe a little more precise. Is this tied in with the mechanical advantage the string has over the limbs? If so is it the distance you mentioned compared to???
Since that was my comment, I might as well carry on with it.
I don't know if it always happens at 90degrees or not, but I expect the point it happens is different on different bows because they don't all bend the same when drawn. You could likely do a decent comparison by taking good pictures of bows on a tillering tree, taking a picture for every inch of craw and then comparing to an FDC to see what the string angle is when stack begins. I don't have a large enough selection of bows to do much with this, though.
It is definitely the leverage the string has over the limbs, as measured from the grip location. This is why string tension goes down as you draw, at brace the string only has a moment arm length equal to brace height. As you draw the moment arm (ie - leverage) increases quite rapidly, giving lower string tension even though the force required to bend the limbs is increasing. At some point in the draw the string to handle distance will hit a maximum. Drawing further will cause the distance to go down and string tension to begin rising again.
EDIT - Further thought has me wondering how this interacts with the string angle at the fingers and the draw force we feel because the angle is combining with the string tension to produce the actual draw force. Let me look at some numbers and maybe do up a spreadsheet to compare tension to draw force and string angle. I think I can get most of the info off Super Tiller output, but I haven't explicitly looked at this before to know for sure.
Mark