> ... because of some of the asymmetries ...
Have you tried to *correct* those, one by one (or so),
to see if that leads to success? Only scanning the start,
I see that for one end's three marks the other has only
two *un*-marks (or something like that), which I take
it what you're pointing to. (Sorry, but I don't feel like
chasing this one, in detail, myself.)
--dl*
====