Knot 17 slips at the same point as my knot. The knot with the longest tails won. I think it is worth further testing.
The end-2-end knot I recommended above (post #47),
Ashley's #1408 (and simiarly his #1452, which is
more commonly presented)
with the extended tucking
I describe should hold. Or else it will exhibit a new level
of rope fluidity, I think! The draw of the two S.Parts,
turning jointly in one direction, should bind the tails
ever more tightly together; that even this structure
(which, note, stands in contrast to the opposite-directions
rotation in
Shakehands, the #18 you tested)
could slip is understandable; but I think that one further
tucking of the tails through that heavily loaded central
nipping circle should see the knot draw up and lock.
(Maybe it will lock to the point of being hard to untie.)
Note that the aforecited on-line document gives
Ashley's #1452 / Ashley's Bend as its #3. BUT
it gets the dressing wrong, or of an inferior form:
the tails as show in the red & blue diagram should
be dressed & set to
cross in that relation,
not align abutting each other and not crossing
as the photograph of the completed knot shows!
(That version is also secure, but IMO produces a
less appealing curvature of the S.Parts, and will
not so readily feed into the extended tucking
that I recommend.)
--dl*
====