I don't think any of our hypothesis will be proven any soon. However, I feel a stonger knot is the one with
1. with the least curvature (as has been mentioned in the documents),
and less tight nip.
As we speculate, the 3 rope diameter will do just that. SO I''ll want to have it.
It depends how the 3 diameters are disposed;
one needs to examine the SPart when it's heavily
loaded, nevermind anything else.
Doing that with a
fig.8 might lead one to question
the broad-turn=more-strength theory, as there often
isn't such a broad turn to be found; rather, IMO, there
is something to the constriction around the SPart of
the eye-legs' U-turns (collars).
First objection is re "strength" : ropes don't break at knots,
experience shows, so how can some supposed strength be
important? (And I recall an observation by one mariner about
damage after a hurricane severely stirred up a marina of boats:
some lines were in fact broken, but not at the bowlines (or eye
splices) (so, likely cutting effects came into play). (Strength is
especially not relevant for abseiling --and yet some do mention
it!)
Furthermore, one must ask
What is *strength*?!--what is measured by a slow-pull device?
--or by a drop test?
--or by cyclical loading until abrasion proves fatal?
My surmise is that the last condition can prove a bane
for some knots that win on the first : precisely because
their SPart curves gradually over longer distance and
so has more movement during loading, there is more
friction-generated degradation in cyclical loading,
and ... the test device is shown to be a poor model
of practical concerns.
NB!2. Also ...
My hypothesis is that having the standing part out-side of a first loop
will help to increase the strength.
//
The claim is that with outside curve of standing part (point x) for a fig 8,
the knot is 15% stronger (limited test of course).
Because ... ?!
The image you show is one I regard as absurd :
a) the supposed "outside" for the
dbl. bowline is hardly
that --the doubled loop lies adjacent to, not around--;
b) the
fig.8 image is ridiculous --and, yes, so broadly
used (what does that say about the users, AND readers?!)--,
as round-crosSection material will not hold such a pictured,
on-top-outside-of orientation.
Rather, in a properly dressed
fig.8 one has one of
the parallel/twin parts reaching to the axial end of the
knot and when loaded bearing against its twin, or the
other part reaching less far and pulling away from its
twin and bearing into other parts of the knot (and it
seems to make a broaded turn --but look closely!).
(Most presentations of this knot family don't even
recognize the difference (and leave ambiguous which
end is loaded).)
I'll have to check what KM#007 (or nearby) says and who ...,
but my best information comes from Canadian climber and
IGKT founding member Rob Chisnall, who by illustration DID
seem to recognize the real difference in loading (and of the
proper setting --which he nicely illustrated), and his assertion
from some limited testing was a difference of --MIND THE
UNITS : % POINTS, not %-- 10 %-pt.s e.g. 70% vs. 60%,
and that for the loading of the end reaching farther and
pressing into its twin (which loading, btw, makes the knot
body assume more of a 45-degree angle to the axis of
tension; the other loading makes it more parallel).
Lyon Equip. did testing with this recognition but not so
much smarts about it in setting --i.e., I would want to
haul hard on the tails to impart more curvature to the
bearing-into-them SParts, for when push comes to shove,
the load on such rope will be WAY more than my setting
and should well challenge the curvature-inducement.
Lyon found some benefit to the
overhand by this orientation,
but not with the
fig.8; IIRC, the
fig.9 might've been
slightly favoring the other way (but much variance),
and the
fig.10 was maybe back to this way?
AND, to this, Dave Merchant opined that these more
complex knots (8-9-10) suffered more from dynamic loading
than did the simple
overhand, which he presumed was
from the movement of rope and friction.
... don't find the EBSB+Y very practical since the tail is not easy to pass through the both loops]
I think that having the minimal body and then
TWO extensions to it makes a tough knot to tie.
Prefer the
dbl (or water) bowline with double collars
--"Janus'd". The 2nd collar helps keep the eye leg up
snug to the body (the collar of the SPart should be
left not-so-snug, to not so bend it), and arrests the
tendency of any rope from this side to creep in;
and without this assisted loosening, the feed into
the knot of the SPart, w/o ally, doesn't go far in
loosening : the knot stays tied, to hold the next fall.
(But each fall does move some rope, and this is something
to examine with some test of repeated loadings.)
Re this last point --of an untightly gripped SPart enabling
repeated movement in heavy loading, consider another
of my ventures to getting a secure-in-kernmantle tie-in:
the
locktight loop, in which a structure somewhat
like the
(multiple) strangle knot grips very tightly,
BUT has a *back door* that remains loosenable (whew!).
Let's see if this URLink works, to this forum:
OK - original sketches to get us started:
Russ
Understand that the exact number of overwraps is a matter
of judgement per circumstance, but these show reasonable
starting points. Version II was hoped to be better with stiff
rope. These knots are meant to be set TIGHT (in contrast to
the
mirrored bowline which is shown too tightly set by Mark
--constraints of image space, I think--, and fares well enough
rather loose.
--dl*
====