On a separate issue - why are you testing one knot against itself?
Do you not realise that you are only testing one of those knots - the weakest.
You determine nothing about the other knot other than that it is stronger.
//
The equivalence of this is a test designed so precisely that it gives totally
accurate results about a situation that bears no relation to reality.,
Putting it another way - variance is GOOD. The more variance your
test shows the greater the opportunity you have of understanding reality.
[/quote]
Well, we've been at this debate before.
You determine that TWO tokens of the knot are
as strong as the rupture --and you have one as
an examinable, or further testable knot, in one
run of the gear --all of machine/time/material.
seriously skewing your data by failing to determine the MBS
of the stronger knots and so skewing your measurement of
variance in favour of an artificially low standard deviation.
Firstly, if my reference values are to be skewed,
I'd prefer them to be so in this direction, where
one is more sure of "at least..." and not guessing
on which side of the mean/median the particular
knotted rope at issue might fall.
Secondly, if the variance is all so great as to
"seriously skew" results, well, maybe there's
something seriously wrong with our material
or tying skills or ... --why the variance?
(In this regard, it's interesting to see in the
Dave Richards tests of kernmantle ropes
that the least variance (in 5 tests per) came
with the
grapevine bend in one of the ropes,
and not with the wrapped-around-bollard test
to determine a baseline tensile strength!
I think that some testers might have concerns
about test specimen length --that there is a "too
short" length whereby imbalances aren't able to
get ameliorated over length of material and so
aggravate the weakening --esp. of back-2-back
specimens as we here discuss.
But one needs to take a hard look at what the
presumed *purity* of testing and running
statistical anaylsis amounts to, in the grand
scheme of things. *I* will tie my but not *your*
knots, and in THIS not that material, and ...
all in all, some test data from hither & yon
can only mean so much.
.:. In say, X tests of dual-knotted specimens
one will have 2X knots having been put to high
forces, half of them breaking (possibly some of
the survivors will show damage --saw that in
one case of tape knots so tested!).
As opposed to having only X knots tested and broken.
As the breaking point is more the focus,
I think that the 2-specimen test results give
a better view of where that might come.
And among the usually not-recorded, not-noted
data is WHERE the break comes. There was one
testing that put that for the
fig.8 eyeknot in the
SPart where eye legs turn around it; and in the
bowline somewhere around the nipping turn
(contrary Dick Chisholm's assertion, note).
Lyon Equip. testing of OH, F8, F9 eye knots
noted interior/exterior loading for the SPart
(i.e., was it the twin strand that pulled AWAY
from, or bore INTO its twin, resp.),
and concluded --not w/great amount of data--
that :
better to be exterior for OH,
interior for Fig.9,
and either way for Fig.8;
and I think Fig.10 might've been there, too,
but now I'm unsure of their conclusion.
Mark shows exterior-8, interior-9, btw.
--dl*
====