What are Vert.Sect.Board's issues w/D.Rchards's Report?

[knudeNoggin]

I tried to refer to Dave Richards's report Knot Break Strength vs Rope Break Strength
today and found only the following placeholder note:

This article has been removed pending further review by the Vertical Section Board.

My curiosity's piqued : what is the Board's concern with the article?
(Search didn't turn up any discussion about it, here.)

*kN*

[knudeNoggin]

A Summary of (missing) Dave Richards's 2005 test report
Knot Break Strength vs Rope Break Strength

It has come to my attention that the NSS Vertical Section Board has
lately stopped their on-line hosting because of some unspecified concerns
about the accuracy in the report. There is no apparent discussion of
the report's alleged failings in this NSS forum OnRope!, where matters
might have seen good review and comment. It is surprising that this
action comes approximately 5 years after initial publication! A query to
the Board has brought one response giving some general indication of
the concerns.

So, here, let me give a summary of the report, and comment on it and
on one critic's complaints.

The subject test report by Dave Richards presented results of testing
three types of kernmantle rope knotted in several common knots.
The ropes, all nylon, are:
S. 12.5mm (~1/2") "static" (low-elongation);
D. 10.5mm (~13/32") dynamic (rock)climbing; and
A. 7.0mm (eh~) low-elongation "accessory cord".

These respective strengths (% efficiencies) of these ropes when knotted
with the listed knots are reported in Fig.s 1..3, tables of absolute break
forces, as (mean of 5 tested samples) :
S ..... D ..... A ....
63.3 63.1 67.1 bowline,
75.3 69.4 73.3 fig.8 eye knot (tied by tracing/"rethreading"),
77.5 69.9 74.7 fig.8 eye knot (tied with a bight),
80.6 70.8 72.0 butterfly eye knot,
51.1 49.9 61.1 sheet bend,
54.7 54.6 57.0 double sheet bend,
53.1 60.3 59.7 fisherman's knot (single),
78.2 73.4 81.0 grapevine bend (aka double fisherman's knot).

An ordering of the ropes when knotted by strongest->weakest
shows the dynamic rope at the weak end (in two cases, nearly
equal to the "static" one => "s/d"), for whatever it's worth.
A-s/d
S-A-D
S-A-D
S-A-D
A-S-D
A-s/d
A-D-S
A-S-D

Richards notes that there was evident slippage in the cases of rope-S
with the fisherman's knot and the double sheet bend, rope-D with the
(single) sheet bend, and rope-A with both sheet bends (single & double);
he thus tied off the tails with either 2 half-hitches or an overhand stopper.

Below the tables, Richards presents three "Comparison Charts" --bar
graphs (horizontally plotted). However, these exchange figures with
the tables for ropes S & D : Chart S matches table D's data, and vice versa;
rope A's table & graph match. Below the Charts per rope is a chart (Fig.7)
that plots bars for each rope per knot; this matches the prior charts and so
mis-matches as to they with the tables (i.e., for ropes S & D).

Finally, in Fig.8 there is a curve plot of the data (supposedly), and this
continues the mis-match of the bar graphs with the addition of an error
in the values for ropes knotted with the butterfly knot --only one rope's
table value exceeded 80% with the others nearer 70%, but all three are
plotted here in excess of 80% (actually, of less than 20% "strength lost").
Moreover, the key for the graph is confused with an apparent plot-line
for "Rope Type", and duplicate lines for ropes S & D (none of which exists).

- - - - - - - - - -

The criticisms stated or suggested by one response to a query about
the report's removal are:

| The biggest problem ... was that ... data graphs had very little
| consistency with associated data information. The numbers or
| graphs got mixed up. Then [Richards's] primary graphs were
| embedded into other graphs compounding the problems
| and then there was a third embedding which was the ultimate
| meltdown ...

?! This describes something other than the simple mistake
in labeling the data-tables & bar graphs --the question is simply
Which is the correct one?, and then the correction is easily made.
("embedding" sounds deep!) And just eliminate the curves graph,
where there are errors in the plot for the grapevine bend (and
the basic question as to how a *curve* is meaningful representation!).

But, beyond this, Richards has carried the confusion into his
concluding "Comments". This section's value is dubious, and
could be omitted --it seems to do little beyond state in words
what the preceding presentations of data *should* make clear.
And drawing conclusions from these data is a dubious endeavor.

| Beyond this, ... someone who is talking about knots should
| call them by their correct names and spell them correctly.

?! Knots are called by many names, and knot names can apply
to more than one knot --and that's all within "a common language"
and even more narrowly, with localizations. Knots nomenclature
isn't pretty, consistent, or entirely sensible; "correct" is a debatable
assessment to make. Richards uses names that don't surprise me,
but for the awkward indications to discriminate how the Fig.8 eye
knot is **tied** --but his meaning is clear enough, though I here
question the point of distinguishing (which isn't his, alone : the
CMC Rope Rescue Manual included testing for both tying methods,
as though the test device should care?!)
I don't understand the criticism about spelling : the spellings are right.

| What befuddled me was that his premise was completely
| inconsistent with any testing that had gone before his work.
| Comparing 12mm static with 10.5mm dynamic with 7mm
| low stretch has little relevance to the amazing work of so many
| that have gone before. His chosen knots had little relevance
| with individuals who use these types of ropes.

?! All of these rope types are of a nature used by one or another
kernmantle-rope user. And the knots are not exactly exotic; one
can question the inclusion of the two sheet bends and the single
fisherman's knot, but it is instructive to see that these venerable
knots from maritime history fare poorly in these ropes --and not
from merely loosening when not in tension, but from slippage
in tension! The Fig.8 eye, bowline, and butterfly eye knots are
commonly employed, as is the grapevine bend; these knots,
among others, are e.g. included in the Lyon Equipment report
on life-safety ropes, for the UK HSE, contract research report
361/2001.

- - - - - - - - - -

My take on this all is this. Richards has done one more set of tests
similar to ones done by others, and has data worth tossing into the
knowledge pot for consideration. We need to determine what is the
correct labeling for the data tables & graphs --my surmise is that the
former are correct, and thus three graphs (re ropes S & D) need to
be corrected, swapping labels, and much of his "Comments" need
removing (they don't add anything). For the interim to a corrected
paper, this can be effected by a note to the correction.

Sadly missing from this, and from every other test report that I've
ever seen, is a clear indication of what exact knot geometry met
the test --such as photographs of the knots when tied & set, and
then when, say, 20% & 50% tensile strength force is achieved.
With the Fig.8 eye knot, we don't even know which of its tails
is loaded; with the butterfly eye knot, which is asymmetric, in
addition to having some various dressings, we don't know which
tail is loaded, either. We might even wonder if Richards used
same-side vs. opposite-side sheet bends, although that is the
more commonly presented version, or concordant vs. discordant
(same- or opposite-handedness) fisherman's knots. But, as noted,
this failing is common to test reports; the aforementioned Lyon
report at least gives voice to some of these differences, although
it is less than perspicuous about exact tying, with some confusing
expressions.

It is interesting to read Richards's observation that the butterfly
knot was "almost impossible to untie after a significant load of
1,000# was applied," as one rationale for using the knot is its
supposed ease of untying. It makes one wonder which version
was used, and which tail was loaded --it is an *a*symmetric knot.


.:. The Richards report could be re-done to correct the mis-match
between data tables and the graph(s); the several graphs could be
reduced to a single bar graph presenting the average efficiencies
for all of the ropes when knotted (also, do we need bars out to
100% for "break strength" ?), dispensing with the like graphs
per rope, and the curves graph. Except for the notes about rope
slippage in certain knots, of results for testing the "cowboy bowline",
and the difficulty of untying some knots, there isn't much more to
say in the concluding, "Comments" section.


*kN*