Author Topic: a Summary of Rappel Knots  (Read 261 times)

Alan Dunham

  • New Member
  • *
  • Posts: 2
a Summary of Rappel Knots
« on: January 12, 2022, 04:29:39 PM »
The Offset Overhand Bend is the standard rappel knot used by climbers, apparently because it is simple to tie.
This report presents some alternatives. No knot is perfect, but some are more suitable than others.

One thing I am lacking is testing to failure in 9mm climbing rope. Any info on resources would be appreciated.

I have been climbing since the mid 70s, in Canada & the USA.


  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 1976
Re: a Summary of Rappel Knots
« Reply #1 on: January 12, 2022, 05:24:30 PM »
Nicely done! Thank you for your work.


  • Sr. Member
  • *****
  • Posts: 4083
Re: a Summary of Rappel Knots
« Reply #2 on: January 14, 2022, 02:48:33 AM »
This report ...

... is awesome --thanks beaucoup!!
(Now how will I have time for dinner & sleep?)

The Offset Overhand Bend is the standard rappel knot used by climbers,
apparently because it is simple to tie.
Prefer "offset water knot" --that'll spare any sidExcursions
into all that "overhand bend" can denote (as "water knot"
is relatively unencumbered).
Also, aim for consistency in naming with "offset" :
e.g., "fig.9" isn't that, but "offset fig.9" --though, yes,
one can see *issues* with that, as the unqualified,
"linear" F9 isn't much of common parlance; and then
IT would beg the question as to which end of it to
offset.   :P

Please do away with the --yes, sadly popular-- "flat"
moniker.  Soles's "offset" hits the key ingredient on
the head.  And those "linear" knots asserted to be "flat"
are not so : they have parts that can hang up on edges.
The Butterfly, SmitHunter's (Rigger's), ... aren't flat (and
of course not offset).

One thing I am lacking is testing to failure in 9mm climbing rope.
Actually, not really.  Testing to failure --by which I presume
you to mean "testing to as high a load as the knot can take"--
isn't relevant.
Rather, what you are lacking is some sort(s) of low-load
(as in the expected application of abseil) varied/cyclical
loading; the problem one seeks in that is any "ratcheting-out"
of material, slowly untying the knot.  (I once saw this happen
in some 6mm kernmantle I was just fiddling with manually,
so hardly much force.)

FYI, Lehman discovered the better version of SmitHunter's
bend a lonnnnng time ago; and he simply was surprised
that Asher had dismissed the knot (A New System...) !
One can SEE that it both gives a nicer curvature to the
SParts AND stuff some material into the collars so to
keep them open and the knot loosenable.  (But in a test
of 1 spec. of each, 1/4" laid nylon common rope, the
strengths were close.  And for practical concerns, irrelevant.)

(Somewhere I think you mentioned Offset 9-Oh,
but Search isn't finding it.  That knot points to the
things wanted for an abseil-ropes joint --cutting
out the irrelevant "9" to an "OH"verhand.)

Thanks much,
ACK --look at the time!


Alan Dunham

  • New Member
  • *
  • Posts: 2
Re: a Summary of Rappel Knots
« Reply #3 on: January 16, 2022, 10:36:33 PM »
Thanks SS & Dan.
Yes the nomenclature is wrong, but climbers are fixated on the flat name as they want a knot that is flat on one side.
You & I both know that the strength is irrelevant, but the first comment I got on an ice climbing site was that the whole study was useless because I hadn't done pull tests on a climbing size rope. Never mind that I'd pull tested 9mm to 1 kN for the untie times.
Anyway, I'm glad you appreciate the study. It gave me many happy hours fiddling around with a pair of cords.


  • Sr. Member
  • *****
  • Posts: 4083
Re: a Summary of Rappel Knots
« Reply #4 on: January 20, 2022, 09:28:20 PM »
... climbers are fixated on the "flat" name as they want a knot that is flat on one side.
. . .
... the first comment I got on an ice climbing site was that the whole study was useless ...
Al, don't feed the trolls!   ;D

Don't fret the edits you get,
as they can be met
with a sternly said "STET",
without regret;
forgive but forget
and all will be fine
(unless the comments are mine)
--Anon. II
[this was more "et"-rhyming in the first draft, alas]

I'd pull tested 9mm to 1 kN for the untie times.
Nice, that.  It should be well greater than any actual
forces.  --contrary that Baillie assertion of 350kg (!!?),
which strikes me as incredible!! ?  I couldn't find this
source via on-line search.

What is more relevant is the ARJoint knot's behavior
on real rappel-force varying loading --something more
"cyclical"--, which conceivably would show some knots
ratcheting out material with each surge of force, not
being retracted back into the knot on relaxation!?
(I witnessed this one time when playing with 6mm
kernmantle "accessory" cord; but I've not seen it
in some other such playings --oh, and that one time
was with manual not real-abseil force.)

Now, about that 9mm rope : I'm surprised that you
used your small cord for "bulk" measurements --thinking
that larger would be both easier to tie off and measure,
and errors in measuring --bit of inexactness-- would loom
larger in smaller stuff!?
And in checking just two of your "Bulk" measures
--for viz., RIGGERSX & FLEMISH joints--
I got quite different values --about 10 greater in each
case, IIRC (I do recall my 34ish for your 25 of the first
knot; I think that the margin was about the same in
numbers for the latter).

Did we measure differently?
I tied dressed & set the knots to a reasonable force
--and I don't tighten the SmitHunter's collars so much
as you do (too much, IMO)--, and marked the points
where the exiting strands had a perpendicular-line
(imaginary) reach on the outside of the strand they
crossed on exit.  (tied off with tea-bag string --yeah,
doesn't everybody save this stuff for small whippings?--
and venetian blinds cord (nice fine braid-over-laid);
and just one strand, multiplying by 2 for these symmetric

((Hmmm, it occurs to me that this "simple" measure
  becomes problematic if sizing it for different-sized
  ropes --what "diameter" is to be used; both?!))


[sending this much, pronto, as I've now TWICE
 been timed-out all of a sudden on prior attempts!]
« Last Edit: January 20, 2022, 09:29:20 PM by Dan_Lehman »


  • Sr. Member
  • *****
  • Posts: 4083
Re: a Summary of Rappel Knots
« Reply #5 on: January 20, 2022, 10:55:56 PM »
Here I think I'll just rattle off remarks as they occur
--and before another time-out!

1) Your strength for SmitHunter's bend (aka "Rigger's") quite
surprises me, as various other testings have shown in to be
not-so.  IMO, "RiggersX" shows a better graduated curve for
the SPart, and so should be expected to be the stronger of
these two --and pretty good beyond.  But you got what you
got, in that particular cord; be nice to understand why!

2) I simply do not accord the non-offset knots with
meaningful "flat"ness --they all have parts that will
abut and snag and hang-up on edges, rough surfaces.
(Thrun's bend --aka "zeppelin"-- can be oriented so
that it has a better vs. worse approach; but needing
to understand/set/retain such orientation is beyond
reasonable for practical use!)

3) My idea of key desiderata / criteria for an abseil-ropes joint
(ARJ) is :
a. security with mixed-dia/-nature ropes
b. easy *flow* over edges & rough surfaces
c. easy to tie
d. easy enuff to untie

4) Perhaps the most remarkable one
 --which you I think allude to but Search isn't landing it--
     is the (to coin a name) "EDK-backed EDK".
"remarkable" IMO as quite simple and I think that it can
survive quite a variety of bad tying,
and so might become the go-to-2BSure knot say in bad
conditions --think high-altitude fatigued white-out and
mitten'd hands!  It is nice in that the 2nd of the 2 overhands
can be put in on either side --behind ("backed") or before (!)
the initial knot (should one tie that in haste and then think
"dang, tails are too short for making another...") !
No, it's not elegant --esp. if sloppily tied--, but one isn't
entering an arts contest.

SOME tester (I forget who) managed to get it to "roll";
I'd sure like to see this occurring --and it surely did not occur
at abseil forces.

5) My "offset 9-Oh" shows a point worth taking : it's not
a so-simple matter to just apply offset loading to a knot;
having so loaded, there is likely a twin strand in this knot
that needn't --maybe SHOULDn't-- be so entangled, but
can be spared that, as it will be the partner twin that effects
the "choke" on the two SParts at their entry.
Thus, my naming "9" & then "Oh" (for "overhand") signals
the primary/choking strand to have a Fig.9 knotting, and
--not needing this, and not contributing it to possible lessening
of the other's work(!)--this *inner*-path strand makes just
an overhand.
(Another example would be taking the offset grapevine
into a (er) "offset strangle-Oh" joint --point being that
the strangle in the choking strand gets the extra surrounding
at the SParts-entry point, and all that's needed in the other
strand is stopper effect against this first knot (in the pull-together

Your "Pivotal bend" is a case for this :: put in a full
("round") turn for the choking strand (and cross those
tails* for less hard bending => easier setting), and you
get the wanted security-stability of the offset knot.
*Oh, I see, done for "PiX".

I once dismissed the "End-Bound Offset (Water) knot"
as weakening the point of extra wraps --which are better
made all-at-once in the choking strand, not spaced apart
by having a twin part running alongside--, but later came
to see its good effect coming not from that so much as
from simply securing the choking strand's tail against pulling out
--which can also be done by tying IT off with an overhand stopper
(esp. easy of a thinner rope around a thicker).

6) Oh, your DoB-3/4 photos could be changed (ONE
of them, i.e.) so that the same color rope is doing the
choking --this will also highlight my point about the
gratuitous knotting of the interior strand, the exterior
one handling the choke with the extra turn and the
interior not needing it.  (Ockham's Razor)   ;)

7) Among my initial explorations for improved ARJoints,
I had some lovely, beehive-looking structures; but these
were not simple to tie, not so easily loosened, and not
really what one wants --they were symmetric!
Hmmm, I think it might be your "Helix3" --lovely looking
surface-bottom, yes!  But not so great in mixed ropes,
et cetera.
Similarly for the Longhorn, another idea we share!   ;)

8 ) Dressing & setting :: most importantly, IMO for most
of the knots, is setting qua stopper knot --not ends joint!
You nicely show the offset fig.8 in 3 orientations;
but the 88t ought not to roll is one has set it as
I recommend : pull tails to tighten the choke;
THEN set knot qua stopper, hauling on SParts
to bend down the nipped tails.
IMO, the problem comes w/o the above method
as setting in pulling SParts apart will occur as there
is untight knot to yield material into the should-be
choking part as SParts pull it open.

AXIS OF TENSION.  (cf. Mark Gommers's work
on "Analysis of Offset Joining Knots" at the PACI
site, and (currently) p.20 of 33 shows what I mean
by this.  It's like "dialing" : have the offset knot
rising up upon a desk before you and just reach
out and wrap a hand around the knot and one can
sort of "dial" it through a range of about 180 degrees.
The SParts thus can go from making (in the OWK)
a forward arc to a backward loop, with the mid-range
orientation maybe the most vulnerable to flyping
--current conjecture w/o much testing.
His site ref.:

<whew, head spinning, maybe with some
confusions above, sorry>

Thanks again,
« Last Edit: January 26, 2022, 01:55:24 AM by Dan_Lehman »


  • Sr. Member
  • *****
  • Posts: 286
Re: a Summary of Rappel Knots
« Reply #6 on: January 25, 2022, 02:25:09 PM »
Hi Alan,

Nice work.

I have found these for your cross reference. They are easily to be found in my folders in the links below.

Octopus Bend = Side-by-side Fig.8 Bend
Longhorn Bend = Hive Eight Bend
PiX Bend =Neat & New Crossed A Bend
Infinity Bend = Rigger Cross East Bend

"My New Bends" :

"My Other Tying Methods of Some Known Knots" :

Some other tying methods of the bends may alter your found gradings.

Happy Knotting