Build My Own FeedQC Lab: What is the strongest rappel knot?
I was climbing in Yosemite last summer and, while at a belay, was talking to a party from Bozeman, Montana. They noticed all the proto gear on my rack and deduced that I was from Black Diamond. My partner and I bailed (typical) and they were watching me like a hawk. I asked what they were looking at, and they said, "We want to see how the QA guy from BD raps.” In particular, they were eyeing my rap knot, since knot choice (Double Fisherman’s, Ring Bend or Euro Death/Overhand) has always been up for such a large amount of debate.
There are advantages and disadvantages to each type of rap knot, but is there one that is superior in strength to the others? I put all three knots (with three different rope diameter scenarios) through a quick-and-dirty series of pull tests and have provided below some basic strength testing data based on my limited testing of the three most common rappel knots. (Note: only two data points per configuration.)
THE OBSERVATIONS
• The Double Fisherman's and Ring Bend had similar strength results
• The Euro Death Knot was the weakest: ~20-30% less than the Double Fisherman's and Ring Bend.
• The Euro Death Knot slipped a bit before failure at ~4000 lbf with the 10.2 and ~2000 lbf with the 8.1 in the system.
THE BOTTOM LINE
The reality is that all three of the methods for joining two ropes for rappelling that I tested were PLENTY strong for the forces seen during a typical rappel (i.e., bodyweight-plus, taking into account some shock loading when bouncing around, jiggering with tangled lines, not smooth rappelling technique, etc).
For what it's worth, when I started climbing I always used the Double Fisherman's, but now I've fully converted to the Euro Death Knot—it's fast to tie, plenty strong, less likely to get hung up when pulling and easy to untie. And finally, no matter what rap knot I tie, I always leave long tails (like at least 12 inches).
Climb safe,
Kolin Powick
-----------------------------------------------------------------------------------------------------------
Kolin Powick (KP) is a Mechanical Engineer hailing from Calgary, Canada. He has nearly 20 years of experience in the engineering field and has been Black Diamond's Director of Global Quality since 2002. Kolin oversees the testing of all of Black Diamond's gear from the prototype phase through continual final production random sample testing. If you have a technical question for KP, please email him at askkp@bdel.com and he will TRY to respond.
United States / English 
9 Dec 2011, 6:28PM
December 10, 2011
Philippines
sharing my experiences in anchoring and riggings, fisherman's knot application is more efficient during joining two end knots using multiple overhand like in rescue operation it is easy to tie and untie. but joining two different sizes, sheet bend is more efficient backed up overhand in long tail ends.
6 Jul 2011, 5:04PM
I've been using the [SmitHunters'] bend a while
and it seems like a real solid alternative.
You need to consider the criteria for an abseil-ropes-joining knot:
1. ease of tying (in sometimes stressful conditions --e.g., late/fatigue)
2. capability to join dissimilar (within usual range) ropes;
3. security (NOT strength);
4. ability to flow smoothly over rock faces (NOT strength);
5. ease of untying (easy enough) (still NOT strength).
Where do you see these criteria tested? --what are the units reported?
(strength has such nice-looking simple numbers)
Any knot that isn't *offset* (the EDK is) is prone to snagging on things. The SmitHunters' bend (and Zeppelin) with ends at right angles to the mainlines are if anything more prone to mischief.
I've seen people use and swear by the figure 8
to combine two ropes. Any data on this how it stacks up
against these knots that you tested?
Some people were ready for The Rapture, too.
Do these people distinguish between the orientations/loadings
of the fig.8? (I.e., which of the twin strands is loaded, which not?)
It's a knot likely to hang up, and possibly vulnerable to untying when
pulling a thin line, perhaps (i.e., when something impedes the movement of the knot, and the thin line pulls without reciprocal tension in the (following) thicker line --just a from-the-keyboard worry w/o any experimentation, at the moment).
I'm a bit surprised that a ring bend is even considered
for tieing a rope, like DF already mentioned: it can only
be used for webbing. Even then it's not secure, there are
a lot of documented accidents caused by a ring bend
(source: Sicherheit und Risiko in Fels und Eis by Pit Schubert)
The ring bend (aka water knot) has a long history of use first
in ropes. Like other twin-parts knots (fig.8, asymmetric fig.9, ...),
there's a choice of which ends to load (as I discussed below, in reference to this site's images). I'm unaware of any security issues (in normal cordage (not HMPE or other hi-mods)).
And while I've heard of Pit's concern --and some reported fatality--, I share others' observations that a weight-set water knot in tape often requires tools for untying (or a lot of grunting). Prohaska even went so far as to devise a variation with a twist in it to guard against the supposed vulnerability, but this has been the de facto standard tape-joining knot in use SOOO much, I'm curious as to evidence of insecurity of this sort.
(Tom Moyer et al. have done some testing of cyclical LOW-load creeping out of the exterior-side tail --something deemed maybe most likely a risk with stiff tape.)
*knudeNoggin*
12 Jun 2011, 6:48PM
I've seen people use and swear by the figure 8 to combine two ropes. Any data on this how it stacks up against these knots that you tested? Is the figure 8 an acceptable knot?
Thanks is advance.
10 Jun 2011, 1:24PM
Thanks Kolin,
Back in the last milleneum, I liked to climb a fair bit and spent some time teaching it, doing SAR and living to climb in Europe. Many of climbings spookiest moments are rapelling .I came across a book called Mountain Skills out of the UK. I learned a very reliable rappel system which I did see once in a U.S. Climbing rag but otherwise I don't see it anywhere.
The idea is to put your device on a shoulder length runner. This way both hands are brake hands and it is much easier to add and maintain a structural prussik back-up. Most importantly it allows the rappel leader to have subsequent rappelers attached to the rope before he departs the belay. Another benefit is the quick availability of a tie-in at the next belay. Safe and efficient.
There is something inherently safe and dangerous about conservatism in climbing theory. On one hand you are slow to take up new good ideas. On the other you may not rush headlong into foolish behaviors. Do you guys ever talk about those Euro buckles that didn't thread? Man I've seen some horrifying and laughable stuff. The best ever was the first time kid bouncing on his buddies homemade daisy Single strand 6 inch pieces of 5 mil, 5 feet long. each string was tied to the next with a square knot with 1/4" tails and......wait for it...... melted together with a lighter! (he was trying to make six inch quickdraws with nylon string and then decide to make a daisy because they were too short and then the knots kept coming undone so he used a lighter to fuse the 10 or so knots) His apprentice was tied in to the observation platform above waterfall cliff at Castle Rocks and just bouncing away enjoying the exposure. 100% deathfall potential at each bounce. Awesome! Maybe I saved a life that day.
10 Jun 2011, 1:18PM
I've always been taught to use a double fisherman's, however I've seen the figure 8 used to join two ropes for raps etc.
What is the disadvantage of a fig8?
Thanks in advance.
10 Jun 2011, 10:54AM
I am pretty sure that the appellation "Euro Death Knot" started back on the late lamented rec.climbing. And it was a joke. Lots of us were already using it, and calling it that just to freak our partners out the first time they saw it.
That's what I remember anyway.
6 Jun 2011, 4:37PM
I'm a bit surprised that a ring bend is even considered for tieing a rope, like DF already mentioned: it can only be used for webbing. Even then it's not secure, there are a lot of documented accidents caused by a ring bend (source: Sicherheit und Risiko in Fels und Eis by Pit Schubert)
I always use a EDK without backup, just long ends (~10 in.), because it offers the smallest chance of the rope getting stuck. Everybody i know also uses the EDK but considering that i'm from Europe that can't be much of a surprise.
26 Nov 2010, 6:29AM
You guys ever investigate less common knots for rappel?
I've been using the hunters bend a while and it seems like a real solid alternative.
http://en.wikipedia.org/wiki/Hunter's_bend
It's a bit tricky at first, but if you untie either side of it, you're left with an overhand.
Packs tight, won't invert, fairly light on the bulk.
Love to see some test results on this and maybe its variants, Eg http://en.wikipedia.org/wiki/Zeppelin_bend
Thanks
28 Oct 2009, 6:56AM
The reason that it's called the EDK is because of paranoia. The complacency of adapting good technical methods stems from isolationism and inbreeding of climber education.
I was using it after having gone through an AMGA rock guides course and was climbing with my friend, Tom Moyer in Salt Lake City circa 1998. Tom (and crew) chided me for using the EDK and so I explained the rationale behind it. Tom was curious enough to get Chris Harmston to test it in the lab at BD under cyclic loading and compared it to the "flat fig 8" (one more twist than a EDK). Yopu can check out Tom's report on his web site. It's still a good knot to use, and Kolin has now repeated this testing with some other good insights, further substantiating its use.
Having said that, I agree with some of the above statements that it takes a certain kind of rope to use the EDK and certain conditions may warrant not using it. However, many more partied have had epics / tragedy from getting ropes stuck when using a knot that got hung-up because of its higher profile than the EDK, than have died from using the EDK. Think of how many user days the EDK has and how few incidents there actually are. You'd spend your time worrying about if you had your seat belt on while driving your car than if the EDK is a good enough knot to use to rappel with.
Remember that there are no black and white protocols when climbing.Get some instruction if you don't know...don't guess.
Great thread !
Marc Beverly
9 Sep 2009, 12:39AM
[WARNING: rockclimbers can post a gazillion pages on musings about equalizing anchors, but the sight of more than 5 lines about knots makes them queasy. And more than just climbers; the stupidest things can be said -- and printed/echoed/re-printed -- for decades and even centuries about knots.]
On the testing reported here:
0) Good to see it! But, OTOH, the relevance of knot STRENGTH to abseiling is pathetically small; the testing that should be done concerns security and ease of moving over rough surfaces, of any vulnerability to being untied in hauling. NEVER is strength an abseil issue; unfortunately for that, it is an easily generated datum (at least if superficially regarded).
1) How were the ropes/knots configured? E.g., for the two photos center & right, it COULD be the case that the knot joined a single rope into a sling and was pulled as such between pins; this would, btw, tend to exagerate the strength of the knot, as material fed out of it during compression would leave the unknotted side to assume more load/tension -- i.e., it really could be the case of adding 100% to and dividing by 2 ! (Which I surmise is how there are some reports of Grapevines or Dbl.Grapevines (Trip.Fish) being "stronger than the rope".)
But for the mixed-ropes cases, clearly one cannot just make a simple sling.
So, how were the test specimens configured?
2) People don't usually realize this (and books ignore it), but there are FOUR possible loadings of that "Ring Bend" structure, two of which fall under that name: the one shown here, where the loaded strands reach to the far sides of the knot; and the reverse where the loaded strands turn interior to the far ends -- here, if the ends were loaded.
The other two loadings make the knot "offset", and note the difference between the loading of the upper two strands vs. the lower two strands **in this case shown of diff.-diameter ropes**: in the former case, the thinner rope makes the choke at the entry point; in the latter case the thicker rope (blue, here) makes the choke. It should be easier for loading to pry the thicker rope (possibly more elastic, in dynamic + haul-line pairing?) around the thinner than vice versa.
3) The orientation of the 8mm+10mm ropes in the case of the Offset Ring Bend are not revealed -- no photo, no report wording. I can only guess that this (important!) aspect wasn't considered. It should be.
It should also be noted that upon tying the ORB/EDK/OOB and setting it lightly, one can rotate the knot over a range of about 180deg in plane perpendicular to the offset axis of the knot (i.e., stretch ropes so knotted across a desk/table-top and the knot can be *dialed* a half rotation from where one rope will be entering and making a backwards circle to where that same rope will be making a forwards arc to the opposite rope's backwards loop. The difference can be noticed in which direction the ends point.
In the image shown for the EDK, the position is about the middle of this 180deg range: grasp that knot by the left hand and turn it knuckles away, and the ends will come towards perpendicular w/axis of tension, towards the viewer; and one could rotate it 90deg in the opp. direction, with ends pointing away from viewer. I've not played around with the different orientations enough to know which might be more vulnerable to "rolling". I'm unaware of anyone who has remarked at this orientation variance -- but that doesn't make it unreal.
Yes, ages of practical usage suggest that likely all orientations have seen duty and survived.
(Regarding the left photo, of the Ring Bend, as given offset loading by pulling the 8mm end to the left --as it points-- and the blue 10mm one rightwards, you can clearly see the 8mm making a "forward arc" in the knot, and the blue rope making a backward loop.)
-------- comments on comments
Re "Why it's nicknamed the 'Euro Death' knot is anyone's guess"
Because that must've been the reaction to those Americans who first encountered it in use by Europeans, and felt --naturally, IMO-- that it was a most ungainly, dubiuos knot (a Ring Bend gone bad). And one still gets this sentiment, hence the efforts to remove doubt by some securing mechanism, such as a simple Overhand stopper.
As for " Same with the 'Yosemite Death Triangle', that comes in part from a simplistic view of mechanical advantage (seeing "2:1" pulley effects where in actuality the effects are well less). This is a structure I would like to see some calibrated testing on, just to see how much force amplification DOES occur IN THE USUAL SITUATIONS (tape or cordage around bolt rings). I think that such testing might reduce the alarm about it, but that's just my surmise.
Re the Blood knot: Why ... ? You have the Grapevine for ropes. And how could a Blood knot need any backup -- getting it UNtied would be the challenge.
UNLESS ... : there are many depictions of this knot carelessly showing it formed by what Stanley Barnes (1951) called "the outcoil" method -- the ropes respectively coil around the other going away from the center, and then the the ends are brought back in a big arc to dive through the center and be nipped. One well-known knots site, Grog's Animated Knots, has this mistaken imagery. In nylon fishline, this form during setting will transform into the "incoil" form -- ropes running past each other to the ends of the knot body and then wrapping back until meeting at the center and being tucked. Damn secure in the incoil form, but if tied in climbing rope by the outcoil way, YMMV: the transformation might occur, or might occur partially occur; where it is resisted, it is so by the friction-gripping mechanics of e.g. Blake's hitch or French Prusik and this impedes the setting/tightening of the knot. Why deal w/this?
The "Triple-T Fisherman's" -- Jost's (& Prohaska's) alternating Offset Fisherman's knot with Overhand tie-off (or guard, depending on perspective: as putting this *extra* one in front or at end amounts to similar results) is hardly simpler or even as simple as the Offset Ring Bend, and if anything it's a bit less stable in holding a good offset form. (Here is where some will want to take ONE testing of the knot as PROOF ... !) Tying off the ORB as I suggest below in reply to Tom can be seen as just one of the several ways that three Overhand knots in the two ropes can be arranged (that tie-off can be cast to come before the ORB).
To those wanting "scientific" over-analyzed statistical number crunching, you need to first consider how many factors can be at work here (lots of "unknown unknowns", in Rummey speak); banging the heck out of a few simple tests is of scant value, esp. when considering how much the testing might model actual usage (slow-pull vs. dynamic, or one-time vs. multiple loadings, say). We could report the data to 20 decimal points, for show ... .
NTJones wrote: "I'm surprised you show (and ... USE) a "single" EDK. Tie a ... backup ... and it blocks the knot from rolling out."
I'm surprised Tom wrote this, as he once showed (and presumably used) the notoriious Offset Fig.8 bend (but was intimately aware of the Mt.Zion tragedy) ! touche', Tom ?!
-- and sans backup on that.
As for "rolling out", has that ever happened, IN PRACTICE (vs. at absurdly high loads from a test device)?! In any case, Tom, I'd in mind suggesting that your Canyoneering site present the slightly trimmer (think "obesity is bad") single stoppered knot. There is some good reason for this: you show gold & patterned ropes joined in the OOB; the patterned rope's end will be what is pried out by loading, so IT is the end that should be tied off with an Overhand around the gold end -- no need to bulk up the knot further (but for user simplicity I'll admit that repeating the tying done initially (making the OOB (ORB) is simple).
It matters another way: in joining different-diameter ropes, you want the thinner rope to be in the position of the patterned one AND stoppered.
www.canyoneeringusa.com/cusapress/knots/index.php
Re "If the ropes are similar in size and character, then the backup knot is not required,""
Not so fast. The factors that make the Offset Ring Bend (and some other such offset knots) vulnerable to failure (coming untied) are stiffness & elasticity -- the former frustrates getting the knot snugly set, the latter provides for greater deformation on loading -- ; differences in diameter at least among the likely ropes employed by rockclimbers / cavers / canyoneers should make little difference. Also, a rough / frictive (think old & gnarly) rope can give more grip to prying parts than a slick one, and be resistant to setting tight ; then, slickness has some of its own risks.
------- re terminology:
"Offset", that's what the EDK is an example of: the knot is offset from the line of tension; and here we can see it as an Offset Ring Bend -- the commonly known knot loaded in an offset way. The site referred to for that mis-named "Triple T-Fisherman's" shows e.g. an Offset Grapevine, something Franz Bachmann thought worth trying.
7 Aug 2009, 9:50AM
Nice Q&D, Kolin. I'm surprised you show (and more surprised you USE) a "single" EDK. Tie a friggin' backup knot on there, and it blocks the knot from rolling out.
If the ropes are similar in size and character, then the backup knot is not required, but if the ropes are dissimilar (or not nice, soft climbing ropes), then the backup knot is pretty much required for security and peace of mind.
Keep up the good work.
Tom
3 Aug 2009, 4:19PM
I still think the best not for rappelling is the Triple T-Fisherman's Knot. It is easy to know and is always in a good position if you pull down the rope.
Take a look at http://www.gudelius.de/spst.htm
Maybe you want to make some tests again.
23 Jul 2009, 12:05AM
I like these QC Lab posts -- thank you!
Ring bend = water knot, right? I thought that wasn't a good knot for a rappel -- insecure in rope, only good for webbing. I thought the 3 common choices were double fishermans, EDK, or square knot with backups. Maybe I'm remembering incorrectly. I switched from double fishermans to EDK and never looked back.
14 Jul 2009, 11:17AM
Good to see some common sense overcoming cheese-headed paranoia. Why it's nicknamed the "Euro Death" knot is anyone's guess. Same with the "Yosemite Death Triangle" when attaching two perfectly good bolts - it's quick and saves some sling. The overhand knot has been the worldwide standard for decades for tying rappel knots. The "best" is definitely not equal to the "strongest" in climbing applications. Scrape off the cheese and do what works.
Just don't use the overhand for tying two ropes together for a two-rope lower where the knot will run over rock or you might wear through the sheath or worse...
13 Jul 2009, 5:31PM
Thanks for the tests - always good to see some quantitative data to back things up. Some other things to think about as well:
1.) Ease of getting your knot undone once you've done the deed.
2.) How many flakes/slots/etc.... will the knot get hung up on on retrieval? Some do better than others.
Maybe you can come up with some quick and easy tests for these issues. Thanks again. Cheers!
13 Jul 2009, 1:01PM
All of us in the QC Lab would love to develop this test and other "curiosity tests" further, but as I wrote in the original post these tests are done "quick and dirty" and mostly to satisfy my own curiosity. I agree that climbers would benefit from the additional understanding of the results which come from larger sample groups. Lack of time, not interest is the main reason we can't. We use statistically significant sampling in all of the testing for Raw Material, Work-In-Process and Batch Testing of our gear and we statistically analyze and report results with regard to mean, standard deviation, and 3 sigma for all our production testing. What all that means is that in most cases, we just can’t devote all the time we would like to these other interesting experiments. My goal with these QC Lab posts is to get some basic ideas of the answers to common climbing questions, usually focusing on the breadth of possibilities without drilling to great depth on any particular one. Thanks for your interest and understanding!
12 Jul 2009, 3:19PM
First off I'd like to say that it's great to see some scientific testing being brought to the debate about the strength of climbing knots. However, with only two data points it's difficult to say that there is any difference between any of these knots. And this is made even more difficult for the reader as there is no mention of the variability in the results. It would be nice to see reports like this with a statistical work-up of the confidence in the results. If this became the norm, climbers would undoubtably become used to interpretting these results. Besides, the majority of climbers I know actually have a technical background and would expect this type of data in a report they read outside of climbing.
Again thank you for the testing and for posting.
10 Jul 2009, 4:05PM
This is the first time I've ever heard of a Barrel Knot, or Blood Knot for use in climbing applications. Usually this is for fishing line and very small diameter cord. I have used this to tie tippet to leader for dry fly fishing applications with very small (7X) tippet. We can sure give it a shot when we have time and see how it stacks up against more typical knots for climbing/rappelling applications. Stay tuned for a future post on this test. There are some knots that benefit from back-up knots. The EDK benefits from another back-up EDK (overhand) right behind the first knot to prevent the "un-rolling" of the knot that occurs under high loads. In most cases, however, the back-up is merely for management of the tails and redundancy at any load ranges that are foreseeable in most climbing applications. Tails slipping through the knot are never the failure mode for a properly tied knot with long tails. In all cases we've tested, the knots fail rope @ knot where the rope cinches down on itself. With proper length tails I don't foresee a tail slipping through the knot making a back-up overwhelmingly necessary from a strength standpoint, but rather for the redundancy factor. Cheers
9 Jul 2009, 3:33PM
have you tested the barrel knot? how does it stack up?
with the barrel knot and the others, did you ever take into consideration the effect a back-up knot has on preventing slippage? for instance, i often back-up a barrel knot with a euro death knot.