How strong is your knot? | Hacker News Reader

How strong is your knot?

news.mit.edu

286 pointschmaynard132 comments

My favorite, as many others are sharing here, is the figure 8. For climbing, it's a super versatile knot and there's lots of variations for specific use cases. It's also a great knot for discussing the properties of a good knot.

I also have found the trucker's hitch really useful from time to time. It allows you to tie down a load, for example a kayak on top of a car. You tie the initial hitch, and then you get to crank down on the hitch before tying the final knot. It's like a mini pulley system without a pulley.

The knot I use every day is a modified shoelace knot. You do an overhand, and then the first loop. Before you make the second loop though, you make two wraps around the first loop instead of just one [0]. Once I started using that knot, I don't think I ever had a shoelace come undone.

[0] Third video here, "Two Loop Shoelace Knot": https://www.sunnysports.com/blog/3-best-knots-tying-shoelace...

As for climbing, don't forget the double-fisherman's... To tie two ropes together for a double-rope rappel. It's the knot equivalent of a wrist-handshake.

You should use an EDK (despite the name sounding scary) if you're anywhere with cracks or areas that a double fisherman's will get caught. The EDK is asymmetric so it's less likely to get stuck, and is generally faster to tie (and untie). Fears of it rolling out are a bit overstated - just leave a solid 6 inches of tail or more.

Or just use the Alpine Butterfly. Stronger, easy to tie, optionally asymmetric, usable as a loop or a bend, quite secure (slightly less so than Ashley's #1452 but still quite good), and non-jamming (unlike the figure 8).

I disagree that it's easier to tie, especially if you're rappelling at night. The EDK is dead simple, it's literally an overhand knot on two ropes. You can do it with your eyes closed. It's also the easiest knot to quickly inspect for correctness by your climbing partner.

Insightful, thank you.

The trucker's knot is awesome. Hugely practical. I use that knot on every load.

https://www.animatedknots.com/truckers-hitch-knot

And in musical format (from the same dude that did 'what does the fox say')

https://www.youtube.com/watch?v=TUHgGK-tImY

Check out the Versatackle[1]. It's very similar, but self-locking so it's easier to tie off at the end: you don't need to maintain tension on the working (free) end for it to maintain tension on the load!

http://useknots.com/other/versatackle/

The two loop shoelace knot almost qualifies as life changing. Such a small change to something I learned as a child makes such a big difference. Not only do your shoes not come undone, but they don't loosen.

I have to agree. There was so much I didn't know about lacing a hiking shoe but I must admit, that is the part of it that has been most useful outside of hiking shoes.

I use "Ian's knot" to tie my shoes [0]. I was so excited when I first learnt this, it is very quick to do and I like that the two loops of the tied shoelace sit flat and straight on the show. Works great for short laces, and never had a problem coming undone.

[0] https://youtu.be/_O-xaJrao1w

If we're including hitches, the progress capture clove hitch is amazing.

If you use a taut-line hitch or equivalent, you can crank down on the "pulley" after tying the final knot.

Doesn't he tie the loop the wrong way in both cases? The (regular) shoe lace knot should be like the reefer knot, not the granny knot.

I've had less problems with simple shoe lace knot after I made sure the loops are correct.

I have a knot problem that I think pretty much everyone has encountered. How do you tie a bundle of branches or a cardboard box or something compressible as tightly as possible? The first way everyone tries is to cross the rope like you would for shoelaces, then putting pressure on the crossed rope with your finger, and trying to cross the ends of the rope with your one free hand to finish the knot. Inevitably the knot is way too loose. Surely there must be a better way?

EDIT: the word "ratcheting" nicely conveys the idea I was looking for.

I would have said the trucker's hitch, but there are other "ratcheting" knots that would work well in this scenario.

https://goneoutdoors.com/rope-knots-give-ratcheting-effects-...

I would use a trucker's hitch, personally.

Instructions: https://www.animatedknots.com/truckers-hitch-knot

Unrelated music video about it: https://www.youtube.com/watch?v=TUHgGK-tImY

A constrictor knot like a boa knot.

https://en.wikipedia.org/wiki/Boa_knot

If you tie an alpine butterfly somewhere in the rope, and then pass the other end through it, you can achieve a 2:1 pulley system in the rope. This greatly increases the compression of the lashing. Hard to explain, but you're basically using a fixed loop in the line as a pulley to increase the force applied.

You can do the same thing going around twice. It's simpler, and gives the same mechanical advantage.

Much more friction and stretch in the system. The less rope, the less stretch.

The standard ratcheting knot is the rolling hitch, although several people on this thread mentioned potentially better knots for this application. My general trick is to use rolling hitches together with the bow-and-arrow technique. The concept behind bow-and-arrow is simple -- if you pull on a segment of rope perpendicular to the rope, it gets tighter.

If I'm tying something to my roof rack, I'll get is as tight as I can going around a few times with a rolling hitch. Then I'll get a piece of rope to pull on that one. How you do that depends on domain.

maybe Cabestan (reinforced with a few more loops) or hangman's knot

The knot your looking for is the gleipnir knot:

https://m.youtube.com/watch?v=Kxxhg-tf8Ug

It’s a recently invented knot that works extremely well for bundling irregular objects

I second the suggestion of a trucker's hitch. A ratcheting knot can be useful in this situation, but IMO what you're really after is a mechanical advantage. The trucker's hitch gives you a mechanical advantage by using a loop in the rope as a pulley. There are a number of ways to make the loop and it doesn't really matter which one you use. This video gives a pretty nice explanation.

https://www.youtube.com/watch?v=raTYwsfnsXg

Once you cinch the load down using your mechanical advantage, you can use any number of knots (ratcheting or not) to secure it.

I also love the truckers' hitch. The key benefit is the tautline hitch that makes it easy to tighten. But sometimes I use an alpine butterfly loop and tautline hitch. I think that it's stronger.

Whoa, I had no idea the trucker's hitch was a real knot; I thought it was an Ylvis joke.

I think what you're looking for is a constrictor-type knot. There are a few options here: https://www.animatedknots.com/?s=constrictor

I am sure the other knots people are posting work well, but most people will never remember them. I suggest a surgeon's knot. It does the job well enough and is simple enough that anyone can use it.

https://en.m.wikipedia.org/wiki/Surgeons_knot

My list of favorite knots:

1. Square Knot / Reef Knot -- This is the shoelace tie, and its one of the simplest knots. Learn this knot backwards and forwards you'll use it a lot. Avoid the granny knot (a mistaken reef knot). Beginners will often tie the granny knot instead of the square knot. If you can do a square knot, the "Surgeon Knot" is a useful variant which is more secure.

2. Figure Eight -- Superior to the overhand knot in almost every way imaginable. Never use an overhand knot, always use figure eight instead.

3. Bowline Loop / Sheet Bend -- When a loop must be sturdy and unmoving, the bowline is one of the best. The Sheet Bend is nearly the same, except you tie two different ropes together. I keep a bowline loop on any "pocket twine" because making a bowline loop is the start of many advanced knots. Such as...

4. Package Tie -- Presents obviously look nicer with a ribbon in package tie form, but any large, box-ish object can be secured with the package tie. Really good for lunchboxes / bags in daily usage. Add another Bowline loop on the end for a convenient handle. https://www.animatedknots.com/package-tying-knot

5. Cable Lacing -- When working with wires and electronics, the cable-lacing knots are great for organizing wires. Its quick to tie and organize, and looks great. https://www.flight-mechanic.com/lacing-tying-and-cutting-wir...

6. Poacher's Noose -- When a loop needs to "collapse" to bind objects together, various nooses come in handy. Poacher's noose is my favorite, ties easily and fewer people recognize it compared to the hangman's noose. Both nooses have political implications however so be wary of who is watching you when you tie a noose.

7. Butterfly coil -- When storing most paracord, twine, or ropes... butterfly coil is fast and effective storage.

8. Over-under coil -- Wrapping up thicker wires / cables (especially electronic cables / wires, but also garden hoses) is best with over-under coil instead of butterfly coil.

9. Half-hitch -- Its the knot you've been using all the time, and you probably don't even realize it. Its called a "half-hitch", because you need two of them to actually make it secure. Two-half hitches solves many problems.

I can think of at least one situation where an overhand knot is better than a figure-eight, where speed and ease is more important than strength.

Figure Eight knot is tied almost identically to an overhand knot, except you "turn left" instead of "turning right" into the loop. Under no circumstances is a figure eight knot harder to tie than an overhand knot.

https://www.101knots.com/wp-content/uploads/2016/11/How-to-T...

In fact, I've practiced the figure 8 knot so much, that it takes more effort for me to do an overhand knot these days. Figure 8 is really easy.

Just as the "Granny Knot" is a mistaken square knot / reef knot... the "Overhand Knot" is basically a mistaken figure-8 knot. Just turn the other way, and the knot will be far, far stronger.

Forgive me, I thought we were talking about with the follow-through. In this case, I guess it's just as easy if you practice, which most people don't. People do practice the overhand knot though.

So, an overhand knot is likely easier for most people, where a figure-eight would be better for those who practice.

They're pretty identical in terms of ease of tying, it's just one extra flick of your wrist before you pass the working end back through.

I think the difference in speed would be measured in milliseconds. You'd need a slow-motion camera to spot see any difference at all.

The double overhand knot is better than a figure-eight when used as a stop-knot at the end of a rope that you want to prevent running out through a block or clutch, because it's easier to untie after being jammed.

Other than aesthetics (and perhaps as a hobby) are there circumstances in which cable lacing would be preferable to cable ties?

Cable ties / zip ties are often better when you need to change the configuration of wires (cutting the zip tie, moving the wires around, and setting new ties up).

Zip ties are more prone to overtightening, but as long as you are aware of the overtightening problem, there's probably no real reason to avoid them.

The main advantage of knowing cable lacing is that it generalizes. There are many long-objects that benefit from the cable-lacing knot. For example, a rope can secure a bunch of 2x4 together using the cable-lacing technique. Any "long lengthwise bundle" of objects can benefit from cable lacing.

Zip ties also have hard corners which can cut into cables in high-vibration environments. Even with correct tensioning (using a cable-tie tool[1] of some sort) they're a potential hazard. Wrapping the cable with a shielding material first can mitigate this, but cable lacing is more reliable.

[1] https://www.mcmaster.com/zip-tie-tools

You might find these interesting:

Cable lacing on the Curiosity rover: https://news.ycombinator.com/item?id=4472312

The Lost Art of Lacing Cable (2018): https://news.ycombinator.com/item?id=20582844

2. False. (Double) overhand aka euro death knot is superior for joining two ends of rappel line due to its smaller profile - lower probably of getting jammed on pull

Knots getting stuck is super annoying (and potentially life threatening) on rappel.

In addition, it is generally believed to be less liable to rolling off the ends than the flat figure-of-eight bend (not to be confused with the standard figure-of-eight bend/flemish bend)

Yeah but edk rolls too, that’s why you need to leave a foot long ends at least

Don't solo overhand knots significantly weaken the cord they're tied into?

It’s two lines and also forces are very small on rappels

All knots significantly weaken the cord they're tied into. A very significant margin of error is built into the ropes to account for this.

Rappel only exerts a force equivalent to your weight, plus a little extra because of bouncing that naturally occurs during descent. 1kN = 224lbs force, so if you weigh 200lbs you can expect to exert about 1kN of force on a static rope. About the weakest static rope people frequently rappel on is a 7mm line, and a typical one[1] is rated to 12.4kN. So abrasion, knots coming undone, losing control of the brake strand, etc. are much larger concerns than static rope breaking due to knots.

Static ropes are ideal for rappelling, but climbers will often use the same rope they lead climb on to rappel on, and for lead climbing applications your rope must be dynamic. The weakest ropes you would usually lead on would be something like 7.5mm half ropes. A typical example[2] doesn't list a maximum breaking strength, because this statistic isn't relevant for dynamic rope. The purpose of a dynamic rope is absorb force during a fall: if you take a lead fall on a static rope, it's unlikely that the static rope will break, but hitting the end of a static rope is just as bad as hitting the ground with the same force. When you hit the end of a dynamic rope, it stretches to absorb this force, giving you a gentle fall. The maximum force on the rope listed is 9.9kN, which is from a UIAA-standard test fall: these falls are far more violent than anything you are likely to experience in actual usage. Ostensibly the rope can handle more force than that, but 9.9kN of force would probably kill you anyway. Luckily the rope stretch means it's unlikely you'd get anywhere near that level of force: a typical lead fall is in the range of 2kN. So again with dynamic rope, margins of error are very high, and abrasion, knots coming undone, losing control of the brake strand, etc. are much larger concerns than static rope breaking due to knots.

This is all fine and good on paper and in a lab, but does it work in the real world? The answer is yes: real world accident statistics[3] indicate that the most common accidents on rappel are simply rappelling off the ends of your rope. I wasn't able to find a single accident caused by rope breaking due to knots. Ropes do break, but this is typically attributed to abrasion over sharp surfaces, or long-term wear, and even these accidents are rare.

[1] https://sterlingrope.com/store/climb/cordage/tag-lines/7mm-t...

[2] https://www.mammut.com/us/en/p/2010-02762-11190/75-twilight-...

[3] http://publications.americanalpineclub.org/about_accidents

EDIT: Just to be clear, although there is a correlation between thickness and strength, the thinnest ropes aren't necessarily the weakest. The examples I give are examples of some of the weakest ropes I have used. This also shouldn't be perceived as a criticism of these ropes: as I said, these are more than strong enough. Thin ropes are certainly more susceptible to abrasion and aren't as durable, but this is a tradeoff for having a very light rope for moving fast. I used these ropes because they are the best ropes for the situations I planned for.

> All knots significantly weaken the cord they're tied into. A very significant margin of error is built into the ropes to account for this.

I'm not a climber, what I usually use is cotton, jute, or sisal twine in a day-to-day basis for various tasks. Cotton twine is very easy to keep in your pocket and very clean, but its the weakest by far. Sisal and Jute are scratchy and messy, but they have superior strength.

Lets say you've got a 40lb cotton twine in your pocket. You can either use an overhand knot and drop the strength to ~20lbs, or you can use a figure-eight knot and instead have ~32lbs of strength.

Both the overhand knot and figure eight knot weakened the twine. But overhand knot is among the weakest of knots, with a 50% drop in performance. Having only a 20% drop from figure-eight makes a huge, noticeable difference in strength.

Sure, those might be the numbers in your use cases, with your materials, but I'd be very careful about claiming those numbers if the situation is changed even slightly.

1. Change in strength is dependent on the cord material and weave. Stiffer cord will lose more strength when tied, while more supple cord will maintain its strength better. The tradeoff is that supple cord is more susceptible to tangling. Citing one number as a strength loss for different knots is sure to be false. I would guess, for example, that your cotton twine sees less percentage loss in strength in both knots than your jute.

2. The change in strength is also dependent on the sharpness of the folds in the material. A sharper turn is weaker than a rounder one. If you tie a knot around something, that decreases the sharpness of the turns, which is going to increase the strength of the knots. We aren't typically just tying an overhand or figure 8 midline in a knot and pulling on the ends--I can't think of ANY place where that's useful. The closest thing I can think of is a stopper knot, and neither of these knots is ideal for that application. I'd expect to see a higher loss of strength with flat overhand/flat figure 8, than with a ring bend/flemish bend (which correspond to the overhand and figure 8). This is because the weakest points in the flat knots is the turn entering the knot, whereas the retraced bend versions of these knots, the cord enters the knot in a straight line.

3. Breaking is often not the largest source of failure. A flat figure 8 will roll and come untied long before the cord breaks in most materials, while the flat usually binds within 1-3 rolls. The breaking load of a flat overhand is much, much higher than the rolling load of a flat figure 8 in climbing ropes I've tried it with.

> We aren't typically just tying an overhand or figure 8 midline in a knot and pulling on the ends--I can't think of ANY place where that's useful. The closest thing I can think of is a stopper knot, and neither of these knots is ideal for that application.

I'll admit that I've mostly used this "useless knot" test methodology you're criticizing here. But in its defense, it is an easy to do micro-benchmark with consistent results.

I guess I could setup a more cohesive test to measure the strength in more "realistic" scenarios. That's really all it comes down to: the strength of the rope in whatever specific situation you're doing.

I'll note that 40lb-strength rated cotton twine is just strong enough to be useful, and just weak enough that a bad knot or two will cause the twine to snap under moderate usage. (Ex: Package Tie to support a box containing 20lb).

> I'll admit that I've mostly used this "useless knot" test methodology you're criticizing here. But in its defense, it is an easy to do micro-benchmark with consistent results.

That's fair, and it does prove that different knots affect strength differently. That's a base principle that is useful to know even if it doesn't get you far on its own.

With regards to your application: while it lacks the "finesse" of tying knots that give you just enough strength, simply getting stronger twine would solve this problem without the complexity.

In general, there's a whole lot of ways knots can fail. If you can solve the problem by just using stronger cord, that frees up mental cycles for dressing, avoiding slippage, effective communication, ease of untying, etc. If you're just tying up boxes do whatever you feel like doing. But for climbing, I'd rather have the rope be more than strong enough to do its job, and have those brain cycles for focusing on preventing other system failures.

For what it's worth: lots of knot literature and lore cautions people about overhand knots as weaking the line more than other knots.

That's why I asked.

The overhand [1] and double overhand are quite different knots. Actually, since you specify that you mean the "Euro Death Knot", it is actually a "flat overhand bend" [2].

The parent's statement is not false, it's just that one must be precise about naming, because casual use leads to incorrect assumptions by others.

Other than that, yes, you are correct.

* [1] https://www.animatedknots.com/overhand-knot * [2] https://www.animatedknots.com/flat-overhand-bend-knot

The "overhand knot" is the basis of many other knots. For example, the half-hitch and double-overhand knot.

While double-overhand and half-hitches are very useful, the "vanilla" overhand knot is among the weakest of knots and should be avoided in almost all circumstances. I can't say that I've ever found a good use of the "vanilla" overhand knot.

I'm not a climber. I just happen to have some twine on my often for daily use.

Regarding cable lacing, it seems more complicated to add or remove cables at different points of the lace, at least when compared to velcro cable ties. However, it is cheaper and does save you time for when you don't have cable ties but do have some versatile string laying around.

I think about the only place that cable lacing is still actually used is in space craft wiring harnesses, where the harness is generally not modified once built. Even then, NASA at least doesn't use the single continuous strand method of lacing. We tie "spot ties" which function identically to Velcro with the exception that you have to cut them if you do modify the harness.

The zeppelin has always been my favorite "knot" (the zeppelin is a bend), but one knot is not the "best"... the alpine butterfly they compare it with is a loop. They're used in different contexts.

I'm just knot layman, but other knots I like are: the icicle hitch, bowline (+yosemite stopper), trucker's hitch, siberian/evenk/halter hitch, tumble hitch, and the chinese button knot.

A Trucker's Hitch is a _super_ useful thing to be able to rig. There are dozens of ways to tie it, and just as many situations where it is useful.

The knot I use most frequently is the taut-line hitch. This is an adjustable loop knot that holds its position under tension. It works well with the trucker's hitch or to set up tents and other temporary tension rigging.

For joining two lines, I like the Zeppelin hitch.

THANK YOU! The Zeppelin bend is my favorite knot, simply due to the name.

A useful, if relatively unknown knot, is Ashely's Stopper Knot, great for the little strings that go into hoodies.

https://en.wikipedia.org/wiki/Ashley%27s_stopper_knot

I am surprised to see no one has put in a good word here for the buntline hitch[1], which is very handy for tying a rope to something fixed (and also happens to be a standard necktie knot).

1: https://www.animatedknots.com/buntline-hitch-knot

I was highly disappointed.. this article had so much potential.

Really it's great but it was such a let down to only see an animation of one knot with their special cord in the article.. I was really hoping to see animations/gifs of a bunch of interesting knots!

For standard knots the Klutz book of knots is really excellent! https://www.amazon.com/Klutz-Book-Knots-John-Cassidy/dp/0932...

The article doesn't mention fishing knots; in case anyone's curious anglers care a lot about knot strength but strength changes based on whether you lubricated the knot when tying and what kind of line you used to tie it (braid vs smooth fluorocarbon for example).

My favorite fishing knot is the uni knot, which is a slip knot that can be used to join lines (double uni) or tie things to line. It's super strong but fishing knots aren't designed to be untied and always cinch down (vs climbing knots which can't collapse / strangle a person). Slipknots are very useful: I use a 3-wrap uni when I'm trussing thankgiving fowl for the oven!

There are some really cool fishing knot strength testing videos on YouTube. The line-to-line joining knots are my favorites -- people get heated debating the merits of the FG knot vs the alberto vs the double uni.

Indeed. Fishing knot strength is always a hot topic of debate, particularly during the slower, off-season winter months when everyone's hunkering down and doing tackle maintenance. Here's a pretty good breakdown of the line breaking strength of various fishing knots across monofilament, fluorocarbon, and braided line: https://www.knotsforfishing.com/knot-strength-chart/

I personally prefer the Palomar Knot as it is probably the strongest knot that is also easy to tie. The Improved Clinch Knot and its brethren are also handy to know since there are so many variants that have high breaking strength; I typically teach the Improved Clinch Knot to folks new to fishing.

I love the palomar too! The only issue with it is that it makes tying pretied leaders tricky since you have to pull the whole terminal end through the loop. Not as versatile as some other knots.

I find it surprising how many people use clinch knots vs the uni in real life. It's certainly easier to tie clinch knots but I've never found the ease worth it, since the uni is so much stronger and can be used for more things. Once I got fast at uni the clinch knots went out the window.

I was recently home and found a classic Klutz book we read growing up called Earthsearch. It has an aluminum cover and talked about tons of worldly topics.

Recycling, race, population growth, poverty. All done at an elementary school appropriate level.

I thumbed through the book and read some excerpts and was surprised at how good it still is. So many times books from 20-30 years ago cover things wildly differently than we would today, but nothing was that way.

I'm thoroughly impressed with Klutz and will definitely make sure some of those books end up with my nieces and nephews.

Joining two lines, (of the same or different diameters) is very much of interest in climbing, and these knots often do "collapse." The simplest/most common knot used for this purpose is even named for fishing- the double fisherman's knot.

It is used for joining the ends of a cordelette, combining ropes for rappelling or ascending, or in a single ended version, backing up (collapsing on) a knot used to tie in to the rope, such as a figure-eight follow through or bowline.

I think you will find that equal to your desire to debate which is the strongest knot is the desire of a climber to not have to think about it- when 100 feet off the deck and on thin footing.

In climbing, the offset overhand bend (EDK) is also very commonly used. It's arguably simpler than the double fishermans.

I'm a huge fan of the zeppelin bend - it's a bit exotic, but is catching on in some climbing circles. Competitive tree climbing is carefully adjudicated for safety... in the PNW at least, judges have absolutely no problem with it being used as a joining knot on a lifeline.

Just a plug for Clifford Ashley, the god-father of knots and their tying. His book has some of my favorite knots in it and is a joy to have received as a gift when I was younger. His stopper knot is a pretty cool one.

https://en.wikipedia.org/wiki/The_Ashley_Book_of_Knots

ABOK is a wonderful book, but I wouldn't recommend it to someone who just wants to learn a few more knots - books by Geoffrey Budworth or Des Pawson are much clearer in terms of how to tie (though far less encyclopedic than Ashley).

This is also a very popular knot reference, although limited to shoelaces: https://www.fieggen.com/shoelace/index.htm

When I was in the 8th grade we were rappelling from an abandoned railroad trestle using some scavenged rope we found. While tying knots and having our own Indiana Jones adventure, a farmer came along and said we were trespassing but more importantly were tying our ropes incorrectly - that "more people die from square knots than any other kind of knot each year".

While I figure this is some passed down knowledge, i have always wondered if this was semi-true. Square knots are terrible.

Practicality is also big one. There are tons, but I've learned bowline over 10 years ago and use it every day when sailing. It's easy to tie and can be taken apart by hands.

The bowline is fine for non-safety-critical uses, but cyclical loading causes it to slip, so I'd suggest anyone who's learning the bowline to learn a "safer" variant and just always use that one (examples are Yosemite Bowline, End Bound Single Bowline, Double Bowline, Bowline on a Bight).

Been using bowlines for safety(fire dept) and sailing 35yrs. Never ever ever seen a proper bowline "shake loose". Can tie it weakside one-handed, upside down, behind my back, in the dark, underwater, while brushing my teeth.

You use a bowline for life-safety in rescue operations? That's really unusual, in my experience (firefighter/rescue tech in upstate New York)

I volunteer for a lifesaving organization in Germany and the bowline is one of the most important knots there. One example is diving, where the diver is secured by a bowline around his hip.

Interesting... Around here if divers are working off a line, they're tagged in with a short lead to the mainline. The lead will have carabiners on either end that are either factory stitched in place, or clipped into a figure-eight. Those aren't meant to be under any tension though, they're just for maintaining orientation.

If a diver is attached to an actual haul line (not common, but happens in swift-water recuse scenarios), then they're going to be clipped in via a carabiner, which is tied on with a double-loop figure-eight. The double-loop figure-eight is the standard "if your life is gonna depend on it" knot around here.

(I am not a rescue diver, but do support dive operations from time to time)

I left firefighting behind in the late 1970's. Bowlines were tied properly then.

It's the "proper" part that's the problem. Even the Yosemite finish is problematic. Google for "deaths from climbing with bowline" and you'll find lots of veteran climbers who have died or been seriously injured from their bowline coming undone, either because they did it messily and it capsized, or they screwed it up and didn't realize it until they took a fall.

Yup, that's why climbers replaced it with the figure-eight loop decades ago. Which is a bitch to untie after loading, but at least it won't self-upset.

Circling back late on this one. Lots of article, many rehashing the previous articel. Common themes seem to be "improperly tied" and "magically untied itself" which to me implies "improperly tied" again. No fault with the figure 8 tho. When we rappelled, and in ROTC mountaineering we used bowlines and Army Gold Braid lines. That should date me.

I use the Water Bowline (in which the "loop" is a clove hitch) in situations where the knot won't be continuously loaded and might flog.

i've always thought it woudl be interesting to do large-scale simulations of knots using a physics engine- generate random topologies, pull both ends, compute some physical parameters such as "force required to pull apart". Then sort of walk around knotspace using genetic algorithms to evolve new knots. THere is probably a much more practical RL approach that doesn't require GA>

Knots aren't super difficult to 'program' in the first place. The first computers were for looms and their weft patterns, after all. It's always been strange to me that no one has decided to set up the math of knots. Most knots are just subtle changes to the same basic elements (bite, over/under, loops, etc). I'd figured there are enough Eagle scouts with PhDs in math to have done this already, but no.

What do you mean? Knot theory is a field.

knot theory is abstract- I care more about the material sciences aspect.

Yes, thank you. Actually going about tying a knot would be much more ... useful? Knot Theory is well and good and amazing, but the physics of knots is also needed.

Knot theory

https://en.wikipedia.org/wiki/Knot_theory

I meant more like how to actually tie a knot, not just topologies.

So... the math of knots is... knot theory, and you're correct that it is topology.

How to actually tie knots -- in a practical sense -- would be knotcraft [1] -- not sure what kind of math that would entail? What kind of connection did have in mind between that and the math of knots and programming? (in your original statement)

If it's a mathematical treatment of the strength of knots (physics), I believe that's what the article is about, which is quite interesting.

[1] https://store.doverpublications.com/0486245152.html

Well, since this theory does not actually exist yet, and I only have a 'feeling' that it should, I'll have to speculate very hard here:

Something where I have some sort of representation of the physical knot that is much less than just a drawing of the knot, but more than just saying 'this is a reef knot'. Something you can do operations on.

Say you are trying to tie a reef knot. You can tie a reef knot many ways. Say you want to write down how to tie it one way.

So you take 2 strings. Say you represent that by the number 2.

Then you have to twist them together, so you write down something like 2T.

Then you have to pull a bite on each of the 2 strings string. So you write down 2T + 2B.

Then you have to retwist them, but the opposite way. So you write down something like 2T + 2B + 2(-T). Then you have a reef knot.

But the thing is that since there are a LOT of ways to tie a reef knot, that 'formula' should be able to be manipulated to produce the other 'formulas' that also produce a reef knot. Also, a reminder, this 'formula' thingy I just made up should not be taken as any sort of gospel.

We all know that when you pull hard on some knots, that they just decompose into a simpler knot. So this representation of knots should be able to be worked though on paper and give the right result. Or with slip knots and hitches, knots that are designed to decompose into just a string when you pull hard on them or something. So when you take the knot tying 'formula' and then work through it, it just returns a string.

I think you could then apply the physics to these 'formulae', like, each twist you do reduces the rope strength by X% and increased friction by Y%.

Some sort of thing like that is what I am thinking, I guess.

I'm somewhat fascinated by people's ability to tie knots that are more complicated than the basic shoelace one. I've watched hundreds of videos on knot-tying and, even with basic step-by-step instructions, can't manage one properly. Is it just the fact that this stuff is better taught practically or am I just somehow the worst knot-tier of all? Even that reef knot gif eludes me...

> I'm somewhat fascinated by people's ability to tie knots that are more complicated than the basic shoelace one.

> Even that reef knot gif eludes me

--------

These statements are in contradiction. The reef knot IS your shoelace knot. If you can tie your shoelaces, you can tie a reef knot, by definition.

Specifically, a "shoelace tie" is the slipped reef knot. "Slipped" means that you leave little loops so that its easier to untie later. Get rid of the "loops" in your shoelace tie, and you'll make a fine (more permanent) reef knot.

I think many use the granny for their shoes.

Shame on them.

I prefer the granny knot because it rotates the shoelaces so they're less likely to get caught on things. I've never had any trouble with it coming untied. https://www.fieggen.com/shoelace/grannyknot.htm discusses the difference in detail, but fails to acknowledge this benefit of granny knots.

Or wear slip-ons.

Yup, I do. Never was taught a different method and this one works well enough but I'd really prefer not to spend the rest of my life being so inept at a basic skill.

I couldn’t remember them until I started needing them. Figure eight, clovehitch, barrel knot, overhand, and fisherman’s are my daily drivers for climbing. Prussiks are useful, but there are more efficient ways of doing progress capture.

Bowlines can be used instead of figure eights, but they can shake open if you aren’t careful, which is obviously bad.

You can add a stopper knot to the bowline I think.

It really takes practice. I can tie a bowline or a figure eight in my sleep, because I used those for climbing.

The taughtline hitch is an amazingly useful knot... but I flub it most of the time. Similarly, the square knot -- even if I take my time, it's a granny with about 45% probability. These aren't knots my life has depended on, and I don't have the muscle memory.

So instead of trying to repeat the steps of the knot, form a set of verbal directions in your mind. Then close your eyes and picture yourself doing them. Then grab some rope and try with actual rope, again with your eyes closed. Repeat the same knot 10 times a day for 10 days.

This method works really well for most because closing your eyes removes all extraneous info. You see the model of the knot and the instructions for tying it, not the messy tangled up physical rope that you aren’t even using yet.

"can't tie a knot? Tie a lot!"

I should learn more knots, but I so seldom need any. I look it up, generally get confused by the diagram, eventually work through it, and then forget it. I joined two pieces of rope last year and secured one tarp. I've tied up laces on my shoes a few times last year too.

If you want a fun one, look up Somerville Bowline. Great take on a bowline that is non-directional, non-collapsing, very hard to get wrong, small, quick to tie, and cannot be undone under tension. Oh and it’s completely symmetrical so you can pull on either end.

Fun, heh.

It does look like it would have a tendency to jam and be hard to untie. One of the key points of a good knot is that it be easily untied once the strain has been removed.

Depends on the load of course, but in my experience having tied it a lot, it only jammed up once, taking an extra thirty seconds to undo. This was when I put about 200 lb on it and swung around a bunch. Jute rope also has a ton of friction, synthetic being smoother would be even less likely to jam.

It's an interesting and decorative knot - I've not seen anything like it before. I'll have to give it a try. I'm trying to think of another quality bight knot that can be tied with the object in place. Not the bowline, not the butterfly.

If you want another one, check out the WykD Fast Bowline. It can be tied as both a regular version and a quick release. It is even simpler but has less symmetry. Interestingly, the quick release is more structurally sound with this one. There is also a variant of the Somerville where the loop is made backwards, which seems to go by Struggler’s knot. It purports to be nearly as strong, and supposedly can be tied with one hand.

Also, there is a method of tying the Somerville called the reverse Somerville. It’s the same knot but tied inline rather than with the bight.

The downside to the Somerville is that you generally can’t get it really tight around the object. A variant of it exists which allows you to do so at the cost of a bulkier knot: before making the loop, cross the bight over all the rope, go under it and come back to where it just was, similar to how you would start a reef knot. I have used that technique before with some success but not enough to recommend it beyond messing around.

I’m curious what this would look like tying dress shoes with a Berluti knot.[0] (I don’t think this knot is in the ABOK, but I could be wrong.[1])

0. https://www.hangerproject.com/blog/2018/01/12/berluti-knot-t...

1. https://en.wikipedia.org/wiki/The_Ashley_Book_of_Knots

I've tied my shoes very similar to that for many years. The only difference is I only pull the bunny ears through the hole. Not the lace ends.

https://www.fieggen.com/shoelace/secureknot.htm

and? Wheres the research data and test results? or are we just supposed to take you at your word?

the news release does mention that the research was published in Science. I agree that it wouldn't have taken much to add a link, but here; https://science.sciencemag.org/content/367/6473/71

IMO, the 2 most useful knots I know (aside from the shoelace knot) are the Carrick Bend and the Bowline.

The Carrick bend is especially useful - it's basically a square knot that doesn't come undone on its own.

I'd put in a vote for the sheet bend, although it's kind of isomorphic to the bowline. The "free" ends are in different places, and it's a joining knot.

The bowline is rather special as it

* doesn't tighten the big loop when you pull on it.

* it can be done with one hand (with lots of practice!)

I wonder how strong the bowline is compared to the others though.

When I was in boy scouts, at one of the camps one year one of the people running the thing had this little game. The idea was that two people would each tie a bowline around their waist, and connect the two ropes with a square knot. Then do it as fast as you could.

Two of the kids in my troop got it down to two and a half seconds. It looked like one of those magic tricks where you flip a rope around and it knotted itself.

I did eventually learn the one handed bowline, but it still takes me twenty seconds to do either that or a basic square knot.

My vote for most useful knot is the figure-8. Can be used to secure something to a fixed anchor, as a bend, or to put a loop in the middle of a line (something like an Alpine Butterfly is a better option there, but if you only know the figure-8, it will get the job done).

I'd love them to do a break down on common knots as I trust my life to figure 8s and fisherman knots pretty regularly.

get into alpining and trad multipitch. Best way to learn a ton of knots, rope systems, and cool tricks!

Yeah, me too, but I have never read of an actual knot failure causing an accident outside of user error (wrong knot, not tied correctly, incorrect application, etc).

edit: in reference to rock climbing accidents...

Look for older reports from rec.climbing. For example: http://canyoncollective.com/threads/some-knot-testing-data.7...

And https://user.xmission.com/~tmoyer/testing/EDK.html

Some of the testing folks at black diamond did a bunch of tests when we were all jabbering about knot safety about 20 years ago. Kinda sad the old posts are getting so hard to find!

We need a search engine focused on organizing the world's information and making it accessible and useful, instead of advertising.

There are two basic classes: knots on slippery material like monofilament and knots on ordinary stranded or braided line.

Fishermen have worked out good knots for monofilament, but they jam when used with stranded line.

The strength of a knot or bend in stranded line can be predicted without technical analysis by looking at the curvature of the line as it passes through the knot. The tighter the curve, the lower the breaking strength. The strongest bends, like the Zeppelin, have very soft, symmetrical curves. They are also the easiest to untie.

I like things like this appearing on HN, it brings all of the HN climbers out of the woodwork :)

My favorite knot which continues to confound MIT to this day is called the Gordian Knot.

link to the Science article: https://science.sciencemag.org/content/367/6473/71

knots are beautiful, useful. Just note that a knot creates a weakness, in some specific cases, if you need weight and resistance gain, a rope splicing is much better

Does anyone know what the best knot is for stitching?

Obligatory link to this handy reference of different knots and how to tie them: https://www.netknots.com/rope_knots

knot very strong at all

Can we get the link replaced with an HTTPS one (https://news.mit.edu/2020/model-how-strong-knot-0102)?

Reply on news.ycombinator.com