Methodology

Why the knots are 3D models, and why flat drawings were abandoned

Every knot on TyingPoint is shown as a rotatable model of the rope itself, with the written steps kept underneath it. That is the third attempt at this problem, and the first two are worth describing because they explain the choice.

An earlier version carried per-step line drawings. They were withdrawn in August 2026 because a knot is defined almost entirely by which strand passes over which at each crossing, and those drawings rendered the rope as a single unbroken line with no over/under breaks at all. However carefully drawn, they could not depict a knot. An overhand appeared as a plain arc with no tuck.

The replacements were no better. Drawn in September 2026 with proper over/under breaks, they still shipped an overhand knot with two crossings where a trefoil needs three — physically not the knot the page taught. The frame they were drawn in was also too small for the strokes, so 39 of 64 steps had the rope sliced flat at the card edge.

The pattern was the problem, not the effort. In a flat drawing, which strand is in front has to be *decided* by whoever draws it and encoded by hand, and that decision is what kept being wrong. So the drawings were abandoned rather than attempted a third time.

A 3D model does not have that failure available to it. The rope is a path through space and the renderer resolves every crossing from the geometry, so a strand cannot be drawn passing the wrong side of another one. The steps are the same single curve revealed progressively, which means a step cannot contradict the finished knot either. And you can turn it over — which settles the question a flat picture never can: did that end go in front, or behind?

Each model is checked by machine before it ships: that the rope never passes through itself, that the knot shows the crossings that knot actually requires from most viewing angles, that anything meant to thread a loop genuinely passes through it, and that no two knots are the same shape under different names. Nine models failed those checks and were rebuilt or discarded. 99 of the 100 knots carry one today; the marlinspike hitch does not, because it only exists while a spike traps the rope against itself and the model cannot hold that. Nothing is traced, scanned, or adapted from any other knot reference, in print or online.

How knot facts are researched and checked

Each knot's history, mechanics and strength notes are checked against established published references — Ashley's Book of Knots and the material derived from it, the International Guild of Knot Tyers, Animated Knots, published climbing, arborist, sailing and angling manuals, and documented testing data where it exists.

That checking is worth describing honestly, because it did not all happen before publication. This site's knot pages were first drafted in July 2026 and audited against published sources in August 2026. The audit found real errors, including several tying sequences that would not have produced the knot they named, a few origin stories with no basis in any source, and strength percentages stated to a precision no published test supports. Those were corrected, and the corrections are sourced.

Two consequences of that audit are visible across the site. Where no reputable source publishes a strength figure for a knot, this site now states that plainly or omits the number, rather than printing a plausible-looking one — a surprising number of well-known knots have never been systematically tested, and an invented percentage is worse than an absent one. And where a knot's behaviour is genuinely disputed between sources, the page says so instead of picking whichever version sounds most confident.

If you find something here that contradicts your training or a manual you trust, the contact page is the fastest way to flag it, and corrections to technique go to the top of the queue.

Strength and security notes: what they mean and don't mean

Every knot's retained-strength figure is built from published testing and documented knot mechanics, expressed as an approximate share of what the rope itself is rated to hold — never treated as an absolute, universal number independent of the specific rope it's tied in. These figures should inform a safety margin, not replace one: rope age, dressing quality, material, and shock loading all affect a knot's real-world performance in ways a single percentage can't fully capture, and the site's own blog post on knot strength explains this in more depth.

What this site is, and isn't, a substitute for

TyingPoint is a learning reference. For any knot used in a genuinely life-safety context — climbing, rope access, rescue, or load-bearing rigging — qualified in-person instruction, practiced partner-checks, and your own inspection and testing of every knot before trusting it with real weight are not optional, and no page on this site replaces them. Knots flagged as life-safety carry an explicit warning to that effect wherever they appear.

Corrections and updates

Knot-tying knowledge is fairly stable, but not entirely static — testing methodologies improve, and genuine mistakes do get caught. Where a specific fact on this site is found to be wrong or meaningfully out of date, it gets corrected directly on the page in question rather than left standing. If you spot something that looks incorrect, the contact page is the direct way to flag it.