Which Fishing Knot Holds Best, and Why Every Published Table Disagrees

What the calculator actually does

Three inputs and three outputs. You give it a rated strength, a unit, a line type and a knot; it returns an efficiency, a breaking strength in pounds and the same figure in kilograms. Inside, there is one lookup table of twenty-three knots, one table of three line adjustments — monofilament 0, fluorocarbon minus 2, braid plus 1 — and one multiplication. There is no model of diameter, no model of the hook or swivel the knot is tied to, and no memory of how the knot was tightened.

The stored percentages take nine distinct values: 95, 92, 90, 88, 85, 80, 75, 65 and 60. Three knots tie at the top — the Palomar, the Snell and the double Palomar — and eight sit together at 90 %, which means the tool cannot separate a uni knot from an FG, an Orvis, a nail knot or a loop-to-loop connection. That is not a fault of the arithmetic; it is what happens when a ranking is stored to the nearest five points. Read it as a coarse ordering and you will not be misled.

Two details found by running the code rather than reading the screen. The efficiency is clamped between 30 % and 98 %, but the highest value the table can ever produce is 96 % and the lowest 58 %, so the clamp never fires — it is dead code. And if a knot is ever added to the dropdown without an entry in the efficiency table, the calculation silently falls back to 85 % while the label goes on advertising whatever percentage it was written with. The code carries a comment warning about exactly that, which is the sort of honesty worth noticing.

Why the published figures disagree, and by how much

A knot efficiency is a ratio between two destructive tests: the load at which a plain length of line parts, and the load at which the same line parts with a knot in it. Everything that varies between those two tests ends up inside the percentage. The polymer, the diameter, the age of the spool, how carefully every strand was laid before tightening, how fast the pull was applied, whether the knot was lubricated, and which of the two ends the machine gripped. None of that is standardised across the tables you find online, which is why the same knot is quoted at 80 % in one place and 95 % in another.

The clearest published demonstration of that comes from fishing gear rather than angling. A 2024 study in Scientific Reports tied the same double weaver's knot in two monofilaments and pulled both to failure. Conventional polyamide 6 lost 30 % of its strength to the knot; a biodegradable polymer tested alongside it lost 62 %. Same knot, same rig, same laboratory. The paper also reports why the biodegradable material did worse: it cracked before the knot could finish tightening, where the nylon stayed ductile enough to bed in first. If a single knot can cost 30 % or 62 % depending on what the line is made of, no single table can be right for everyone's spool.

There is a second, quieter problem: the denominator. A 2012 study on arXiv tested eight leaders that their makers rated near 15 lb, each with and without a knot, precisely because so few published tests check a manufacturer's claim at all. Its most useful result is not a number but an ordering: the leader that was strongest without a knot was not the one that was strongest with a knot, and the product that broke lowest once knotted was a different brand again. That is the case against picking a line by its label alone and then applying a percentage to it.

Where the line actually breaks

The usual campfire explanation is that friction heat generated while cinching burns the line, and that this is why you wet a knot. The measured picture is different and simpler. In the Scientific Reports work the failures happened at the curvature peaks — the places where the strand bends most tightly inside the knot — and a majority occurred in the upper monofilament, which is the one carrying the sharpest bend and several such peaks. A knot weakens a line because a bend forces the outside of the strand to stretch further than the inside, so one face reaches its breaking strain while the rest of the cross-section is still comfortable.

That reframes the advice rather than cancelling it. Wetting and slow, even tightening still matter, because both help the strands slide into their final position instead of biting into each other while they are still crossed badly — which is what dressing a knot means, and a badly dressed knot pinches the line at a tighter radius than the design intends. But this article carries no number for what wetting is worth, because no primary measurement of it turned up. If someone quotes you a figure for that, ask them which line, which knot and which testing machine.

Matching the knot to the job, which the tool cannot do

The dropdown mixes four different jobs into one ranking. Tying line to a hook eye or a lure is one job, and it is where the Palomar, the improved clinch, the uni and the Snell live. Joining two lines of different material or diameter — braid to a fluorocarbon leader, most often — is a second, and it belongs to the FG, the Albright and the double uni. Making a loop that lets a lure swing freely is a third, which is what the non-slip loop knot and the Rapala are for. Attaching line to the reel spool is a fourth, and that is the arbor knot, ranked last at 60 % in a ranking that does not matter, because if you are ever pulling on your backing knot the fight was already lost.

This is where the flat line adjustment does real harm. Braided line is slick and has almost no stretch, and a knot that grips beautifully in nylon can creep and pull through in braid; the FG knot is a braid-to-leader connection built around that fact, and the received advice for the clinch family in braid is to add turns or change knot entirely. The calculator hands every knot the same plus one point on braid, which reverses the warning it ought to give. Read the line-type control as a small nudge and take the choice of knot from what the connection is for, not from the ranking.

The other knot page on this site, and what it does not contain

Allin also carries a knot-tying reference, and it is worth saying plainly what it is not. It holds exactly ten knots — bowline, rethreaded figure eight, taut-line hitch, trucker's hitch, clove hitch, sheet bend, reef knot, double fisherman's, prusik, and round turn with two half hitches — and none of them is a fishing knot. There is no Palomar, no clinch, no uni, no FG. It also carries no strength figure of any kind: not one percentage anywhere in its data. What it does carry is a purpose and a when-not-to-use-it for each knot, numbered steps, and a set of checks on the finished knot, which is the part a percentage cannot give you.

One defect in that reference is worth reporting, because it concerns safety rather than accuracy. Its dictionary defines four strings in all six languages that the component never renders: a banner reading that the page is a reminder sheet and not a course, a warning that the reef knot is not a bend and spills when the two ropes are pulled apart, and a warning that the climbing and rescue knots demand training, a backup and a check before every use. Twenty-four translated strings sit in the file and reach nobody. The safety-critical knots still get a small red chip, but the sentence explaining what the chip means never appears.

Frequently asked questions

Is the Palomar really the strongest fishing knot?

In this calculator it ties for first place with the Snell and the double Palomar, all three stored at 95 %. That is a stored opinion, not a measurement, and it is stored to the nearest five points, so the three are indistinguishable by design. What is genuinely in the Palomar's favour is not the percentage but the procedure: it is tied with a doubled line, has few steps, and is easy to inspect once tightened — three properties that reduce the chance of the mistake that actually costs you a fish. A knot you tie correctly at eighty per cent beats a knot you tie badly at ninety-five.

Does wetting a knot before tightening really make it stronger?

No figure is offered here, because no primary measurement of it could be found — and this article does not repeat numbers it cannot trace. What can be said from the measured work is that knots fail at the points of tightest curvature inside them, so anything that lets the strands settle into their intended path before they lock is working in the right direction, and lubrication does that. The common claim that friction heat melts the line is not what the failure analyses describe. Wet your knots; just do not put a percentage on the habit.

My line is rated 20 lb — will it break at 20 lb?

Not reliably, and the record books assume as much. The IGFA sorts records into line classes — 2, 4, 8, 12, 16, 20, 30, 50, 80 and 130 lb — and tests the line submitted with a claim; if the line over-tests its stated class, the claim is moved up into the next class. A rule like that only exists because labelled ratings and measured breaking strains routinely differ. So the calculator's output is a percentage of a number you have not verified: if the label is optimistic, the answer inherits the optimism, and if the line is old, sun-bleached or nicked, the real starting point is lower again.

Why does the dropdown say 95 % and the result box say 93 %?

Because the label is fixed text and the result is computed. Each knot carries a hand-written percentage in its menu entry, and the calculation then adds the line-type adjustment on top: minus two points on fluorocarbon, plus one on braid, nothing on monofilament. So a Palomar reads 95 % in the menu and returns 93 % on fluorocarbon or 96 % on braid, and the same two-point gap appears on every knot in the list. Nothing is wrong with the arithmetic; the label simply describes the monofilament case and never says so.

The tool says braid improves every knot by a point. Is that right?

It is the weakest part of the model, and it points the wrong way for a whole family of knots. Braided line is smooth and barely stretches, and knots that rely on the strands gripping each other can creep under load in braid where they would hold in nylon; the FG knot, which wraps braid around a leader rather than knotting it, was developed precisely to get around that. The plus-one is applied identically to all twenty-three knots because it is a property of the line type in the code, not of the pairing. Take the ranking from the job, take the line adjustment as decoration, and if you fish braid, use a connection that was designed for it.