So rather than telling you which knot to use I’m hoping that I may inspire some to do testing of their own. You may or may not replicate my results, however at the very worst you’ll go into the new season with a new found confidence in your own knots. Just the other night I had a ‘knot off’ with Tracy which I won 5 – nil, confirming the need for us to change tippet knot as suggested by my previous testing (Tracy was tying 3 turn water knots, I was tying J-knots). This home test was using the same nylon but ‘snatching’ the knots to failure by hand, I will of course repeat this using the material we intend fishing with.
In order to test hook/fly attachment knots I used sized 4 hooks, this proved to be a bit of a mistake. The Instron is routinely set up with a pair of grips top and bottom which can easily be used clamp on to anything that fits. This clamping action will damage soft nylon significantly to the point where failure is certain to occur within the grip, I therefore chose to tie hooks on both ends of the line using the same knot and grip the bends of these in the machine. Thus for every test result generated, one of the two knots survived although it was subsequently snipped off and a new piece of nylon attached for the next test. Tightening some of the knots when one end is already attached to a hook proved to be particularly hazardous (made worse by the short length of material used) – all of the knots were lubricated with saliva, but there may have been a bit of blood on some too! I’ll remove the hook-points before I do another set of tests.
So on to the results. The control group is the same as that for the tippet knot testing, being as I was using the exact same material I didn’t see the need to repeat these. First up was the Kreh loop, my normal knot for attaching bonefish flies. I must admit I was a little disappointed with the results, although the average was a respectable 73% (of the control group average) the range was less impressive with the lowest result at just under 60%. That said, I’m anticipating my tippet knot to be in the region of 70% (J-Knot) so perhaps this isn’t so bad, after all if I get a breakage I’d prefer it to be at the fly. I also repeated the Kreh loop with a variation suggested by Graeme on the board – a figure 8 initial knot rather than an overhand knot. In my tests this faired marginally worse (65% average) although this could just be a statistical variation due to the smallish sample size.
The Orvis knot and the tucked ½ blood proved to be similar to the Kreh loops with averages of 70% and 73% respectively, again with quite large ranges. When viewing these results it’s no wonder that the question of ‘what’s the strongest knot’ throws up so many contradictory opinions – any one of these four knots could beat the others in head-to-head tests, and unless the sample size is large enough it could be very easy to get a false impression of what beats what due to the vagaries of statistics. This is where seeing the range of failure figures, as produced by a mechanical tester, offers a distinct advantage.
One knot that did produce a distinctive result was the uni-knot (or grinner), not that it was especially strong at an average of 70%, but because the range of results was extremely small. This super consistency could, of course, just be a statistical aberration or it could be real, without testing a lot more samples it would be hard to be certain. What this result did get me thinking about though is what’s important to me? Is it the highest knot strength I could potentially produce or is it in fact knowledge of the lowest? I suspect that what really matters is the latter.
The two stand-out knots of this series were the 10 turn blood knot and the Eugene bend, with averages of 94% and 91% respectively. Even the lowest result from these two was over 83%, impressive I’m sure you’ll agree.
I do have a confession about the 10 turn blood knot though – the first result was just over 50% but I had real difficulties in tying it, well not so much in the tying – more the ‘snugging down’. Try it to see what I mean, there’s just so much friction I found it almost impossible to tighten, this knot definitely accounted for the most blood spilt! On subsequent tyings I decided to make some minor changes, firstly I allowed the tag end to wrap around the 10 twists on the way back to the loop nearest the eye of the hook. In this respect it started to remind me of the beginnings of a Bimini twist, and we all know how strong that is. Secondly, I decided that if it didn’t want to ‘snug down’ then I wouldn’t force it, I also wanted to avoid a full piercing of my hand. Again this is reminiscent of a Bimini in that here you essentially have a braid of line where the load is shared between the individual strands rather than a barrel knot of one central load bearing strand with other non-load bearing strands wrapped around the outside. With these changes the results were very impressive, in fact I had two where failure occurred in the line away from the knots. As such, I decided to do some extra tests and deleted the initial poor result as a bad tie on my behalf.
Anyway, I hope there is something of interest here and that you’ll think about doing some of your own tests.