Thank you!

[Paul Arden]

I would like to thank all Board members - and particularly those of you who post here in the technical analysis section - for the marked increase in politeness! I’m sure you all agree that it makes for a much more pleasant experience if everyone is respectful to one another, even on those rare occasions when we might disagree

So thank you all very much for making the Board a much friendlier place.

Cheers,
Paul

[Merlin]

Paul

May I take the opportunity of opening this new thread to start a discussion about the TA forum (TAF, a new acronym) itself. And I would also like to thank Vince and Mark for sharing their views about the way to analyze technical difficulties (and well written, unlike me with my non native language).

We all have our own professional experience in various domains, which corresponds I think to the various approaches we are using on the TAF. In the high tech world of aviation, there is no place for errors and all must be carefully checked and challenged. This does not preclude some bad mistakes with severe consequences in case some step was neglected for various reasons, including economic ones.

I am more familiar with the automotive world. In this world, there is a failure tolerance of 0.5% in service. This means that if you get claims for 0.5% of a given car model on the market, then you are in serious trouble. There is little chance (not nil however) to get a car crash, the failure is rarely immediate and you can usually park your car before being stuck on the road. Today computing holds a very large place in car conception (e.g. the gasket of the door of your car is computed), but there are areas where computers are of little use. An example is injection system breakdown. Years ago we faced a crisis with one car model. It was about the 12th version of powertrain for this car (drivers are testers but they do not know it). Although the problem did not come from the fuel (for us, not for the motor manufacturer and the injection system manufacturer), we defined a practical solution within a couple of weeks. We implemented it all over the market in one and a half month. Failures begun to slow down but it took about one more year until the failing piece of the injection system was changed to get rid of the problem. No computing, tough discussions to say the least, and just practical solutions.

Read across exists in the automotive world with a different meaning: you can "read across" a technical required performance (on an engine or a test rig) if the relevant characteristics of your new product is recongnized to be better than another already approved product. In other words, without actual testing, your product is said to pass the test. That is far from modeling stories.

In the fishing world, I am not aware of a large use of computing design. Computing assistance exists, for checking some points, but complete computed design doesn’t as far as I know. The first difficulty is to input in a computer what you want to get from the use of the rod, and something like feel is not accessible to a robot, for the time being. Most progress is being made by testing and trial and error. Most rodmakers are not familiar with engineering and mechanics of the fly cast. It is a pragmatic world. This is why IMHO, looking after massive computing weapons to get the truth about mechanisms in this small world is a kind of holy grail, which makers don’t believe in.

That difference of approach in between such worlds has likely contributed to the foundation of the “swamp”. It would then be nice to define our expectations, and if we cannot agree on a single approach, anyone always has the possibility to express his own views by opening a thread corresponding to his level of requirement on the subject he wants to discuss. My personal view is that the best is the enemy of the good, and that we should stay as close as possible to the lowest level of technique to give clear messages to laymen. But it is only my view.

One thing I have experienced in the rod design domain is that the more you want to refine the analysis, the less you capture the underlying mechanisms. I remember a paper from scientists where they could not explain why the tip of a rod was going back first by comparison to its motion. A very simple analysis of a spring model can answer this question. Many people have used Abaqus for example (a powerful software for analyzing mechanical systems), but could hardly explain the consequence of their modification of input, I mean the way they were acting to change the output. The bigger the black box is, the less we understand from it. This is not applicable to any system of course; sometimes complexity is required to get the best of behavior predictions, even if we do not fully understand the details of what is going on. That means that the goals are just different. And speaking of aviation, I do not think we need such a level of technicality and consolidation of results to get the main trends explaining most of observations. The fact that we shall not be able to explain all observations with a simplified approach is a prerequisite, and not being able to explain what happened to Erikson’s cast is secondary (for me at least) for example. Now if someone thinks he can go a bigger step beyond with powerful computing systems, he is most welcomed, but don’t ask me that, I am using Excel.

Let’s take the case of the influence of the taper of the line for example. We start from an explanation of a practical approach by Bruce with pretty crude references to physics, and what I done is to try rising the level a little bit to see if we could be able to give the major trends associated with tapers, something that skilled casters know by experience. Then playing with basic tapers would be quick (and dirty and cheap by the way). Thanks to Lee and his amazing videos, we got a first example with the Scandi lines. The simulation appeared to be reasonable in terms of trends, but soon we got a problem with the Spey lines where something happened which was unexpected. Scientifically speaking, we are a bit short since we did not control the line tapers, which is a minimum if we want to make the best possible estimates. It is bad luck indeed. Imagine the trend would have been closer to expectations with the Spey lines, nobody would care. But bravely we go on and try again, thanks again to Lee and the faithful videos analyzers.

The purpose of the thread on “engine and brakes” was to draw our attention on the main cause of line acceleration, in other words the relative loss of mass of the rod leg, and not the loop itself, which behaves globally like a brake. Using this representation you can figure out that the relative change of mass of the fly leg for a TT is larger than the one for a reversed TT. One type of taper favors acceleration (diminishing taper), another one limits acceleration (increasing taper). This is a simplistic explanation, helping to understand the trends in line taper design. There is little need to get six ODE and a RK4 converging solution to get the trends. The time will come to scrutinize the loop and its morphing trend, and then maybe the current basic approach and mathematical tools will not be adequate. If there is a volunteer, welcome to the jungle.

Merlin

[Paul Arden]

For me it’s always been an exploration. Way back at the beginning when videos showed the haul continuing to RSP and not finishing at maximum rod bend as we believed back then. I don’t think you need a reason to explore; you only need curiosity. It’s always been about discovering what you don’t know, not what you do know. And believe me, there’s a lot I don’t know!

The recent videos showing an acceleration in the fly leg after the loop has been formed I think is a significant step forward. In fact this past month has been one of the most interesting times here in quite a few years. Tremendous videos from Lee and I hope I can add a few when Graeme is over.

I’m very pleased to see this part of the Board functioning properly again.

Thanks,
Paul

[gordonjudd]

Thanks to Lee and his amazing videos, we got a first example with the Scandi lines. The simulation appeared to be reasonable in terms of trends, but soon we got a problem with the Spey lines where something happened which was unexpected.

Merlin,
A very thoughtful post, and I agree with the viewpoints you presented. At the same time I am glad to disagree with you on subjects where I see some chicken and egg tradeoffs a different way. We can disagree without being disagreeable, and I think is a hallmark of technical people with an open mind who are happy to admit they were wrong about something.

The main problem that I see is that the physics of casting is complicated, and its gets more and more esoteric the more you dig down on a subject. I have always viewed the technical forum as a place where complex subjects can be discussed and if it involves complicated analysis such as the physics of distributed, variable mass systems and if lay readers with no technical background are put off by it, then so be it.

Feynman once said:

Hell, if I could explain it to the average person, it wouldn't have been worth the Nobel prize.

Our discussions are not Nobel worthy, but I personally do not have the deep understanding or communication skills to explain the physics of casting to the non-technical readers at this forum. As a result I seem to just piss people off as they have no idea of what I am talking about.

The basics of determining the fly velocity history for a given cast have been known since Dr. Greg Spolek's article written in 1985. Yet it seems to still be a controversial subject at Sexyloops 34 years later. I don't understand it, but then most of the readers of this forum have not read (or understood) the rich library of papers dealing with the physics of casting.

There are things like what causes the Dolphin nose shape and morph in loops that are far from being understood. But in my experience smart people around the world will figure it out. The analysis of the physics of the string shooter is a good example.

The difference in the velocity history of the Scandi and Spey lines is a good example that any number of experiments agreeing with an expected result can not prove the analysis is correct or complete. However, one counter experiment can prove it wrong. That is the way science should work, and hopefully will guide our discussions in the future.

Gordy

[Viking Lars]

the rich library of papers dealing with the physics of casting

I for one wouldn't mind a thread with links to online ressources covering the above. Could that be an idea?

Lars

[Walter]

the rich library of papers dealing with the physics of casting

I for one wouldn't mind a thread with links to online ressources covering the above. Could that be an idea?

Lars

I second that. I know many of these papers aren’t free but maybe SL could start a library with paid access...

[James9118]

The basics of determining the fly velocity history for a given cast have been known since Dr. Greg Spolek's article written in 1985.

I'm glad that a spirit of cordiality has broken out in the tech forum, as such I'm sure what I have to write is taken in the spirit in which it is intended.

All models are wrong, but some are useful. I'm sure we've all seen that saying before, and having spent the best part of 30 years as a scientist in one of the top UK research facilities (including some time in an equivalent US laboratory), with an operating budget of getting on for £1 billion a year, I have no evidence to suggest it isn't true. So just to re-emphasise the most important bit - all models are WRONG.

As such, stating that Spolek is absolutely correct grinds on me slightly. Same with Gatti-Bono and anyone else you care to mention.

I'm also reasonably handy with a fly rod. I've won more BFCC competitions than anyone else (by a significant margin) - admittedly because I've probably attended more events that anyone else, but you can only beat those who are there to compete. As such, I have a significant problem with many accepted papers and the conclusions that can be drawn from them. For example, the US lab I worked in was in Albuquerque at 5000ft above sea level. On arrival there I was instantly able to cast much further than I could in the UK - in fact I added over 15ft to my #5 best. Now this is totally at odds to a peer reviewed, published paper on lift. I also know that other casters report the exact same result and you have to question why that is - my only conclusion is that the model is wrong - but we already know that don't we?

I guess the point I'm making is that I really don't care if my views are at odds to Spolek et. al., because my views are very much based on an emperical view point - i.e. if the model doesn't reflect reality, then it's rare that reality is wrong.

So can I ask that we treat all the published papers in the manner that they went through the peer review process - i.e. as the hypotheses that they are - and not as theorems?

James

[Paul Arden]

I know many of these papers aren’t free but maybe SL could start a library with paid access...

That might be possible I don’t know. I’m certainly happy to approach the relevant publishers if it makes sense. I can also approach the authors. If we list publications with links I shall start a Sexyloops page that can be linked to from the Board header.

I’m also an empiricist; I think anyone who gets involved in competition casting has to be (and in fact probably any ‘elite level’ casting). We try everything and measure the results. I’m not quite sure if it’s the cart or the horse that’s in front here but what I do know is that the stopless cast came first, from the field (or in fact a hockey rink) and the physics came after.

When ideas come from the theoretical side, it’s always the field/river/clifftop that gives them value or not. Case in point hitting the boing/LSP/capturing kinetic energy. Theory says it should go further. Practical casting shows that the loop tails/collapses.

And what I’m seeing in the two loops thread is totally unexpected. Isn’t it fun?

Cheers, Paul

[Graeme H]

In the newly sprung spirit of cordiality, I'm keeping my mouth shut.

Cheers,
Graeme

[Walter]

I know many of these papers aren’t free but maybe SL could start a library with paid access...

That might be possible I don’t know. I’m certainly happy to approach the relevant publishers if it makes sense. I can also approach the authors. If we list publications with links I shall start a Sexyloops page that can be linked to from the Board header.

I was thinking that for the non free papers that SL could buy a copy and create a private library for members who are willing to pay for access to that library. I really don’t want to pay individually for all of the papers so if some sort of shared access was available that would be great.

I’m kind of with James in that no model is perfect but I don’t think that means they are all worthless. Newton’s laws are an example. They are an oversimplification but without them we would find it difficult to understand motion.

In that spirit I don’t want to pay full price for a paper only to find out it really isn’t worth it. I would rather rent it for a few days.

A reference page of available papers both free and paid would be a good start.

Maybe someone could act as librarian and scout out useful papers?