Single CTS or follow the Courseline?

[goboatingnow]

But no electronics are to used. That is what traditional basic navigation is about.

IT doesn't matter, Dockhead is doing a comparative model , anyway CTS applies using GPS or not.

[goboatingnow]

I think you might profit from reading the 16 pages of comments

If after reading all that, you still believe in doing it that way, then I'll race you from Needles to Cherbourg at spring tide for pink slips You will use your art form, and I will do a precise CTS calculation. I'd love to have a 65' Van Hellman schooner

Dockhead, poster child , poster child

Dave

[LJH]

Originally Posted by jackdale

Over the SWL.

You argument is only one that an economist would make.

Old Arabic expression "He who foretells the future lies, even if he tells the truth."

Both methods are riddles with parameters and assumptions that likely will not pan out. SWL keeps assuming no leeway, which I do not think is valid. I would also gather that an accurate deviation card is available - how many of those have you seen.

Both will give a good estimate. I will apply Occam's razor.

The RYA calculates a CTS in True as does SWL. After you take the heading from the map, in either method, you need to apply variation, leeway and deviation to get a compass heading to steer at the helm.

[Dockhead]

Originally Posted by goboatingnow

DockHead,

The real test ( of any CTS method) will be seeing if the EP error workup, which results in a error circle at the end of the single CTS , encompasses both calculation methods, if it does then its a precision versus accuracy argument.

I would wager that the error of position circle resulting from simply sailing the same exact CTS (+- helm errors) and allowing for speed variations, and leeway, will be far greater then either theory error circle. Good luck with that.

If however the error of one or the other methods lies outside the EP error , then the one that lies inside is better.

The difficult of course is how can you model a real crossing simulating maintaining a real single CTS and mathematically derive any meaningful result.

of course , we can initially compare the two theory methods to closer inspection by simple comparion of a 5 minute CTS workup compared to both methods

I strongly disagree with SWL comments that one should pick extreme situations, initially we should look at common sitiuations, Reductio ad absurdum can come later ( many models work fine for general use but fail to explain certain things at the extremes , see Newton)

Right now we can deal with the "Ad ignorantiam" issue.

Dave

Well, I agree with most of your logic. I agree with the approach of working error circles in order to put all this in perspective. I'm not sure what the right criterion is for saying whether other errors wash out the difference in accuracy between RYA and SWL. We will definitely see a much small error circle for SWL method, than for RYA, for very simple reasons -- average of the last hour will be much closer to reality than average of the whole passage, as I have written a few times. I don't think anyone can deny that SWL is a better approach.

As to helm, leeway, and speed -- I think we can leave out helm -- a reasonably effective autopilot will hold an accurate course, plus or minus a couple of degrees which will cancel itself out. I have never been able to detect any long-range helm error, at least with my pilot (an old big hydraulic Ray). Leeway and speed we can put aside for motoring in calm weather, which is the acid test of these methods. I don't think anyone will buy an argument like "My method gives acceptable results IF you're hard on the wind in rough weather and hand steering, because under those circumstances, all the errors will wash out the inherent errors in my method." This is not impressive. The methodology will be put to the test in calm weather motoring at a constant speed, and we'll see the difference.

Errors from speed we will use for a different purpose -- to find out how quickly the advantage of CTS over rhumb line crabbing breaks down as a result of input errors. I am predicting (and I am sure that I am right, but we will test it) that it will be remarkably robust in dealing with input errors, when used as we use the method in reality (with a mid-passage correction or two).

We can then go on to test at what extent of input errors the advantage of SWL over RYA is washed out. I guess that's a good third test.

I suppose for errors we can use random numbers applied to heading (for mistaken estimate of leeway) and to speed.

As to extreme scenarios -- I guess there's no need to discuss it anymore -- we've already seen the results. Now we will use some real data from real crossings (I will be using spring tides, however! ).

[Dockhead]

Originally Posted by goboatingnow

Dockhead, poster child , poster child

Dave

Shhhh!

[Dockhead]

Originally Posted by LJH

The RYA calculates a CTS in True as does SWL. After you take the heading from the map, in either method, you need to apply variation, leeway and deviation to get a compass heading to steer at the helm.

Correct.

All of our scenarios have assumed motoring at a constant speed in calm weather, so no leeway anyway, in order to remove as many superfluous complicating factors as possible.

In real life, if you are sailing, you will of course add (or subtract) for leeway (and deviation and variation) as the final step in getting your course to steer. This is uninteresting except to the extent that it introduces more errors into the calculation (i.e. you assume 5 degrees of leeway but it turns out to be 8 degrees, so you are steering 3 degrees off -- does that knacker your passage?). We will be testing for that.

[goboatingnow]

Originally Posted by Dockhead

Well, I agree with most of your logic. I agree with the approach of working error circles in order to put all this in perspective. I'm not sure what the right criterion is for saying whether other errors wash out the difference in accuracy between RYA and SWL. We will definitely see a much small error circle for SWL method, than for RYA, for very simple reasons -- average of the last hour will be much closer to reality than average of the whole passage, as I have written a few times. I don't think anyone can deny that SWL is a better approach.

As to helm, leeway, and speed -- I think we can leave out helm -- a reasonably effective autopilot will hold an accurate course, plus or minus a couple of degrees which will cancel itself out. I have never been able to detect any long-range helm error, at least with my pilot (an old big hydraulic Ray). Leeway and speed we can put aside for motoring in calm weather, which is the acid test of these methods. I don't think anyone will buy an argument like "My method gives acceptable results IF you're hard on the wind in rough weather and hand steering, because under those circumstances, all the errors will wash out the inherent errors in my method." This is not impressive. The methodology will be put to the test in calm weather motoring at a constant speed, and we'll see the difference.

Errors from speed we will use for a different purpose -- to find out how quickly the advantage of CTS over rhumb line crabbing breaks down as a result of input errors. I am predicting (and I am sure that I am right, but we will test it) that it will be remarkably robust in dealing with input errors, when used as we use the method in reality (with a mid-passage correction or two).

We can then go on to test at what extent of input errors the advantage of SWL over RYA is washed out. I guess that's a good third test.

I suppose for errors we can use random numbers applied to heading (for mistaken estimate of leeway) and to speed.

As to extreme scenarios -- I guess there's no need to discuss it anymore -- we've already seen the results. Now we will use some real data from real crossings (I will be using spring tides, however! ).

A convergence on the test citeria is emerging.!!.

I do think simulating speed errors is important as I think this is the biggest factor ( aside from incorrect tide data) in determining the final position error.

Firstly I think it would be useful ( and you ll forgive me "asking" given your doing the work) just to do a 5 minute CTS, obviously electronically determined, compared with a manual exercise on each method using hourly tides.
I actually think for that you should compare hourly tide data, from an hourly atlas, rather then using the centre hour point of the 5 minute data. To some extent I think you need to compare data extraction methods.


what I was saying , is that my practical experience suggest that the in complex multi hour CTS, the resulting final EP error circle is very large. For that I did a workup some years ago, where I took the method of computing the ground track, adding a EP error circle and inflating the circle each hour, The end circle was very large ( I was demonstrating to a class the problems with single CTS and why I dont like it).

Hence I would contend that, perhaps aside from extreme situations, the resulting EP from either method lies within the error circle of the real world situation. Hence any increase in precision, does not in fact result in a real world increase in accuracy.

However initially I think it would be interesting to see the correlation between the RYA's assumption that taking a 1(2 ,3,4 etc) hour tide and extending it to say 1 hour 20 produces less or more errors then computing exact tide quantum's.

Good luck and thanks for the expected work load.

Dave

[Dockhead]

Originally Posted by goboatingnow

A convergence on the test citeria is emerging.!!.

I do think simulating speed errors is important as I think this is the biggest factor ( aside from incorrect tide data) in determining the final position error.

Firstly I think it would be useful ( and you ll forgive me "asking" given your doing the work) just to do a 5 minute CTS, obviously electronically determined, compared with a manual exercise on each method using hourly tides.
I actually think for that you should compare hourly tide data, from an hourly atlas, rather then using the centre hour point of the 5 minute data. To some extent I think you need to compare data extraction methods.


what I was saying , is that my practical experience suggest that the in complex multi hour CTS, the resulting final EP error circle is very large. For that I did a workup some years ago, where I took the method of computing the ground track, adding a EP error circle and inflating the circle each hour, The end circle was very large ( I was demonstrating to a class the problems with single CTS and why I dont like it).

Hence I would contend that, perhaps aside from extreme situations, the resulting EP from either method lies within the error circle of the real world situation. Hence any increase in precision, does not in fact result in a real world increase in accuracy.

However initially I think it would be interesting to see the correlation between the RYA's assumption that taking a 1(2 ,3,4 etc) hour tide and extending it to say 1 hour 20 produces less or more errors then computing exact tide quantum's.

Good luck and thanks for the expected work load.

Dave

Good, I think we agree about the testing.

For the one hour data there are two approaches -- 1. just use the UKHO tidal atlas; or 2. take an average of the one hour's 5 minute tides.

The former will be interesting because that is what real sailors really use in real life. The latter is interesting because it eliminates any random variation between what's in the tidal atlas (and how we read it; which diamond we choose), compared to the 5 minute data.

I guess the latter is probably the right approach.

[Wotname]

Originally Posted by Dockhead

PFM = Pure Freaking Magic

Thanks
So it's AFA; I had better get UTS with TLA and FLA.

[LJH]

Originally Posted by Dockhead

Correct.

All of our scenarios have assumed motoring at a constant speed in calm weather, so no leeway anyway, in order to remove as many superfluous complicating factors as possible.

In real life, if you are sailing, you will of course add (or subtract) for leeway (and deviation and variation) as the final step in getting your course to steer. This is uninteresting except to the extent that it introduces more errors into the calculation (i.e. you assume 5 degrees of leeway but it turns out to be 8 degrees, so you are steering 3 degrees off -- does that knacker your passage?). We will be testing for that.

But if I can trim an extra .25 Kts when I adjust my heading...experience will tell. It will be interesting to see the results.

[goboatingnow]

Originally Posted by Wotname

Thanks
So it's AFA; I had better get UTS with TLA and FLA.

where I come from and like all sailors lingua , Dockhead PFM of course is sanitised for the sensitive ears of CFs readers

I was first introduced to the term , doing some work for Boeing and whn I asked a wing designer how it really worked , he looked at me and said, PFM

Ive since discovered that it applies to many things in the physical world

dave

[LJH]

Originally Posted by goboatingnow

where I come from and like all sailors lingua , Dockhead PFM of course is sanitised for the sensitive ears of CFs readers

I was first introduced to the term , doing some work for Boeing and whn I asked a wing designer how it really worked , he looked at me and said, PFM

Ive since discovered that it applies to many things in the physical world

dave

Lol, it especially applies to electronic boxes.

I bet Boeing is wishing they had better PFM with the Dreamliner. Something to be said for engineering done with a slide rule, says he with an older boat!

[goboatingnow]

I bet Boeing is wishing they had better PFM with the Dreamliner. Something to be said for engineering done with a slide rule, says he with an older boat!

My other good Boeing one was many years ago I was flying up the East coast of the US going home to Syracuse, we were in a storm and the wing was flapping around a bit .There was a guy in the window seat staring intently at the wing, I was next to him and eventually my curiosity got the better of me and I asked what was so interesting.

"Oh the wing shouldn't be doing that in these conditions", how do you know I said , "I designed it " he said ( deadpan). Lets say the next hour or so wasn't the most comfortable Ive ever been on in a plane!.

Dave

[LJH]

Originally Posted by goboatingnow

My other good Boeing one was many years ago I was flying up the East coast of the US going home to Syracuse, we were in a storm and the wing was flapping around a bit .There was a guy in the window seat staring intently at the wing, I was next to him and eventually my curiosity got the better of me and I asked what was so interesting.

"Oh the wing shouldn't be doing that in these conditions", how do you know I said , "I designed it " he said ( deadpan). Lets say the next hour or so wasn't the most comfortable Ive ever been on in a plane!.

Dave

ROTFLMAO...If the wings didn't flap how would you get there! Seriously though if the wing didn't flex it would crack somewhere like the wing root and all would not be good!

[goboatingnow]

ROTFLMAO...If the wings didn't flap how would you get there! Seriously though if the wing didn't flex it would crack somewhere like the wing root and all would not be good!

Oh I know that , but he was concerned about some aspect of it. !!!!!