The "Making Trees" fallacy appears again...

[Ronny170]

Not sure if anyone saw the boat collision video from the end of last year where a supposed boat captain claimed that you could use stationary landmarks in the same way as CBDR to visually assess collision risk with another vessel. His (incorrect) claim was that if the other vessel was "making trees" against the background landmarks that it meant the other boat was going to pass in front of you, and therefore no collision would occur; and if the other boat was 'stationary against the land' then you were definitely on a collision course with that vessel. This video was posted to all of the major social media platforms, and had literally millions of views last time I checked.
A rebuttal video was then posted a few months later, highlighting the errors in these claims made by this boat captain. It looked convincing enough to me. Did anyone catch this back and forth, or have any thoughts on the apparent controversy that followed? The rebuttal video (which includes the original video in it) was posted to Reddit, IG, and Rumble I think. Part 1 can be seen here on Reddit-
https://www.reddit.com/user/Sivadthe...n_course_with/

Thoughts on the matter??

[requiem]

Oh god, thank you for reminding me of this... I think I still owe one guy there a reply.

Anyway.... it's all about the relative distances. The issue with the rebuttal video is that the ratio between the target boat and the background reference is generally 1:1, which is why the rebuttal appears to work. The original video had problems for a similar reason: even the landmarks were moving relative to other landmarks.

The gold standard for a CBDR (constant bearing, decreasing range) case is using a compass to monitor change in bearing. However, all your compass is really doing is measuring the progress of the observed boat against a stationary landmark in the distance. With the compass that "stationary landmark" just happens to be really, really far away in the extreme North of Canada.

So long as the relative bearing of the selected landmark will not noticeably change during the observation window, it's no different from using a compass and will work perfectly fine. However, the nearer the landmark, the shorter the observation window.

Suppose you pick a landmark that won't noticeably change bearing over the course of 60 seconds, but will over the course of 20 minutes. If you take a quick 10 seconds to assess the target against it and observer your target moving a great deal relative to it, your assessment of "safe" is likely valid. In contrast, if you take a quick look and come back in 20 minutes to see the target has moved relative to it, you cannot make a meaningful assessment based on that.

[Ronny170]

I'm not sure I follow you on the first part. You mentioned that "the ratio between the target boat and the background reference is generally 1:1" and that's the only reason the analysis works. What ratio are you referring to exactly? Are you talking about some distances that are set to 1:1? Didn't see any mention of taking ratios with the measurements.



For CBDR, not sure what you mean when you say that "all your compass is really doing is measuring the progress of the observed boat against a stationary landmark in the distance." The compass is used to measure the relative bearing between a point on the observing vessel and the target boat in CBDR. Stationary landmarks in the distance aren't relevant or used in CBDR, so I don't follow you on that either.


The amount of change in bearing of a landmark from the perspective of the green observation boat doesn't really matter much. What matters as far as the landmarks is concerned is whether the target boat makes the same amount of headway (making trees) against the landmarks within a given amount of time. If it's the same on two or more occasions with decreasing range then a collision would be likely. You can actually make a meaningful assessment that way you just need two timed assessment periods to compare to.


What was really interesting was looking at how the target boat would have always change speeds in order to appear to remain stationary against the background. I actually thought that was the most interesting part of the whole thing.

[El Pinguino]

I use that method quite often. Its all about knowing its limitations.... if you can make out trees on your reference point the reference point is too close.
In that reddit example the best thing to have used would have been the clouds. Distant peaks are good.
Also works OK at the speeds I poke around at - say 6 knots. Higher the speed more distant the reference has to be.

[s/v Jedi]

The method described by the so called ships captain is valid. The rebuttal (I haven’t seen either video) is designed to create an argument by creating a situation in which the method doesn’t work.

The background objects must simply be much, much further away than the ship that is being watched.

I’ve done this a million times competition sailing and it is accurate enough for that, which means it’s accurate enough for cruising sailors

[sailingharry]

Like Jedi, I've used the method many many times. It is exceptionally useful. Of course, there was that time in Newport harbor as officer of the deck on a Navy ship where the captain insisted I get a maneuvering board solution out of Combat and wouldn't take my assessment...

I haven't seen the videos either. But only a fool would use a visual tool like this, come to a conclusion, and never track it again! It does give you the likely answer in 10 seconds, where a hand bearing compass would take a minute or two or three. But you have to keep watching! Besides, over the next 10 minutes, both the target boat and your boat will change course and speed anyway.

[s/v Jedi]

Also, compare this to using the radar with a radar bearing line (EBL). You set the line on the approaching target and if the target moves forward of the line they will pass your bow, if they fall behind they will pass your stern and when they stay on the line they will hit you.

[Lodesman]

Originally Posted by requiem

The gold standard for a CBDR (constant bearing, decreasing range) case is using a compass to monitor change in bearing. However, all your compass is really doing is measuring the progress of the observed boat against a stationary landmark in the distance. With the compass that "stationary landmark" just happens to be really, really far away in the extreme North of Canada.

That's entirely incorrect. That constant bearing (the CB part of CBDR) is a relative bearing, not fixed to a stationary landmark.

While the rest of your argument has some merit, that is if the bearing of the fixed landmark is changing so slowly as to present a fixed point, it can give a rough estimate of there being a risk of collision. The trouble is in knowing when those conditions are being met - as pinguino says, you have to know and understand the limitations.

I certainly don't see why you would use this method over the tried and true method of taking compass bearings.

Not to mention, I'd be a little leery of any advice from someone who on speeding up to create a risk of collision then states that the other vessel "now has to give way."

[Ronny170]

Originally Posted by s/v Jedi

The method described by the so called ships captain is valid. The rebuttal (I haven’t seen either video) is designed to create an argument by creating a situation in which the method doesn’t work.

The background objects must simply be much, much further away than the ship that is being watched.

I’ve done this a million times competition sailing and it is accurate enough for that, which means it’s accurate enough for cruising sailors

You might want to watch the videos then. The method offered up by the ships captain is entirely incorrect. Another vessel "making trees" does not indicate that a collision will be avoided, but in fact if that vessel is making trees at a consistent rate, it means a collision is actually imminent. He has it completely backwards. The rebuttal videos show that this is the case in the majority of situations, and not an exception.

[requiem]

Originally Posted by Ronny170

I'm not sure I follow you on the first part. You mentioned that "the ratio between the target boat and the background reference is generally 1:1" and that's the only reason the analysis works. What ratio are you referring to exactly? Are you talking about some distances that are set to 1:1? Didn't see any mention of taking ratios with the measurements.

From what I remember of the rebuttal video, distance from the observing vessel to the target vessel has been roughly the same as from the target vessel to the land mark. Even the large scale one, e.g. observer 200 miles off the coast, target 100 miles off, "landmark" on the coast.

Originally Posted by Ronny170

For CBDR, not sure what you mean when you say that "all your compass is really doing is measuring the progress of the observed boat against a stationary landmark in the distance." The compass is used to measure the relative bearing between a point on the observing vessel and the target boat in CBDR. Stationary landmarks in the distance aren't relevant or used in CBDR, so I don't follow you on that either.

The compass measures the absolute bearing of the target boat relative to a fixed landmark: magnetic north. Imagine if the only landmark on an otherwise featureless shore was a large pinnacle, but the other boat wasn't aligned with it. So, you find a protractor or other object to measure just how far the other boat is "offset" from the pinnacle, and watch to see if that offset changes. That's how a compass works.

(To get a relative bearing you can either use a pelorus to get a direct measurement, or use the compass to similarly determine your heading and then take the difference of the two to determine a relative bearing.)

Originally Posted by Ronny170

The amount of change in bearing of a landmark from the perspective of the green observation boat doesn't really matter much. What matters as far as the landmarks is concerned is whether the target boat makes the same amount of headway (making trees) against the landmarks within a given amount of time. If it's the same on two or more occasions with decreasing range then a collision would be likely. You can actually make a meaningful assessment that way you just need two timed assessment periods to compare to.

.

Rarely are shore landmarks equally spaced for convenient measurement. This is the more complicated part, because the rebuttal is looking at a constant rate of change against a moving background scale, whereas the way most people use the method is measuring any change against a fixed background, and the two concepts are subtly but sufficiently different.

[requiem]

Originally Posted by Lodesman

That's entirely incorrect. That constant bearing (the CB part of CBDR) is a relative bearing, not fixed to a stationary landmark.

Now we get into semantics because I know you know what you mean, and hopefully vice versa. For reference:

"Such risk shall be deemed to exist if the compass bearing of an approaching vessel does not appreciably change."

To me, I consider this an absolute bearing rather than a relative bearing; it can remain constant even when the observer's boat is yawing about and making a relative bearing similarly swing about. But for practical purposes the difference is minimal; if you're maintaining a true course in spite of such yawing, and are effectively averaging measurements from a pelorus (let's avoid the "I just use a stanchion" debate) or unstabilized radar, the two methods are effectively equivalent.

Originally Posted by Lodesman

While the rest of your argument has some merit, that is if the bearing of the fixed landmark is changing so slowly as to present a fixed point, it can give a rough estimate of there being a risk of collision. The trouble is in knowing when those conditions are being met - as pinguino says, you have to know and understand the limitations.

I certainly don't see why you would use this method over the tried and true method of taking compass bearings.

It makes for a rapid assessment. For me, I have mountain ranges on the horizon that are many miles distant. There's no waiting for a needle to settle down, and then waiting to take a second bearing, when the eye has an angular resolution of (clickety-click) an arcminute.

[Ronny170]

If your suggestion is that magnetic field lines are sufficiently close to parallel around the Earth's surface, we can concede that. And a distant landmark such as the Sun or stars is sufficiently far away so that marks made over time can be considered parallel. Moving past landmarks that are within visual range, this is not true, or not true enough in most cases. That's where it gets interesting.

Originally Posted by requiem

From what I remember of the rebuttal video, distance from the observing vessel to the target vessel has been roughly the same as from the target vessel to the land mark. Even the large scale one, e.g. observer 200 miles off the coast, target 100 miles off, "landmark" on the coast.

Sure, that's probably because it's easier to fit on the screen or on a diagram that way. But it makes no difference if the distance to the land is greater than the distance between the boats, at least up until we get close to massive or infinite distances away as mentioned. This was brought up in the second video. Doubling, tripling, quadrupling, etc doesn't much matter. A boat can be alot closer to the land than to the other boat too, and it would still work, you would just have a lot more translation against the land by the target boat that way, but the math doesn't change. Shoreline can vary in distance too, doesn't really matter.

Originally Posted by requiem

(To get a relative bearing you can either use a pelorus to get a direct measurement, or use the compass to similarly determine your heading and then take the difference of the two to determine a relative bearing.)

A compass can also be used to establish and check the relative bearing of another boat quite easily. You don't need any other special equipment. But again relative bearing isn't what the videos are showing. He's talking about the rate of change of the target boat against the shoreline. Finding relative bearing isn't needed, even though of course if you measured it between the 2 boats it would remain constant on a collision course.

Finding relative bearing between the observing boat and a landmark on the shore isn't part of EITHER method, with CBDR or the visual method mentioned, so I'm not sure that's even relevant here.

Originally Posted by requiem

Rarely are shore landmarks equally spaced for convenient measurement. This is the more complicated part, because the rebuttal is looking at a constant rate of change against a moving background scale, whereas the way most people use the method is measuring any change against a fixed background, and the two concepts are subtly but sufficiently different.

You'll have to explain this one-- why would landmarks on the shore need to be evenly spaced?? The method in the video doesn't use or require that. It DOES involve measuring changes against a fixed or stationary background, at evenly spaced intervals of TIME though. That's the entire point of the method in fact. Whatever landmark is "behind" the target boat after X number of seconds is "the landmark" in question. You take a second mark after X number of seconds again and compare the 2 apparent distances traveled by the target boat. Consistent distances traveled means collision is likely, with decreasing range.

[Chotu]

Hard to imagine the point of this when we already have a way that works perfectly that doesn’t require a background. How often is there a background to look at?

Just look at your own rigging and parts of your boat.

If the boat isn’t moving its angle with respect to some landmark on your boat, you’re on a collision course.

I don’t understand why all of the complexity.

[Ronny170]

Originally Posted by Chotu

Hard to imagine the point of this when we already have a way that works perfectly that doesn’t require a background. How often is there a background to look at?

Just look at your own rigging and parts of your boat.

If the boat isn’t moving its angle with respect to some landmark on your boat, you’re on a collision course.

I don’t understand why all of the complexity.

There isn't really any additional complexity, despite what some might suggest. It's just an alternative way of visually assessing collision risk with another moving vessel, and it's really quite straightforward. The controversy came about when someone posted a video a while back showing an incorrect way to use landmarks to make this assessment, and that is what generated the interest and discussion again.
To your point, if you are in an open boat with no vertical elements to sight from, of if the seas are tossing you around a bit, the method you mentioned becomes impractical or ineffective. If there are landmarks visible behind the other boat, then this other method works just as well if not better in those circumstances.

[Chotu]

Originally Posted by Ronny170

There isn't really any additional complexity, despite what some might suggest. It's just an alternative way of visually assessing collision risk with another moving vessel, and it's really quite straightforward. The controversy came about when someone posted a video a while back showing an incorrect way to use landmarks to make this assessment, and that is what generated the interest and discussion again.
To your point, if you are in an open boat with no vertical elements to sight from, of if the seas are tossing you around a bit, the method you mentioned becomes impractical or ineffective. If there are landmarks visible behind the other boat, then this other method works just as well if not better in those circumstances.

Yeah. I guess I never really pictured doing it in a dinghy or a rowboat or something. That doesn’t really come to mind very often.