The "Making Trees" fallacy appears again...

[requiem]

Originally Posted by Ronny170

The technique works by observing the landmark which is behind the target vessel (M) at some initial point (in your diagram would be T0). Then at a set time later (T+1), you observe (a now different) landmark behind the target vessel M, let's call that second landmark L2. You gauge the apparent distance traveled by M with some reference (ruler, etc). THEN you repeat-- allow a set amount of time to go by, same amount as before. The M boat will now appear to be in front of a new third landmark, L3. You do the same measurement between L2 and L3. The angle to L1 is now irrelevant.

Ok, I have added points L2 and L3 to the diagram. Each is behind the target vessel at the proper time.

Comparing L1-L2 and L2-L3, I see the apparent distances traveled subtend 8° and 7°. Did I simply measure wrong, or is this a hint that something else is going on? The theory of the video is that these values should remain the same, yes?

Consider if instead of L2 we chose a different landmark at that time; instead of a large tower ashore perhaps there was a small lighthouse on a rock at point '?'. Since it's inline with E1 and M1 it should be perfectly valid to use. However, this completely changes the L2-L3 measurement: instead of 7° using L2, using the '?'-L3 distance gives you 14°.

Thus the method in the rebuttal video only works when the shore landmarks also happen to be in a neat, tidy line, such that the E1-L1-L2 triangle is.a near copy of the E2-L2-L3 triangle. But, as any point on the E1-M1 line works for an 'L2', and any point on the E2-M2 line works for an 'L3', those triangles can end up being very differently shaped, with very different angles.

(It also assumes you'll have a clearly distinguishable landmark at each of the proper timepoints for taking measurements, which is not often the case.)

[sailingharry]

So, in theory, as painstakingly detailed, trees doesn't work. And some here are firmly, passionately convinced.

Several here, myself included, use it regularly, and find it exceptionally helpful. If there is significant motion, the boats pass as initially assessed (unless something changes -- and with sailboats especially, something is ALWAYS changing). If it's not significantly changing, a bouncing compass will confirm that it's not significantly changing -- and action will be required.

Racers use it all the time to asses the frequent close-in passings. It's equally useful in crowded harbors. And for initial (and continuing) assessment of distant and closing contacts. If it takes 3 seconds to asses tree motion, it's a collision risk.

So, a question for naysayers. Have you frequently seen a vessel that is clearly taking (or giving) trees that suddenly becomes an imminent collision threat? Putting aside for the moment that an imminent collision threat required ignoring it for far too long! In thousands of times using this technique, I can't rightly recall it failing me.

Just for the record, my binoculars do have a compass and I have been known to use it on occasion. I also have the slick, but out of production, Ray compass "stick".

[Lodesman]

Originally Posted by sailingharry

So, a question for naysayers. Have you frequently seen a vessel that is clearly taking (or giving) trees that suddenly becomes an imminent collision threat? Putting aside for the moment that an imminent collision threat required ignoring it for far too long! In thousands of times using this technique, I can't rightly recall it failing me.

I don't want to be rude by calling this a stupid question, but let's say it's ill-considered. Why would "naysayers" use the system that they say 'nay' to? And why would you then presume, they would allow themselves to get into an "imminent collision threat"?

That said I know using the "right way" you frequently see vessels that aren't on constant bearing appear stationary against the background, and vessels that are CBDR with the background flying by. I also have a better-than-average understanding of relative motion(IMHO), so understand where making trees works and doesn't work. There are many variables dependent on many unknowns, that play into whether or not you should trust the method. Hence my wonderment at why one would use this rather suspect method, when there is a tried and true, easy-to-use, reliable method?

[Ronny170]

Originally Posted by requiem

Attachment 277298

Ok, I have added points L2 and L3 to the diagram. Each is behind the target vessel at the proper time.

Comparing L1-L2 and L2-L3, I see the apparent distances traveled subtend 8° and 7°. Did I simply measure wrong, or is this a hint that something else is going on? The theory of the video is that these values should remain the same, yes?

Consider if instead of L2 we chose a different landmark at that time; instead of a large tower ashore perhaps there was a small lighthouse on a rock at point '?'. Since it's inline with E1 and M1 it should be perfectly valid to use. However, this completely changes the L2-L3 measurement: instead of 7° using L2, using the '?'-L3 distance gives you 14°.

Thus the method in the rebuttal video only works when the shore landmarks also happen to be in a neat, tidy line, such that the E1-L1-L2 triangle is.a near copy of the E2-L2-L3 triangle. But, as any point on the E1-M1 line works for an 'L2', and any point on the E2-M2 line works for an 'L3', those triangles can end up being very differently shaped, with very different angles.

(It also assumes you'll have a clearly distinguishable landmark at each of the proper timepoints for taking measurements, which is not often the case.)

So I went back and watched parts 1 and 2 again of the rebuttal videos. I then ran a bunch of calculations, which I can share what I found tomorrow. I also went back over all your posts and tried to look past things that were confusing such as using 'angular motion' when only straight lines were used or taking compass measurement differences to stationary landmarks which aren't part of the measurements taken or needed, to try to see what your contention(s) actually are. With your last diagram, I think you're getting closer to an important point.

With regard to the two rebuttal videos, it seems 2 distinct but related questions were brought up or argued there. Question #1 was, on a collision course, will the other target boat 'move forward' and have apparent forward relative movement against the shoreline, or will it appear to remain stationary, as claimed in the original video, at least for non-infinite landmarks? I think the rebuttal videos make a very clear cut case that the target boat when on a collision course WILL 'make trees' or move forward against the land as a general rule. If it appears to sit still against the land, you will pass in front of it and not collide, again for non-infinite objects. The original video as posted is incorrect, and the rebuttal videos effectively end that debate.

Question or issue #2 was that, given that another boat is in fact 'making trees' and therefore might be on a collision course, how can you tell if it really is? Or in other words how can you tell if the other boat's apparent forward movement is constant against the land.
This issue I don't think was handled that well or completely in the videos. There were at least two tacit and critical assumptions made in those calculations for determining if the relative apparent movement of the target boat was constant or not. These were not mentioned, but should have been brought up at least, because they affect the distance measurements.
This I think is at least part of what you have been trying to express in your previous comments. I'll post what I came up with tomorrow and you can let me know. I can also go over why small immeasurable angle differences are not a limitation to the method, as you brought that up as a separate issue with using the landmark method.

[requiem]

Ok, I'll make a few very brief comments and wait for your results...

1. From the perspective of a moving observer, the relative bearings of fixed objects rotate astern (remember, eyeball observation doesn't give ranges, only bearings):


The speed at which they do so is a function of both the observer's motion vector and the (unknown) distance to the object.

2. This applies to both "infinite" and "non-infinite" landmarks; there is no hard cut-off.

3. What makes a fixed object count as "infinite" is a function of both its distance and the amount of time for which it's under observation.

[s/v Jedi]

Originally Posted by requiem

Ok, I'll make a few very brief comments and wait for your results...

1. From the perspective of a moving observer, the relative bearings of fixed objects rotate astern (remember, eyeball observation doesn't give ranges, only bearings):
Attachment 277346

The speed at which they do so is a function of both the observer's motion vector and the (unknown) distance to the object.

2. This applies to both "infinite" and "non-infinite" landmarks; there is no hard cut-off.

3. What makes a fixed object count as "infinite" is a function of both its distance and the amount of time for which it's under observation.

Exactly. A heading is nothing but a bearing on a distant object.

[Ronny170]

Originally Posted by requiem

Ok, I'll make a few very brief comments and wait for your results...

1. From the perspective of a moving observer, the relative bearings of fixed objects rotate astern (remember, eyeball observation doesn't give ranges, only bearings):
Attachment 277346

The speed at which they do so is a function of both the observer's motion vector and the (unknown) distance to the object.

2. This applies to both "infinite" and "non-infinite" landmarks; there is no hard cut-off.

3. What makes a fixed object count as "infinite" is a function of both its distance and the amount of time for which it's under observation.

Ok, almost all of this is not true or incorrect, so let’s get this cleared up first. As said before, there is no ‘angular motion’ here, only linear motion or travel is present in the hypothetical set-ups for CBDR or the ‘trees’ method from the videos. Angular motion (or movement or velocity) is the motion seen when a body moves along a curved path, as when moving in an arc around a fixed point. That type of motion isn't at play here.

And so relative bearings do not ‘rotate’ from the perspective of an observer, either fixed or moving in a straight line. Nor do the fixed or moving objects to which you are measuring your relative bearing ‘rotate’ regardless of whether your perspective is fixed or you are moving in a straight line. There is no angular or rotary motion with objects traveling in straight lines from any perspective. Relative bearing simply changes as the relative positions of two objects change, with respect to some third point. It’s not a linear rate of angular change with respect to distance traveled by the moving object, that’s for sure, but that in no way implies ‘rotation’ of the relative bearing or ‘rotation’ the object.

Eyeballs don’t give ranges, but neither do bearings, relative or otherwise. Distance to target is the range. A bearing helps locate where that object is at that range.

Finally, the type or degree of change in relative bearing does NOT apply equally to both infinite and non-infinite reference points and landmarks. With infinitely distant reference points, that point will maintain its relative bearing as long as you continue to travel regardless of speed and distance, and not change. This is entirely different for non-infinite landmarks, which may change in relative bearing rather quickly in some cases, depending on speed and distance traveled by the observer, and the distance between the observer and the landmark, and the observer’s heading or orientation relative to the landmark.

Arguably there is no hard cut-off, yes, but the relative bearing to a star, a distant lighthouse, and a buoy at 100m will not change at the same rate as you travel by, and in the case of the star will not change at all (of course stars ‘move’ in the night sky, but that’s nothing to do with what we’re discussing here). “Infinite” is determined by distance, and is a functional designation, as nothing that we can see is technically infinitely far away. But for the distances we might travel on here on earth, ‘very very far away’ objects beyond our atmosphere for example can be treated as infinitely far away, and the relative bearing won’t change as you move, doesn’t matter what your ‘motion vector’ is. ‘Very far away’ objects can be treated like infinitely far off for short periods of time, and the relative bearing shouldn’t change appreciably in that time. Anything else and the relative bearing will change as you travel, sometimes very significantly within a short time, unlike 'infinite' reference points.



So now to the issue of measuring apparent relative movement of the target boat against the land. I did a number of hand calcs for this, not a formal proof, but I think the following is correct. In order for the ‘trees’ method to work strictly as advertised, the math requires parallelograms in order to be perfectly happy. The sight lines taken from the observing boat to the target boat on collision courses will produce the first pair of parallel lines of the parallelogram.
The other pair of parallel lines required are A) the course line segments of the observing vessel as it travels along it’s straight course line, and B) that of the land or specifically of the landmarks on the land. That means two things. #1 that the observing vessel has to be traveling parallel to the shoreline or landmarks, and #2 that the landmarks themselves are in a line parallel with the observing vessel’s course line or route. The speed and orientation or heading of the target boat during the process is irrelevant and doesn’t affect these calcs.

Now I played around with variances from this ‘perfect’ mathematical scenario just described. For most real-world type scenarios where one might use this technique, it remains remarkably accurate despite a fairly wide range of variations from the ideal. Where I found the technique to break down most significantly is when the observing vessel is close to shore sandwiching the target boat and simultaneously traveling very sharply or obliquely away or toward the land. These scenarios aren’t really where you‘d tend to use this technique anyway. But with most other ‘non-ideal’ situations it seems to hold up well enough. Longer times between measurements of apparent movement exaggerate any errors of course.
As for the landmarks, the math was pretty forgiving if the shoreline was sufficiently far from the observing vessel. As long as you don’t use a condo along the beach, and then a distant mountain top situated 5 miles behind the shoreline, and then a buoy a mile out from the coast, it still works well enough and produces fairly consistent apparent distances.


To your earlier point about landmarks not being evenly spaced, well I don’t think that’s really the case. There’s always some shiny structure or unusual building, or a defect or imperfection in the landscape, or patch of color on something that can serve as a spot to mark. Obviously sometimes there isn’t even land to use, but traveling along most coast lines there’s always something to mark at any given time, allowing you to measure and compare the amount of ‘trees made’ by another boat. If you are traveling even somewhat parallel to the shoreline, the 'trees' method should work well enough to gauge collision potential. In that case, constant rate of apparent forward movement of the other vessel against the land = high collision potential.

[Lodesman]

Originally Posted by Ronny170

If you are traveling even somewhat parallel to the shoreline, the 'trees' method should work well enough to gauge collision potential. In that case, constant rate of apparent forward movement of the other vessel against the land = high collision potential.

Why would anyone do this? It seems a lot more involved than just taking successive compass bearings.

[Kelkara]

The maximum rate a target boat on a collision course can "make trees" is the speed of your boat along the shoreline being observed. That is if it is off your beam, for any other angle it will be less.

My boat travels at about 4 knots, or a slow jogging pace, so I just imagine a jogger on the beach. I observe the boat for 3-5 second, and in that time a jogger might pass a couple of trees max. On a near shoreline that might be a significant distance, but if there is a boat between me and a near shoreline I'm paying more attention to the "decreasing range" part of CBDR and already planning my give-way move if necessary, not taking bearings or watching trees. But on a distant shore a mile or two away 2 trees is nothing (and for a shoreline 5 or more miles away barely even discernable), I'm watching to see if the boat moves a distance impossible to jog in 5 seconds, such as half a mountain ... if so then it will pass significantly in front of me.

Additionally if the target boat is "losing trees" against the shoreline, then it will pass behind my stern ... I think that holds true for all geometries.

In either case a 3-5 second glance has determined that there is no need to get the compass out and take successive bearings over a number of minutes ... My only concern now is if they change course.

I'm not using the trees to determine whether a boat IS on a collision course. but to make sure that it IS NOT on a collision course ... if there is any ambiguity then it desrves more than 5 seconds of my attention.

[Ronny170]

Originally Posted by Lodesman

Why would anyone do this? It seems a lot more involved than just taking successive compass bearings.

Well I'm not sure if you saw the beginning of this thread, but it was about the validity of using the apparent movement of another boat against the background or landmarks on the shoreline as a way to determine collision risk, as a quick visual assessment, instead of using a compass to take bearing / relative bearings measures to the other boat, as in CBDR. The 'making trees' method.

[Lodesman]

Originally Posted by Ronny170

Well I'm not sure if you saw the beginning of this thread, but it was about the validity of using the apparent movement of another boat against the background or landmarks on the shoreline as a way to determine collision risk, as a quick visual assessment, instead of using a compass to take bearing / relative bearings measures to the other boat, as in CBDR. The 'making trees' method.

Yes and the argument for using the 'trees' method was it gave an instantaneous assessment, as opposed to "needing time and successive compass bearings" as with the traditional method.

Where you describe measuring a constant rate against the trees, requires time and successive measurements - it's the worst of both worlds. Judging apparent lateral distances by eyeball is at best rough, and most likely misleading.

So if it's going to take time and you keeping your eyes locked on target, why would you do that over taking a couple bearings?

[Ronny170]

Originally Posted by Lodesman

Yes and the argument for using the 'trees' method was it gave an instantaneous assessment, as opposed to "needing time and successive compass bearings" as with the traditional method.

Where you describe measuring a constant rate against the trees, requires time and successive measurements - it's the worst of both worlds. Judging apparent lateral distances by eyeball is at best rough, and most likely misleading.

So if it's going to take time and you keeping your eyes locked on target, why would you do that over taking a couple bearings?

Perhaps someone made the "instantaneous" argument you mentioned in quotes there, but I don't think that or the other in quotes was from me, because that's certainly not correct. It's not an instantaneous method, and I don't know how looking at another vessel in an instant can tell you anything about collision risk, regardless of method. Of course it's not instantaneous, but neither is finding your compass and then aligning it with the target vessel and then taking a reading. And then doing the same over again after to get a another reading to compare to.

It's not a 'this method is better than the other' type of argument. In fact the original question was, does the 'trees' method even work the way some claim, and if not, then how? There may be situations where the visual assessment method might work better for some, and vice versa, and the point here was only to try to understand how that method actually works, since it's obvious many people don't. If it works, dismissing it altogether seems pointless, the same way someone might suggest that using a hand compass to assess collision risk is also pointless, because why not use radar instead?
It's actually been an interesting discussion, always something lo learn. Will it replace a compass or radar? Of course not, but it is still an interesting alternative and valid in certain situations if properly understood.

[requiem]

I think I was the first to use the phrase "the idea behind the method is to obtain an instantaneous assessment", but Jedi mentioned an "immediate assessment", and we're largely talking about the same thing: being able to determine if another vessel is a hazard with only a brief glance (say, no more than 2-3 seconds).

The other methods described here, whether by compass or that of the rebuttal video, require taking two (compass method) or three (rebuttal method) separate observations to arrive at a similar conclusion. Reading the bearing off the compass is the more efficient of the two, as it only requires two observations, doesn't need any background landmarks, and you don't have to worry so much about some of those special cases mentioned.

Returning briefly to the topic of objects at "infinite" vs. "non-infinite" distances. The quote "‘Very far away’ objects can be treated like infinitely far off for short periods of time, and the relative bearing shouldn’t change appreciably in that time." is the key one here, because doing such immediate assessments doesn't require more than a few seconds. So, of the objects listed, how much might their bearings change?

Assuming the observer travels 10 meters with the various objects on their beam, where the bearings will change the greatest, here's how they change:

• Polaris, at 323 light years: 0.00000000000000019°
• Buoy, at 100 meters: 5.7°
• Lighthouse, at 5 nautical miles: 0.062°

So, for me it's reasonable to conclude that a landmark over five miles distant may be treated as a fixed reference point for the duration of a short glance.

This relates to making trees as follows: if the target vessel is CBDR, then a compass bearing on it would read the same each time you took a measurement. Suppose it happened to be in line with Polaris: take a bearing 10 meters further along your route, or even 1000 meters, both the bearing to it and Polaris would be the same.

Now, suppose instead it's aligned with the 100-meter buoy when you take the bearing, and you take another bearing after you proceed another 10 meters. The bearing of the target vessel will be the same, but the bearing of the buoy will be almost 6° further aft. This is consistent with the rebuttal video's "making trees" indicating collision risk, yes?

Finally, consider a five-mile-distant lighthouse. It will have moved 0.06° aft during that observation period, and thus is behaving a whole lot more like Polaris than the buoy. Suppose you wait another 30 seconds before taking another bearing: 30 seconds at 6 knots moves you another 90 meters along, for a distance of 100 meters and a bearing change of a whopping 0.6°.

Thus if the target boat has significantly moved with respect to that lighthouse, you know it's not on a constant bearing and thus not CBDR.

[s/v Jedi]

No, if you are going to use tools (compass, binoculars etc.) then a radar is by far the best tool to use and this is the primary method in use.

You set an EBL on the target and you set a VRM ring on the target. Now you can watch both range and bearing to target, continuously and in real time.

The target is “making trees” when it gets ahead of the EBL. When it stays on the EBL, the VRM shows if it’s coming at you or if it’s moving away from you.

[El Pinguino]

I can't remember when I last used an HBC to watch the bearing of a nearby ship.

And a pair of human eyes should be able to detect whether an object is getting closer or receding. Not so easy if you have only one eye.

Even approximate range is 'trainable'.
There was a wonderful little table published in PBO years ago that was a guide to rangefinding by eye.
From seaward
Range at which you can differentiate between forest and pasture, range at which you can see houses, vehicles on road, beach, camels on beach, people on beach, sex of people on beach.
Wish I still had it but you can train yourself.