Collision Avoidance, Cones of Uncertainty, and Appropriate CPA

[Dockhead]

Originally Posted by ramblinrod

. . .

Somehow, the old man has managed to stay alive all these years.

Amazing!

Just luck, I'm sure. ;-)

Actually not luck, but the skill of the mariners on the bridges of ships you have encountered.

They generally assume that we have no idea what we are doing, and maneuver far ahead of an encounter with a WAFI. Their maneuvers are generally calculated to take them far enough away from us that no dumb sudden move we might make -- like a 90 degree turn, 1/4 mile away -- can cause a collision.

If you will talk to commercial mariners about collision avoidance, as I have, you will hear this story over and over again.

One of the really ingenious things about the COLREGS is that the system works and collisions will be prevented if even only one vessel understands what to do.

[El Pinguino]

Originally Posted by Dockhead

Actually not luck, but the skill of the mariners on the bridges of ships you have encountered.

They generally assume that we have no idea what we are doing, and maneuver far ahead of an encounter with a WAFI. Their maneuvers are generally calculated to take them far enough away from us that no dumb sudden move we might make -- like a 90 degree turn, 1/4 mile away -- can cause a collision.

If you will talk to commercial mariners about collision avoidance, as I have, you will hear this story over and over again.

One of the really ingenious things about the COLREGS is that the system works and collisions will be prevented if even only one vessel understands what to do.

Well... that didn't work when people who knew what to do encountered the Fitz and the McCain.

[El Pinguino]

Originally Posted by Dockhead

How do you "plot a solution to miss the nearest part of the ship"? You are describing something which is impossible. What kind of "plotting" are you talking about? You have no way of knowing where the "nearest part of the ship" will be at CPA. We've discussed this; I'm surprised you still haven't caught on.

As to not getting any closer, if you merely turn 90 degrees -- this is utter nonsense. The only thing I can guess, as to how you could think something like this, is that you are imaging a highway, and you stop at the edge of. But in open sea, there is no highway and no edge to stop at -- there is just the ship towering over you, with no change in bearing you can see, 40 seconds or so from impact. You cannot see what's going to happen. He is barrelling right at you and getting closer by the second, even if there actually is a positive CPA. Will he pass you by, or run you down? You can't know this -- at 2 1/2 cables out, you are a pure sitting duck. The best information you have is the CPA indicated on your AIS, but even this is not accurate enough to tell you whether you will get run down or not, at only 2 1/2 cables away. At 5 knots, you can only move 100 meters or so in the time it will take him to run you down. You should never get that close to a ship in open water!

As I said -- forget the dimension of the ship, its "closest part", and all that. Draw a cone based on your reasonable knowledge of where he will be at CPA. You are not safe unless you are outside of that cone, and you're not completely safe even outside of it. You cannot do collision avoidance without margins for these things you can't know.

Maybe I am missing something here.... how do you draw a cone ( which is 3 dimensional) on a plane surface?

[El Pinguino]

Originally Posted by Dockhead

OK, now we're getting somewhere! This is already a more reasonable conversation.

One thing we can agree on is the fairly obvious point that the greater the variation of our speed, the less knowledge we have about where we will cross. And at some point, in really gusty wind say, normal collision avoidance procedure simply doesn't work. In such a case, Rod has correctly stated that passing ahead of a ship is really dangerous and must be avoided. But in most cases, particularly if this is the ocean and not a lake, particularly if the sailboat is larger, the variation of speed with the wind does not mean we have no idea where we will be in 15 minutes, so normal collision avoidance procedure applies.

Where we don't agree is what to do with this. If the wind is gusty and our speed varies a lot, that does NOT mean we just sail on blithely into close quarters with a ship. On the contrary, the less steady or speed and course, the further we have to stay away. The fundamental concept is that we need to prevent intersection of the zone where ship is likely to be at CPA, from the zone where we are likely to be at CPA, and if we are give-way, this maneuver must be done in one large, visible course change which makes it obvious to the other vessel that we have taken control of the crossing and have resolved it. Early enough that the other vessel will still have a chance to maneuver itself in case it is not satisfied with what we have done.

If we are give-way, then we can maneuver prior to the risk of collision arising, and that's a good idea if we can manage it. But we can't just make up our own ideas about when the risk of collision exists.

Court cases say that a "risk of collision" exists at different ranges, depending on the circumstances, but there are some cases which say that at the very least, it exists at 5 miles. Other cases say that the give-way vessel should maneuver at no less than 3 miles off. What they teach professionals is, as I mentioned before, detection of dangerous targets should be completed by 10 miles off, and maneuvering normally by 4 miles off (but perhaps less if circumstances of multiple targets etc. require).

It's crucially important to coordinate your maneuvers, with what ships you encounter are doing. So if you are stand-on, and at 4 miles off you still have a small CPA (less than 5 cables is per se too small in open sea, and as I mentioned, one mile is the usual minimum standard), then you need to be getting ready to maneuver yourself. You have to keep in mind that your ability to resolve a risk of collision situation is proportional to your speed. So you must not leave it until too late. 4 miles is probably not too early to take action yourself in most cases in open sea, if you are stand on, and certainly by 3, or at the very least 2, you need to act.


You have recommended just turning 90 degrees, 1/4 miles out, if you still have a problem. I'm sorry I can't think of a gentler way to say this, but that is simply nuts! I really don't understand why this is not obvious to you. At 1/4 mile off, you are already invisible below the bows of a medium sized ship, so now only you can do anything. But you can't do anything! 1/4 mile off is less than 44 seconds from impact! 5 knots is only 2.57 meters per second. Even if you could turn instantaneously, there is no way to get out his way -- that is, out of where he might be in 44 seconds -- which you cannot know +/- 100 meters. His beam might be 50 meters!! If you wait until 1/4 mile off and find yourself there with a 0 CPA, you will live or die only by sheer luck.




Good practice in open sea is to keep 2 miles away from other vessels, at least, from fast moving ships. One mile is pretty much minimum, although you might pass a bit closer behind once you can see his transom. The slower you are, the less you can afford to cut these distances.

So combining this '2 miles' with your oft quoted 'alter course at 5 miles' how does one deal with a ship right ahead on a reciprocal course when your ship is under power?

Lets say a closing speed of 20 knots ...14 for her... 6 for you... at 5 miles CPA is only 15 minutes away....

To get 2 miles to starboard of your track in 15 minutes while you are doing only 6 knots requires an alteration of somewhat more than 90*....

This is why there are no hard and fast rules as to 'when' and 'how much'.... every 'risk of collision ' situation is unique ... and outside of TSS situations rarely involves right angles.....

Knowing 'when' and 'how much' is down to experience... not book learning or hard and fast rules....

Oh look.... passing with 100 metres in open waters on a reciprocal course!!!!!! .... we is all going to die...!!!!!!!!

[Dockhead]

Originally Posted by El Pinguino

Well... that didn't work when people who knew what to do encountered the Fitz and the McCain.

True. I didn't say it's perfect.

But dealing with a crazy yank DDG capable of 30+ knots is a bit of a special case.

[Dockhead]

Originally Posted by El Pinguino

So combining this '2 miles' with your oft quoted 'alter course at 5 miles' how does one deal with a ship right ahead on a reciprocal course when your ship is under power?

Lets say a closing speed of 20 knots ...14 for her... 6 for you... at 5 miles CPA is only 15 minutes away....

To get 2 miles to starboard of your track in 15 minutes while you are doing only 6 knots requires an alteration of somewhat more than 90*....

This is why there are no hard and fast rules as to 'when' and 'how much'.... every 'risk of collision ' situation is unique ... and outside of TSS situations rarely involves right angles.....



Knowing 'when' and 'how much' is down to experience... not book learning or hard and fast rules....

Oh look.... passing with 100 metres in open waters on a reciprocal course!!!!!! .... we is all going to die...!!!!!!!!

With that aspect, of course you're fine. But that was not the case we were discussing -- rather a 90 degree crossing with 0 CPA. At 1/4 mile!

That looks more like this:

[ramblinrod]

Originally Posted by Dockhead

With that aspect, of course you're fine. But that was not the case we were discussing -- rather a 90 degree crossing with 0 CPA. At 1/4 mile!

That looks more like this:

Attachment 157075

Incorrect.

According to the scenario I set the we were crossing at 90 degrees, me at 5 knots he at 20, me at 1.25 miles from the intersect point, he at 5.

With a closing constant angle where at the course intersect point, my bow was going to collide with his nearest transom edge, as he passed in front.

If he keeps his current course, (which I assume he will, if he has held that course since I've been watching him), and I turn 1/4 mile from the intersect point, I am not dead ahead of him.

I am 1/4 mile (2.5 cables) away from the original intersect point. If I turn 90 degrees, to the opposite direction he is heading, I am on a parallel course, 1/4 mile away from his course, and if neither changes course, we will maintain that distance for a few minutes until his transom passes the intersect point. I return to my original course, and by the time I get that 1/4 miles to cross his original course , (3 minutes), I will be 1 mile astern of his transom.

[Dockhead]

Originally Posted by ramblinrod

Incorrect.

According to the scenario I set the we were crossing at 90 degrees, me at 5 knots he at 20, me at 1.25 miles from the intersect point, he at 5.

With a closing constant angle where at the course intersect point, my bow was going to collide with his nearest transom edge, as he passed in front.

If he keeps his current course, (which I assume he will, if he has held that course since I've been watching him), and I turn 1/4 mile from the intersect point, I am not dead ahead of him.

I am 1/4 mile (2.5 cables) away from the original intersect point. If I turn 90 degrees, to the opposite direction he is heading, I am on a parallel course, 1/4 mile away from his course, and if neither changes course, we will maintain that distance for a few minutes until his transom passes the intersect point. I return to my original course, and by the time I get that 1/4 miles to cross his original course , (3 minutes), I will be 1 mile astern of his transom.

"Intersect point"? Where did you learn navigation? What access do you even have to that information? Sailors talk about range and bearing. I thought we were clear that "1/4 mile off" means range to target.

But in any case, this is not the scenario which we were discussing.

I'm not really so much up for starting all over again, with yet another scenario. I will only note that you continue to talk about things you cannot possibly know -- like the position of his "transom edge". Likewise -- "intersect point" -- this is a fantasy, particularly the word "point". You could plot it (if you knew how), but it is not a POINT on the chart -- it will be a circle.

One last thing -- the photo I posted is not necessarily "dead ahead of him". You cannot see at that distance where you are with respect to his course line, +/- 100 meters. You can't tell, and you won't know which way to turn.

[ramblinrod]

Originally Posted by Dockhead

"Intersect point"? Where did you learn navigation? What access do you even have to that information? Sailors talk about range and bearing. I thought we were clear that "1/4 mile off" means range to target.

But in any case, this is not the scenario which we were discussing.

I'm not really so much up for starting all over again, with yet another scenario. I will only note that you continue to talk about things you cannot possibly know -- like the position of his "transom edge". Likewise -- "intersect point" -- this is a fantasy, particularly the word "point". You could plot it (if you knew how), but it is not a POINT on the chart -- it will be a circle.

One last thing -- the photo I posted is not necessarily "dead ahead of him". You cannot see at that distance where you are with respect to his course line, +/- 100 meters. You can't tell, and you won't know which way to turn.

Well ya see, this is what happens when you have your head stuffed in nav electronics too much.

If you are crossing the ships path at 90 degrees, and you are planning to pass astern, do you consider his transom centreline, or the nearest edge of his transom? (Answer: the latter).

When my AIS alarm goes off, and I look to verify what the electronics are telling me, I see a boat (in reasonable visibility).

I see the outer surface of the boat.

I don't give a rats @$$ where the centreline or the AIS antenna is.

I want to miss the boat, and I adjust my course to do that.

I keep eyes on that boat until I pass.

I've already objected to your use of the worst possible scenario to support your point, when my statements were based on consideration of a totally different less dire scenario. Yet you keep reverting to that.

I was the one that introduced the 1/4 mile distance into this discussion, and I clearly defined it was from the potential collision intersect point between my bow and his nearest transom edge as he crossed in front of me.

I know precisely where that is in relation to me, as I have a pretty keen focus on it. That is what I have to miss.

This needs to be clear because the vessel is 700 feet long. I'm certainly not going to worry about missing his AIS antenna. I want to miss the boat that antenna is attached to.

When you take a compass bearing on a vessel, do you make it at the AIS antenna? When you check for a constant angle, do you reference some portion of your boat and the ships AIS antenna?

BTW, I think I have as much right to create a scenario to make a point, as you do.

[Lodesman]

Originally Posted by ramblinrod

Note that one does not need to know the colour of the waterline stripe.

Didnt' say that you did.

Try again.

Standing at the helm, in flat water, while sailing flat, my eye will be about 7 feet off the water.

Seriously?!? So my estimate was off by a foot? You already agreed that your statement about seeing vessels at 20 miles was, let's say, exaggerated.

Sitting on the high side, when heeled 20 degrees, it will be about 9 feet.

Assuming your sitting position is 3 ft above your standing position, and assuming height of eye is 3 ft above that, so the same as standing at the helm, which you claim is 7 ft, when level. To rise by 2 feet in a 20º heel, your beam would need to be at least 12 ft (which it is not).

Sitting on the high side, when heeled 20 degrees and in 6 foot waves, will be about 12 feet (on the crests).

From this vantage, something large enough, assuming clear conditions, can be seen from 10 miles if it is more than 12 feet off the water.

Let's agree with your somewhat suspect 12 ft height of eye. The distance to the horizon is 4.1 miles. To see something at a range of 10 miles, assuming your HoE is 12 ft, the object would need to be at least 25 ft high (again from Bowditch table 12). I agree that it is possible to see large vessels at that range btw, but I wanted to correct your misconceptions.

[El Pinguino]

Originally Posted by Dockhead

"Intersect point"? Where did you learn navigation? What access do you even have to that information? Sailors talk about range and bearing. I thought we were clear that "1/4 mile off" means range to target.

But in any case, this is not the scenario which we were discussing.

I'm not really so much up for starting all over again, with yet another scenario. I will only note that you continue to talk about things you cannot possibly know -- like the position of his "transom edge". Likewise -- "intersect point" -- this is a fantasy, particularly the word "point". You could plot it (if you knew how), but it is not a POINT on the chart -- it will be a circle.

One last thing -- the photo I posted is not necessarily "dead ahead of him her". You cannot see at that distance where you are with respect to his her course line, +/- 100 meters. You can't tell, and you won't know which way to turn.

That photo you just posted of the handy sized Salen reefer bows on?..... assuming the photo was taken with a standard lens you are +/- 50 metres ahead of her ... and you are directly ahead of her as can be seen by the way her Panama lead and foremast are in line.... assuming you are crossing you will have an actual CPA of somewhat less than 50 metres.... given that she is about 150 metres long your AIS will say you are passing about 200 metres ahead and your radar will have been telling you about 125 metres..... which is why one should get one's head out of one's ****instruments and start using one's eyes.

And yes with 'big' ships you should be watching the bearing of her bow... not the bridge as it will most likely be abaft your beam when you spear under her bow...

Am I to understand that this entire thread has only been about the situation when two ships are 'meeting' at right angles? ... golly....

[ramblinrod]

Originally Posted by Lodesman

Let's agree with your somewhat suspect 12 ft height of eye. The distance to the horizon is 4.1 miles. To see something at a range of 10 miles, assuming your HoE is 12 ft, the object would need to be at least 25 ft high (again from Bowditch table 12). I agree that it is possible to see large vessels at that range btw, but I wanted to correct your misconceptions.

Hmmm, the line of sight calculation depends on which on-line calculator I use, and seems to vary from 8 to 10 miles for both points being 12 ft above sea level.

[Dockhead]

Originally Posted by El Pinguino

Maybe I am missing something here.... how do you draw a cone ( which is 3 dimensional) on a plane surface?

To be precise, it's a section of a cone. The concept is plotting the accumulating uncertainty about the positions of moving vessels. It's analogous to the cone of uncertainty used in hurricane forecasts. See here:

https://en.wikipedia.org/wiki/Cone_of_Uncertainty

It's a powerful concept which is essential to extracting useful information out of imprecise data.

[Dockhead]

Originally Posted by El Pinguino

That photo you just posted of the handy sized Salen reefer bows on?..... assuming the photo was taken with a standard lens you are +/- 50 metres ahead of her ... and you are directly ahead of her as can be seen by the way her Panama lead and foremast are in line.... assuming you are crossing you will have an actual CPA of somewhat less than 50 metres.... given that she is about 150 metres long your AIS will say you are passing about 200 metres ahead and your radar will have been telling you about 125 metres..... which is why one should get one's head out of one's ****instruments and start using one's eyes.

And yes with 'big' ships you should be watching the bearing of her bow... not the bridge as it will most likely be abaft your beam when you spear under her bow...

Am I to understand that this entire thread has only been about the situation when two ships are 'meeting' at right angles? ... golly....

The entire thread is about how to avoid getting run down by ships. By learning why you need a certain CPA in order to be reasonably sure of not ending up under his bows.

As to the photo -- I was looking for something roughly 1/4 mile off in front of a ship at sea speed, but couldn't find it -- no one crazy enough to be in such a place to snap photos.

As to your analysis of the photo I did find -- yes -- I completely agree with you. Depending on relative speed, you are probably dead. And YES! Your AIS would have told you that you had a couple hundred meters.

I wrote about your point -- position of AIS antenna! It's just one of many reasons why you can't pass that close (200 meters by AIS, which is much futher than Rod's "180 foot safe pass" ). All of them having to do with what you don't know or can't know.

Eyeballs are essential reality checks, of course, but they will not save you here if you are sailing into a risk of collision situation, and think you know to 180 feet where he will be at CPA, and that therefore you can plan to pass that close. By the time your eyeballs tell you have a problem, you're screwed. But I know you know that instinctively, Ping. But some people have no idea what appropriate CPAs are, and that's what I'm fighting with. Surely you encountered some of those sailors when you were on ships.

[conachair]

Originally Posted by ramblinrod

I was the one that introduced the 1/4 mile distance into this discussion, and I clearly defined it was from the potential collision intersect point between my bow and his nearest transom edge as he crossed in front of me.

One point there, if you are looking at a minimum CPA of 0.25Nm then you will seriously freak out a lot of ships crew who will have standing orders for a minimum of 1Nm. Not uncommon for a ship in the english channel to tweak it's course 5 odd miles away to give a CPA of almost exactly 1Nm for a yacht, which would mean you crossing the bows more like 4Nm in front if passing ahead for a 20Kt ship & 5Kt sailing boat crossing perpendicular. Winds are rarely so flakey that the average speed varies too much.