Collision avoidance at night

[Lodesman]

Originally Posted by Martkimwat

My phrase to Tack away meant any from the situation, I am quite aware what a tack is and for you to pick up on that shows you are being obtuse and looking for an argument, rather than trying to understand the situation being explained, which you patently don't since it is not possible boat B to turn towards you without either already being on starboard tack or for her to Gybe on to Starboard tack, If you don't believe me, get yourself a piece of paper and a pencil and draw the scenario out, you will see it is quite impossible without being on starboard tack long before the vessel is anywhere near turning into you unless that vessel has found a way of sailing backwards towards you. If you cannot see what is plain to see, there is no point in trying to explain it to you further, I just hope I don't come across you sailing in the Solent or the South of England generally, or anywhere for that matter.

Your phrase "tack away" has no meaning, so whatever it is that you think you're saying, is not being conveyed so as to be understandable. I suggest you take out your pencil and start drawing - vessel B is on the port bow of vessel A. If vessel B is on a port tack say a beam reach, and alters 45º to stbd, it will still be on a port tack, albeit a broad reach. If A stands on, B passes astern of A as would be required by the rules. No need to sail backwards. As it is night-time and it is not possible for A to know that B is not Stbd, A's best option is to fall-off, or turn away from the situation by turning to stbd. Once you figure out where you are wrong, I'll be waiting for your apology.

[waterman46]

Originally Posted by Lodesman

You've hit on the weak point of that particular rule - you have to make an assumption about the status of the other vessel, ie. assume it's on stbd tack, and give way. 50/50 chance it's on port and being so, will give way. The better answer is to assume stbd and take avoiding action by falling off.

I think falling off for a while, to allow time for a better observation of the other boat and to see if you now have a constantly changing bearing, is a valid choice, but not the best one.

I diagrammed the situation myself and I can see only one instance where the other boat could be on starboard tack (hence having right of way) and still showing a green light with unchanging bearing. This is when the other boat is just fine off the port bow and running close to dead downwind, perhaps even a little by-the-lee but on the starboard tack. All other points of sail of the other boat showing a green light with unchanging bearing, require the other boat to be on port tack, and as leeward boat you have right of way. In daytime, the proper action would be to hold your course and let him avoid you, but at night you can't be sure he is on port tack if he is close to dead ahead. The question implied that, but not very clearly.

It is good to think how the other boat sees you, in these situations. If you are taking action to avoid collision, you want your lights to clearly signal that change.

Since at night you can't determine which tack he is on if he is nearly dead ahead, you should either fall off or tack immediately. Tacking has the advantage that it suddenly and clearly shows him that you have made a very large course change (he sees your red light change to green) and that he should hold his course.

The danger in tacking is that if he at the same time decides to head up, to starboard to go behind you before you tacked, you may now be on a new collision course. You need to be prepared for that possibility.

[Lodesman]

Originally Posted by waterman46

The danger in tacking is that if he at the same time decides to head up, to starboard to go behind you before you tacked, you may now be on a new collision course. You need to be prepared for that possibility.

I'm not sure what course of action, you think is the best? Standing on, and betting he's on a port? Or tacking with the danger you rightly assessed?

Yes, I suppose you could work out with some certainty that the other vessel is not on a stbd tack, if it's on the other side of the true wind - but for most "yachts" that is a fairly large arc, roughly 45º from your bow.

And then there's the question of how you determine the range based on 1 green light. You could assume legal minima, trusting there's no mist or other obscuring phenomena. Minus time to assess steady bearing. Is there any danger in falling off?

[Wotname]

Originally Posted by Lodesman

........ Is there any danger in falling off?

I have pondering this all thought the thread.

Does one tack or fall off?

It seems to me that falling off achieves almost all of the requirements of the colregs except that the course change from say close hauled to a beam reach won't be apparent to the other vessel (in the scenario presented in the OP's question). Thus the other vessel will not know for awhile that I have manoeuvred as required. It will only become apparent when the relative bearing opens.

Tacking on the other hand gives the other vessel an immediate visual conformation that I have manoeuvred to avoid an collision.

What would I do - I guess it might pragmatically depend on where my next way point was. If I was close hauled beating directly to windward, I would probably tack. At least that keeps me going towards where I want to end up. However if I was close hauled and laying my course line, I would more likely fall off - right through to a broad run to show the other vessel my stern light for a minute or so before returning to a beam reach until the relative bearing had opened enough to resume coming back up hard on the wind. By showing my stern light for a short time, the other vessel should realise I am manoeuvring to avoid him.

[StuM]

Originally Posted by Wotname

I have pondering this all thought the thread.

Does one tack or fall off?

It seems to me that falling off achieves almost all of the requirements of the colregs except that the course change from say close hauled to a beam reach won't be apparent to the other vessel (in the scenario presented in the OP's question). Thus the other vessel will not know for awhile that I have manoeuvred as required. It will only become apparent when the relative bearing opens.

Tacking on the other hand gives the other vessel an immediate visual conformation that I have manoeuvred to avoid an collision.

What would I do - I guess it might pragmatically depend on where my next way point was. If I was close hauled beating directly to windward, I would probably tack. At least that keeps me going towards where I want to end up. However if I was close hauled and laying my course line, I would more likely fall off - right through to a broad run to show the other vessel my stern light for a minute or so before returning to a beam reach until the relative bearing had opened enough to resume coming back up hard on the wind. By showing my stern light for a short time, the other vessel should realise I am manoeuvring to avoid him.

Since there is an immeidate change in the light the other vessel sees, tacking complies best with "Any alteration of course and/or speed to avoid collision shall, if the circumstances of the case admit, be large
enough to be readily apparent to another vessel "

Bearing away is not necessarily apparent to the other vessel for some time.


Tacking also puts you on a heading to clearly pass behind the other vessel.


Showing a stern light jst tells the other vessel that he s in an overtaking situation. Now he will possibly start to maneuver to give way adding to the confusion.


Even if laying my course line, I wold still tack. I'd much rather lose a bit of CMG than get into a confused crossing/passing situation.


And falling off and subsequently maneuvering doesn't necessarily mean that the same situation won't repeat when I turn back onto my close hauled course, When I tack, I know that I will be clear of him when I tack back again.

[Dockhead]

Originally Posted by StuM

Since there is an immeidate change in the light the other vessel sees, tacking complies best with "Any alteration of course and/or speed to avoid collision shall, if the circumstances of the case admit, be large
enough to be readily apparent to another vessel "

Bearing away is not necessarily apparent to the other vessel for some time.

Tacking also puts you on a heading to clearly pass behind the other vessel.

. . ..

I agree with all of this analysis, and I would probably tack, too. It's a big advantage that he sees the other light immediately, and I think the geometry will be better if you can get onto something like a reciprocal course. But don't forget the big drawback of a tack in this situation -- it's a turn to port ("Turn to port, see you in court"). You will get a heading to "clearly pass behind the other vessel" -- only if he doesn't maneuver himself!

Turning to port is dangerous when the other vessel might turn to starboard at the same time.

In light of all the other factors which you analyzed, it's probably a risk worth taking, but with care. To mitigate the risk you can fall well off after your tack.

[StuM]

Originally Posted by Dockhead

Turning to port is dangerous when the other vessel might turn to starboard at the same time.

If he does, you will immediately see his light turn to red and can take appropriate action by tacking back and passing red to red.

[Dockhead]

Originally Posted by StuM

If he does, you will immediately see his light turn to red and can take appropriate action by tacking back and passing red to red.

Yes -- if you aren't too distracting from handling the ship through the tack.


But what if at the very same moment he sees your light turning to green, and he does the very same thing???


This dangerous situation is exactly what the Rules are intended to prevent -- if someone fails to stand on and you get into simultaneous maneuvering, them at least if both vessels are turning to starboard, it's less risky in most cases.


I think it's a challenging situation actually, this hypothetical. Like you, my first instinct was to tack, but I start to wonder.

[Lodesman]

Originally Posted by StuM

If he does, you will immediately see his light turn to red and can take appropriate action by tacking back and passing red to red.

You will have enough way on to make back-to-back tacks? You're lucky. I know if I tack, I'm going to lose a whole lot of speed, and wouldn't get another tack in until I picked up some speed. Depending on the wind, I'd risk winding up in irons. Falling off, I'll maintain my speed and rudder authority. I certainly wouldn't want to be faffing about turning inside, when it is likely that he would turn inside to avoid me.

There seems to be a big concern about showing the other light, but that is not at all required by the rules. The alteration need only be large enough and the bearing change will be readily apparent. As Wot suggested, falling off enough to show the sternlight would make it blatantly obvious - it would not then become an overtaking situation. Risk of collision would cease, or if B was on port tack, it may persist, but he would have been and would continue to be windward and required to take avoiding action.

[Dockhead]

Originally Posted by Lodesman

. . . Risk of collision would cease, or if B was on port tack, it may persist, but he would have been and would continue to be windward and required to take avoiding action.

If you take action and if this action fails to eliminate the risk of collision, what was the point of it? Now you're relying on him to take avoiding action. And how much time is there for that? We can't know with only a single nav light visible (in real life no way would I do this without radar).



This geometry is exactly why I would consider tacking even though it requires a turn to port, bringing other risks with it.

[Lodesman]

Originally Posted by Dockhead

If you take action and if this action fails to eliminate the risk of collision, what was the point of it? Now you're relying on him to take avoiding action. And how much time is there for that? We can't know with only a single nav light visible (in real life no way would I do this without radar).



This geometry is exactly why I would consider tacking even though it requires a turn to port, bringing other risks with it.

From the single light you have no idea what his course is, other than within a 135º range. But from closing on a steady bearing, a right turn on your part (with no change on his) will in most cases cause the bearing to draw left and you'll pass red to red or you ahead of him. In some cases it would be possible that on whatever course you steady up on, you would remain on a steady or near-steady bearing (say parallel course or slowly opening or closing). In any of these cases, B would definitely be port-tack windward and be required to keep out of your way. Point is you would have a lot more time to sort out the situation, establish radio contact, sound 5 short, whatever. Turn left towards B, you close the distance and reduce the time to deconflict any issues, especially if he does the expected thing and turns to stbd as well.

[Dockhead]

Originally Posted by Lodesman

From the single light you have no idea what his course is, other than within a 135º range. But from closing on a steady bearing, a right turn on your part (with no change on his) will in most cases cause the bearing to draw left and you'll pass red to red or you ahead of him. In some cases it would be possible that on whatever course you steady up on, you would remain on a steady or near-steady bearing (say parallel course or slowly opening or closing). In any of these cases, B would definitely be port-tack windward and be required to keep out of your way. Point is you would have a lot more time to sort out the situation, establish radio contact, sound 5 short, whatever. Turn left towards B, you close the distance and reduce the time to deconflict any issues, especially if he does the expected thing and turns to stbd as well.

This is an interesting and useful discussion.

What you say will be true with some crossing geometries and not with others. Also it depends on range -- if he's quite close, a changing bearing may not guarantee safety. The problem is hard because you don't have information so you don't know which it is. But I think in most cases a tack to port will deconflict faster -- you turn towards B only momentarily, and a diverging course will appear sooner.



In this configuration, as was posted in Post






you will create a meaningful CPA faster with a turn to port, and any speed you lose in tacking will further improve CPA. That will work even in close quarters. Whereas a turn to starboard might well not get you across his course line if he's close and especially if he's much faster than you.



But in other configurations it could be different, and the hell of it is, you don't know, because you can't see anything except one nav light.


Working on very little information and not knowing how much time I have, I would feel safer seeing green to starboard, than green to port. I think if you see his green to starboard you can't collide (provided he has any way on). Whereas seeing green to port is only safe if you see a significantly changing bearing, and in close quarters maybe not even then. If you turn to starboard you may have to get across his course line; if you turn to port until you see his green well to starboard of his bow, you'll pass behind him for sure. That would be really valuable to me working under pressure not knowing his range and having no more information than the one nav light. The turn to port will resolve it for sure in almost any circumstances; the turn to starboard -- not.



For me that's the crucial bit, and I now feel more confident in the choice of tacking -- if you don't know whether you have room or time to pass ahead of him, then turning to port will allow you to SEE when you are safe, much faster, and will get you out of the danger zone even if he is quite close. If B were a fast moving large vessel, it may be too late to pass ahead, if he's very close.



It's a bit like the old "point the bow at his stern" crude but remarkably effective rule of thumb, which is really really valuable too, when you're in a crowded harbor, say, and have too many targets to analyze each one of them.

[Lodesman]

Originally Posted by Dockhead

In this configuration, as was posted in Post



Attachment 214154


you will create a meaningful CPA faster with a turn to port, and any speed you lose in tacking will further improve CPA. That will work even in close quarters.

Unless he turns to stbd, as would be expected in most situations.
Maybe not the diagrammed situation as he is stbd, but this isn't the situation being discussed. The dynamic is different, and I might choose to turn up rather than gybe if the conditions merited that.

Whereas a turn to starboard might well not get you across his course line if he's close and especially if he's much faster than you.

This is where you are mistaken. The steady bearing that indicated a risk of collision puts the collision point at the intersection of a defined pair of vectors - you swing your vector away from that point and past his course-line as soon as you put the wheel over.

if you don't know whether you have room or time to pass ahead of him, then turning to port will allow you to SEE when you are safe, much faster, and will get you out of the danger zone even if he is quite close. If B were a fast moving large vessel, it may be too late to pass ahead, if he's very close.

Unless he changes speed suddenly, which is unlikely, you will have immediately cleared ahead of his bow from what would have been the point of impact. Depending on the geometry, you would either cross his bow quickly, diverge from his course, or (worst case) parallel him, which delays the time to collision significantly, and gives you room to fall off further.

So assuming you have to tack through 90º and he's somewhere in the middle of that, you'll spend half the time in the turn getting closer without showing him your intentions. If he is much faster, and alters to stbd as expected, while you bleed off speed tacking, he could run you down - what's your "out"? You'll be running towards each other, unsure of which action the other will take - continue the turn or reverse.

It's a bit like the old "point the bow at his stern" crude but remarkably effective rule of thumb, which is really really valuable too, when you're in a crowded harbor, say, and have too many targets to analyze each one of them.

You know how much I hate that saying In some situations you can't see the stern, and where there is a likely chance that he's going to turn - then "point at the stern" means you should have gone the other way.

If the choices are:

1) I might not get an immediate resolution, but there's no chance of a collision; or

2) I'll get to see that it worked out sooner, as long as we don't collide.

I'm going with 1).

[waterman46]

Originally Posted by hamburking

The question never said the other boat was a sailboat, and it does not matter.

Steady bearing...change course.

If it were heading away from you, then the white stern light would be seen.

Just remember these poems:

If to starboard red appears, its your duty to keep clear!
Green to green, red to red, perfect safety, go ahead!
If two lights you see ahead, starboard your wheel and show your red!

It does not matter that you are on a sailboat...you must avoid collision. Considerations of point of sail, or starboard tack over port, are only for sailboat races, not COLREGS.

To be clear, your boat actually has the right of way, and the other boat should "starboard their wheel and show their red". But my experience is that I'm in a little sailboat and the other guy is a mega ton freighter who likely can't even see me...so I'm going to turn and get the hell out of there!

You have made several statements that are untrue. Actually it does really matter if you are on a sailboat, and on which tack each of you are, and if on same tack, whether you are leeward or windward. And most important that these apply whether racing or not. Don't depend on some little catch phrases. Read and understand Section 12 of the Colregs, available at

https://www.navcen.uscg.gov/pdf/navRules/navrules.pdf
And BTW the racing rules regarding collisions are the same. We would have a H of a mess if they were different.

[Dockhead]

Originally Posted by Lodesman

You know how much I hate that saying In some situations you can't see the stern, and where there is a likely chance that he's going to turn - then "point at the stern" means you should have gone the other way.

Well, if you see a green and no red, you should be able to see his stern. It's dark here and you can't actually see it, but you know it's there, and so if you get that green light well off your starboard bow, you KNOW you're safe. I think that's the crucial point here. This is the only maneuver which allows you to see for sure that you are safe (since a changing bearing won't tell you that if you're close, if you can even discern the changing bearing).



As to the likely chance that he will make a turn to starboard -- that is definitely a huge drawback of the turn to port. I agree.

Originally Posted by Lodesman

If the choices are:


1) I might not get an immediate resolution, but there's no chance of a collision; or

2) I'll get to see that it worked out sooner, as long as we don't collide.

I'm going with 1).

I don't see how you get to "no chance of a collision", with a turn to starboard. Depending where he is, and his course, you may find yourself staring at a red AND a green. Or you see his green, and the bearing is now changing, but you have no way of knowing whether it's changing enough. That's what I have against this maneuver in a nutshell.

Originally Posted by Lodesman

This is where you are mistaken. The steady bearing that indicated a risk of collision puts the collision point at the intersection of a defined pair of vectors - you swing your vector away from that point and past his course-line as soon as you put the wheel over.

Well, I respectfully disagree with this. Because you don't know whether you've swung the vector enough or not. The steady bearing is with a side light, not with his bow. If you're very close and/or he is much faster than you, you might very well not make it across. This is less of a problem with a small vessel, of course, but with a big fast moving ship it's terrifying.



Furthermore, you need to take into consideration the accuracy of your bearings. If you're using a HBC from the cockpit of a small boat with a sea running, in the dark, there are a wide range of possible pairs of vectors which can LOOK like a steady bearing. But you might actually be passing behind somewhat, and falling off puts you right onto a collision course -- you might not be able to perceive it. Furthermore it takes TIME to take those bearings -- and you don't know how close you are. I just wouldn't want to be in a position, dealing with a target at unknown range, trying to figure out whether I'm safe or not by such a time consuming and inaccurate method. I'd rather know for sure by having his green sidelight far enough off my starboard bow to know that I cannot be pointing higher than his stern, no matter how close he is or big he is. No bearings required to figure that out.


But again, I agree, that the possibility of his starboard turn is a huge disadvantage of this maneuver.

Originally Posted by Lodesman

Unless he changes speed suddenly, which is unlikely, you will have immediately cleared ahead of his bow from what would have been the point of impact. Depending on the geometry, you would either cross his bow quickly, diverge from his course, or (worst case) parallel him, which delays the time to collision significantly, and gives you room to fall off further.

This assumes perfect knowledge of the bearing to him, and assumes that he is a point in space rather than a vessel with dimensions, such that any change of bearings results in "immediate" clearing ahead of his bow. I think this is a very dangerous assumption, for the reasons stated above. You have no good or fast way of determing whether you will clear his bow or not. That's the whole issue here.

Originally Posted by Lodesman

So assuming you have to tack through 90º and he's somewhere in the middle of that, you'll spend half the time in the turn getting closer without showing him your intentions. If he is much faster, and alters to stbd as expected, while you bleed off speed tacking, he could run you down - what's your "out"? You'll be running towards each other, unsure of which action the other will take - continue the turn or reverse.

Well, I agree, that if he turns to starboard while you're tacking to port, you're stuffed. That's a nightmare scenario.



But whether while tacking you're getting closer without showing your intentions -- maybe that varies according to the boat. If I have a decent amount of way on, I can tack through 110 degrees and harden back up on my course in about 10 seconds, losing less than a knot of speed. At least on my boat, that would be the quickest way possible to show my intentions -- suddenly flashing the green, where only red was visible before. I see however that on a slower or less responsive boat, the tacking option might be less attractive.


Interesting discussion. It's kind of a nightmare to imagine suddenly spotting a single nav light on a dark night, with no other means of observation, no radar, no AIS. I hope this will remain purely hypothetical for me -- on dark nights I'm usually at the nav table with a plotting sheet and the radar on the big screen, while someone else steers.