Collision avoidance at night

[Dockhead]

Originally Posted by Wotname

I do agree that any small change in either course and/or speed may not create a sufficient CPA however I am seeing this as an exam question and we are given as a fact that the relative bearing is constant - there is no ambiguity here, unlike what might arise in real on water conditions.

Although not mentioned in the quoted post, my previous posts have described an initial heading change of ~70 degrees (from close hauled to broad reach- deep enough to show my stern light to the other vessel) before coming back up to a beam(ish) reach. i.e. still a course change of 45 degrees. This degree of course change coupled with the significant increase in speed achieved from falling off from being close hauled to reach would most surely provide a solution of an acceptable CPA.

Not saying by falling off to the degree I have described is the answer wanted by the RYA but I do suggest it would the satisfy the requirements imposed by the colregs - but as always, I am happy to be shown where my understanding is in error.

I think you're right that this is outside the scope of the exam question.


But I think it is a crucial, a central question of collision avoidance PROCESS -- how you get from an apparent collision course, to a safe resolution of the potential collision. "Change course and/or speed" is just not a satisfactory answer to that question precisely because of what we don't know.



A really big change of course and/or speed might do it, but how do you know? Just to take an extreme case to illustrate the point -- suppose (contrary to the hypothetical here) that you are dealing with a very large ship making 24 knots, which you for whatever reason only noticed on an apparent collision course at 3 cables of distance, on a perpendicular crossing. Will 70 degrees of course change, in either direction, be enough?



The answer is that if your means of taking bearings is only accurate to say 5 degrees and without a consistent error (typical HBC off the deck of a small moving boat with the sea up), there is NO action you can take, which will guarantee safety, not even a 180 degree turn and scamper off in the other direction.



Other cases -- slower or smaller vessel, more distance, better bearings -- will reduce the uncertainty about where is safe and where is not, but this uncertainty ALWAYS exists to some extent or another, so a crucial part of collision avoidance process is figuring out what action will be SUFFICIENT to achieve safety.



This is why I like the tack vs. falling off despite significant other drawbacks of this maneuver -- you can instantly see when you have achieved safety. Plus your own maneuver is better visible to the other vessel. These are powerful arguments.

[Wotname]

Originally Posted by Dockhead

I think you're right that this is outside the scope of the exam question.


But I think it is a crucial, a central question of collision avoidance PROCESS -- how you get from an apparent collision course, to a safe resolution of the potential collision. "Change course and/or speed" is just not a satisfactory answer to that question precisely because of what we don't know.



A really big change of course and/or speed might do it, but how do you know? Just to take an extreme case to illustrate the point -- suppose (contrary to the hypothetical here) that you are dealing with a very large ship making 24 knots, which you for whatever reason only noticed on an apparent collision course at 3 cables of distance, on a perpendicular crossing. Will 70 degrees of course change, in either direction, be enough?



The answer is that if your means of taking bearings is only accurate to say 5 degrees and without a consistent error (typical HBC off the deck of a small moving boat with the sea up), there is NO action you can take, which will guarantee safety, not even a 180 degree turn and scamper off in the other direction.



Other cases -- slower or smaller vessel, more distance, better bearings -- will reduce the uncertainty about where is safe and where is not, but this uncertainty ALWAYS exists to some extent or another, so a crucial part of collision avoidance process is figuring out what action will be SUFFICIENT to achieve safety.
.............

I believe we are both on the same page here!

Originally Posted by Dockhead

This is why I like the tack vs. falling off despite significant other drawbacks of this maneuver -- you can instantly see when you have achieved safety. Plus your own maneuver is better visible to the other vessel. These are powerful arguments.

However just to be the devils advocate for a moment re tacking with respect to the OPs question, why is the tack so much better? Is it not just a course change of 90 degrees and with the same speed. Whereas falling off which has an initial course change of 70 degrees and coming back up by 20 degrees (net of 50 degrees) plus a significant increase in speed surely is similar. Both tacking and falling off (as described) give the other vessel a visual indication that I have made a clear heading alteration.

Neither gives an immediate indication of safety. The only time one can tell if a safe CPA is being achieved is by the change in the relative bearing and both methods will show a change in the relative bearing when it happens.

[Dockhead]

Originally Posted by Wotname

. . However just to be the devils advocate for a moment re tacking with respect to the OPs question, why is the tack so much better? Is it not just a course change of 90 degrees and with the same speed. Whereas falling off which has an initial course change of 70 degrees and coming back up by 20 degrees (net of 50 degrees) plus a significant increase in speed surely is similar. Both tacking and falling off (as described) give the other vessel a visual indication that I have made a clear heading alteration.

Neither gives an immediate indication of safety. The only time one can tell if a safe CPA is being achieved is by the change in the relative bearing and both methods will show a change in the relative bearing when it happens.

I don't say that tacking is "so much better". I am not dissing falling off at all, and I don't think anyone here is wrong. I hear all the reasons put forward by you and others for falling off -- in my opinion they are not wrong.



Only quibble I have with what you wrote is about the "immediate indication of safety". I contend that when you see his green (and no red) at least 3 or 4 points off your starboard bow, that you know that a collision is impossible, and you need no other data, not bearings, or anything, to know that.



That is the crucial advantage of the tack, in my opinion. In less than a minute, maybe less than half a minute -- and the timing may be crucial if you are close -- you can be sure. You can't even take a proper series of bearings in the amount of time you can achieve that with the tack.


How one FEELS about this might be much influenced by the boat. With a boat which is very cumbersome to tack and which loses way and is at risk of getting into irons -- you might be less attracted to this than I am. My boat is really easy to tack -- although she is cutter rigged, the staysail is self-tacking, and with slippery dyneema taffetta on both sides of my jib, and I can pull the jib across in seconds. Showing the other side light in seconds.

[lars]

You are already close hauled falling off would take you away from your destination and you could just end up playing chicken with the other guy for the rest of the night.you should tack because
There is no reason for you to think you are not the burdened vessel
Tacking will show him your green and this make your action clear green to green
You are not giving up hard won ground to windward that you are obviously trying to get
You can get back on course quickly and clear his stern

[Lodesman]

Originally Posted by El Pinguino

So she can be heading from South anticlockwise through 112.5º (NB not 135º as stated somewhere above )

That wasn't a reference to the arc of the sidelight; I was referring to range of possible headings of the other vessel given the broad positioning "on the bow." I should have said "roughly"; sorry if that wasn't clear. With the arc of visibility, there will be overlap with the port and stern lights at either end, so where you see just green is, I think you'll agree, is an arc of somewhat less than 112.5º. Put the definition of "on the bow" as being the 30-35º centred between "ahead" and "abeam" - that gave the course of the other vessel being roughly limited to an easily-visualized chunk of the circle.

[Lodesman]

Originally Posted by lars

You are already close hauled falling off would take you away from your destination and you could just end up playing chicken with the other guy for the rest of the night.you should tack because
There is no reason for you to think you are not the burdened vessel
Tacking will show him your green and this make your action clear green to green
You are not giving up hard won ground to windward that you are obviously trying to get
You can get back on course quickly and clear his stern

OK you tack - as your bow is coming past him and he should be seeing your red turn to green, you see his green turn to red. What do you do?

[lars]

You should assume he is waiting for your action. Also we dont know how far you two are seperated could be miles or 10 feet. He has been looking at your red so if he was on port tack you would be priviledged. He thinks he has right of way and you have no reason to doubt it

[Dockhead]

Originally Posted by Lodesman

OK you tack - as your bow is coming past him and he should be seeing your red turn to green, you see his green turn to red. What do you do?

You're not allowed to make the same strong point twice. It's not fair!


There's no good answer to this -- this is the nightmare scenario. Tack back (and what if he ALSO changes back?), or fall way off. And pray.

[Lodesman]

Originally Posted by lars

You should assume he is waiting for your action. Also we dont know how far you two are seperated could be miles or 10 feet. He has been looking at your red so if he was on port tack you would be priviledged. He thinks he has right of way and you have no reason to doubt it

Presumably your light came into view for him, at the same time as his came into view for you. It would take as much time for him to determine risk of collision, and make a decision to take action. You have no way to tell if he is port and therefore the give-way vessel. He would only wait for your action if he was stbd and therefore the stand-on vessel.

[Lodesman]

Originally Posted by El Pinguino

Pedantic semantics....... 4 points is the cut off between 'on the bow' and 'broad on the bow'..... a poorly worded question.

I see 4 pts as being the middle of "on the bow". I think within a few degrees or 1/2 point of the bow, you would simply say it is "ahead"; then out to 30º would be "fine on the bow"; roughly 30-60º is "on the bow"; from there to about 1/2-1 point ahead of the beam is "broad on the bow"; and "on the beam" covers the rest. Similar on the quarter. I think everyone has their own take on those definitions, but they should be roughly equivalent.

[Lodesman]

Originally Posted by Dockhead

A really big change of course and/or speed might do it, but how do you know? Just to take an extreme case to illustrate the point -- suppose (contrary to the hypothetical here) that you are dealing with a very large ship making 24 knots, which you for whatever reason only noticed on an apparent collision course at 3 cables of distance, on a perpendicular crossing. Will 70 degrees of course change, in either direction, be enough?

Other cases -- slower or smaller vessel, more distance, better bearings -- will reduce the uncertainty about where is safe and where is not, but this uncertainty ALWAYS exists to some extent or another, so a crucial part of collision avoidance process is figuring out what action will be SUFFICIENT to achieve safety.

You raised this point earlier and I could think of no way to illustrate the point without, er, illustrating it. Please excuse the crudeness of the drawings.

1st drawing - steady bearing scenario; think of each vessel as having a vector defined by time and speed. The speed might differ between the vessels, but the time component is equal. Collision would happen at time X:



2nd drawing - so the same scenario with a faster vessel makes its vector longer, but collision would still occur at time X:



3rd drawing - Back to the original scenario, a large alteration to stbd swings your vector away:



4th drawing - not only have you avoided the collision, but you will be across the other vessel's bow sooner than the collision would have occurred, for example at 50% of time X. Not only that, the other vessel would have only advanced 50% of her vector, so would be an equivalent distance back from what would be the point of collision, which you are already avoiding by a healthy margin:



5th drawing - a faster vessel, having a longer vector will be even further back from what would be the collision point. It may seem counterintuitive, but it's actually safer crossing a faster vessel.




You brought up the point that it could be a particularly large vessel and a light doesn't indicate bow, mid or aft, but it's still a sailing vessel and I don't think they've built any the size of supertankers yet. A larger vessel would have higher visible minima for its navlights, so should be seen farther away. If you were so close that you could collide with the other's bow while the bearing of his light was drawing left, then I submit you've already gotten to extremis, then 17(b) exists and you have to do whatever to avoid collision or mitigate it. I again would rather bear away and reduce the relative speed than turn toward and increase it. But of course at that range you may be able to make out more than the sidelight, and a turn to port might be the best option.

[Dockhead]

Originally Posted by Lodesman

You raised this point earlier and I could think of no way to illustrate the point without, er, illustrating it. Please excuse the crudeness of the drawings.

1st drawing - steady bearing scenario; think of each vessel as having a vector defined by time and speed. The speed might differ between the vessels, but the time component is equal. Collision would happen at time X:



2nd drawing - so the same scenario with a faster vessel makes its vector longer, but collision would still occur at time X:



3rd drawing - Back to the original scenario, a large alteration to stbd swings your vector away:



4th drawing - not only have you avoided the collision, but you will be across the other vessel's bow sooner than the collision would have occurred, for example at 50% of time X. Not only that, the other vessel would have only advanced 50% of her vector, so would be an equivalent distance back from what would be the point of collision, which you are already avoiding by a healthy margin:



5th drawing - a faster vessel, having a longer vector will be ever further back from what would be the collision point. It may seem counterintuitive, but it's actually safer crossing a faster vessel.




You brought up the point that is could be a particularly large vessel and a light doesn't indicate bow, mid or aft, but it's still a sailing vessel and I don't think they've built any the size of supertankers yet. A larger vessel would have higher visible minima for its navlights, so should be seen farther away. If you were so close that you could collide with the other's bow while the bearing of his light was drawing left, then I submit you've already gotten to extremis, then 17(b) exists and you have to do whatever to avoid collision or mitigate it. I again would rather bear away and reduce the relative speed than turn toward and increase it. But of course at that range you may be able to make out more than the sidelight, and a turn to port might be the best option.

The drawings are really useful


The issue, again, is UNCERTAINTY. The bigger the speed difference, the less effect your maneuver has, and the bigger piece of sea where he COULD be at CPA, so the further away you have to be from the borders of that piece of sea to be safe. The bigger effect even a slight deviation of his course, will have on where he will be at CPA. That's the whole issue.


I don't get how you can think that crossing AHEAD of a much faster vessel, could be safer. On the contrary, it seems to me, this is more and more dangerous, require greater and greater margin of error, the bigger the difference in speed. No? You have less and less control over a crossing, with a bigger difference in speed, with a faster vessel. At some point there is nothing you can do (encoutering Condor ferries in the Channel making 45 knots, for example) to prevent a collision. I have always endeavored to avoid passing ahead of faster vessels unless I have really good information (AIS or a really good radar plot) and a sufficient distance, preferable couple of miles CPA. I would hate to do it with any vessel where I lack any information other than a glimpse of a single side light.



Bearing off in our hypothetical means passing ahead (if you cross at all) -- this goes wrong more easily than passing behind. You can't always put on more speed and rarely can put on a lot more speed, whereas you can always take off way or stop, to increase CPA in a passing behind posture. When you initiate a bearing away maneuver you cannot know how you will cross with him -- the most you can know is that you have a changing bearing, which may or may not indicate safety -- either that you're not crossing with him or that you are crossing ahead, but you can't know by how much. I guess that's really what makes me instinctively not want to do it this way.


YMMV.

[Lodesman]

Originally Posted by Dockhead

Bearing off in our hypothetical means passing ahead (if you cross at all) -- this goes wrong more easily than passing behind. You can't always put on more speed and rarely can put on a lot more speed, whereas you can always take off way or stop, to increase CPA in a passing behind posture. When you initiate a bearing away maneuver you cannot know how you will cross with him -- the most you can know is that you have a changing bearing, which may or may not indicate safety -- either that you're not crossing with him or that you are crossing ahead, but you can't know by how much. I guess that's really what makes me instinctively not want to do it this way.

You're quite right that there's more than one way to skin this cat. A lot really depends on other factors that aren't listed in exam questions.

StuM pointed out there's no requirement in the Rules to avoid turning to port and while that's true I think one should generally be circumspect about doing so, for the same reasons it's discouraged/verboten in those rules.

If as someone suggested, I was planning a tack anyway then I might be more inclined to - with my big old heavy ketch a tack requires forethought and strategy, it's not a snap decision. I'd also be more inclined to go that way if I had radar or AIS telling me he was 2 miles away. If the light was drawing right, or fine on the bow/ahead then turn to port would probably be preferable. Likewise if it was broad on the bow, drawing left, or visually/by radar/AIS verified on a port tack, then I'd stand on with my bail out plan to haul out to stbd.

[StuM]

Originally Posted by lars

You should assume he is waiting for your action. Also we dont know how far you two are seperated could be miles or 10 feet. He has been looking at your red so if he was on port tack you would be priviledged. He thinks he has right of way and you have no reason to doubt it

After all this time and all these discussions we STILL get people thinking that there are such things as "privilege" and "right of way".


Please read COLREGs!

[Dockhead]

Originally Posted by Lodesman

. . . StuM pointed out there's no requirement in the Rules to avoid turning to port and while that's true I think one should generally be circumspect about doing so, for the same reasons it's discouraged/verboten in those rules.. . .

For all my great respect for StuM, I totally agree with this. It's an important part of collision avoidance -- the instinctive turn to starboard. When everything turns t*ts up and no one knows what to do, if AT LEAST everyone is turning to starboard, then there is a much better chance of avoiding tragedy. Always turn to starboard absent some compelling reason not to.