4 bearing method. Questions to you, captains.

[Von Due]

Did another test and differently this time in that there was no actual ship within range but what I did was

I drew an arbitrary line representing an imaginary ship's course.

Decided it was traveling at 8 knots and marked on the line where the ship would be at 20 minutes intervals. Did that as close as the pixels and eyesight would allow.

Drew another line representing the sub travelling at 2 knots and marked where the ship would have been at 20 minutes intervals.

Drew lines directly bewteen corresponding marks on both lines to get bearing lines. These would be the accurate bearings.

Measured the angle on these bearings then using the protractor, creating lower limits and upper limits for each measured bearing angle accepting that there is no way to tell, without hydrophone lines, where within the limits the true bearing would be.

Based on these 3 pairs of lower and upper limits, I created 4 red lines.

1st and 2nd red line had the symmetry points based on 2nd bearing, upper limit, 1st bearing lower limit.

3rd and 4th red line had the symmetry point based on 2nd bearing, lower limit, 1st bearing upper limit.

1st and 3rd was further based on the 3rd bearing upper limit.

2nd and 4th was further based on 3rd bearing, lower limit.

The 2nd and 3rd red lines were the ones furthest away from the ship's true position, on the course line I drew first, some 2700m apart.

I chose a 4th position for the sub to do the triangulation, based on what a reasonable speed would allow over 20 minutes.

Result was, without further conciderations that the uncertainty of ship's position in this particular plot, on doing the final 4th bearing, was more than 20 km, approaching 30 km*.

Further considerations: It is valid to anchor each extreme red line with the 2nd anchor of the other extreme red line. This results in 2 even more extreme red lines further apart than 2700m, resulting in even greater uncertainty in pinpointing the ship's position.

However, by using the compass one can find that some combinations of upper and lower limits for each of the 3 bearings result in invalid solutions.

For example, if the ship was at bearing 1's upper limit then it could not possibly have been at the 2nd bearing's upper limit as that would mean it could not have made it to the 3rd bearing in time. In fact, it could not have been at the 1st bearing's upper limit at all.

One can investigate each possibility and eliminate invalid solutions, thus reducing the possible angular span for each bearing but that will take a lot of effort to do just once.

Incidentally, for the 3 true bearings
1st true bearing was very close to halfway between the lower and upper limits.
2nd and 3rd true bearings were both very close to the 1/4 degree mark from the lower limit.

*The 4th bearing for triangulation was the true bearing. I did not create upper and lower limits for this. With upper and lower limits, the uncertainty would be even greater.

[Pisces]

Von Due:

I commend you documenting your efforts, but without a visual representation I find it hard to comment on it or give suggestions. Screenshots would have been nice.

[Von Due]

Originally Posted by Pisces

Von Due:

I commend you documenting your efforts, but without a visual representation I find it hard to comment on it or give suggestions. Screenshots would have been nice.

I didn't take any screenies of this but I could do it again and take screenies if that is necessary. I was thinking a description would at least allow someone else to do the same test with other parameters and get new results, be it worse or better than mine. Again, if necessary I will make another test and make screenies. Might take a while to get it done though.

One more thing I found.

The hydrophone guy sometimes have trouble keeping up with the ship, especially if the change of bearing happens fast, that is less than about 15 seconds apart. On a few occations he would report a bearing of say 44 degrees when the bearing line clearly was at 45 degrees. This complicates things even further. A 1 degree error either way from true bearing would give an unacceptable 3 degree margin for error.

I haven't looked into this possibility but could it be that the accuracy of the hydrophone reports depends on the rank and specialisation of the operator?

[Pisces]

I've noticed the crew report timing is related to the duration of the sound clips (in particular the bearing numbers) they are uttering. That also causes the bearings given to be different from what they are supposed to be at that time. I think it is game lag induced.

I never noticed much difference due to experienced crew. But I never really tested that either.

[Von Due]

Originally Posted by Pisces

I've noticed the crew report timing is related to the duration of the sound clips (in particular the bearing numbers) they are uttering. That also causes the bearings given to be different from what they are supposed to be at that time. I think it is game lag induced.

I never noticed much difference due to experienced crew. But I never really tested that either.

Thank you! The sound clip length makes sense and it made me look into the sounds folder a bit harder. Curiously enough I see speech clips for functions I don't think is in the game, like "Weather report sent". Weather reports were present in SH2 but I never saw any of that in SH3. There are others there as well.

Anyway, back on topic. Looking through the folder now trying to find the set of clips that would make the bearing report.

EDIT
Digging around in the forum I found some posts suggesting that Ubi did indeed leave out a number of planned features in the final product. Wonder if it's perfectly safe to remove these sound clips then or if there are remnants of code that would throw a fit if those files aren't there.

[Sailor Steve]

Originally Posted by Von Due

Digging around in the forum I found some posts suggesting that Ubi did indeed leave out a number of planned features in the final product. Wonder if it's perfectly safe to remove these sound clips then or if there are remnants of code that would throw a fit if those files aren't there.

I've removed sound files with no problems, but the safer way is to use a sound editor such as www.goldwave.com to make the files blank. That way the file is still there, but there's nothing in it.

[Von Due]

Originally Posted by Sailor Steve

I've removed sound files with no problems, but the safer way is to use a sound editor such as www.goldwave.com to make the files blank. That way the file is still there, but there's nothing in it.

Gracias, Sailor Steve. I'm considering the possibility to cut down on these clip lengths but it would take quite a lot of testing to see how it would all work. Would have been nice of Ubi though if they didn't decide the sound guy would constantly remind us that the merchant we are tracking is indeed a merchant and not a warship. Reminds me of the tank sim Steel Panthers*. Get hit by a rifle bullet and for the remaining mission the crew will constantly bellow out We're Hit! Some things are not really necessary. Anyway, I'm rambling.

*Steel Fury, rather. My memory is utter crap sometimes.

[Quinte]

Well, I went ahead and made the experience again, taking screenies this time.

Here is my initial setup: Target is doing about eight knots, on a regular course, sub is doing about two knots on a straight line as well (never mind the fact that this is, from a tactical point of view, stupid: I can't ever hope to intercept my target if it's steaming fast).


First time around, I've drawn my bearings simulating what bearing lines would give me (i.e not quite perfect, but almost).

The Spiess Line is pretty much bang on.



I've selected a fourth position from my sub, doing around 7 knots to get there. From ther I draw my last bearing, and obviously, the calculated position of my target, course and its speed are spot on.




Now, I've just changed my last bearing to introduce a 1° error. The result is an error of about 100 meters i the estimated position, the course calculated is still way good enough to take a blind shot.





Okay, so now I've gone and redone the experience tracing my bearings with the stock protractor.
My Spiess Line cuts the actual course with an error of about 500 meters.



But my estimated position after the fourth bearing is way off.


2400 meters off, actually.


The course I calculated, and the speed, are not exploitable (the angle between the actual course and the predicted one is about 19°).



So I guess this settles it for me, the four bearings method doesn't seem even remotely useful without using fractions of degrees when taking the bearings.
That said, this was an interesting experiment, and I'll keep playing with it to see how you can maneuver your sub during the setup to minimize the error.

[Von Due]

Thank you tons for bothering doing this! Yeah while the 4 bearing method is, if I may say so, a rather beautiful method and use of geometry, its usefulness without having at least the 10ths of a degree seems to me to be questionable.

EDIT. It should be mentioned that when I plotted this using the exact bearings, the red line was about 5 meters off the ship's exact position and the triangulation resulted in no more than 20 meters error. So, if one had the exact bearings, the 4 bearing method works.

[Sailor Steve]

Originally Posted by Von Due

Would have been nice of Ubi though if they didn't decide the sound guy would constantly remind us that the merchant we are tracking is indeed a merchant and not a warship. Reminds me of the tank sim Steel Panthers.

It's worse than that. They magically report the Auxiliary Cruiser as a warship, when it is in fact a merchant with some guns fitted. Same with the Armed Trawler - it's still a trawler, armed or not. How does a guy listening to the propellers know that? Now that I think of it, the 'Flower'-class Corvette is based on a large trawler design.

Get hit by a rifle bullet and for the remaining mission the crew will constantly bellow out We're Hit! Some things are not really necessary. Anyway, I'm rambling.

Please, don't get me started on the crew knowing they've been spotted by a merchant 10 km away and start crouching down and whispering. Same with hydrophones. They know you've been heard at the same time the enemy does.

In my own translation of the Das Boot Sound Mod I tried to change all the Speech files so the only time your crew whisper is when you order Silent Running.

[Von Due]

It certainly looks like the game master side of the game is quite a busybody passing on information neither the player nor the AI should have access to.

[Von Due]

Back on topic:

One thing that might work although it will increase the workload by quite a bit is to note down the time it takes, in seconds, for the bearing to be reported on the neighbouring degree then estimate how close to either extremes true bearing is. It will not completely solve it but it may be one way to make it more accurate.

One problem I see is that one would almost certainly be forced to pause, especially if one uses 3.15 intervals. In the real world, one can expect more than a single person keeping track of it all as it happens so pausing could be used to fake 2 people working simultaneously. Not ideal but it might be a workable solution.

[Aquelarrefox]

Quinte: good job! very interesting. As i said, that is enought good to panning intersection, and get a enought good course to a perpenducular aprouch, with this also is posible with pericope readind of one or 2 distance estimate the distance of target in a defined position (using sen equa) for a blind shoot but im not sure how acurrate it would be in medium range. I should test it with a convoy (better oportunity)