Isn't the TMA a little too perfect?

[MBot]

Originally Posted by Destex

And that is, completely wrong.

I am not sure why you think this is wrong. This statement sounds absolutely correct to me.

[shipkiller1]

Originally Posted by MBot

I am not sure why you think this is wrong. This statement sounds absolutely correct to me.

Because in the real world of passive TMA (without any range data) you cannot determine the course, range and speed of a contact without (driving) changing the bearing rate.

TMA is a geometry problem. You can find the third leg of a triangle if you know the other two legs. In passive TMA, you are starting out with only one leg of the triangle.

[FPSchazly]

Originally Posted by shipkiller1

Because in the real world of passive TMA (without any range data) you cannot determine the course, range and speed of a contact without (driving) changing the bearing rate.

TMA is a geometry problem. You can find the third leg of a triangle if you know the other two legs. In passive TMA, you are starting out with only one leg of the triangle.

Wow, I never thought about it that way Cheers!

[MBot]

Originally Posted by shipkiller1

Because in the real world of passive TMA (without any range data) you cannot determine the course, range and speed of a contact without (driving) changing the bearing rate.

TMA is a geometry problem. You can find the third leg of a triangle if you know the other two legs. In passive TMA, you are starting out with only one leg of the triangle.

Right, but what is wrong about the statement?

An accurate firing solution is not affected by your change in course, as your sensors detect the track from the anticipated bearing

Changing your course changes the bearing rate. If the target still lines up with the anticipated bearing from your solution, the solution is good. If the target moves away from the bearing anticipated by the solution, the solution was incorrect. That is what the statement says, and isn't that correct?

[shipkiller1]

You could interpret it that way, but it keeping with the flavor of the entire text, it is more likely interpreted the other way.

[SeaCadt07]

I would say the TMA is both too perfect in some ways and not perfect enough in others.

Too perfect
- it's pretty quick to get to 95% if you are proactively driving the problem. Not really realistic in that regard most of the time. And in reality, you will never really know if your solution is that good anyway.
- visual contacts are instant 95% solution which is not really true. It takes at least 2-3 observations for even an experienced approach officer.

Not perfect enough
- Drastic changes to solution anytime you lose contact. Existing solution should just extrapolate out.
- There should be at least a best guess for course much earlier. I was trained to always assume something closing until the data says otherwise, especially right after initial gain.

[Shadriss]

Originally Posted by SeaCadt07

I would say the TMA is both too perfect in some ways and not perfect enough in others.

Too perfect
- it's pretty quick to get to 95% if you are proactively driving the problem. Not really realistic in that regard most of the time. And in reality, you will never really know if your solution is that good anyway.

Untrue. Proactively driving for a solution is the quickest and best way to obtain a solution. Putting a percentage to it is complete garbage, but as I've mentioned elsewhere, a really on-it Fire Control Team (SONAR, Control, FTs, etc) can create a shooting solution with NO maneuvers in some circumstances.

- visual contacts are instant 95% solution which is not really true. It takes at least 2-3 observations for even an experienced approach officer.

Partially true. You make the assumption that there is no other data to rely on for refinement. If it's the first data at all, you would be correct. For most surface contacts, there is already sonar data and a fire control solution to use for refinement, so a single observation can provide a nearly perfect solution if the periscope operator is any good.

Not perfect enough
- Drastic changes to solution anytime you lose contact. Existing solution should just extrapolate out.
- There should be at least a best guess for course much earlier. I was trained to always assume something closing until the data says otherwise, especially right after initial gain.

Agree with both of these points. Relative motion is very obvious very early on in any TMA problem... direction of motion should be indicated well before the point it actually does.

[Julhelm]

Consider that the dot stack shoes error between detected bearings and generated bearings. Our percentage is kind of like an inverse of that, and much easier to code.

[Destex]

TMA algorithms produce all kinds of quantified parameters to indicate the accuracy of the solution. They should be referred to with suspicion, of course, but for a rather simplified game such as CW, I don't see anything drastically wrong with representing solution quality with percentage.

[banryu79]

Originally Posted by Destex

TMA algorithms produce all kinds of quantified parameters to indicate the accuracy of the solution. They should be referred to with suspicion, of course, but for a rather simplified game such as CW, I don't see anything drastically wrong with representing solution quality with percentage.

Exactly...

[Shadriss]

Originally Posted by Julhelm

Consider that the dot stack shoes error between detected bearings and generated bearings. Our percentage is kind of like an inverse of that, and much easier to code.

Originally Posted by Destex

TMA algorithms produce all kinds of quantified parameters to indicate the accuracy of the solution. They should be referred to with suspicion, of course, but for a rather simplified game such as CW, I don't see anything drastically wrong with representing solution quality with percentage.

Originally Posted by banryu79

Exactly...

All fair enough... I look at it as the AO's estimate of the solution's accuracy. But that it's a percentage isn't the point - the point is that tying certain elements to that percentage (IE, plot tracking at 95%, course estimation at XX%, etc) is an issue.

At any point, I should have all three parameters - it's a solution. Not a good one (percentage applies here!), but a solution. It's the Captain's job to weigh all that's going on, how he feels about that solution, and attack when he's ready. Under this system (the CW one, to clarify), I feel like I have no choice but to wait to shoot on a 95% solution because it's the only one that provides me any feel for relative motion since I can't look at PBB data.

There are times when a bearings only shot is needed, I don't argue that point. But for any type of deliberate attack, this system almost forces you to wait longer than you might really need to.

[Julhelm]

You should get speed and course already at around 50%, with range being the final factor. There's no real provision for plotting 'wrong' solutions, so instead we assume that only data which your TMA team is confident about gets plotted.

If you remember the first two Silent Hunter games, they also used a 0-99% solution indication, where the solution for your torpedoes built up over time as you made sonar/radar/visual observations on the track. In those games though, the map was still the 'all or nothing' realtime plot. We just use the solution to drive the plot display instead and let you aim your weapons as you see fit.

[Shadriss]

Originally Posted by Julhelm

You should get speed and course already at around 50%, with range being the final factor. There's no real provision for plotting 'wrong' solutions, so instead we assume that only data which your TMA team is confident about gets plotted.

If you remember the first two Silent Hunter games, they also used a 0-99% solution indication, where the solution for your torpedoes built up over time as you made sonar/radar/visual observations on the track. In those games though, the map was still the 'all or nothing' realtime plot. We just use the solution to drive the plot display instead and let you aim your weapons as you see fit.

The problem then lies in the assumption - all data gets plotted regardless of confidence, simply because what you aren't confident about now may well turn out to have been correct all along. A solution has three parts, and you never pass along a solution without all three of them, no matter how you or your team may feel about individual portions of them. The Captain and the rest of the control team has to steer off of something, and a single data point on the plot that never moves doesn't cut it. Even a vector arrow indicating direction of motion would be a huge improvement. As it is now, it is very difficult to determine the difference between a lag and lead geometry in the early going, and the vector arrow would help in that.

Speed and Course may well show up in the data block in the corner at 50%, but I rarely look at that block because it SHOULD be represented on the plot in a quick visual reference for use in tactical decisions.

As to Silent Hunter games, I've only played the 4th and 5th (and still play them quite a bit). It's hard to compare the two, though, since the systems that drive the plots are very different, as you point out.

[banryu79]

Well... "possible course interval" could be indicated by drawing a "cone" rapresenting a continguos range of possible courses... until you have a 95% quality solution and you got an etremely narrow cone = a line/arrow.

But I do not know if it will improve the gameplay... Maybe a simple arrow plotted on the contact along the perpendicular of the LOB to just indicate if the bearing is drawing left or right could be useful?

[I think that if you are looking for the kind of feedback on sensor data that you describe in your post then Dangerous Waters comes to mind!]

[TigerDude]

I joined to participate in this discussion.

The TMA is way too easy and way too perfect. Real world contacts at max detection distances are tenuous, with not good bearings. Towed array bearings were always worse (and not clearly on one side off the ship vs the other).

A whole team worked on this. It was hard. The data was imperfect. And surface contacts were actually worse (non-intuitively). Fast moving, so less time to maneuver, and periscope observations are typically terrible (the difference in .2 divisions and .3 divisions is a 50% range error).

And active search torpedoes aren't the end-all solution. It's a big ocean.