TMA Logic

[Adam106]

I'm having a blast with Cold Waters so far, but can anyone answer how the TMA logic works?

Let's say that you've had a contact for a while and generated a TMA solution of maybe 50-55% and then sensors loose that contact, maybe due to ownship speed increase to chase. After a while your sensors regain the same contact - it seems that the TMA solution has to start again from scratch? This seems very counter intuitive since the solution was pretty good beforehand and logic would say your TMA team would start from the last known solution.

There needs to be a way of maintaining target tracks with TMA solutions even if you loose sensor contact. This should be displayed as calculated position on the map. The solution could degrade slightly over time, but you shouldn't be forced to start over when regaining contact.

This is how real fire control equipment works is it not?

Or am I wrong and in game losing a high sol. contract does indeed mean quicker TMA next time?

[jenrick]

I'd also like something from the Dev's on how my AI crew constructs the TMA plot. I've done the usual things I would do to personally improve my plot (increase bearing rate, multiple course changes, etc), and gotten very minimal improvement in a solution. I've also simply driven straight at a contact, who was driving straight at me, and my solution went to 95% pretty fast. Both of these are not common occurrence's, I'd just like some guidance from the devs on how the AI constructs the TMA plot.

-Jenrick

[Julhelm]

Combination of bearing rates and signal strength. Generally it's a good idea to put the contact on the beam on a lag LOS. But distant or very quiet contacts may take a long time to TMA.

[jenrick]

So one thing I noticed just a minute or two ago.

I had the "sink the tender" mission. I had established contacts, identified them, an went to periscope depth to get a visual confirmation to maximize my solution. Came up, marked the three surface ships (got the 95% solution for each one), and then down scope. The wife asked me about something, and for maybe about 2 minutes I wasn't paying attention. When I looked back I was back to a 60% solution for all three at a radically different range (say the visual range was 20ky the new range was 35kyd or something similar). Additionally they targets headings had all changed.

I get sometimes tossing out the TMA and starting over, but doing that shortly after a visual? What is the AI protocol for that?

This seems to be an offshoot of what Adam106 is referring to in his post.

-Jenrick

[PL_Harpoon]

If a target makes a turn / change speed it can mess up TMA solutions.
On the other hand if you already knew it's exact position before I think it shouldn't suddenly change it's mind so drastically. But since I already know where they are and can deduce the rest I don't consider it that much of a problem.

[jenrick]

It's not so much an issue of weapons employment, I know where I'm gonna launch those TLAM's. To me it's just more of an issue of how the AI does the TMA. I have no issue with some heading changes and a bit (please note a bit, not 10kyd) of distance jumping if i have sketchy contacts.

"Fire control, lets call the distance to master 1 as 20,000 yards. Down scope."

"Sir, the plot says its gonna be closer to 19,000."

However what we have is:

"Fire control, lets call the distance to master 1 as 20,000 yards. Down scope."

"Sir, we just threw away the old plot and we're going with 42,000 yards."

It just reminds me of DW and SC's aTMA just trying every solution possible rather then using an actual methodology like a human does. In DW and SC it wasn't as big a deal, as you could always step in and do it. Here there's no option, you have to wait for the aTMA to catch up.

-Jenrick

[Julhelm]

It acts just like it did in Red Storm Rising.

[wathomas777]

It acts just like it did in Red Storm Rising.

What he said.

In Cold Waters, TMA is not very accurately portrayed. Actual TMA is very accurately portrayed in both Sierra's FAST ATTACK, and The Sonalysts sims.

TMA is a pretty tedious process.

In reality, with all sensors in passive mode you can only get a bearing from which the sound arrives. As your target and own ship manuever, this bearing changes. TMA is done by taking a history of these bearings and estimating course and speed to the target

As you plot these bearing lines you began to notice a pattern. On a plot table, you begin to guess the target speed and course by comparing the plotted bearing lines with little rulers called speed strips. By rotating these strips and comparing them against the bearing lines, you begin to suss out a speed, course and range.

You can also maneuver your own ship to force the bearing lines to "cross". If the target has not changed course, these crossing lines can represent either a minimum range or maximum range which helps you bracket the target

An FT also does the same thing but he does it on the FC station. He can adjust speed, course and range settings on the computer to "stack dots" on the FC station. These dots represent the historical bearing lines on the chart. When the dots are lined up vertically with little or no deviation, it means it's a good solution. Only one course, speed and range will allow all the dots to stack properly.

The FT and people tracking on paper often "compare notes". The chart table will often ask the FT to enter "their solution" to see how it looks. We also chart the FT's solution to see if it makes sense on the charting paper.

Tracking a contact is truly a team effort with at least 4 people involved.

Now if only I could tell my Chief of the watch to make turns for so many knots and maintain a desired depth....yeah, that'd be great..... :-)

[jmr]

4 people for one contact. I bet it gets super crazy in a high contact density are like a shipping lane near a busy port.

[wathomas777]

4 people for one contact. I bet it gets super crazy in a high contact density are like a shipping lane near a busy port.

Yes it's crazy. But you always work on your closest contacts. And once you nail the solution for a contact, then you don't have to pay him much mind anymore, until he changes course or speed.

But yes, it is organized chaos. I'm pretty sure that it's a bit different now. Computers can probably do all the manual stuff pretty easily now.

But back in the 80's that's how it was done.

The shocking thing was I expected 688's to be marvels of modern technology, only to find out our laptop computers were more powerful than the ones running our fire control stations.

[jenrick]

wathomas777:

I know that both CW and RSR don't simulate TMA accurately. My only major issue is that sometimes it seems to provide very fast TMA (even with the fast pace of gameplay) compared to what I would expect based on the geometry. I am basing this off of a lot of time in DW (though I did just get Fast Attack of Ebay, great game!) which I know certainly doesn't mean much compared to real world experience. I don't have any problem with the TMA plot taking a while even when doing the right things (reasonably high bearing rate delta, lagging the target to help with the first, etc), that's just gonna happen sometimes. Maybe the other skipper is really good and doing all kinds of things to mess up the plot. However sometimes it will seem to jump from 40% or so to 95% for no apparent reason, or counter intuitively. As I mentioned I had one engagement where I turned head onto the target, and as far as I could tell they did the same. Several non time compressed seconds later, I have a 95% solution off a situation that as far as I know should have just given me a bearing an no useful information regarding range.

The other part of what I would like to see changed, is how the tac map and solution sometimes jumps a large distance even when a pretty solid datum was given to use as the basis. Visually ID'ing a ship and ranging it with a laser range finder, gives bearing and range. I'd think in the real world you'd at least be able to put an X on the plot with where the heck he is right now, and then start working. Where I see it the most often, is when I have a real shaky contact, and I visually ID and range them well before I have anything close to a solution. The tac map will display their current true position, while the scope is up. The second you go scope down, they will immediately go back to an guesstimated position, that may be 5-10kyd off from where they were plotted just seconds prior. I have no issue with the solution degrading, potentially rapidly, and that fix getting less and less useful over time (based on target speed, etc). However to have that data point of a true location immediately go out the window is... vexing.

-Jenrick

[wathomas777]

wathomas777:

I know that both CW and RSR don't simulate TMA accurately. My only major issue is that sometimes it seems to provide very fast TMA (even with the fast pace of gameplay) compared to what I would expect based on the geometry. I am basing this off of a lot of time in DW (though I did just get Fast Attack of Ebay, great game!) which I know certainly doesn't mean much compared to real world experience. I don't have any problem with the TMA plot taking a while even when doing the right things (reasonably high bearing rate delta, lagging the target to help with the first, etc), that's just gonna happen sometimes. Maybe the other skipper is really good and doing all kinds of things to mess up the plot. However sometimes it will seem to jump from 40% or so to 95% for no apparent reason, or counter intuitively. As I mentioned I had one engagement where I turned head onto the target, and as far as I could tell they did the same. Several non time compressed seconds later, I have a 95% solution off a situation that as far as I know should have just given me a bearing an no useful information regarding range.

The other part of what I would like to see changed, is how the tac map and solution sometimes jumps a large distance even when a pretty solid datum was given to use as the basis. Visually ID'ing a ship and ranging it with a laser range finder, gives bearing and range. I'd think in the real world you'd at least be able to put an X on the plot with where the heck he is right now, and then start working. Where I see it the most often, is when I have a real shaky contact, and I visually ID and range them well before I have anything close to a solution. The tac map will display their current true position, while the scope is up. The second you go scope down, they will immediately go back to an guesstimated position, that may be 5-10kyd off from where they were plotted just seconds prior. I have no issue with the solution degrading, potentially rapidly, and that fix getting less and less useful over time (based on target speed, etc). However to have that data point of a true location immediately go out the window is... vexing.

-Jenrick

You are 100% correct. CW doesn't do a good job integrating solid data into the solution. RSR was better at it. Also the game does accelerate TMA to keep the action moving. Plotting dots was long and tedious. Sub warfare was always 99% boredom followed by 1% sheer terror.

I absolutely loved fast attack, but it had a 2 hour cap on each mission. You could easily spend 3/4 of that time getting a good solution. This became a fatal flaw, especially on the TLAM missions where flight time was significant. To get to your launch point you had to manuever in such a way to endanger the ship and if you hit 2 hours, even if you had birds in the air, you failed.

And the worst part, was it wasn't even something stupid like ("x" is going to be gone in 2 hours) it was simply that was a built in artifical wall.

My personal feelings are as such.

RSR was the best overall game in my opinion. It was not a good "systems" simulator, but it nailed it as a "tactics" simulation and I know quite a few sub drivers on my boat who used it to brush up on their tactics as to where to put own ship to maximize effectiveness.

My perfect subsim would be the authenticity of the displays and systems modelling shown in Fast Attack, with the playability of RSR/CW.

However, if CW simply rises to the level of a graphically superior version of RSR, I'd be fine with that as well. it's really close right now. A few more tweaks and I think we'll get there.

Call me old fashioned but I hate 3d view. It's something that you never had the advantage of. I'd love to disable completely and have your 2D view (plot) be default view. I'd also like the event camera to interrupt 2d view momentarily to see your kill or miss and then go back to main plot. That's what RSR did and yes it was inaccurate, but it was OK for playability

However I know that too many people love to camera out from the sub so 3d is here to stay. But it would be nice if it weren't default and if it didn't take 95% of screen.

[Ansgar Burkhard]

Yes it's crazy. But you always work on your closest contacts. And once you nail the solution for a contact, then you don't have to pay him much mind anymore, until he changes course or speed.

But yes, it is organized chaos. I'm pretty sure that it's a bit different now. Computers can probably do all the manual stuff pretty easily now.

But back in the 80's that's how it was done.

The shocking thing was I expected 688's to be marvels of modern technology, only to find out our laptop computers were more powerful than the ones running our fire control stations.

Hehe. There is a reason the military is using more and more COTS in their platforms (F-15, Virginia class submarines, M1 tanks). Especially in this day and age of cyber resilience it is super important.
For instance, a smartphone will have more computing power than the F-22 computers. Sure they are super specialised but if you want to DDOS a military craft it doesn't take too much.

It is a combination of open technology (as in there are international conferences and It is discussed publicaly) in the civilian sector and the vast amounts of money spent on said technology. Last year alone, more money was spent in the gaming computer market (gaming peripherals like, certain CPUs, GPUs and soundcard for some reason) than tge entire UK defense budget. If you look at computing market and smartphone market combined you get almost as much as the global defense budget. New times... And these big ticket military purchases take time to build and be deployed. For instance, the new dreadnoughts form the Royal Navy are 2006 tech. By the time they are deployed it will be the 2030s and they are to remain in service until 2070 or so. Can you imagine using a computer from 2006 nowadays? I have a work laptop from 2009 (ThinkPad W510) and CW is one of the few games I can play in it. If I need to process satellite images in it it takes ages and it is generally not the fastest at doing anything (still runs though).

Anyway, found your remark interesting and kind of turned into a rant...

[wathomas777]

Hehe. There is a reason the military is using more and more COTS in their platforms (F-15, Virginia class submarines, M1 tanks). Especially in this day and age of cyber resilience it is super important.
For instance, a smartphone will have more computing power than the F-22 computers. Sure they are super specialised but if you want to DDOS a military craft it doesn't take too much.

It is a combination of open technology (as in there are international conferences and It is discussed publicaly) in the civilian sector and the vast amounts of money spent on said technology. Last year alone, more money was spent in the gaming computer market (gaming peripherals like, certain CPUs, GPUs and soundcard for some reason) than tge entire UK defense budget. If you look at computing market and smartphone market combined you get almost as much as the global defense budget. New times... And these big ticket military purchases take time to build and be deployed. For instance, the new dreadnoughts form the Royal Navy are 2006 tech. By the time they are deployed it will be the 2030s and they are to remain in service until 2070 or so. Can you imagine using a computer from 2006 nowadays? I have a work laptop from 2009 (ThinkPad W510) and CW is one of the few games I can play in it. If I need to process satellite images in it it takes ages and it is generally not the fastest at doing anything (still runs though).

Anyway, found your remark interesting and kind of turned into a rant...

Yeah, considering that 688's were designed in the early 70's I should have figured as much.

One of the big criticisms against the film Hunt for Red October was how "modern" the 688 was. All the stuff was in the right location, but really "sexed up" technologically speaking.

During an interview one of the producers said that they toured a 688 and at first they were going to duplicate it exactly, but after seeing how "old" everything looked they figured no one would buy that our first line subs had such ancient tech, so they sexed it up with color displays and digital readouts.

I laughed because that was my first thought as well.