View Full Version : [WIP] Prolonged DC Evasion
Purpose: To create a more intense, slow and methodical escort pursuit and submarine evasion.
Why? I like to sweat.
Before the how and what I'm doing, I think a little background can help. From information I have gleaned from dozens of books and interviews with WWII Submarine Veterans and a couple Soundmen from Destroyer Escorts, I have come to the realization that what I really want in a subsim is not only the intensity of attacking. In fact, what I really want is an atmosphere that is even more intense while I'm trying to escape. And it doesn't feel right to me if an escorts attack (and my evasion) feels rushed.
Japanese Sonar in WWII had serious limitations--the most serious being the loss of contact with a submerged target during the attack run. This was because during the run, a deep target would would pass beneath the escorts sonar cone, creating a time when the sonar was blind. A submarine at 300 feet would typically pass beneath this cone at a range of approximately 500-600 yards (in real life). That gave the submarine time to evade.
Sonar Range was not the issue. The Japanese Type 93-1 Active Sonar could detect a submarine at 2500 yards if sonar conditions were good. This range remained basically unchanged throughout the war. The main upgrade the Japanese had in Active Sonar technology came from the Germans who gave them their S-Gerat sonar in late 44.
What I have done/learned: I've increased the sonar range to 2400 meters. I changed the sonar cone elevations (up and down angle) to 92-102 degrees. I then decreased the sensitivity of the sonar to 0.03--and I'm now tweaking it by smaller amounts. It is the sensitivity that is so important to creating a good balance.
What happens: I've now been testing this for over 20 hours. The scenario is simple: Gato Class vs Subchaser (Elite) in 0 wind/clear conditions. The sonar cone seems to be about right--ie the SC loses me on the way in while about 500 meters away. Now--you may be thinking that this will make evading easy--just go flank and turn. But it isn't easy. He still has time to pick up my noise passively and if he has the angle, leads me very well. Sometimes I am pinged for about a minute and he doesn't contact me. It is possible to crawl away and think you've made it only to picked up 20 minutes after he'd last had a firm contact. And I think the little guy learns though I won't say more about that for now.
The Downside: It can take 3-4 hours to get away. You can get sunk.
The Upside (to my way of thinking): it can take 3-4 hours to get away. It's all about "Angles and Dangles", when to pour on the coal and when to cut speed to 1 knot.
What I'm looking for (and why I'm posting this now): I want to run a couple more tests to tune this a bit more. But by this weekend, I should have a decent test set up with a couple small scenarios. One of them will probably be the one I'm using now. I'd like about 4 people who are interested to run this test and provide feedback. I would prefer old-timers who've been playing subsims since Aces of the Deep--because this feels a lot like I remember evasion felt in that sim. I would also prefer that this test is done with external view disabled.
Well--this is a long post. If I've bored you to tears send me a bill and I'll mail you a box of kleenex ;). My closing thought is:
This is the most fun/challenge I've had evading escorts in any subsim I've played. But the proof will be in the pudding.
Cheers!!!
Peto
Questions Welcome.
panthercules
11-27-07, 01:58 AM
Count me in. I haven't been challenged by the escorts in this sim yet, and I miss the nail-biting times from SH3. One question - will I be able to test this using my current NSM/ROW/RFB/RSRD install, or will I need to skinny down to stock or something else to give this a test? As easy as JSGME has made it to deactivate all the mods and get them back in again, it probably doesn't matter, but if I'm gonna be spending a few hours getting dogged by destroyers I'd just as soon have my nice, tricked out boat to spend that time in if I have a choice :lol:
BTW - when you say below that "The sonar cone seems to be about right--ie the SC loses me on the way in while about 500 meters away.", I'm curious how you know that they have lost you (I've never really spent any time analyzing the escort behavior in SH4 - always playing with external view off (except when I'm testing the weather or something else like that), so the only way I can tell they've really lost me is when I eventually get away :D )
Well--for testing I ned external view on so I can check ranges. It feels like cheating :lol:. I know he loses me on active because his pinging fades and stops while he's running in.
As far as mods (Great question)--you should be OK. I'm using TM and others. The only files that will be changing are Sensor_AI and a couple sound files. The sound files are to make your sub more quiet so you can hear the escort's approach better--Very Important!!! I'll have it set for JSGME. NSM won't matter when you're trying to get away :hmm:.
Thanks Panther! I'll be looking forward to your remarks--and if you don't like it, that's what I want you to tell me. Just tell me why. I don't expect it to be everybody's cup o' tea.
:up:
Bill Nichols
11-27-07, 08:56 AM
I'm totally in favor of this :up: :up: :up:
I've reached March, '45 in my current career and I'm frankly disgusted with the escorts. True, they've forced me to break off an attack a couple of times, but evading them has been a cut-and-dry process of going deep, slow and ultra-quiet. They've yet to cause me to break out in a sweat, and they give up too easily.
:arrgh!:
I'm totally in favor of this :up: :up: :up:
I've reached March, '45 in my current career and I'm frankly disgusted with the escorts. True, they've forced me to break off an attack a couple of times, but evading them has been a cut-and-dry process of going deep, slow and ultra-quiet. They've yet to cause me to break out in a sweat, and they give up too easily.
:arrgh!:
Bill: Thanks for the Post! Can I ask what mods (if any) you're using? My experience was much like yours but TriggerMaru definitely raised the bar of escort ferocity ;)!
Very interesting.
Gets me thinking... What about an active sonar with a HUGE range, but low sensitivity. I'm talking 20km type ranges.
Would that work? IJN escorts were known to drive around banging away with the sonar pinging. You could make a "bungo pete" clone of a few escorts, and the only change is a super range sonar... If the sensitivity is low enough it just might work.
AVGWarhawk
11-27-07, 09:11 AM
Peto!!!!!!!!! This little ditty is long over due!!!! Often when I'm detected and attacked I will go silent and deep. As I wait it out (about 15 minutes before the DD leave) I leave the computer to do other things. If I have to go silent for 4 hours I can enjoy the realistic counter attack and go do other things around the house. This opens up other aspects of the counter attack that can become a concern....battery life, CO2 levels. These two aspects just do not matter at the present moment. 15 minutes and they are gone, I surface and start the chase. Nice work:up:
Bill Nichols
11-27-07, 10:12 AM
Bill: Thanks for the Post! Can I ask what mods (if any) you're using? My experience was much like yours but TriggerMaru definitely raised the bar of escort ferocity ;)!
I'm a big fan of RFB, that's what I always use.
Ditto for AVGWarhawk's comment, my experience has been identical to his.
Very interesting.
Gets me thinking... What about an active sonar with a HUGE range, but low sensitivity. I'm talking 20km type ranges.
Would that work? IJN escorts were known to drive around banging away with the sonar pinging. You could make a "bungo pete" clone of a few escorts, and the only change is a super range sonar... If the sensitivity is low enough it just might work.
What I've been thinking about doing to recreate "always active" escorts is creating a radar which operates under water. It would have a sweep of 270-90 degrees. What I'd like to figure out is how to make it "Ping" audibly. Thinking about working on that after I get this one balanced out.
This project will take time--when 1 test can take 4 hours and (adding in luck's X factor) it's difficult to know how close I am to achieving what I'm looking for. But I believe we're on the same page regarding escorts that run active all the time.
Thought: What would an escort that only had active sonar do? It would be easy to strip the passive off and see if it would then depend on active.
But for now--I'd better go work on that computer on the bench behind :shifty:...
Cheers!
AVGWarhawk
11-27-07, 10:22 AM
For once I would like to be held down long enough were my men are slumped on the deck plates and I open the oxygen tanks and open CO2 absorbant waiting out the DD....maybe even get the message, 'we are out of air...' Watch my battery level slowly diminish. Make that ultimate Skipper decision.....stay down a bit longer and hang on or duke it out on the surface. If you can create this effect....you will be my hero:rock:
Just don't make all the DD tough as nails;)
The sound nodes for the ping are there, actually. The sonars have them, the other sensors do not. Removing them is trivial with S3D, adding them likely is as well...
Would be kind of cool because hearing the ping actually drew US boats to convoys.
After 1.4 (and ideally after the new S3D) I was planning on also trying some new DC racks out that will help in your quest. Accurate DC loads that I added make the ships run out faster. I was thinking of making a thrower that points at the rack model, and only throws the DC 0.1m or something (with no thrower effects). The benefit here is that throws, unlike racks, have a reload time. By setting it to 90 seconds or so, I could guarantee that it would not drop 2 DCs on a single pass. Replacing 1 roll rack with these would have some normal rolls of multiple charges, and a 2d slow rack that would drop 1, then have to wait for the next pass. If 90 seconds still used em up too fast, I could set it to 300, 600, 900 seconds or whatever.
tater
Accurate DC loads that I added make the ships run out faster. I was thinking of making a thrower that points at the rack model, and only throws the DC 0.1m or something (with no thrower effects). The benefit here is that throws, unlike racks, have a reload time. By setting it to 90 seconds or so, I could guarantee that it would not drop 2 DCs on a single pass. Replacing 1 roll rack with these would have some normal rolls of multiple charges, and a 2d slow rack that would drop 1, then have to wait for the next pass. If 90 seconds still used em up too fast, I could set it to 300, 600, 900 seconds or whatever.
tater
I've been thinking about that to. What I'd like to see is much longer reload times for K-Guns and racks. This would mean that an escort could only make a dc run every 5/10/15/whatever minutes. The better the Escort crew rating, the more sparing they are with dc expenditure--lots of options.
Big question: is it possible to set k-guns so the fire (2 to a side for this example) about 3-5 seconds apart rather than tossing them all off at once? I'd like to be able to create a pattern of DC's rather than escorts dumping most of them in the same pickle barrel.
I think limiting dc's on escorts is good--but I agree with Ducimus' thoughts on the subject: An escort with no dc's doesn't have teeth, so I'd rather they have too many (unhistorical) than have them run out too soon...
AVGWarhawk
11-27-07, 11:04 AM
How about the DD detecting you as you ping away with your sub? Can this be done?
I'd rather have them be historical, myself (in terms of loadout), then tweak the other factors.
Racks have no reload time, so the only way is for me to make a thrower than acts like a roll rack. As for the K-throwers, the AI uses them the way it wants to, we have no control. What you can do is turn the nodes slightly so that they launch in some pattern.
Also, I could make more k-gun variants. I can make some with only a 105m throw range, and others with 75m or whatever. Always put 1 of each, and they will fire at once, but over a wider area. (right now each thrower has 3 ranges it can fire at, so we could set them to 74m, 75m, and 76m for 1 thrower, and 104m, 105, and 106m for the other (those being the 2 RL ranges they threw at)).
I'm very interested in your ultimate goal, once 1.4 is out I'll help with DC launchers if you think it's useful.
tater
I'm very interested in your ultimate goal, once 1.4 is out I'll help with DC launchers if you think it's useful.
tater
:rock:
How about the DD detecting you as you ping away with your sub? Can this be done?
I'm not sure I know what you mean... Are you referring to when the sub is using active to determine range or an escort using active location before it has located you? (I'd like to accomplish both of these btw).
Changing the reload times to get longer "hold downs" was on my list of things to do anyway. Given that the interaction between sensors and AI behavior is so very critical, this is one of those cases where modding 1 thing (DCs and launchers) seriously interacts with another (underwater sensors). We'd be doing each other a favor :D
tater
Ducimus
11-27-07, 11:59 AM
First i'll have to say that my thinking here, is admitidly two dimensional.
If you want a longer evasion period, the answer is to quite simply increase the "power" of the detection gear used by the AI. However, the "bark" if you will. The "bite" is the depth charge. In most cases, the "bark" makes the "bite" worse because they AI becomes more accurate. So the trick in my mind, is to make them less accurate, to make it safe for them to be super sniffers.
That said, after thinking about it for a minute, (as i type this in fact), you could probably just extend the maximum range on the active sonar considerably, without altering its down angle. That would (in theory) make them both more active (no pun intended), and not super accurate.
Part of my adjustments to the AI were as follows:
- Type93_1A -> Max elevation from 108 to 106
- Type93_3A -> Max Elevation from 112 to 112
- Type93_5A -> Max Elevation from 120 to 118
Thats the "down angle", gives them a bit more depth probing, and makes them a bit more accurate. The more max elevation they have, they longer they can keep contact. By default, i think the max elevation is 100 in the stock file. The values i list above, are just me fine tuning it.
As an aside, i think it bears mentioning that active sonar, is very mechanical in its use. From all ive ever been able to discern from the AI in SH3, and Sh4, the AI will only start pinging when three conditions are met, as part of its hardcoded AI routine.
1.) You are within the sonars geometric cone.
2.) you are presenting a favorable aspect to the AI, your surface factor being an X number or greater.
3.) You are doing both 1 and 2, for the time period specified by the sonar Detection time parameter as specified in the sim.cfg
Another thought that just hit me is if you really wanted to soup up the AI detection, just try lowering its detection time to a ridciously small number, but then you want the player to have a chance to evade as well, so this might not be possible unless you want to latch that AI unit on the player like a pack of dogs on a 3 legged cat.
AVGWarhawk
11-27-07, 12:09 PM
How about the DD detecting you as you ping away with your sub? Can this be done?
I'm not sure I know what you mean... Are you referring to when the sub is using active to determine range or an escort using active location before it has located you? (I'd like to accomplish both of these btw).
What I mean is my sub sends a ping for range and the DD can hear it. Currently, it is just to easy pinging away without recourse.
Ducimus
11-27-07, 12:13 PM
I think its silly you can use your active sonar without recourse as well, but thats something that lays in the land of hardcoded. Nothing anyone here can do about that.
Thats the "down angle", gives them a bit more depth probing, and makes them a bit more accurate. The more max elevation they have, they longer they can keep contact. By default, i think the max elevation is 100 in the stock file.
First off--I want you to know that this project has actually been most influenced by your Trigger Maru mod. Your excellent work has set me --again--on the dark path of modding ;).
The reason I am decreasing the angle of the sonar (more shallow) is to create a blind zone. I've found that with the deep delving sonar, small escorts like SC's and MS's could just whip around an immediately reaquire me. Adding the range created a situation where (in ideal sonar conditions) I was always being located and attacked again immediately. Too fast for my taste. By shallowing it up, I now have a blind zone ~500 meters out from the escort. Outside of being more historical--it also creates a slower pursuit by the escort. The deeper delving sonar was something the allies did (Q apparatus) but the Japanese never figured out. If they had, we would have lost a lot more submarines during the war.
What you've said is all correct--no argument--you know your poop :lol:! What I'm pursuing though creates a scenario where attacking a convoy is merely an introduction to a slow tango to follow. And it's working pretty well so far but as I've stated, needs a little more tuning.
Your three points for getting the escort to go "Active" are also my experience. I plan to look into that later to see if I can change it by making a suedo radar/sonar. After all--their radar is always on.
THanks for your input! Your thoughts on the subject are Very Welcome!
Sailor Steve
11-27-07, 06:14 PM
Something similar was done a while ago for SH3. This lead to battles in which the sub could survive hours of terrifying depth-charging. It's definitely a good idea for SH4.
Don't forget the other half: the kill radius should not be more than 7.5 meters (25 feet) for any standard depth charge.
Something similar was done a while ago for SH3. This lead to battles in which the sub could survive hours of terrifying depth-charging. It's definitely a good idea for SH4.
Don't forget the other half: the kill radius should not be more than 7.5 meters (25 feet) for any standard depth charge.
Agreed. It's much more fun to be pummelled into submission than to get that straight out kill shot :lol:! Ducimus has done good work with his DC's--they have a much larger Xmeters range where they explode. It compensates well for the AI always knowing how deep you are.
Ah yes. Ask any submariner. Nothing beats a good pounding by depth charges ;)...
The kill radius, and the damage radius in general are related, but not the same.
There is a range of damage the DC can do. Say MinEF 100 to MaxEF 200 damage points.
There are 2 ranges, Min, and Max. At Min, the DC does 100% of the random value generated (between 100 and 200 in this example). In addition, at MinRadius it does general hitpoint damage to the sub. Between the min and max radii, the damage drops from 100% to 0%. The threshold for sub damage has to do with armor level factor (zones) and armor. For a typical sub this means 100 damage points. If the example above did 200 damage, then at MinRadius it would do 200, well above the 100 needed and would seriously mess with the sub. Halfway between Min and Max Radius it would drop below 100, and the sub doesn't take damage. Somewhere in that range (probably armor related for the guns themselves) you see the deck gun taking damage.
I've seriously upped the damage values of the DCs, and I have yet to get reliable 1 DC kills at any range, even in contact. The reason the stock DCs are pretty nasty is that they have a huge max radius. The plus to this is that they damage multiple compartments on the sub, even though each takes lighter damage than you might expect even with a contact detonation. I think the basic problem is that subs are grossly too strong.
tater
There is a big problem with the depth error setting being high, though. It's the reason the bloody escorts blow their sterns off.
In reality, they should only blow at 30, 60, or later 90m for Type 95 DCs, and at ~7.5m intervals for Type 2 DCs. The AI overrides the detonation depth setting, so that doesn't work. (I had a way worked out to make them work right had the damn setting worked).
Depth error cuts both ways though, if you are diving hard when they drop, some will still detonate at your lower depth.
It's tough to balance, that's for sure.
tater
Ducimus
11-27-07, 06:56 PM
Very tough to balance.
Tight depth accuracy?
Pin point dropping AI.
Loose depth accuracy?
AI blowing his own stern off.
Large damage radius?
DC's become very deadly.
Small damage radius?
Not enough component damage
Smaller still damage radius?
DC's become nothing more then noisemakers.
I went larger than TM for the larger type 2 DCs than TM (MaxRadius). I used 25m as I recall.
The Max radius can actually be somewhat large because below 100 damage, nothing happens. If max is only 200, and the max radius is set to 10m, min to 5m, the DC will only do ANY damage at 7.5m, farther than that, it does nothing at all. I think they should do SOME damage out to between 10 and 15m, so if the min is ~5, the max should be ~20-25m.
Upping the EF values helps as well.
Getting a single DC to kill you is damn near impossible, even if it blows up in contact or actually inside the sub (they have no collision node, so they pass through the sub).
Adding a collision node so that DCs will bounce off your hull is on my list of things to do... imagine the fear hearing THUNK! Worse, imagine it falls into the bridge and gets stuck there, lol.
tater
Ducimus
11-27-07, 07:26 PM
:hmm:
I need to rethink my AI schema maybe. Im just afraid of opening up the max radius because im making the AI a bit more touchy. If i up one, i have to dumb down the other to offset it to so speak.
Adding a collision node so that DCs will bounce off your hull is on my list of things to do... imagine the fear hearing THUNK! Worse, imagine it falls into the bridge and gets stuck there, lol.
tater
I know someone who experienced this tater :o. He tells a good story.
This conversation you guys are having about DC's tells it all. Frankly--I like the variable depth that Ducimus uses except that it causes them to premature when they hit the water once in a while--would armoring the DC Racks help that? I don't like that escorts know your depth so well. They never guess. DC damage rating is always a touchy subject and one that is very subjective. We know that in reality--to sink a sub a dc had to go off within 25 feet of the hull. Maybe a work around would be this--create 3 levels of dc mods--wimpy/moderate and nasty. If they are stand-alone mods, people can easily make up their own mind. And if anyone would complain about that, it would be easy to ignore ;)!
Well--I'm about ready for another test. I got sunk in my last one (and that is a rare thing).
Cheers!
Ducimus
11-27-07, 08:21 PM
They key in all of this, is making the AI responsive, sans pin point DC's.
The main variables to accomplish this is a looser accuracy with smaller damage radius.
What if you if you made the sink rate a little slower? The forward momentum of the player sub (at extreme depth) would (in theory) offset a tighter accuracy rating.
My DC mod changes the type 95s to 1.9m/s like they were in RL.
Standard in game is 3m/s.
My DC mod changes the type 95s to 1.9m/s like they were in RL.
Standard in game is 3m/s.
I thought they were slower. I used to count to 10 for every 100 feet. Glad to know it's not my counting ;).
I just finished a test. I took me from my last post until now to get away (subtract 10 minutes for a smoke break--I go outside for that). Setting I used this time were: Elevation 92/100, Range 2400, Sensitivity 0.025. I think it was a bit on the easy side considering elite escort and perfect sonar conditions for him. I'll try another run with sensivity set to 0.027--starting to narrow it down.
Note: It wasn't real easy getting away but I think the elite should be a little more tough. 2 elites probably would have been a serious buggar though...
Edit: I just went into edit the sensor file and noticed that Sensitivity was set to 0.03. Apparently 2 is lowest decimal places allowed.
Double Edit: Just finished another test and the .028 stuck this time. Think I'll take up another notch or two yet.
Depth Charges: With the larger blind zone in this test, I think it might be ok to expand the damage radius a little. By playing your cards right, you have a little more time IF you're deep.
Well--I haven't patched to 1.4 yet. I want TM installed for my testing so I'm waiting (Take your time Ducimus--don't forget to eat :up:). But I've learned a few things in the last couple test runs.
For the record, setting changes I'm currently using for Type 93-3A sonar:
Min Range=200 (was 0)
MaxRange=2400 (was 1600)
MinElevation=92 (was 90)
MaxElevation=100 (was 112)
Sensitivity=0.028 (will robably use .026 on next test)
0 wind
Getting away from 1 escort is tough with these settings. Getting away from 2 Elites is a nightmare. Tried a team of 1 Elite DD and 1 Competent SC this evening and it was very challenging. Still tweaking Sensitivity for the magic number.
After the 1st run or 2, the escorts lengthen their run after dropping dc's. Apparently they know they have to be farther away to locate me and so they compensate for it. They definitely coach each other IF they are part of the same group. I tried 2 separate escorts in a test that were independant. 1 lost me and never found me again even though the other still had me located. Possible it was a "glitch" but I knd of doubt it.
Depth Charges still work :shifty:. I think pulling in the +/-17 a couple meters might be OK and possibly extending the MaxRadius for damage might be better too. But that can all come later after I'm satisfied with the sonar model (I can tweak those for testing myself). It's still very dangerous though and they manage to rock the boat fairly often.
The way this mod changes the sim has also changed my tactics. I won't say how because others may find their own way of doing it. At this point, I don't want to influence that aspect.
I've played a test where I attacked a small convoy. I now find myself planning an escape route as much as I think about sinking ships :hmm:.
I'll keep testing some more and wait for the falderal from the 1.4 release to subside--and give Ducimus a chance to grab a sandwich. When any of you want me to post a small test package, let me know. I'd appreciate some input at this stage but can certainly wait. The mod will include the following file changes:
AI_sensors
amb_subinterior
Creaking (1, 2, 3 & 4)
DepthChargeFromSubmarine
Sonar
The Audio Changes are important. Please try them. They really help too, as you'll be able to hear the escorts better during their attack runs.
And AVGWarhawk? Be careful what you wish for. You might just get it ;)!
Peto
BTW, the DC shake is totally unrelated to damage. It's set to ~60m I think.
Yep--Thank Heavens or I'd really be toast :lol: !!! DC settings are kind of secondary for me right now except I've put a lot of thought into your ideas for K-guns to replace racks, patterns with Kgun angles and ranges. Is there anyway to set a randomizer that could give a depth charge loadout a +/-10? I think I know what it would take to do it but it's beyond my capability right now.
I can build you a great computer--just don't ask me to program it!
The only way to randomize the loadouts is to make many launchers, each with a different load, and use them on different ships. SHIPS can have loadouts, too, you know, like planes (planes have a choice of bomb loads). I can make Minekaze1, Minekaze2, etc. They need to be explicitly placed in the editor, but they take up no new rec manual places, and they are identical other than their loadouts. A player will never know which one they face.
tater
The only way to randomize the loadouts is to make many launchers, each with a different load, and use them on different ships. SHIPS can have loadouts, too, you know, like planes (planes have a choice of bomb loads). I can make Minekaze1, Minekaze2, etc. They need to be explicitly placed in the editor, but they take up no new rec manual places, and they are identical other than their loadouts. A player will never know which one they face.
tater
I hadn't thought of that and that's exactly what I wanted. I'd like to Not know exactly how many dc's an escort has. I'm also moving closer to your camp for keeping the number of dc's carried closer to historical.
Early War: Destroyers carried about 18, later upped to 36 and sometimes as high as 50. I'm not sure when the numbers changed but I think we could make a good guesstimate. Destroyer Escorts (Etorofu Class) Carried about 60 and could be as high as 120. Kaikoban (entered service last half of the war) carried 120. Subchasers (limited space) started at 18 and later 36 (50 max for these guys I think). A little creative thinking (which you excel in :up:) can fill gaps for Etorofus and Kaikobans. Your Minesweepers and Auxinlary SC's leap to mind.
These numbers represent the 250 pound variety. The larger "600 pounders" used later in the war meant less DC's on the racks. K-guns always used the lighter dc's. Can the game handle using 2 different depth charge types on the same escort?
Personally, I think the minimum reload time for escorts should be 10 minutes. The max should probably be 20.
Depth Charge Patterns: Thinking of an escort with 4 K-Guns and 2 racks--Port side KGuns: Forward aimed to 300 range 150/Aft aimed 260 range 75 (bearings relative). Starboard: Forward aimed 60 range 75/aft 100 range 150.
Escorts with 2 Kguns: 270/90 Range 150.
And that could change by the loadout too. I think a pattern should cover as large an area as possible while trying to stay with historical weapon capabilities.
Just some thoughts. I'm certainly open to ideas.
From navweaps, it appears that the heavier (162kg) type 2 could be thrown since it lists the range "When used with the Type 95 DC..." that implies use with another DC type, and the only other type used was the type 2, though I suppose it could be the 105kg and 110kg type 2s, not the 162kg version.
Perhaps I need to look into that.
K guns should probably only fire to 105m with a ~100kg weapon (perhaps slightly less with the 105 and 110). The Y guns have the 75m range only when tossing 2 DCs, and they are broken in game and only fire one direction.
It appears you've been thinking about this considerably more (and longer) than I have. So I think I'm just going to leave the dc's to you and pitch in ideas if you ask for them.
Frankly tater--I have complete confidence that your solutions would be great even without my imput :up:! I'll keep tweaking sonar for now. It's just that tweaking sonar inevitably results in thinking about depth charges :shifty:.
Cheers!
Any and all comment is great. The more people looking at things, the more likely we are to come up with novel (and useful) ideas.
tater
Agreed. And I think after a couple more test runs I'll be looking at things other than sonar a little more critically. I'm old and don't multi-task as well as I used to :88).
Just finished a test and have found a setting that has allowed me to escape from a SC twice now (elite/ 0 Wind). It took between 2-3 hours both times. It requires a lot of manuevering to keep the aspect in my favor. And in the last test I lost both of my periscopes and radar before getting away :shifty:.
Sensitivity set to 0.026.
Going to try that setting with 2 escorts and wind 5 in a few minutes.
How does AI skill alter that? Will be interesting for you to test once you have an elite baseline.
tater
Ducimus
11-29-07, 04:34 PM
I've been thinking about the following changes to DC from whatever they are now
-Depth accuracy, i think its at 17 or 18, im thinking 10, or maybe the stock 5.
-Sink rate from 3 meters, to taters suggested 1.9 MPS
-Max radius from what is it.. 17 or 18? to taters suggested 25.
No other changes to the AI are occuring since last posted version of TM.
I've been thinking about the following changes to DC from whatever they are now
-Depth accuracy, i think its at 17 or 18, im thinking 10, or maybe the stock 5.
TM is indeed set to 17. I've been uising 7 with no blown escorts so far.
-Sink rate from 3 meters, to taters suggested 1.9 MPS
3m/s is the actual sink rate for the type 2 DC in use starting in 1943. The 95s were 1.9m/s. My parachute retarded version is set to 1m/s right now.
-Max radius from what is it.. 17 or 18? to taters suggested 25.
In TM the min is 4.5, the max is 14.5. I used TM as my benchmark, really. I currently have the 100kg versions (type 95) set to 16, and the 110kg type 2 at 16.5---that's ~100 damage at ~10m. I used 25 for the 162kg type 2---that's ~100 damage at 15m. Note that my 100kg versions are equal in power to the stock DCs in TM (damage wise).
tater
Back from another test. I think the magic number is Sensitvity 0.027. At least it's a good baseline to start with IMO. Gor away from an Elite DD and a Competent SC in about 3 hours. Wind 5. It was a little too easy IMO although I did have external view on. I cheated :yep:. It's quite a bit tougher without (as you know).
AI Skill: I believe this works on a percentage basis. Elite are the only ones that get 100% of rated range. Veteran = 90% etc. I haven't looked this up to be positive but I'm pretty sure I played around with that setting in SH3 and it's likely the same type of formula in 4.
I think the main thing that should make escorts tougher as the war progresses is not sonar but escort experience and DC loadout. When the 600 pounders (Type 2) come in I think they should be Bad news. They were in RL. But I think their faster sink rate will do much to accomplish that.
The main sonar upgrade the escorts should get is the Type 3 and that could be "a little" deeper probing but (I'm pretty sure) have a shorter range. The system that probed deeper were more low frequency.
Depth Charges: I think we make Small moves at first so we can kind of sneak up on the best setting. They're probably powerful enough now but we could try narrowing the margin of error like you were talking about in the prior post. Doing more than that would make it hard to measure the difference.
btw: I would kind of like one of you guys to front this mod to the community. I'll accept a by-line in the credits ;). The reason for that is you have more clout than I do (deservedly). It'll get more attentiion if your name is attached to it IMO.
this is really interesting stuff.....looking forward to the finished product.....I would love a 3-4 hour standoff....
Wilcke
Ducimus
11-29-07, 05:31 PM
Im a work.. tired, and not much sleep, don't have the files in front of me, but feel i need to alter the DC's in next TM version a little bit. Im just afraid to over do it because i dont have time to adquetly test any changes.
edit:
Maybe up the radius to 16
tighten accuracy to 10 , 7 or even 5.
decrease sink rate from 3 to....2.
I know the DC sink rate is not historical, but im using this, in conjunction with players forward movement and depth , as a supstitute for a loose depth accuracy .
edit:
or maybe
15 max radius
15 or 10 accuracy
2.5 sink
Peto, this is great work, and speaks for itself. If it gets bundled, it'll get bundled.
My rule is throw stuff out and see what sticks :D
tater
Well, it's accurate as long as the 95s were used. The type 2s were introduced in 43, so some boats would still have the 95s til at least the end of 43, IMO. Maybe some later in backwaters.
tater
tater: Thanks for the encouragement! :yep:
Ducimus--don't worry too much about the depth charges yet IMO. I think you'll find the mod I'm working on has a serious impact on what needs to be done with them--and maybe less work than you are currently thinking. Take a nap when you get home :up:!
Wilke: In certain conditions you may wish it was a 3-4 hour stand-off. It has the potential to go much longer at times ;).
I am liking the direction this is heading.
In general, from my reading, it seems like ASW prosecutions in ww2 on US subs could be very cursory, or affairs that tested the ability of the boat to remain submerged. Even with the absurdly high stock DC counts, they go through them VERY fast in stock, which is perhaps why they loaded so many on the ships.
Holding a sub down, but not killing it, is very historical, too.
So a mod, or set of mods that makes it possible to have long prosecutions, but not always, would be a great thing. The blind spot makes their DC run dash possibly less accurate, making survival better for the sub, but not a walk in the park.
In your testing, you need to try 2 tests, too.
In one test, your escorts are the leaders of whatever group they are in.
In another test, make some merchants, and a merchant is the leader. The AI behaves differently for TFs vs Convoys. A TF is ANY group where the lead ship is a warship. 100 merchants with a PT boat as the leader is a task force. 100 CVs with a fishing boat as the leader is a convoy.
The escorts will be more persistent in a TF than a convoy.
tater
The second test you mentioned has already been done tater.
I'll try a task force scenario later this evening.
After that, I'm going to try it out in real application--ie, in career mode--along with a little SD radar thing I'm working on that'll make planes less likely to be detected all the time ;).
Edit: I'm also working on 2 versions of a fairly large Lingayen Gulf scenario. About 80 ships were involved in the invasion force (3rd Fleet plus AP's and Auxilerys).
ReallyDedPoet
11-29-07, 06:22 PM
Guys this stuff sounds fantastic, well done :yep::up:
RDP
panthercules
11-29-07, 09:07 PM
along with a little SD radar thing I'm working on that'll make planes less likely to be detected all the time ;).
Oooh oooh - I'm liking the sound of this - almost all the RL patrol reports I've read so far seem to be full of entries mentioning that their radar got knocked out for hours or a day or two at a time by water coming down the hatches or some other reason, so something that would in effect simulate periodic outages and let some planes sneak in undetected by radar would be a great addition :up:
Oooh oooh - I'm liking the sound of this - almost all the RL patrol reports I've read so far seem to be full of entries mentioning that their radar got knocked out for hours or a day or two at a time by water coming down the hatches or some other reason, so something that would in effect simulate periodic outages and let some planes sneak in undetected by radar would be a great addition :up:
And that goes without saying that SD radar just wasn't reliable in the 1st place. It had 4 weak spots where planes could get in closer without detection and was very bad at picking up anything at low altitude.
Not making promises on the radar yet. Just saying that, "I'm starting to work on it."
:yep:
1 Sub Chaser, Competent crew, 0 wind--Still a serious challenge getting away.
TF with DD as group leader: It's up to you to get away. They don't leave :nope:.
Elite SC, wind 10--still difficult.
I didn't play through any of these all the way--just over an hour each to get an idea of how it would go.
I'll try a couple large scenarios tomorrow and probably try career mode for a while to see how it goes.
Something I have noticed: Results are inconsistent even in the same scenario. I think this is a good thing because it indicates that the X factor is still important. To say that on the average, things are more challenging would be a serious understatement IMO.
An interesting effect that I don't think is in game is how sonar is FUBARed by DC explosions for a while. This is a good reason to prevent the escorts from turning around and dropping again right away.
I still need to try making a Kgun that points at a rack model so I can set a reload time...
The new S3D will probably help rather a lot.
tater
Yep. I wish dc noise was a factor for escorts. It feels wrong to hear another escort pinging away while the ocean is roiling from explosions. The larger blind zone doesn't "fix" that but it decreases how often it occurs.
Just patched up to 1.4 and added TM. Did a complete reinstall of the program 1st and then defragged. I almost did an OS reinstall but--well--I hate doing that if no one pays me to :shifty:.
I'll be testing again in a few minutes. I'm cautiously optimistic about how this is going :hmm:...
Just about ready for Beta. Will release this mod later today (perhaps sooner than later ;)).
Regarding Depth Charges: Something I've been meaning to mention but keep forgetting--One of the reasons I like the +/-X setting for depth of detonation is because it addresses the fact that in real life, deoth charges were inaccurate. Drop a barrel in the water and it is almost certain to not sink straight down but it will typically veer off at an angle as it sinks. The allies compensatedfor this with the Tear-drop design but I'm not certain the Japanese did this (maybe with the type 2?). Maybe tater knows.
I think the allies used hydrostatic fuses. The IJN used a simple displacement system I believe. A hole allowed water to enter, and when it filled up, BOOM. The chamber had 2, and later 3 settings. Those would indeed veer slightly as they were "ashcans" not the streamlined ones the US had.
That said, I think the effect on depth accuracy would be very minor. The real issue is that IJN DCs (the early type 95s) could ONLY be set to detonate at 30m or 60m. If the sub was at 45m, the charge would always be 15m off in depth. Later 95s had a 90m setting, and the Type 2 DCs seem to have had 7.5m settings. Not sure if there was a min depth though (ie: it's possible it was every 7.5m AFTER 20m or some safety factor).
Right now I have the depth error set at 7m, FWIW. Stock is 5m.
The air-dropped DCs actually use the detonation depth setting, so perhaps it might be possible to have a DC rack drop a "bomb" instead... that's another big pain in my ass to work with, lol.
tater
I concur with the PIA!!! :lol: Let me know if you want any help with this DC modding you're doing. I've just released the sonar mod and plan to wait for responses before doing anymore with it now. Actually, I think I'll celebrate with a shower LOL.
I plan to work on SD Radar too. I have some ideas. BTW!! Do you know an "Average Altitude" for randomly generated aircraft in the sim?
Cheers!
Thank You. SD radar rarely picked up anything under 1000 feet so putting those in may require some scripting but that wouldn't be anything worse than time spent. And I have much more time than money. In fact, that's probably why I don't have more money :lol:!
AVGWarhawk
12-01-07, 05:19 PM
Follow link to Peto's released mod of this.
http://www.subsim.com/radioroom/showthread.php?p=707909&posted=1#post707909
Let him know what you think:up:
vBulletin® v3.8.11, Copyright ©2000-2025, vBulletin Solutions Inc.