Japanese ASW

[inqistor]

If all ASW upgrades are simply adding more DC mounts, it seems easiest way of solving it, will be to add different devices. Double, triple, or quadruple launchers.
If DD have, for example, 9 DCs it should be modified to have, like 3 TRIPLE launchers (DC device, with higher accuracy, but similar damage). That could solve the problem.

[JWE]

ORIGINAL: inqistor
If all ASW upgrades are simply adding more DC mounts, it seems easiest way of solving it, will be to add different devices. Double, triple, or quadruple launchers.
If DD have, for example, 9 DCs it should be modified to have, like 3 TRIPLE launchers (DC device, with higher accuracy, but similar damage). That could solve the problem.

This is not the way the algorithm works. It is not linear. If you don't understand how it works and what the implications are to making a unit change, please refrain from making uninformed comments. Thank you.

[witpqs]

ORIGINAL: vettim89

My initial testing suggests that yes that is true

ORIGINAL: witpqs

ORIGINAL: vettim89

The whole problems with mount numbers and ASW is self compounding. Ships with high ASW strengths (which is saying nothing more than ships with a lot of ASW mounts) have an inheritantly higher rate of successfully detecting and engaging subs. Then when the do engage they have vastly higher chances of securing a hit based on number of evolutions through the ASW routine because each moutn fires separately. In RL ASW mounts fired as a group as patterns. Higher number of mounts allowed larger patterns. Larger patterns would more likely cause damage to the sub but still only one DC was likely to actually get a "hit". So higher percentage to hit but no greater effect. ElCid did some opening work on this in RHS and I am trying to resurrect it. Alas there are only so many hours in a day.

Are you certain about the part I marked in bold? Do you know for a fact that such is in the code?

My initial testing suggests that yes that is true

The reason I asked is that going way back to developer comments my recollection is that a) the ASW rating is merely the sum of the number of ASW devices and b) it is used for nothing other than displaying a number in the ship display. Further, the chance of detecting and engaging a target took into account 1) nationality, 2) year, 3) ship type, 4) crew experience, 5) leader stats, 6) other relevant things like detection levels of both sides, and I don't know what else.

Ship type meaning that a DE gets a higher chance than a DD, which gets a higher chance than an APD, and so on.

As far as hits and kills, I'm sure some of the items above play a roll, plus add in the weapons own stats (effect or whatever).

I was wondering if you had gotten any indication from a developer on the ASW Rating (which is just number of devices) affecting detection and engagement.

[Bullwinkle58]

ORIGINAL: witpqs
Are you certain about the part I marked in bold? Do you know for a fact that such is in the code?

After sniffing more I found the thread I was remembering.

tm.asp?m=2452466&mpage=19&key=&#2705803

It's very long and winding. JWE's comments on ASW are on Page 19 Post from 1/27/2011.

The post in particular:

"quote:

ORIGINAL: FatR
1)Japanese ASW. Holy ****. You can't believe it until you see it. The best result I noticed was kaibokan No.205 sinking one sub and near-sinking (damage into 80s) another in one phase. In the deep water She ran out of DCs for the main array too, otherwise it would have been two clear sinkings in one phase. And that with crew EXP less than 50. While American boats are quite accurate with their attacks, they don't seem to have much of advantage in avoiding detection, compared to early war. And if you get spotted by one of these E-class monsters, you are toast.
Now, I'm all for giving JFBs the ability to conduct an effective ASW campaign, if they invest appropriate effort and approach the whole thing meticulously. But "Just send your E-class ships to sea!" strikes me as an unsatisfyingly simplistic solution.

What can we do about it? We don't know how chances for detecting a sub are determined, so I propose reducing the accuracy of Type 2 DCs to 8 overall (other ships that carry them aren't as lethal as kaibokans, but still can sink subs even in deep water more often that I like) and introducing a subset of Type 2 DCs specifically for E-class escorts, that has accuracy further reduced to 7, to reflect that simple increase in the number of devices does not give an arithmetic increase in effectiveness. I mean, air combat suffers from this problem too, for certain planes, but at least there the problem doesn't break the whole equation. /quote


JWE replies:

"There's a three part answer to this. Unfortunately, all the parts have to play together, so it's not simple, but it is doable with some thought. It's a combination of a math thing and an engine/data interaction thing.

1) math thing - yes, lots of launchers show a power law result in the chance to hit. The more launchers, the more times the hit routine is called, so the waaaay better chance that one of them will get a hit. In DaBabes, we reduced the # of launchers from 12 to 6 for the late war C and D 'E'-types. Indeed, they had 12 launchers historically, and they were pretty bad juju, but as far as we can determine, they popped those off in salvos of 6 (right half, or left half, or front half-both sides, or back half-both sides). So we cut the launchers in half, but doubled the ammo. So they get exactly what they had in terms of total DCs, but the sequential hit probability is now based on a smaller number - makes it better, lots better, but still doesn't quite get to the brass ring.

2) engine/data interaction thing - you may have noticed the 'messages' that say 'so-and-so sub dives deep and escapes'. Ok, cool; it happened. But there is an interaction between the "depth" of a DC (the DC Range parameter) and the "depth" a sub can go (10x its Durability parameter). Now, in stock, later war Japanese DCs (like the Type-2) have "Depths" of 475', while the best US subs can only dive deep to 300'. In DaBabes, the later war US subs can dive deep to 410', while the Type-2 DC bottoms out at 357'. That doesn't mean those subs are immune, it just means that if they can avoid the first stonk, they have a better chance of avoiding the rest. They are still vulnerable to the first stonk, which is why we decreased the launcher numbers to compensate.

3) accuracy thing - the simplest, but also the hardest to dink with. Accuracy is a small number, so when you roll dice against it the values converge rather quickly. When you divide by 2 and take an integer, there's not much difference between 2 and 5; and 1 (Exp)4 is still 1. Probably best to leave accuracy alone, in this case.

Now, I understand that what I'm saying is a bit vague, and does not answer the question specifically. I'm skating on some thin ice about not revealing algorithms here, but the hints are a synopsis of what a good math person could eventually figure out, so why not provide a short cut.

Hope these comments are helpful. Ciao. "

----
This post suggests a couple of crude ways for us non-math majors to fool the unseen routine. FWIW, late war USN fleet subs were not restricted to 300 feet, or even 410 feet, although the latter is close to the warranty papers that came with the boat. Not a skipper in the USN who wouldn't have taken a Balao-class to 550-feet to live to fight another day, and have had good confidence that Rosie did her job back home.

I really like the "survive the first stonk" midset because that's what often happened. If the boat got below 250 feet or so with some decent way on the first salvo ensonified the water and the sub could change course, go silent, go deeper, and had an excellent chance to be outside the PK circle on salvo two. The IJN usualy got one kick at the cat, then the boat's deck log reported "distant charges drawing astern" as she moved away.

Given these broad hints from JWE, I may try to cobble this into my own tweaks for my next game versus the AI.

Anyway, I think it's important to try to recall ground already covered. Getting harder as time goes on and these debates get deeper and broader with playing experience.

[crsutton]

ORIGINAL: vettim89

ORIGINAL: Bullwinkle58

ORIGINAL: bradfordkay

I know that the number of launchers is the key variable, but we can't change that as the ships historically had them. The kicker is the difference in ASW electronics and the crew's experience with them. The electronics are not extensively modeled in the game, but I think that there is some sort of positive modifier for allied ASW to simulate this (somebody can verify this?). Perhaps a similar negative modifier can be applied to the IJN and the allied modifier reduced somewhat? 

I hope an editor guru can answer this.

I'm also interested to hear from anyone, in a general way, who knows the workings of Da Babes mods, what JWE was speaking of when he said Japanese ASW had been dealt with/modified/toned down, etc. in those mods.

The whole problems with mount numbers and ASW is self compounding. Ships with high ASW strengths (which is saying nothing more than ships with a lot of ASW mounts) have an inheritantly higher rate of successfully detecting and engaging subs. Then when the do engage they have vastly higher chances of securing a hit based on number of evolutions through the ASW routine because each moutn fires separately. In RL ASW mounts fired as a group as patterns. Higher number of mounts allowed larger patterns. Larger patterns would more likely cause damage to the sub but still only one DC was likely to actually get a "hit". So higher percentage to hit but no greater effect. ElCid did some opening work on this in RHS and I am trying to resurrect it. Alas there are only so many hours in a day.

Yes, I don't really know if the solution is in the mounts and attack resolution. The solution might be better dealt with by altering the dectection. Simply put, Japanese escorts and Air should not detect Allied subs much of the time. This would reflect both Allied air search radar which many times alerted them to patroling planes before being spotted and attacked by the aircraft and the fact that Japanese escorts with inferior radar and sonar simply failed to find Allied subs after being attacked. (Or did get get a good enough fix to launch an accurate salvo).

I am not an computer programmer so have to defer to those who actually put the hours into fixing the game for us. (And I can't be more thankful for the efforts) But a few suggestions that may or may not be feasable.

1. Lower the maximium training level that a Japanese pilot can obtain in ASW-to say 30.
2. Vastly lower the exp level for non-DD type Japanese warships.
3. Put a maximum CAP on detection levels that can be obtained on Allied subs-especially when radar equipped. It pisses me off that my subs with air search radar are at a 10/10 detection level all the time.
4. Tweak the targeting priority of Allies subs to target merchies over escorts.
5. Find a way to allow Allied subs to attack on the surface at night. This is the real advantage that radar equipped Allied subs had over Japanese convoys but it just is a no show in the game.
6. Find a way to allow more multiple attacks by Allies subs vs a convoy especially if the sub goes undetected during the initial attack.

I have to admit. I love the sub wars in AE over just about every other aspect in the game. However, I embarked on my two campaigns over a year ago in the expection that the sub wars would mirror the actual course of events. (And this is the reason that I so readily agreed to impliment the historical torpedo rule) Now that I am over 600 turns into a campaign, I have to admit that I am growing more and more frustrated with the whole sub thing. In September 1943 I am meeting up with Japanese ships all the time and really not sinking any in numbers. Now I know that I will get some decent torpedoes around Jan 44 only to realize that is when the Japanese players start to get some real nasty sub killing escorts. Oh, what fun.

I am willing to concede a lot of balance issues to allow my Japanese opponents to do better and enjoy the game but I think the game fails totally as a simulation if the Japanese player is not sweating out the loss of valuable merchant ships as the war progresses. As I see it now my whole sub campaign really only provides a mild nuisance, to the point where I would say that if I withdrew my whole sub fleet tomorrow, it would really have no impact on the game one way or another.

I hear the same lame argument that we all should be happy that the Japanese players are allowed to overcome the mistakes of a bad Japanese system and do historically better, however nobody is suggesting that I as the Allied player should be allowed to pay some PPs to say fix my topedoes in 1942 and do better than the Allies did historically. I am learning to live with a lot of quirky things about the game but right now, I would have to say that the sub thing is the most dissapointing discovery for me. I just expected a lot more out of my Allied subs.

[witpqs]

Thanks Moose. I saw that post and it's good. But in terms of vettim's post John's is only dealing with the weapons, not the detection and engagement that both have to happen before the weapons are used.

[Bullwinkle58]

ORIGINAL: crsutton


Yes, I don't really know if the solution is in the mounts and attack resolution. The solution might be better dealt with by altering the dectection. Simply put, Japanese escorts and Air should not detect Allied subs much of the time. This would reflect both Allied air search radar which many times alerted them to patroling planes before being spotted and attacked by the aircraft and the fact that Japanese escorts with inferior radar and sonar simply failed to find Allied subs after being attacked. (Or did get get a good enough fix to launch an accurate salvo).

I am not an computer programmer so have to defer to those who actually put the hours into fixing the game for us. (And I can't be more thankful for the efforts) But a few suggestions that may or may not be feasable.

1. Lower the maximium training level that a Japanese pilot can obtain in ASW-to say 30.
2. Vastly lower the exp level for non-DD type Japanese warships.
3. Put a maximum CAP on detection levels that can be obtained on Allied subs-especially when radar equipped. It pisses me off that my subs with air search radar are at a 10/10 detection level all the time.
4. Tweak the targeting priority of Allies subs to target merchies over escorts.
5. Find a way to allow Allied subs to attack on the surface at night. This is the real advantage that radar equipped Allied subs had over Japanese convoys but it just is a no show in the game.
6. Find a way to allow more multiple attacks by Allies subs vs a convoy especially if the sub goes undetected during the initial attack.

I'm not much of a programmer either, but if you look at the JWE post I quoted while you were typing yours I think you'll see ways to big-time tweak IJN escorts by using data/the editor and not getting into the exec file, which isn't going to happen in any major way now. The suggestions you make above look to me like code changes. Detection levels and targetting priorities are, and they feed (probably) in all directions at once.

My personal #1 wish, if exec changes were in order, would be for #6. That alone would upset the IJN merchant marine destruction rate by a factor of at least x2. But that might really mess with the economy. Reading in some awe the threads on Merchant Ships and The Japanese Economy in the modding sub-forum I'm getting a much better idea of how intertwined this beast really is. And taking out merchants at a much higher rate than now would seriously mess with things.

[Bullwinkle58]

ORIGINAL: witpqs

Thanks Moose. I saw that post and it's good. But in terms of vettim's post John's is only dealing with the weapons, not the detection and engagement that both have to happen before the weapons are used.

Right, because I suspect he better understands the politics with Matrix about the liklihood of code changes, and is suggesting work-arounds in data that might approximate sorta the same level of results, even if you have to go through Hartsfield International to get there.

[witpqs]

ORIGINAL: Bullwinkle58

... even if you have to go through Hartsfield International to get there.

Better than Detroit! [:D]

[oldman45]

Unfortunately I bet detection and engagement are code issues. Looks like the babes mod is the best base line to start with.

[vettim89]

ORIGINAL: witpqs

Thanks Moose. I saw that post and it's good. But in terms of vettim's post John's is only dealing with the weapons, not the detection and engagement that both have to happen before the weapons are used.

Please keep in mind my testing is in its earliest stages. I can tell you that early war USN DD's with ASW of 2 almost never found IJN subs even those right underneath them. There was a pretty marked increase in both # of attacks and DL of the subs after the first upgrade when ASW went to 6

[inqistor]

ORIGINAL: vettim89

ORIGINAL: witpqs
Thanks Moose. I saw that post and it's good. But in terms of vettim's post John's is only dealing with the weapons, not the detection and engagement that both have to happen before the weapons are used.

Please keep in mind my testing is in its earliest stages. I can tell you that early war USN DD's with ASW of 2 almost never found IJN subs even those right underneath them. There was a pretty marked increase in both # of attacks and DL of the subs after the first upgrade when ASW went to 6

But, what was their experience level? Have you increased it at the beginning?

[Nomad]

ORIGINAL: vettim89

ORIGINAL: witpqs

Thanks Moose. I saw that post and it's good. But in terms of vettim's post John's is only dealing with the weapons, not the detection and engagement that both have to happen before the weapons are used.

Please keep in mind my testing is in its earliest stages. I can tell you that early war USN DD's with ASW of 2 almost never found IJN subs even those right underneath them. There was a pretty marked increase in both # of attacks and DL of the subs after the first upgrade when ASW went to 6

Don't the DDs get SS radar in the first upgrade? That might be a factor.

[vettim89]

ORIGINAL: Nomad

ORIGINAL: vettim89

ORIGINAL: witpqs

Thanks Moose. I saw that post and it's good. But in terms of vettim's post John's is only dealing with the weapons, not the detection and engagement that both have to happen before the weapons are used.

Please keep in mind my testing is in its earliest stages. I can tell you that early war USN DD's with ASW of 2 almost never found IJN subs even those right underneath them. There was a pretty marked increase in both # of attacks and DL of the subs after the first upgrade when ASW went to 6

Don't the DDs get SS radar in the first upgrade? That might be a factor.

Ken, that was my next test. I had run the baseline 12/41 fit with 2 DC racks and no radar (ASW: 2). I then repeated the test with the first refit minus the radar. My next evolution will be to do the first refit with the radar.

Inquisitor, the tests are being run in a sand box scenario with only the DD's and subs in it. I am running it as a 12/41 test so far with early war exp levels for the US DD's and IJN subs. I cloned the ships and commanders so each unit is identical to all others.

Of note through the first evolution of tests not a single IJN sub was still detected when the turn was over. This increased when the ASW level increased to 6 but still was only a few boats.

[Stvitus2002]

they had 12 launchers historically

In the current official patch we now have split-salvo sub torpedo
launches, right? How much code work would be involved to get a split
k-gun launcher? Dependent on crew & captain experience?




WO 0/0

[vettim89]

ORIGINAL: Warrant officer 0/0

they had 12 launchers historically

In the current official patch we now have split-salvo sub torpedo
launches, right? How much code work would be involved to get a split
k-gun launcher? Dependent on crew & captain experience?

WO 0/0

The split launchers for TT was attached to a code change IIRC. With the current system, splitting K-Guns would make it worse not better as the system would count both halves as another launcher and increase ASW value even more

[PresterJohn001]

Japans got to have something to look forward too in 1944 [8D]

[vettim89]

ORIGINAL: PresterJohn

Japans got to have something to look forward too in 1944 [8D]

Keep in mind that the changes would affect USN DE which are almost as deadly as Super "E". It will cut both ways

[Chickenboy]

ORIGINAL: vettim89

ORIGINAL: PresterJohn

Japans got to have something to look forward too in 1944 [8D]

Keep in mind that the changes would affect USN DE which are almost as deadly as Super "E". It will cut both ways

Well then, won't we have an argument from half the players out there pointing out that late war USN ASW is a shadow of itself IRL? Please see JWE's notes earlier in this thread about unintended consequences from 'nerfing', 'denerfing' and finally 'renerfing' these various combat systems.

[vettim89]

ORIGINAL: Chickenboy

ORIGINAL: vettim89

ORIGINAL: PresterJohn

Japans got to have something to look forward too in 1944 [8D]

Keep in mind that the changes would affect USN DE which are almost as deadly as Super "E". It will cut both ways

Well then, won't we have an argument from half the players out there pointing out that late war USN ASW is a shadow of itself IRL? Please see JWE's notes earlier in this thread about unintended consequences from 'nerfing', 'denerfing' and finally 'renerfing' these various combat systems.

Which is exactly why I have not proposed a single change/fix. I want to collect a mound of data before I even begin to look at any modifications. Just from a "feel" standpoint, ASW seems accurate in the early/mid war periods but seems over the top in late war. Just looking right now to see what various setting yield in the system