Matrix Games Forums

Observing that in a test game of AE, with wholly stock values for submarines and AS vessels,
submarines seemed astonishingly vulnerable (as in to be attacked is to be hurt or sunk most of the time,
even with primitive AS assets in 1941), Mifune replied, partly in jest, "yes, submarine armor seems to
be thinner in AE than it was in WITP."

This caused me to remember that submarines, indeed, did have armor values in WITP - and the long
hunt in the database to remove them. But now I wonder - was the presence of that data because someone
who knew how the algorithms worked also knew it would help (and would compensate a bit for the overpowered
AS weapons)? If so, might it not be a good idea to think about that a little more?

My first instinct is that a uniform rating of 5 (mm?) is not a good way to model the differences between submarines.
My second is - why not use pressure hull? Reacting to that, I felt "what a lot of work to find that data! if we even
can find it?" But I looked at it anyway.

Note that in this system durability is related to depth - which is itself related to pressure hull thickness. There are
some inconsistencies in the data - but there seems to be a prettty good research base and it appears that durability
is test depth / 10. Probably what matters is crush depth - and here the problem is that officially different navies use
different standards to relate safe operating depth and crush depth. Notabily the British and Germans use a a more conservative
ratio than others do - so we might want to compensate in those cases. But ignoring that factor - I found a very nice way to
not need any research at all: durability divided by 3 is remarkably equal to hull thickness. Thus, a US late model submarine
rated at 30 durability (= 300 feet) would be rated at a 10, an earlier type rated at 25 (= 250 feet) would get an 8, a still earlier
type rated at 20 (= 200 feet) would get a 7 - and so on. A German submarine rated at 45 would get 15 - and that is exactly
the hull the Hedgehog was designed to crack if it hit it.

So the question then becomes - do submarines in this system survive AS attacks better if they have "deck" and "hull" armor?
Surely they will do better vs bombs (simulating aircraft dc in this sytem) and gunfire - but what about vs DC? It would be much
easier to make this change - addinv armor values - than to change the weapons data - and the relative value of the weapons data
research would be wholly preserved - while submarines might survive long enough to justify their cost (at least occasionally)

Experimenting I have learned a few things:

1) Modified ship data can be applied to ongoing games.

2) German submarines almost uniformly used 18.5 mm plating - the single exception was the Type VIIF - which was able to implement a proposal long considered but impossible on heavily armed boats - 22.5 mm. This permitted it to dive to 690 feet. In my simple system - if 22.5 is represented as 23 - and you triple it - you get a durability of 69 - remarkably close for a rule of thumb. Otherwise German submarines are the deepest diving in the war - and potentially in the game - if you allow the historical boats in theater. [I have them enter at Amsterdam Island - an island without any facilities - and forbid building any - it isn't suitable for an airfield or even a port. This is almost in the SW map corner - and then they must transit to a real base - or not make it - as the case may be. It should be a house rule not to invade this island - and it is so small - a mere dot - it may not be noticed.]

3) I learned some British subs were rated for diving only 200 feet - but were designed (and plated) for 300 feet - so these end up just like US subs - rated at durability = 30 and "armor" = 10.

4) DC attacks rarely - possibly never - achieved kills by hits like shells, mines or torpedoes do. Instead, cumulative damage is the usual (or possibly only) mechanism. The record for DC dropped in a single attack without killing the submarine is 678 - on a U boat which came back to report the details. This makes me think this idea may have more merit than I originally believed.

For the record, long ago, in WITP days, I railed against "armor ratings" I found on submarines in the database. If we were going to have sub armor, why not have it related to something - instead of a uniform value (of 5)? Well - here I have related it to something - hull plating - and I find it was not very hard to put that data in.

However, I found numerous errors in submarine details - speed - range - weapons - all sorts of errors. Some were understandable (RN subs were modified for Pacific duty, and could put fuel in ballast tanks, and so were not any longer using the original design range). Others more mysterous - torpedo tubes facing the wrong direction? Or entirely fictional? So reviewing sub data was worthwhile for more reasons than just to add the armor data to side and deck.

Addiion: The British fleet boat design - with a speed of 22.4 knots - was rated at 14. Maybe it was a typo during data entry?


I found also a German desing with conning tower armor - but it was not built. I think the conning tower should remain at zero armor. Some private mail indicates hope this mechanism helps subs survive. It may - we will see. But we still have to reduce the chances a single attack will actually do damage to something closer to history or even present day reality. We actually have a tested mechanism to use - from RHS WITP - but it takes a huge effort to do all the data entry on every escort on all sides. This essentially reduces every AS attack to one DC attack - or one ahead weapon attack plus one DC attack - per pass - when the detection routine starts over again. Each pattern has the same effect as any other pattern for the same size DC - but the accuracy is a function (square root) of number of DC in the pattern. Effect is also the square root of DC charge in pounds - so bigger is better but not proportionately - a lot like IRL. Hedgehog is different - a higher pk in the form of accuracy value - the pattern was certain to get a hit if the datum point was right - and might get two hits if it was perfectly centered on the target. But each projectile is tiny - so effect is very small per hit. However, it works regardless of depth - so we have to consider if that increases pk (it does - but probably we have to help code - unless special code exists for certain devices). Anyway - the weakest pattern of DC is 2 - and this has about a 2% pk (probably less, but close to right) - so as the pattern goes up - pk increases in the right range. If we have a pattern of 4 pk becomes 3% (1.414 times 2). A pattern of 8 becomes 4% (2 x 2). A pattern of 12 becomes 7% (2 timess 3.4). A pattern of 16 becomes 8% (2 x 4). Another approach might be to calculate the area of the pattern for each size - and multiply the pk of the smallest one by how many times it fits into the larger one? Testing indicates both approaches work fairly well and produce results in the right range - low % per pass. You get "shots" = to the total number of patterns you can drop. I am not sure "range" should be used - but IF it should because code likes a range - then it should be the range of throwers (about 45 apparently).

Our (missing) friend Ron Saueracker was likely the first to employ sub armor in vanilla WITP. The concept was to give those seemingly egg shell tender subs some kind of thickness to simulate how water makes ASW weapons have greater limitations.
Plankowners remember how easy sub killing was in the original version(s), and Ron was really a sub nut.
I was quick to accept his theories.

Every 33 feet is an additional 14.7 lbs per square inch, and somehow it was thought the deeper the sub, the less effective the ASW weapon might be.

I'm not sure if thermals have been simulated in any versions, but if not, I'm sure they could be, even in an abstract way?

The basic problem with ASW is there is no real ASW in the design of the game system. Weapons are modeled on surface weapons - and there is no model for sound detection at all - no hydrophones - no sonar - no sonabouys. [In a way we were better off in WWI - when you could pound the side of a sub chaser with a hammer and time the return echo with a stopwatch - at least you had SOME sound detection!] To the extent ASW is in the game, it is because of DC, and because of possible AS code associated with PC, DD and DE classifications - which seem to be better than putting the same weapons, say, on a CL or on an AK.

Absent some sort of sound detection per se, it is hard to see how to add thermals to the game? It isn't even easy to see how to put DC on aircraft - although I always do that and keep trying to measure the effects - in the hope of showing some. But there may be indirect approaches. Thus - I was able to add a form of ESM to a mod once - basing detection range on the passive rather than active range of a radar set - which in fact was more often used in the passive role by 1944. Here I have experimented with using the code of AS ship classifications as a way to influence "sonar suite" - if it is any good I called it a DE - and if it was minimal I called it a PC - in order to get its AS capability better than being a PB or AMC or whatever. Here I need to learn to understand what new classifications (e.g. Escort) mean before I dare attempt such a thing. But MAYBE we can get some sense of sonar quality by indirect means of that sort.

reviewing submarine class data to deal with adding "armor" (pressure hull data)
I noted errors of in the stock data set of several kinds -
so I did a comprehensive review - and except for the Russians (not yet done)
found virtually all class records had errors, most had significant errors.

The Gato class of almost standard fleet submarines - that is
30 class records of various kinds of Gato/Balao/Tench class submarines
is notable in that ALL of them are rated for 30 torpedoes (2 tubes times 3 shots five times -
three forward - two aft). 30 rounds! Never mind the Gato and Balao really had 24
and the Tench (rarely documented, but see US Submarines Through 1945, USNI)
had 28. On the other hand, the Balao and Tench could dive much deeper than Balao,
but this is absent in the data: officially 400 feet vice 300 feet. In fact, Tench is
even deeper - the exact figure is not available although that it was built stronger in several
respects is stated. I rated it at 420 for convenience as it works out to the minimum incriment
in my protection scheme - and that is likely conservative. Finally - for reasons unclear,
the Gato group is all rated 1 knot slower than their official top speed. These deep diving
submarines (Balao and Tench) are actually better than the Japanese Guppy design (360 feet) or the best of
Japanese fleet boats (330 feet) - although you cannot tell it from the 300 feet rating of stock.

This is typical. Many Allied submarines were given too many torpedoes and/or tubes (or wrong
tube facings). Yet
some have significantly too little range. Even when range is correct, the maximum fuel load
is too small - particularly with the cases where fuel was stored in ballast tanks for long cruises.
US subs available early in WWII typically cruise 18,000 to 21,000 miles - at precisely the time
the US has few forward bases and can benefit from such ranges. The US invested heavily in
development of diesel engines and electric motors (related to RR engines) - and achieved more
range efficiency than any other navy - although the Germans finally managed a diesel electric
plant for a Type IX for Indian Ocean duty that was very impressive - it was exceptional.
For reasons unclear, the USN gets external torpedo tubes described as such, but the Dutch do not.
Yet there is no difference between the torpedoes other than their name - to let players know they
are external. It appears that the concept of including the external tubes is valid. But the data is
not consistently used: sometimes they face the wrong direction, or are mounted in subs that did not
actually get them. Other issues relate to speed. Sometimes subs are given too little speed, sometimes
significantly so. Or depth. After review, on balance, Allied submarines are better off, even without
extra torpedoes.

One fine aspect of the stock data is that the names of submarines, and the number in a class in the
theater, is perfect, or virtually so. Less perfect is where the submarines appear. Japanese submarines
may appear at Haiphong!!! Or Port Arthur. No clue why? And on their date of commissioning. But
Allied subs usually appear in a logical place after time to work up - e.g. Panama. It is a vast data set
and to get the names and numbers and dates right is a very impressive achievement.

The test data on armor on submarines is too small to be conclusive yet. It does not appear to hurt anything
and over time I expect it to result in slightly better survivability. At the same time, I reverted to the original
WITP system to the extent possible re torpedo salvos - favoring the largest possible value consistent with
correct torpedo inventory. Largo salvos ought to result in shorter patrols, which in turn ought to result in
slightly longer service life, statistically. Then too, the new AE code seems superb at "captain elects not to
engage this target" or at selecting the smaller torpedo salvo if available, when appropriate - and larger
salvos should result in fewer cases of firing two at a major target - which seems realistic.

The final dimension of making submarines last longer will require changing the ASW offensive weapons -
a much more time consuming effort. But a review of escorts is probably a good idea anyway - to insure
consistent data standards - and no fictional things like extra torpedoes as we found in the US fleet boats.

Adding Russian submarines to the reviewed list only confirmed the previous pattern:
data needed review in any case. Some of it was very close: depth was uniformly
rated at 30 - but in one case really was 33 and in the other two cases should have been 29.
33 is very good in relative terms - only the best Japanese boats get that - and 29 is quite
competative with the many cases of 24 to 30 in most classes. I quibble with full speeds
not being rated to the nearest knot and main guns not getting the standard 18 shots.
But more significant - why too many mines on the minelayer? Should be 7 shots (=
14 mines) vice 10 shots (20 mines). Why does a class with 10 torpedoes total get 12?
Why is a forward mounting shown aft? And most of all - why cannot we use correct range
and fuel data? Looks like total guessing on fuel - vastly too much is required. Finally -
here all the subs were listed with submerged displacement. Stock uses two systems -
surface for some - submerged for others - but mostly surfaced displacement. I think
that makes sense. In game terms, subs are considered surfaced most of the time - and
fight surface vessels and planes in that state very often. Also - operationally - subs of that
era without snorkels usually cruise on the surface. Finally - the extra displacement when
submerged is already filled with water - and should not be considered in damage control
calculations when the sub will sink (it is "sunk" already being underwaster - by exactly
the amount of the extra displacement). The Russian subs got rated at 10 mm armor -
1/3 of 29 - and 11 mm - 1/3 of 33 - which isn't bad.

Now to see how well all the subs of both sides do in a statistically significant number of
engagements? Time will tell.

"For a number to make sense in the game, someone had to deliberately program code to do that. There are too many significant numbers for us to reasonably expect that. I'm finally beginning to appreciate that it's just a game, and it does what it does. The historian in me is sad, but that's the way it is."

After a tiny number of barely effective ASW battles, we finally have an I boat systematically attacked multiple times by
two ASW task forces - and repeatedly score hits. Each hit caused 1 point of system damage, and many of them caused
a floation or engineering hit. But unlike the attacks earlier in the same game - sans adding armor to the subs - no
catastrophic damage. It is not conclusive - we need more than 30 such attacks to get a vague sense of the statistics -
but it is an indicator that the armor may indeed mitigate the attacks and make for a cumulative damage model rather than
a hit to kill model - which is the design intent.