AArm reflection

[OlegHasky]

The A-arm data of Arm Coys is creating questions..

Mark V -G type, armed with its famous 7.5cm gun, protected by 21-124mm seriously sloped armour
looks to be outclassed by US standards majroly..

14x Panther G type in a Coy providing
123A-Pers
135 A -Arm,
and Armour value 216

US Coy in the other way, with 9x 75 M4, 3x 76 M4 3x "Jumbos" , and M4 105mm providing
a power of
170 A-Pers,
249 Armour
and a monstrous 359 A-Armor value

Now,
A-Pers 123-170 capacity is well clear, ( M4 105mm How´ in a Coy making the proper diffrence. ( open to give even more for the US - when reflecting M4s superior manouverability)

Armour 216-249 is peaked by "Jumbos" and it is also clear and well designed

but I am finding very hard to explain A-Armour 135-359 difference.
Panthers did indeed became a ´diffrent effort´ driving-off from France open-country in to the tight Arden reality. But regarding the power of its AP capacity it is just outclassed by ..Shermans. It is not 135-200 (with all the 76mm´s of M4´s and Jumbo´s )
but a stunning 135-359. it is more than plus 150%

The one explanation I have in mind is the historical stamp on the dificulty of using Panther long range guns in Ardens. But the stamp schould be the same for those long range 76´s as well (?).

Anyway.. This is the question of how the game is designed.

It is designed to just role-play the theater?.. where Panthers by game desing are convicted to be very less usefull.
Or we (as players) have the open way to change the books to some point (as we are doing it just playing out the scenarios 100´s of times trough out the players in our own way)
it is colidating a bit.

I meen here - our hands are in some way tight. And this "future based" reflection is laying limitations.
While correct historical outcomes are touched by just playing out the battles. The tank sector seem to be frozen by hardly writen statistics.
Its a real time game,
And there are possibilities to use armour in the way that could change the books (but not without those limits)

For example: In my online game , my opponent (as US) used a position to support his troops (fighting with my forces for the village) by long range 1,5km tank fire, and it was very influentive

You could be wanted to do the same with German Panthers, but the effort will be straightly damaged by the reflection of the "future" based design - wich is giving the player the Panther abilities from "after the battle" not from the start of the battle (when it was a still a tabnk to be feared.

I think this is a "question" I am still not sure wich way of reflecting is better. One thing I know is that such reflection of AArm is limiting the options for tactics seriously.


----------
hope that Ive made myself clear enough using english[:-]

[Lieste]

AArm is based on ROF x Penetration - so a lot of ineffective rapid fire weapons will have a high AArm value, but will cause no casualties. OTOH, a single perfectly accurate and effective weapon that obtains a kill each time it fired may well have a far lower AArm value.

I think it matters where a unit can be driven away by fire - but the important comparison is accuracy at range, and penetration/armour. A high ROF provides additional chances if the first shots are unlucky, but cannot transform the performance of a totally inadequate weapon.

The German tanks do seem to suffer a little from their high/low armour split - IRL they had some areas which were practically immune from penetration, offset by some weaker areas, in game the whole frontal armour is vulnerable to a wide range of weapons. American armour seems to have more homogeneous armour so isn't as badly affected by this. (Alternatively the thinner areas of the German vehicles are not vulnerable to weapons that could penetrate, due to the averaging using thicker armour elsewhere).

Using ingame data - from the in-demo pages, the Panther can penetrate the Sherman front from 2200m (M4/M4-105) down to 1800m (M4 76W), and not the Jumbo at any range. The Sherman with the 105mm or 75mm cannot penetrate the Panther from the front, and the 76mm only from within 700m (roughly).

Flanking fire is effective at longer ranges - the Panther will penetrate all Sherman marques except the Jumbo at all ranges - Jumbo from 1900m or closer only.
The Shermans can hurt the Panther at long ranges in flanking fires too - the 76mm at ranges to 1700m, the 105mm at ranges to 1000m (max range DF) and the 75mm to around 900m.

[OlegHasky]

ORIGINAL: Lieste

AArm is based on ROF x Penetration - so a lot of ineffective rapid fire weapons will have a high AArm value, but will cause no casualties. OTOH, a single perfectly accurate and effective weapon that obtains a kill each time it fired may well have a far lower AArm value.

Ok Lieste, thx, Youve got a point here.
(I saw You were involved in the most intresting discusion regarding armour in PanzerCommand. So i take You as an expert on the matter[;)]).

If that (what mentioned above) is indeed the case, -I am glad that those statistics doesnt reflect the true nature. And that the game-engine relying majorly on the properties writen deep down below. So the actuall AArm value visible is just a mechanical summary of secondary factors.

I do truly hope , that what youve written is indeed the case.
Because the flashlight of AArm stats pointing in your face in the game main screen can be very alerting. Its desing very missleading.

[Lieste]

The high AArm of US vehicles comes from the 50 cal. It is meaningless for tank vs tank action, but this over-inflation does affect exit conditions - it might also affect morale/decision making.

[OlegHasky]

Yes, very true.

But if so. I think the AArm rate schould be a secondary statistic, not a flag one in your face when not really influenting the true Arm clash.
Your hypothese fits very well. And as stated - I would be pleased if this is the case.
But let us wait for ´the one´to confirm it - same way admit, that AArm flashlight is missleading, when it comes to Arm vs Arm

if.. ´he´ will honour us with his presance that is. because I am far away from kissing the sapphire ring.

[Lieste]

Looking at this again - I've made a variety of 10s, with differing equipment. (note these values will differ slightly from stock ones, due to modified penetration values and armour - most data is not dissimilar to the originals, penetration is usually a bit lower, but armour values tend to be lower as well-Cromwell was however given applique, and some of the inferior German AT developments eg the PzIV70(A) got a modest improvement in average armour, though it is still far worse than the 'proper' version PzIV70(V)).

Ranging from a simple Jeep with a 50cal, to the Sherman Firefly or Jumbo, and from the StuG to the JTiger.

The lowest observed AArm value is for the 10 stuarts:
Veh: Apers-AArm-Arm
M5A1 Stuart: 50-70-52
Jeep 50cal: 17-77-0
M3A1 50cal: 17-77-16
Cromwell: 64-84-97
Chaffee: 64-88-63
Sherman Firefly: 47-92-106
Sherman 105: 93-153-0 (which is obviously wrong!)
Sherman 75: 81-165-95
Sherman Jumbo: 81-165-180
Sherman 76: 80-167-111

StuG IIIG: 46-92-85
JagdTiger: 165-95-260
PzIV70(A&V mixed): 50-96-123
PzIV70(V): 50-96-136
JagdPanther: 51-96-158
Panther: 64-96-131
Tiger I: 69-96-128
Tiger II: 69-96-227

From observing close range contacts, it appears that relative AArm firepower is important in determining retreats. It also has a direct effect on exit requirements for a force - that a Jeep with a 50cal HMG can obtain more exit points than a proper tank (albeit a light one) is an obvious distortion.

The firepower seems to use shell weight (capped at a low level ~1kg?) rate of fire and accuracy to determine the value - penetration may play a factor, but it is also capped well below the values obtained by the heavy weapons (as the identical results from the 75mm L70, 88mm L56 and 88mm L71 suggest). The lower result for the Jagdtiger seems to come from a lower initial accuracy.
Maximum range either has no effect or is capped to a short range also, as the 128mm outranges the other weapons by 1km.

That a standard 75mm Sherman is in any way superior in AArm terms to the Firefly is odd to say the least, but for some 'game' purposes this is strongly true. Also all medium Allied armour is either 'equivalent to' or vastly superior in morale and planning terms to sometimes much more powerful systems on the German force - as demonstrated by kill ratios in actual play - at least until close contact retreat/routs the German forces away without casualties prior to mass surrendering (which is understood to be a bug due to be resolved shortly).

[OlegHasky]

Youre an analysing monster Lieste[;)][:)], let us not take it too sweety, but a real incredible one. Consumes the whole topic in a very lucid fashion.
Surely a base to refer to when it comes for the sector.

I think Pthr should reconsider releasing a whole position, with detailed game co-relations, served just like this.

[Arjuna]

I will look into this once I have the game compiling again under VS2010. Then I will look at some saved games and see what factors are really at play. From that analysis we'll decide what if anything needs to be changed. Thanks for your patience.

[Arjuna]

I have had a look at this issue. Yes the AArmFP values of the 50 cals are providing a high AArmFP value overall for Allied tank units. However, this has no real bearing on tank v tank combat. The anti-armour firepower value is just an indication of capability. The actual capability is determined on a case by case basis such that in case A a 50 cal may be very effective against an enemy halftrack but in case B against a Panther it's ineffective.

It doesn't adversely affect the exit conditions as these are based on a percentage of a side's total Armour value, not the total AArmFP value.

I agree it can be a little misleading. To make it less so we could ignore weapons whose primary role was not anti-armour. But this would then mean that many units would end up showing no anti-armour at all and that would be misleading too. However, this may be the lesser of two evils.

Another option would be to apply a reduction modifier if the weapon's primary role was not anti-armour. Eg if not anti-armopur then aarmFP = wpn->GetAArmFP() * 0.25. However, this may have an adverse effect on 105mm GunHow Shermans. Always trade-offs! [:)]

What do you think?

[Lieste]

If you combine weapons using (W1^n+W2^n+W3^n)^(1/n)  rather than as a simple sum then you will not see very large distortions caused by additional (secondary) weapons.

I think that the 'effectiveness' of a weapon should be related more to the actual observable in-game effectiveness - here neither AAAW or LAW are as universally useful as their AArm FP suggest due to short (or very short) effective ranges - the AAAW also suffer from poor penetration, while the LAW have low accuracy.

Because of this, I find the near-parity of a 50 cal to the significantly more powerful ATG to be misleading - rate of fire cannot really compensate for low effectiveness... that even the KwK 42, KwK 43 and PjK 80 have AArm values only 1/3 more shows that effectiveness is not being clearly displayed.

I'd suggest that improving the method of calculation could provide comparable values for medium calibre weapons of normal performance (ie the 75mm mid velocity weapons), while modestly enhancing the presented values for high-velocity AT weapons, and limiting the value shown for LAW and AAAW types (or obsolete AT guns).

It wouldn't change how the engine handles firing within a firing step, but it would present more useful information to the player (and possibly AI?).

Basically my thinking is that a suitable combination of range, penetration, accuracy, rof and weight of shot could be produced, with the individual terms raised to exponents that gave a useful 'single value' effectiveness.

Poor quality AArm weapons would not have more than 1/2-1/3 of a 'normal' medium calibre gun, and the best weapons would show individual results around twice this value, for example.

Again, combining the firepower (or Armour) into the unit total, rather than simply summing the weapons would better present the quality of the equipment, and the likely results of an open-field battle, but this is less vital than not simply adding the secondary weapon to the tank.

(Eg if a M36 individually scores 32, and a Sherman 75 only 20, you will need more Shermans to "match" the AArm value of the 90mm guns... but to "match" a 5 vehicle TD platoon, you might need more than the 8 vehicles a simple sum would suggest - perhaps 13 or 14... of course it may be that tactical circumstances mean that 5 Shermans would be adequate to do the task, and the excess firepower of 5 M36 is wasted... or that no matter how many 75mm guns you have they will be inadequate...)

(eg from the test data a few posts up...)
50 cal 7.7 AArm
75mm Gun 8.8 AArm (using Chaffee Data)

Simply adding them - 75mm + 50 cal ~ 16.5 (which matches the Sherman 75mm data)
combining using 2, 0.5 as exponent pair:
(7.7^2+8.8^2)^0.5 ~ 11.7 (or the 50cal is reduced to ~ 50% effectiveness when combined with the main gun)

I feel 7.7 for the 50 cal is overstating it's usefulness even when mounted in isolation though - and a tentative estimate of 5 which I derive from the parameters in the Estab file using a sum of the products of accuracy, range, penetration, rof and shell weight (each raised to an arbitrary exponent) gives:

Direct sum:
8.8+5 = 13.8
Combination:
(8.8^2+5^2)^.5 ~ 10.1 (or the 50 cal in combination adds only 26% of the nominal value).

[Arjuna]

Lieste,

Thanks for your reply. I acknowledge that your formula would tend to reduce the importance of weapons with a lower AArmFP value. In fact we used to use virtually the same formula ourselves but discarded it for speed/performance reasons as square/cube etc root functions are very expensive and this is a function that will be called often. Having said that machine performance has increased significantly since the days when we designed this bit of the code, so maybe it or something similar could be used now.

When we calc the AArmFP value for a weapon we already factor in rate of fire, accuracy and a normalised penetration value. The latter is derived by the following formula:
normalisedPenetration = ( armourPenetration / kDefaultMkIVFrontalArmour ) * kDefaultMkIVArmFactor

armourPenetration is derived from the lookup table in the estabs which factors in the accuracy at the given ranges. So the only factor of yours we don't take into account directly is shellweight. But we echewed that because there was no direct predictable corelation between shell weight and penetration. A bazooka round for instance weighs bugger all but can penetrate a good deaql of armour relative to solid shot fired from a 75mm gun. That's why we opted to store a penettration value in the range table itself and thus allow the designer to enter an appropriate value based on what evidence they can glean from various sources. Fopr this we used a plethora of historical sources and averaged these out to produce the values we have in the estabs. That is not to say that they cannot be refined or tweaked to give a more realistic result. But I think it a better method than simply relying on shell weight.

Back to the original issue of discounting the value of those pesky AAAMGs on the Shermans. Another less processor intensive option is to simply apply a modifier to weapons whose primary role is not AArm. Eg if we halved the value of such weapons, then we would end up in your example with:
8.8 + ( 5/2 ) = 11.3

Maybe not as punitive a reduction but still a step in the right direction.

[Lieste]

I was thinking along the lines of weight correlating with behind armour effect - so a light APCR will penetrate more armour than a full weight APC round, but the reduced diameter of the penetration path (which accounts for much of the increase in effective penetration, and reduction in shell weight) will give less spall and kill-probability where both rounds are likely to penetrate. HEAT is also noted as being relatively poor with respect to BAD, but can penetrate high thickness's of armour with simple and inexpensive weapons.

In any case, the individual Estab items can have a single stored AArm/Apers/Arm/Bomb value. This value can be computed using any formula you choose without any performance penalty as it can be pre-computed. There is a cost involved in calculation of unit strength from combination of these sub-elements if a square-root form is desired, but this can be reduced by holding all internal values as k^2, and adding/subracting/multiplying these directly, and only taking a single root function when each unit is displayed.

[Lieste]

Why do all the heavy AT guns share an identical AArmFP? There is a considerable difference in real-time effectiveness observable*, but the 88L56, 75L70 and 88L71 are rated the same. (Even the less capable 75mmL48 is only a few points lower on the AArm scale

*A KT/JagdPanther will deal with allied heavy armour frontally at noticeably longer range and more reliably than the Panther or Tiger will do. The Jagdtiger is slightly below the KT at short range, but the increased ranges and long range accuracy/penetration should (IMO) make it more of a threat (albeit a localised and inefficient one (at ~45kg per round - more when supplied packaged)) - as a practical weapon system it is ridiculous though - far too heavy and large for use on European roads**

** How likely is it that bridge crossing capability can be eventually tied to the crossing and vehicle weight classes? Right now any bridge that can pass a motorcycle will also handle a Jagdtiger.