Question on Estab values

[Lieste]

Looking at the documentation that comes with the Estab editor, the penetration values should be those for 30 degree slope, and the armour should be calculated as if it was sloped at 30 degrees (I assume this means a vertical slope of 30 degrees?).

Comparing my on-hand data to the estab penetration values shows BFTB using a consistently (if slightly variable) higher penetration value - for some of the larger weapons this is particularly noticeable - the 90mm M3 seems to be about 50% higher than the APC round should manage at 1000m, most others are at least 20% high, but a very few (the 2pdr, and the 75mm M3 for example) are near their published figures.

Backing out armour data also appears to show differences to the documented values - here I get for example a value of approx 86mm for the Tiger I front - most of the armour is sloped less than the 30 degree target, so the effective 'thickness' is lower than the nominal 100mm.
Now, I don't claim to have apportioned areas in any terribly scientific way - mostly by eye using published armour values and an average of 5 nominal thickness/slope pairs - each modified to give an "effective thickness" at 30 degrees with a slope exponent of a 'guesstimated 1.415', but the final armour values seem at odds with this rough calculation - most are around 20% higher than expected, with a few like PzIV J and PzIV L70(A) close to the nominal value and others (mainly tank-hunters) up to 50% stronger. (Sherman overestimated by ~30-40% - the Jumbo by perhaps 50%).

What should be going on here, so that additional vehicles/weapons can be added in a consistent manner?

(Edited to correct mistyped exponent)

[Lieste]

Are the firepower values (accuracy and penetration) linearly interpolated or step-wise? If closing from say 501 to 499m, does the penetration rise from (a nominal) 99 to 101mm, or jump from the 1000m value to the 500m one (eg 80mm to 100mm)

[simovitch]

A couple of things may make "book" values appear different from the estab values. If the Gun Penetration data was based on a vertical target plate, we used a lower estab penetration value modified by cos(30). For example if a gun penetrated 50mm of vertical plate, it could only penetrate a plate 43.3 mm thick at 30d sloped.

Target plate thickness got an additional modifier in the form of a deflection factor. The sloped armor not only increases thickness of the plate on the horizontal plane, it also has the tendency to deflect the shot. I have to dig out my data tools from my old computer to get you the formula. I agree it should be shared with modders so that we stay consistent.

As to your second query, Arjuna would have to answer that. I always assumed it was linear interpolation between data points.

[Lieste]

Well that was my point really - the values for sloped armour for most German weapons are given as book-values at 30 degrees, yet the in-game values do not agree closely - 75mm PzGr39 from the KwK L/48 for example is given as 106mm (Doyle/Chamberlain/Jentz) where the in-game value is 124mm.

I already account for the deflection increase with slope by modifying cosine with an exponent:

Armour basis = Armour thickness / cos(slope)^1.415 * cos(30)^1.415
1.415 is a value that makes the theoretical penetration into 60 degree sloped armour around 75% of the LOS penetration of a normal plate.

[simovitch]

I just started taking a look at the values in the estabs. It seems that the Gun penetration values were adjusted higher to reflect the penetration of a vertical equivelent of a 30d test plate. This establishes the base value for the penetration capability of the gun. Even so they still seem a bit too high, so I need to look into this further.

So I would hope that the armor thickness in the estab prior to application of the deflection factor is based on vertical equivelent, i.e. the plate thickness/cos(a) where (a) is the slope of the plate from vertical.

More to come as I dig into this (again).

[Lieste]

I thought that (some) of the data looked like vertical plate (Tiger I, some of the other armour) but it isn't consistently done - some are very close to the 30 degree basis, and some are 20-50% above vertical equivalences.

Not looked at any of the light vehicles - but I have prodded all the tanks and T/D.

[simovitch]

We more or less averaged the turret and hull thickness and angles, and applied the deflection factor 1/cos(a)^1.585

Since BFTB was built off of the HTTR estabs which did not include the deflection factor I went through every HTTR vehicle that was to be used in BFTB and made these modifications. The Tiger I was not modified using the final formula, so I just changed it in the estabs. Here you can see the excel data and results, with the formula (for front armor) highlighted on top.

[Lieste]

You shouldn't average the angles and thicknesses like that - they should be calculated separately and then the results averaged. Usually it won't matter too much, but some vehicles will change a lot.

I note that these represent a best case for the armour of the Tiger - the average slope is less than 25 degrees for the hull - again this doesn't matter too much for this vehicle, but some will be much more variable over their thickness/slope on a single aspect.

[Lieste]

For future developments it might be beneficial to have a 'moderate high' and 'moderate low' armour value for each face - then you get a better representation of eg the PzIVH/J relative weakness in the turret, or the relative immunity of parts of the Panther front armour, but vulnerability to a wider range of weapons for the turret/mantlet.

A particularly important case is the low-end AT guns in the german infantry - here most average flank armour is invulnerable, but the weapons should be modestly effective against large areas of the hull side of the Sherman - but not those areas reinforced with applique or the turret. A huge difference between barely effective and incapable.

[simovitch]

If this was a true tactical simulation I might agree that it's worth another look. In reality there are so many variables effecting the actual angle of impact that it is just not worth it to try to meet technical accuracy in this single situation. I think this method gives a good enough overall representation of the ballistic model for all vehicle types.

[Lieste]

Possibly more important than armour values is target size - I understand this is directly related to hit % in some way - I note that the vehicle length/width/height is a gross value, and except where a vehicle is exceeding boxy, leads to much larger than actual vulnerable sizes. This is especially the case for tanks with HV guns (eg most German armour, and a scattering of Allied stuff). Many of the Allied tanks have little gun overhang though, and therefore much smaller "sizes" for vehicles of similar bulk.

eg Pz IV E @2.68/2.84/5.92m is no smaller in terms of vulnerability than a PzIV F2 @2.68/2.84/6.62m, a PzIV G @ 2.68/2.88/6.62m or a PzIV H @ 2.68/2.88/7.02m

In each case, the additional length comes from gun overhang, and the additional width is track extensions &/or side skirt, which is not really part of the protected space.

The Tiger I is a little shorter than 6.28m (that is for the SturmTiger) rather than 8.45m and other vehicles have similar over-dimensioning from a vulnerability standpoint.

I'd suggest using hull length, rather than gross length as a basic correction, possibly with a front/side prismatic correction factor if it seems necessary (to adjust between eg a box and triangular target of the same 'size' where the box is 2x the area).

[Lieste]

While looking at sizes - it seems that width and height are transposed on many (most? but not all) entries. Is this a genuine transposition, or is it only the label in the Estab? Which way around is "correct".

It should be made consistent, as height should be slightly more significant than width/length, and in any case although the frontal area is unchanged, it does affect flank area (often increased) and footprint (corresponding decrease).

So far I am adjusting lengths listed to give hull length. And will revise my width/height data after giving it some more thought. Some errors with PzII/Pz38 based guns.

Some idea of the possible significance (a sample of some values captured during data collection - not sure of the state of British tanks, I didn't capture the original data before I over-wrote it):

% variation (2sf) +/- pf front/side/top where gun overhang and orientation errors are present in BfTB Estab
PzIIIM: +0.39/+34/-3.5 {BfTB Front same, Flank too large, Top small compared to my revised data}
PzIVJ: -0.01/0/-0.01
PzVG: +1.4/+48/+14
PzVIE: -1.6/+67/+1.42
PzVIB: +0.05/+69/+14
M5A1: +0.2/+0.63/+0.23
M4(75): +0.69/-3.8/+5.6
M4(76): +0.50/+8.1/+33
Pz35S: +3.2/+2.5/+0.29
PzB2: -9.3/-20/+8.2

[simovitch]

From your %'s it does look like turret overhang was used, I will take a look at this.

Looks like most British tanks did not have a turret that overhung the hull, except the Firefly.

What is your source for hull dimensions?

[Lieste]

Do look at the width height relationships also - some like the PzIIIM seem correct, while the PzIVJ, PzVG and PzVIE for example have the larger dimension (width) used as height and vice versa.

I got many from 'earlier models' eg for the PzIV series, where the short barrel guns lacked overhang - there was some growth in the hull, but not much, and before the long guns came in. Where there are ammunition carriers, they give the length approximately for the gun-carrier also.

I used Chamberlain and Ellis and Chamberlain and Doyle for most dimensions, using these as gross length. Examination was used to assess whether this was suitable as a net measure (small variations not significant, particularly for length measures, height being IRL far more significant for accuracy (due to range errors)).

Those not obtainable by inspection and acceptance of the gross values, those not sharing a common hull with a known vehicle and allied a/c not included in the texts I had to search online for.

I used a mix of AFVdatabase, wwiivehicles, wwiiequipment and wikipedia where I found nothing better. A few random finds in forums etc. All values were written into a spreadsheet, to check trends and also to calculate areas and volumes of the gross 'box' encompassing hull & turret/SS but not gun.

I'm also in the process of writing an accuracy spreadsheet to derive hit% for a user specified target, weapon/ammunition and a flat fire approximation, with a direct sampling of 1000 rounds with a specified spread of gun-target ranges around a 'chosen' FCS range, and aimpoint 'correct' or off by a partial form. I think the range for 'good accuracy' is a little low for some weapons, but I want to experiment and evaluate a bit before I change anything else.

Do we know what the 'standard' BfTB targets are for Apers and AArm fire?

[Lieste]

Fuel - as payload, it appears to be listed as a mass - 1kg of fuel or bulk fuel? While in the range/speed panel it seems to be using litres - as fuel weights are 0.75kg/l or thereabouts for gasoline/avgas this might be giving incorrect payload/weight/supply ratings and/or ranges.

[Lieste]

I added a non-LRF version of an M1A1 just for giggles. From my ballistics calculations I obtained a hit % for a centre of mass aimed shot of >99% up to 1600m. Using the ratio of the minimum tgt height that would give 99% to a 'standard' 2m high target I weighted the hit% above 100% for closer ranges (4000% at 100m, 660% at 500m, 220% at 1000m) hit % at 3000m was still at 50%. These values are broadly in line with my experience as a simulated gunner in a manual mode M1 - no difficulty, except where a gross error is made at hitting within at least 1200m, around a 50/50 chance of striking a stationary exposed frontal target at 2.5-3km.

The penetration values were 519-480 over the 3000m range. Armour taken as 400mm frontal, 160mm flank, 100mm rear and special armour neglected, so HEAT weapons are too effective by some extent (but are fairly low risk anyway due to poor values of accuracy and penetration).

Over a multi-day battle, the tanks of a Bn (54) destroyed 270 AFV using APFSDS rounds (HEAT listed as Apers only). Around 5-6 vehicles were lost due to flanking fires from the heavy vehicles of Pieper's KG, and several platoons routed and surrendered where they failed to evade sufficiently prior to rallying. Despite the extremely heavy weighting of hit % above 100% there were a very large number of fired volleys that appeared to have no effect on their armoured targets, and suppression & tiredness reduced the effectiveness on subsequent days to very poor ( a handful of retreats, most with no observed casualties, for high ammunition expenditures (60-80 rounds)).

I found the results interesting, and am looking at revising the break-points of a few sample guns to match outputs from the ballistic calculations. I'll probably start with 75L70, 75MkV and 17pdr APCBC + 17 pdr APDS (looking at deploying these on two separate vehicles - one with acceptable accuracy and good penetration, the other with superlative penetration, but poor consistency).

From this testing it appears that fatigue/entrenched/suppressed adverse modifiers are very onerous. Being able to take 4000% to 'very low'. Too much or ok? I'm not sure, but it seems that the hit% adjustment is too high, and more should be taken from availability/shots taken.

Also I note that 'firing' markers are still given even when all ammunition has been expended, and it is very hard to extract a fatigued, out of ammunition unit (even when facing same) in order to rest and resupply it.

[simovitch]

Vehicle dimensions: I swept through all the turreted vehicles and made adjustments for disregarding barrel length. It was easy once the general chassis length was known. Turns out also that many height and width dimensions were swapped (from httr days) but I doubt this makes a difference in game play, but I verified and fixed these anyway.

Ammo packing weight: All ammunition has packaging weight now considered, based on a generalized formula:
small arms +10%
<49mm +7.5$
<99mm +5%
>100mm +2%

Fuel payload: All fuel related data is based on litres. Dave will have to chime in here but I doubt liters were converted to tonnes for resupply purposes. To me it's very nearly irrelevant and the .25 kg difference could be considered jerry cans, etc.

[Lieste]

So long as you are talking about packing on top of complete rounds that sounds lowish. As the reference weights seem to have been the projectile weight (from the Estab Weapons screen) then they might be too low by a considerable margin.

eg low velocity weapons:
5cm GrW + 25% on complete round/projectile
leIG18 - +20% on complete round, but +30% on projectile
leFH16 +8% on complete round, but +26% on projectile
sIG33 +7% on complete round, +17% on projectile
sFH18 + 4% on complete round, +14% on projectile

eg high velocity weapons:
76 (r) +20% on complete round, +105% on projectile
88 (Flak) +11% or +28% on complete round, +101% or +132% on projectile (packed as 3x or 1x)
15cm K +5.5%* on complete round,  +54%* on projectile (*conservative estimate - no excess packaging on separate charge - assuming 50% (half that of modern M119 charges) this might be ~ +21% and +76%)

These weights are still far lower than the 1950 manuals give for packaged US ammunition, which remains largely the same to the current day. Unsure of wartime practice, but I feel it is doubtful they could have cut the overhead by 3/4 or more...

M107 155mm, M119 Zoned Charge +7% shell alone, +81% including charge in tube (23lb charge/70lb packed weight).

Note none of the rounds above are APCR/PzGr40 or HL, all are unitary HE or AP. The lighter rounds have a higher % packaging.

[Lieste]

Did you look at T/D? - the PzJ series are all longer than their hulls - I know that they can't 'shorten' themselves for manoeuvring, but they are excessively large flank targets at present.&nbsp;

[simovitch]

ORIGINAL: Lieste

Did you look at T/D? - the PzJ series are all longer than their hulls - I know that they can't 'shorten' themselves for manoeuvring, but they are excessively large flank targets at present. 

Yes, I got all the jgpz, Stug's and US TD's as well. I even checked the AC's for the most part.

Regarding packing weight, I assumed the estabs list shell weight, i.e complete round. I'm only adding the weight of the box they are shipped in. Did you verify that it is indeed only projectile weights given in the estabs?