Pz IV tactics?

[Joachim]

Though Thor presents some interesting points, I have to point out that Jens/pyewacket's model does not fit to Dschugaschwilis assumption.

e is fully visible

The 1/3 factor actually states that hitting the upper parts of the tank in hull down-mode is 1/3 while the tanks is still half exposed. Not very convincing:

As the Bell curve modelling the hit prob is dependent on distance, gun type, ammo etc. and we will assume the classic "ceteris paribus" approach, the "bell" we move across the picture of the tank is the same in both cases.

The perfect aim point will maximize the volume of the bell over the upper parts of the tank, and thus at least increase the hit probability for those parts - which was 1/2 in the hull up example.

So to get a hit prob of 1/3 for a hull down tank, one hat to use a less than optimal hit point. This might make sense when aiming for weak points, but CM does not model aiming for weak points.

Aiming for center of mass, however, will at least increase the hit prob for a compact target.

Thus we can assume that D's sentence holds for a gunner with basic training.

@pyewacket:

A good approx for the hit porb is given in CM when you target the tank - see referenced thread in CMAK forum above.

Gruß

Joachim

1/6 lower hull * 1,0 = 1/6

1/2 upper hull * 1,0 = 1/2

1/3 turret * 1,0 = 1/3

For a hull-down tank:

1/4 upper hull * 1/3 = 1/12

3/4 turret * 1/3 = 1/4

[Yggdrasill]

Originally posted by Joachim:

[QB] Though Thor presents some interesting points, I have to point out that Jens/pyewacket's model does not fit to Dschugaschwilis assumption.

e is fully visible

The 1/3 factor actually states that hitting the upper parts of the tank in hull down-mode is 1/3 while the tanks is still half exposed. Not very convincing:

[Joachim]

Pyewacket assumed a certain hit on a hull up tank.

The turret is 1/3 of the hull up tank.

For a hull down tank, the turret is 3/4 while the visible upper hull is 1/4, so it is 1/9 of the total tank thus we have 1/3 plus 1/9 = 4/9 of total. oops... a hull down tank has only 4/9 visible compared to a hull up tank. Somehow I guestimated at 1/2... it's getting late over here

Short version:

If you hit a target of size 9 with a probability of 100%, why should you hit a target of size 4 with only 1/3 probability?

First guestimate would be you hit it with at least a prob of 4/9.

Long version:

For a uniform distribution, if 4/9 of a tank are visible, the probability to hit those 4/9 should be at least 4/9 if before we had a hit prob of 100%. So for a distribution which puts higher probability on a small area, this should hold, too, if we select a "good" aim point aka center of the 2d normal distribution.

If the prob to hit the turrret for a hull down tank is less than those 4/9, the aim point is chosen poorly (-> bad gunner).

If the aim point is chosen well, we have at least a prob of 4/9 to hit the tank.

The example for a hull-down tank gets:

hit prob upper hull >= 1/4 upper hull * 4/9 = 1/9

hit prob turret >= 3/4 turret * 4/9 = 1/3

And we hit the turret at least as often as before...

Gruß

Joachim

[Yggdrasill]

Joachim,

Ok now I understand. Yes, you are correct. I did not think clearly in terms of the relationship between the size and to hit probabilities.

*but* what if you change the ratios for hull down so that, for example, half of the exposed area in a hull position is turret and half upper hull (instead of 1/4 and 3/4). Then the chance of a turret hit does go down.

Do we know how much of the exposed area for a hull down PzIV is hull and how much is turret?

I guess the only way to "know" this is is to run a bunch of tests with immobilized PzIVs in a hull-down position, and see what the relative turret/hull hit ratios are.

thor

[Dschugaschwili]

Sorry, I have to enclose a part of the message in a code block to get the greater than operators to work.

There has been a lot of arguments about the hit probability against a hull-down tank in relation to the hit probability against a hull-up tank in this thread lately.

The problem here is that first, we don't know how much of the frontal area of the tank is visible if it's hull-down, and second, even if we knew that the hit probability is highly dependant on the other factors contributing to the overall accuracy of the gun.

Example 1: We're firing with an extremely accurate gun. So accurate that we not only have almost 100% hit chance, but almost all the shots strike the target within an area that's significantly smaller than the target. That means that we can make the target smaller without reducing our hit chance much.

Or put it another way: we can put the target hull-down and still hit it with almost every shot, so our hit chance hardly went down at all.

</font><blockquote>code:</font><hr /><pre style="font-size:x-small; font-family: monospace;">Example 2: We're firing with another gun. I'll assume that the shot

pattern follows a normal distribution.

We're firing at a square target with 1m sides and we're aiming at the center of it.

For this argument, I'll split the target into four parts (~plates),

each of which 25cm high, and I'll call them P1 to P4 from top to bottom.

We want to fire against the fully visible target once, and we want

to place the target hull-down later.

I'll call the hit probability against a certain plate of the hull-up

target hpu(Px), and against a plate of the hull-down target hpd(Px).

All plates have equal shape, and against the full target we're aiming

at the edge between P2 and P3, so (as I showed above) we get

hpu(P2) > hpu(P1) and hpu(P3) > hpu(P4).

(We also get hpu(P2) = hpu(P3) and hpu(P1) = hpu(P4),

but we don't need that here.)

Now let's put the target in a hull-down position in such a way that

only the upper half is visible, so we have hpd(P3) = hpd(P4) = 0.

Aiming center-of-mass now means aiming at the edge between P1 and P2.

So in relation to the aim point, P1 is now in the position of P2 in the

hull-up target, and P2 takes the place of P3.

All else being equal, we will therefore get hpd(P1) = hpu(P2)

and hpd(P2) = hpu(P3).

We can now put the hit probabilities against the hull-up target (pu)

and against the hull-down target (pd) into relation:

pd = hpd(P1) + hpd(P2) + hpd(P3) + hpd(P4)

= hpd(P1) + hpd(P2) + 0 + 0

= 0.5 * (hpd(P1) + hpd(P1) + hpd(P2) + hpd(P2))

= 0.5 * (hpu(P2) + hpu(P2) + hpu(P3) + hpu(P3))

> 0.5 * (hpu(P1) + hpu(P2) + hpu(P3) + hpu(P4)) *

= 0.5 * pu

* (because hpu(P2) > hpu(P1) and hpu(P3) > hpu(P4))

So we get pd > 0.5 * pu. And obviously also pd <= pu because hpu(P1)

and hpu(P4) are non-negative.

We can do the same calculation with different exposure percentages,

but the result will always be that the chance to hit a hull-down target

compared to the hit chance against a hull-up target will always be

greater than the ratio between the visible areas, so all those people

claiming half visible size equals half hit chance are plain wrong.</pre>

[Pyewacket]

I am not a statistican (not really sure about what you're talking ), so correct me with the following.

The initial question was: Due to the fact the turret of a PzIV is the weakpoint, is it better to stay in hull down or hull up?

So assume there's a gun which is able to penetrate the PzIVs turret only (no other parts).

Assume also the gunner is excellent, means he accurate hits every target with each shot. As somebody said (I think, or not? not sure anymore) the gunner will always aim at the center of what is visible of the the PzIV (is this the case?) . Then, no shot will hit the turret, PzIV alive (hull up)! If in perfect hull down the gunner will shoot at the turret, PzIV dead (hull down)!

In reality the gunner would shoot at the turret no matter how much he sees from the tank. But is there an abstraction in CM?

Isn't the bell you're talking about a prove that the gunner aims always at the visible center of a tank?

can your models give us the final answer what we should do with a PzIV?

[Pyewacket]

Example 1: We're firing with an extremely accurate gun.

oops Dschugaschwili. You're faster...

I have no idea what you're doing there, but the result (if it's right) is great. Is this the prove for:

if you miss the center of hull up , there's a chance to hit the tanks outer part. If you miss the center of a hull down you hit nothing. right?

[Sergei]

Mr. Dschuga, could you please edit that post so that we don't need to scroll the window sideways? The thing with using code formatting is that it doesn't permit word wrap, so a whole paragraph becomes one line.

[Dschugaschwili]

Sergei:

Edited the post. I didn't have to side-scroll even before it, but I hope it's better for you now.

Pyewacket:

if you miss the center of hull up , there's a chance to hit the tanks outer part. If you miss the center of a hull down you hit nothing. right?

Right. Although if you're talking about tank combat, a hit against a plate that you can't penetrate is little more than a miss.

So assume there's a gun which is able to penetrate the PzIVs turret only (no other parts).

Assume also the gunner is excellent, means he accurate hits every target with each shot. As somebody said (I think, or not? not sure anymore) the gunner will always aim at the center of what is visible of the the PzIV (is this the case?) . Then, no shot will hit the turret, PzIV alive (hull up)! If in perfect hull down the gunner will shoot at the turret, PzIV dead (hull down)!

I doubt that CM uses a model this fine. I think that CM first determines if it's a hit, and then pick the plate that's hit on the basis of some fixed distribution. So I don't think that even an elite crew at 10m would always hit the upper hull of a hull-up target. But BFC has stated that guns in CM do indeed aim center-of-mass, which is probably the excuse for having a fixed distribution of hits against various plates. (By the way, I had your above thoughts before, only that in my case, it was a crack Firefly bouncing two shots off a Panther's upper hull before being destroyed. Afterwards I wondered if a green one would have had a better chance of hitting a vulnerable plate).

can your models give us the final answer what we should do with a PzIV?

Short answer:

1. If you're going against a gun that can not kill you anywhere, it's a matter of preference. Going hull-down increases the chance of turret hits, and because of that also the risk of gun hits, but it eliminated the risk of track hits.

2. If you're going against a gun that can kill you everywhere, you should go hull-down to reduce the risk of being hit in the first place.

3. If you're going against a gun that can kill you with turret hits only, you should not go hull-down because going hull-down increases the risk of being hit in the turret.

Dschugaschwili

[Joachim]

Originally posted by Pyewacket:

</font><blockquote>quote:</font><hr />Example 1: We're firing with an extremely accurate gun.

oops Dschugaschwili. You're faster...

I have no idea what you're doing there, but the result (if it's right) is great. Is this the prove for:

if you miss the center of hull up , there's a chance to hit the tanks outer part. If you miss the center of a hull down you hit nothing. right? </font>

[Dschugaschwili]

Originally posted by Joachim:

In CM the likelyhood to hit a hull down tanks is higher than 4/9 of likelyhood to hit a hull up tank. 4/9 are a (theoretical) lower bound for that factor (see some post of mine above).

Actually, one of my previous posts (the one with the code section) also proved this lower bound, and also that the expected hit probability must be somewhere between that lower bound and the hit probability against a hull-up tank, but the exact value is dependant on other accuracy constraints.

Simply speaking, the lower your base accuracy against the full target, the closer you get to the lower bound determined by you. Having a high base accuracy will let your hit chance against the hull-down target rise up to close to the base accuracy.

Dschugaschwili

[Pyewacket]

3. If you're going against a gun that can kill you with turret hits only, you should not go hull-down because going hull-down increases the risk of being hit in the turret.

Dschugaschwili, thanks!!

[Joachim]

Originally posted by Dschugaschwili:

</font><blockquote>quote:</font><hr />Originally posted by Joachim:

In CM the likelyhood to hit a hull down tanks is higher than 4/9 of likelyhood to hit a hull up tank. 4/9 are a (theoretical) lower bound for that factor (see some post of mine above).

Actually, one of my previous posts (the one with the code section) also proved this lower bound, and also that the expected hit probability must be somewhere between that lower bound and the hit probability against a hull-up tank, but the exact value is dependant on other accuracy constraints.

Dschugaschwili </font>

[coe]

well I guess I opened a can of worms: Basically I thought that perhaps the kill model for the hulldown or nonhulldown Pz IV (or any tank) was more like a two step process:

1. Determine if the shell hits the tank at all.

(thus perhaps hulldown lowers the probability of a hit)

2. If it hits, determine what part of the tank it hits (thus hull down increases the probability of certain parts getting hit).

If in general it makes sense for any tank to go hulldown rather than stay out open, then the effect of going hull down must be greater in step 1 than the effect it has on step 2.

e.g. assume that non hulldown is a guaranteed hit and that the only way to kill the tank is the turret and that in non-hulldown the probability (once it is determined the tank has been hit) of the hitting shell to strike the the turret is 1/3 (as opposed to any other part of the tank).

therefore in non hulldown mode:

100% * 33.333% = 33.333% chance of a kill.

but now lets assume that in hulldown mode there is no longer a guaranteed hit but instead only a 50% chance of hitting. Assume however if you hit the tank you can only hit the exposed parts - assume for simplicity sake that the only exposed part is the turret. Thus the probability of nailing the turret given that the tank is hit to 100%.

therefore in in hulldown mode:

50% * 100% = 50% chance of kill.

Thus in this example, going hull down wouldn't pay off. However if going hull down reduced the chance to hit the tank at all to less than 33% it would be better to go to hull down.

Anyways, that's my take on the model.

Coe

[Hat Trick]

In real life, I assume, it always made sense for a tank to be hull down (or at least it never hurt). After all, if an enemy gunner could improve his chances of a kill by adjusting his aimpoint to be centered on the turret, the enemy gunner would do so even if the tank was not hull down. Therefore, a hull down tank should never have an overall higher probability of being knocked out than a fully exposed tank. Or am I missing something?

[Bone_Vulture]

Sigh... This discussion is really revolving around in circles.

To put it short: if a PzIV is hull down, the probability that it'll receive a hit will go down, while the probability that the first hit is lethal (penetrating) will go up.

[Dschugaschwili]

Originally posted by Hat Trick:

In real life, I assume, it always made sense for a tank to be hull down (or at least it never hurt). After all, if an enemy gunner could improve his chances of a kill by adjusting his aimpoint to be centered on the turret, the enemy gunner would do so even if the tank was not hull down. Therefore, a hull down tank should never have an overall higher probability of being knocked out than a fully exposed tank. Or am I missing something?

You're missing the following point:

Adjusting the aim point towards vulnerable plates requires knowledge about which plates are vulnerable. As long as the type of the enemy tank is not identified, every gunner will aim center-of-mass to maximize the hit chance. Only if the gunner knows what the enemy tank actually is can he specifically aim at weak plates. So in your example the gunner will only start aiming at the turret once he has identified the enemy tank as a PzIV.

On the other hand, being hull-down forces the enemy gunner to aim closer to the turret. In your example, if the enemy PzIV is hull-down your gunner would aim at its turret from the first shot on, even without knowing that it's in fact a PzIV. So he has a better chance of killing the hull-down PzIV because he is aiming at the weak plate earlier, although he doesn't know it yet at that time.

So being hull-down can hurt you in real life.

Dschugaschwili

[Abteilung]

Question: Why wasn't the turret up-armored on the Pz.IV? The Pz.III had its turret improved, as did the Pz.II and Pz.38(t). Why not the Pz.IV? It seems counter-productive to me, given its overall production numbers and importance to the panzertruppen.

[Bone_Vulture]

Originally posted by Abteilung:

Question: Why wasn't the turret up-armored on the Pz.IV? The Pz.III had its turret improved, as did the Pz.II and Pz.38(t). Why not the Pz.IV? It seems counter-productive to me, given its overall production numbers and importance to the panzertruppen.

Perhaps the turret wasn't such an easy target in real life. And any increase in turret weight would cause the redesign of the rest of the tank: traverse mechanism, engine, suspension and ground pressure would have to be re-evaluated.

Cosidering that by the time the late PzIV models rolled out, the Germans were already experiencing serious armor shortages, so I don't think they had the time or resources to upgrade tanks on the field (well, ugrade more than what they did).

[ February 24, 2004, 12:14 AM: Message edited by: Bone_Vulture ]

[Abteilung]

Sounds like a plausible explanation Bone_Vulture. Thanks for the reply.

I have found that the Ausf.D - E had a traverse rate of about 11 seconds. The F1 received the 5.0cm frontal and 3.0cm side armor upgrade and altered the traverse rate by two seconds to about 13.

Now my next question is, could the turret front and mantle have withstood an up-armoring to the neigborhood of 7.0 - 8.0cm without overloading the dependent components?

IIRC, the 8.0cm on the hull and superstructure was pushing the envelope insofar as the chassis' ability to handle the extra weight was concerned; then again I think they were referring to the jagdpanzer variants and not the turreted tank. The jagdpanzers seem to be carrying a lot more metal due to the revised superstructure, so I don't know if the suspension limits would necessarily apply to the turreted tank.

I wonder what the weight of additional armor on the turret front would have been.

I guess I'm crazy to be thinking these things, but I can't help but wonder sometimes.....

[manchildstein (ii)]

the thing that is odd with the PZ IV in cmbb, is that we have people saying, 'i didn't go hulldown.. just stayed out in the open where it was safer...' maybe it has more to do with a 'one size fits all' turret than anything else... maybe they've got the armored protection correct, but the chance of a turret hit is waay to high for the PZ IV in particular... perhaps kind of like how the narrow turret front on the t-34 takes more than its share of hits in '41.... anway... for a qb, i'll purchase panzer IVs in '41.. late war though they seem to cost more than they're worth....

[Joachim]

Originally posted by Abteilung:

Question: Why wasn't the turret up-armored on the Pz.IV? The Pz.III had its turret improved, as did the Pz.II and Pz.38(t). Why not the Pz.IV? It seems counter-productive to me, given its overall production numbers and importance to the panzertruppen.

Some other thread mentions that most of the (rather small) turret front was covered by the gun shield, resulting in extra 30mm protection for most of the turret front. The remaining areas with only 50mm total armor were very small.

Gruß

Joachim

[Abteilung]

Since it is so large, I'll provide a link to a particular passage and pictures I found regarding the Ausf.F1. I am now quite interested in purchasing the reference this data comes from as it appears to be the work of Jentz and Doyle (Panzertracts). See what you guys make of this data.

Edit: I just realized the pic was originally hosted by tripod *grumble*. Sorry 'bout that. I xferred it to my old comcast webspace which seems to be active still. It is now viewable.

Link

[ February 24, 2004, 08:43 PM: Message edited by: Abteilung ]

[Keke]

Originally posted by manchildstein (ii):

the thing that is odd with the PZ IV in cmbb, is that we have people saying, 'i didn't go hulldown.. just stayed out in the open where it was safer...' maybe it has more to do with a 'one size fits all' turret than anything else... maybe they've got the armored protection correct, but the chance of a turret hit is waay to high for the PZ IV in particular... perhaps kind of like how the narrow turret front on the t-34 takes more than its share of hits in '41....

Yup. I'm playing "Tank Warning" (great scenario btw) as Germans, and my top priority is to keep Pz-IVHs out of any possible hull down positions...and if at all possible, hit T-34s of my opponent when they are in their temporary hull-down positions while moving over the crest towards me. Sounds like an impossible combination, but with forward slope positions this can be achieved. In one turn I managed to kill 15 T-34s while losing only 1 Pz-IVH...

Unless these tanks were specifically designed to fight on forward slopes, I consider this modelling of hull down, with virtually enhanced turret hits, to be the greatest weakness of the current game engine. IMHO, of course.

[Joachim]

Originally posted by Abteilung:

Since it is so large, I'll provide a link to a particular passage and pictures I found regarding the Ausf.F1. I am now quite interested in purchasing the reference this data comes from as it appears to be the work of Jentz and Doyle (Panzertracts). See what you guys make of this data.

Edit: I just realized the pic was originally hosted by tripod *grumble*. Sorry 'bout that. I xferred it to my old comcast webspace which seems to be active still. It is now viewable.

Link

Interesting... Seems like the equivalent to the 50+20mm armor of the PzIII series.

I wonder how often that "Vorpanzer" actually was used.

Gruß

Joachim