Panther Shot Trap Still Not Trapping

[c3k]

Again, I decry the assumption of knowledge about this phenomenon.

Perhaps that 1^ descent angle is critical, when considering the ogive of the shell, curvature of the mantlet, and location relative to the mantlet.

Ken

[Amizaur]

I don't have ballistic tables for any German or Allied ammo - but the trajectory could be simulated quite accurately using some ballistic simulator tool - if we know V0 and some additional data, like velocity profile or time of flight to known range.

I have ballistic tables of some Russian guns only.

Here is some data for 800m range:

...gun type......shell type.......MV.........TOF....Impact Vel...descent angle

122mm D-25.....BR-471.....781m/s.....1.1s.....680m/s.....0 deg 26'

100mm D-10.....BR-412.....895m/s.....1.0s.....771m/s.....0 deg 22'

85mm D-5T.....BR-365 ....792m/s.....1.1s.....717m/s.....0 deg 27'

85mm D-5T.....BR-365K...792m/s.....1.1s.....653m/s.....0 deg 27'

76mm ZIS-5....BR-412.....662m/s.....1.3s.....583m/s.....0 deg 34'

So on 800m it's usually about 0.5deg

[Amizaur]

Again, I decry the assumption of knowledge about this phenomenon.

Perhaps that 1^ descent angle is critical, when considering the ogive of the shell, curvature of the mantlet, and location relative to the mantlet.

Ken

The mantlet is rounded. No matter what the descent angle is, 0deg, 0,5deg, 1 deg or 2deg or -3deg, there is always some place on the mantlet vertical profile resulting exactly the same hit angle.

1deg of difference in descent angle is corresponding probably to about 0,5cm of mantlet circumference (so probably 0.25cm of mantlet height) the "ricochet zone" on the mantlet is much larger than that so it really DOESN'T MATTER what the descent angle is (the higher, the better) as long as it's not minus 10deg because then the ricocheting shell can miss the roof completly...

[Sgt Joch]

if its always level, it would simplify the calculation. We can assume that shots in the center of the mantlet will not ricochet while shots at the very bottom edge of the mantlet will ricochet, but probably into the turret ring?

The issue is more trying to figure out the area of the "sweet spot" where the shell will be deflected and still retain enough energy/shell cohesion to pierce the roof.

Vanir, were you testing with 75 mm AP or 76 mm? In the Isigny test, the deflected shot was a 17pdr SABOT. The photo you posted in #65 was 76 mm (HVAP?). #73 is a turret ring penetration, is that 75 mm ammo? looks in too good a shape for that.

[Battlefront.com]

I find it slightly depressing that, after all this time, I still have to illustrate why a single real world test can't be extrapolated for all game situations all the time. The more technical the game aspect the more important it is to take into consideration the specifics of a test before drawing conclusions. CM is inherently a simulation, so if you make a mistake on the input then you're going to get misleading output. Therefore it is critical to not be cavalier about the variables as if they don't matter.

It could very well be that a specific Allied gun, with a specific amount of barrel wear, mounted on a specific type of vehicle or ATG carriage, at a specific range, at a specific height differential, at a specific angle of impact, with a specific aim point criteria, with a specific batch of a specific type of ammo, etc. can produce a 1 in 8 shot trap result on a regular basis. Or it could be that if the gun fired another 16 rounds under the same circumstances there would be no further shot trap hits. But we'll never know because apparently this is the only test we know of.

Now, let's say that we accept the one specific test at 1 in 8 results as totally representational (which is a leap). To make sure we're comparing apples to apples, an ingame test would have to duplicate the circumstances as closely as possible to match those of the real world test. That is the *only* way to directly compare a single real world test with a single ingame test. Anything other than that is inherently apples to oranges.

Now, it may turn out to be that the game has it wrong. I never claim we have things unquestionably correct because one shouldn't ever do that. However, after 14 years of doing this I also quite aware that customers can also be wrong. It's nice when both sides can approach a situation like this with that understanding.

Steve

[Offshoot]

The mantlet is rounded. No matter what the descent angle is, 0deg, 0,5deg, 1 deg or 2deg or -3deg, there is always some place on the mantlet vertical profile resulting exactly the same hit angle.

Except that the mantlet in-game is not rounded - the curvature of the mantlet model is made up of 7 or 8 flat planes. And are the shape and size of various amunitions modelled as well? The actual effect of these variables depends of course on how BF do their calculations and whether there are any extrapolations beyond the in-game 'physical' shape.

[Redwolf]

Except that the mantlet in-game is not rounded - the curvature of the mantlet model is made up of 7 or 8 flat planes. And are the shape and size of various amunitions modelled as well? The actual effect of these variables depends of course on how BF do their calculations and whether there are any extrapolations beyond the in-game 'physical' shape.

Excellent question, that.

[JasonC]

What is a probability of occurring for a macroscopic physical event of ordinary mechanics?

Microstate detail questions may involve real probabilities in some pure physics sense. Macroscopic physical events among objects with masses in kilograms and times in milliseconds do not. Every single one of them is effectively deterministic, on those scales.

What we mean in the first place by assigning a probability of occurrence in such cases is that the deterministic outcome depends on the detailed values of the problem parameters, and that the region of that parameter space that results in this class of deterministic outcomes rather than that class of deterministic outcomes, is of this proportional size - combined with an assumption that those specific parameters will take on that full range of values, effectively at random, from specific physical case to specific physical case.

The probability arises, is used for, and means, precisely the *uncertainty* in those initial conditions of the specific parameters of the problem. It is the post-system mapped outcome of a smeared out unknown set of those initial conditions.

It is a complete fallacy to claim that each detail must be known to assign the right probability to the event. If you knew every detail, the probability of the event would be 1 or it would be 0. You are using a probability in the first place because you don't know the initial parameters precisely. The probability is exactly meant to record a portion of the initial ensemble that will result in this rather than that post system outcome.

Why do people hide in these obvious fallacies?

[Battlefront.com]

It is a complete fallacy to claim that each detail must be known to assign the right probability to the event.

Who said you need to know EVERY detail? I'm talking about basic stuff like what the gun was that fired the rounds, what the range was, what was the gunner aiming for, what type of ammo was used, etc. These aren't minor trivialities, they are critical components of physics based simulation.

Or are you saying that you know for absolute certain that there are no effective ballistic differences between various 75mm rounds or that range plays no roll in determining outcome? Or that 8 rounds fired is a large enough sample size to extrapolate to dozens of different guns and ammo combinations in all angles and manner of hits?

Further, I'm saying that comparing Vanir's test with the US Army test can not be done with any measure of confidence unless we're reasonably sure the variables are similar. Or that the dissimilar variables are accounted for and deemed inconsequential. An example of the latter is something like the soft values for the test crew. Poor leadership isn't likely to have influence on the test results. Ammo type, range, gun type, etc. can't be so easily dismissed.

All that US Army test proves, for certain, is that the chance of getting a shot trap hit is greater than 0. Which is good considering we have ample evidence to suggest that's the case from other sources.

Why do people hide in these obvious fallacies?

It would be great if you answered your own question

Steve

[Sgt Joch]

Before getting off on a tangent, it is important to point out that the 1 in 8 WAS NOT A TEST RESULT.

It was achieved in a combat situation, so the real sample is not 1 in 8, but X in Y, with X = number of shells deflected by the mantlet and Y = all hits achieved to the front of a Panther by Allied forces in WW2.

Actually we also have to break that down by type of ammo since obviously every type of ammo reacts differently. That is the reason I asked Vanir what type of ammo was used in the test.

75 mm APC fired at the Glacis plate of a Panther will ricochet almost 100% of the time while 76 mm (17 pdr) APCBC will penetrate same plate 40-50% of the time. Both shells are roughly the same, the big difference is the muzzle velocity. So obviously it is NOT as simple as just figuring out what angle the shell hits the mantlet at.

Is there any recorded instance of 75 mm APC being deflected by the Mantlet into the hull top?

[JonS]

Anecdottaly (see: Wilson "flamethrower") a panther was KO'd by a Churchill via a mantlet ricochet. That was probably a 75mm

[Battlefront.com]

We should probably stop saying "75mm" because there's so many 75mm guns about. A Churhill's 75mm is not the same gun as a Sherman's 75mm.

I can not emphasize enough that I'm not trying to say that CM has it perfect. I'm just saying that I refuse to try and make a case to Charles for change unless there's some reason to indicate change is needed AND the circumstances AND the degree of change needed. If you think I'm being critical of the "evidence" so far, Charles wouldn't even bother looking into it at this stage. At least I'm reading the thread

Steve

[Vanir Ausf B]

In the tests I did about 47% of the hits on the mantlet ricocheted. That means that out of 2310 hits there were about 1086 ricochets, of which only 1 hit the deck plate that is literally inches away. There is no rational explanation for this vanishingly small number.

[Battlefront.com]

Remind me again how many of the 2310 hits were recorded with v2.12. Or at least v2.11.

Steve

[Vanir Ausf B]

All of them were done with 2.12

There were a further 636 done under 2.11 that I mentioned in the first post in this thread, so the total for both versions combined is 2946.

[c3k]

Umm, if half bounced UP, that would be about right for a random distribution of hits, right?

What if the game doesn't randomize the aimpoint? Is it biased towards the center of mass?

How many would hit the lower "sweet spot" band? That band is PURE conjecture. Perhaps it's a precise spot, different for each round/velocity/impact angle combination.

Not discounting the work you've accomplished, which is great, but of the 2310 mantlet hits in v2.12, how many were from a "bad" spot, how many with a "good" combination?

Pouring over the books, I don't see a consensus about where and how to take advantage of this supposed "guarantee" ricochet effect in real life. Only that it MAY be possible and it's the ONLY chance you've got if facing a Panther with certain Allied tanks.

Ken

[Thomm]

Going back to the sketch on the first page (unfortunately the red arrow does not make a lot of sense) and looking at the geometry of the mantlet, I agree with Vanir Ausführung B that more than one in a thousand shots should ricochet downwards into the hull.

Best regards,

Thomm

[Amizaur]

Simple question - what happens with a shell that hits an angled armor (let's say the angle is between 30deg and 85deg) and doesn't penetrate and doesn't barely penetrate (with could result in it being stuck in the armor plate) and doesn't break into pieces?

Every other shell will bounce away in some direction and this is called "a ricochet".

So this is quite normal for a non-penetrating hits against an angled armor plate. No matter what the shell diameter, type, velocity - if the shell doesn't penetrate or desintegrate it has to go somewhere. It won't just stop in place.

Now if we know that shells do ricochet, and most of non-penetrating hits o Panther mantled would ricochet, we need to know what area of the mantlet would give such ricochet angle that the projectile would hit the hull roof. Look at the geometry - IMO almost any shell that hits the lower part of the mantlet and doesn't penetrate, would end hitting the hull roof.

Now the next question is - what part of those shells would ricochet with enough energy to penetrate the hull roof ?

IMO shells that hits the area of about 1/3 height of the the lower part of the mantlet would not only ricochet into the roof, but also could retain large part of their energy after the ricochet. If that remaining energy is enough to defeat a 17mm plate, then we get a penetration.

Of course a 45mm shell is not likely to penetrate the roof even if it retained 30% of it's initial energy after ricochet.

On the other hand, a 600m/s 6kg 75mm shell that retains 10-30% of it's energy after the ricochet is IMO quite likely to penetrate the 17mm roof.

[Sgt Joch]

what I am having trouble figuring out is what angle a shell has to hit to be deflected into the hull top.

If you look at the photo, you can see where the shell gouged the mantlet before hitting the turret ring.

If a shell is coming in level, you would think most deflections would be into the turret ring or bottom of the turret.

I have trouble seeing how it could be deflected into the hull top, unless its coming in with a downward angle.

[Thomm]

Another idea: the "mirror images" of the shells that ricochet downwards are those which ricochet upwards. Perhaps - by monitoring their flight pattern - their residual velocity and their residual angle of obliquity could be reverse-engineered.

Best regards,

Thomm

[Apocal]

I have trouble seeing how it could be deflected into the hull top, unless its coming in with a downward angle.

Or the gun is slightly elevated, the tank is negotiating some slight incline or microterrain, etc.

[Sgt Joch]

Or the gun is slightly elevated, the tank is negotiating some slight incline or microterrain, etc.

ah yes, good point.

[c3k]

Okay, what if it hits the bottom 1/100th of the mantlet? Will it hit the roof? No, it will hit the lower edge of the turret face. What about the lower 2/100th? Etc.

The "band" to ricochet into the hull roof may not exist. It may be a specific spot, only. (Different, based on different shells, etc.)

YES, Panthers took hits which ricocheted and penetrated the 16mm hull roof. Sometimes the driver or radio operator were hit by them. But which ones? (Shots, that is...) And, based on the relative rarity of the chin mantlet and the incredible reputation for frontal invulnerability that the Panther had, it must've been VERY RARE to get a ricochet which put the Panther out of the fight.

Ken

[Thomm]

It has been written before, and I guess it is vital for this discussion. Assuming that Charles decides what happens to impacting shells based on the geometry of the polygon mesh, we get only three or four discrete armor angles for the lower half of the mantlet!!

Perhaps this filters out many of the possible hit vectors.

Again, it would be interesting to observe the trajectories of the shells deflected upwards. If these come in distinct groups regarding their angle of flight, then the coarse discretization of the mantlet indeed counts.

Best regards,

Thomm

[Thomm]

How about this: From a simple sketch with very simple assumptions (and input from some papers about ricochet mechanics) I come to the conclusion that the mantlet area with normal vectors from 15 deg to 35 deg angular offset from the vertical is a good candidate for penetrating ricochets. Areas lower may cause hits to the turret ring, while hits on areas higher up on the mantlet would absorb too much kinetic energy for deflecting the shell.

The area described above covers roughly 5 % of the projected area as seen from the shooter. This is a small area, but not as small as the 0,04 % that Vanir Ausf B sees.

Now we throw in the fact that the angle between the normal vector of the lowest polygon and the vertical appears to be greater, in the range of 40 degrees, and now we have a situation where most of the theoretically possible hits hit the actual polygon mesh at an angle which is unfavorable for an energy-preserving ricochet.

In other words: throwing in a few more polygons may make the tank more prone to penetrating ricochets.

Best regards,

Thomm