Panther Shot Trap Still Not Trapping

[sfhand]

I am not a grog... but I know you are out there

How many panthers were killed by shot trap hits during the war?

[Vanir Ausf B]

How many panthers were killed by shot trap hits during the war?

How many were killed by hits to any other area? This sort of data does not exist.

But if shot trap kills were as incredibly rare in real life as they are in the game then the answer would probably have been 0.

[Ts4EVER]

The problem was big enough for the Germans to take note and fix it.

[MikeyD]

This is increasing starting to sound like a whopping case of 'be careful what you wish for'. If all of a sudden 75mm Shermans and even Daimler scout cars started knocking out your big expensive Panthers frontally on a regular basis players would squeal like a stuck pig over it.

[Battlefront.com]

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

This is where a lot of the ballistics variables come into play. what you're seeing in that shot is that the round didn't so much as ricochet, but rather "slid" from the initial point of impact to another part of the tank. And that's very likely due to the physics we're talking about here. Which is to say the simplistic view of a round being deflected strictly by angles alone is a flawed approach because it operates on an assumption that simply can't be relied upon.

In game feedback terms this complicates things greatly. I do not know how such things are calculated, but I know that rounded surfaces (even as far back as CMx1) have special attributes to influence outcomes due to the unique nature of rounded surfaces. What affect this has on Vanir's test I do not know, but it is in the mix. Perhaps some of the hits were to the lower turret after skidding off the mantlet.

And that's something else to consider. For a round to deflect off the mantlet and into the top hull the round would probably have to be marginal or marginally placed. Otherwise it would have likely penetrated the mantlet itself or skidded into the base of the turret.

Also, looking at the Patton Museum picture you're looking at a hit where the round was damaged upon impacting the lower mantlet, damaged further as it hit lower turret, and still managed to penetrate the SIDE armor at a very oblique angle. That's a pretty tough round with a lot of power behind it.

As for the Germans addressing the problem with the chin armor. As I said earlier, they were 2/3rds through the entire life cycle of the Panther before they made this change, despite having made other significant changes between D, A, and G models. This, to me, shows that it wasn't a big deal at least in 1943 and into 1944. But perhaps some gun/round showed up on the battlefield in the summer of 1944 that changed the equation.

I'd be curious to see what happens when a Firefly is used against a Model A vs. a Model G vs. a Model G with chin armor (which we don't have in the game yet, unfortunately). Maybe that gun, or something on the Soviet side, changed things so that the shot trap went from theoretical problem to significant problem.

Steve

[Vanir Ausf B]

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

This would require the mantlet to be irregularly shaped rather than symmetrical.

[Vanir Ausf B]

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.

Yes, this is something I have been thinking about as well. At the very least, the way the Panther mantlet is modeled in the game will greatly decrease the granularity of results vs. what would be seen with a rounded object.

[sfhand]

The problem was big enough for the Germans to take note and fix it.

Well, according to wikipedia...

"From September 1944, a slightly redesigned mantlet with a flattened and much thicker lower "chin" design started to be fitted to Panther Ausf G models, the chin being intended to prevent such deflections. Conversion to the "chin" design was gradual, and Panthers continued to be produced to the end of the war with the rounded gun mantlet."

the problem wasn't severe enough to warrant an immediate change or even cease production of the "flawed" model. Since there is no evidence this was a glaring weakness in the panther, I'd much rather see BF continue to put effort into developing new features for v3.0 than trying to make something more "realistic" when nobody can document what was "real".

[A Canadian Cat]

This is where a lot of the ballistics variables come into play. what you're seeing in that shot is that the round didn't so much as ricochet, but rather "slid" from the initial point of impact to another part of the tank. And that's very likely due to the physics we're talking about here. Which is to say the simplistic view of a round being deflected strictly by angles alone is a flawed approach because it operates on an assumption that simply can't be relied upon.

Indeed - this not a billiards table.

I'd be curious to see what happens when a Firefly is used against a Model A vs. a Model G vs. a Model G with chin armor (which we don't have in the game yet, unfortunately). Maybe that gun, or something on the Soviet side, changed things so that the shot trap went from theoretical problem to significant problem.

My experience with Fireflies shooting at Panther is they "don't need no stinking shot trap" they just need to get a hit

[c3k]

This would require the mantlet to be irregularly shaped rather than symmetrical.

No.

Imagine that a specific round has to hit the mantlet at a 67^ angle to deflect into the roof. 66^ is non ricochet as is 68^. (Fictitious example. ) The mantlet is symmetrical in this case. Nor does it matter what elevation the Panther gun is at. In fact, it relies upon symmetry for this effect to be a vulnerability.

The imaginings that there was a broad band of mantlet impact zones that could cause the ricochet is baseless. It may have been one specific angle foe each shell.

Further, I would postulate that the ogive of the shell would be critical to this effect. As would L/D ratio, angle of incidence, velocity, alloy, energy, and any compound angle.

It was a rare phenomenom. It was taught as a ray of hope to outgunned allied troops. It existed but was more rare than Luke's ability to destroy the deathstar with a single shot. In fact, that should be analyzed by the statistically mimded amongst us as a comparison.

The ricochet spot may be a lot smaller than many think.

Yhe reality of many engements remain to tell us that the ticochet effect was very rare.

Typos due to smart phone and bumpy road...

Ken

[Vanir Ausf B]

No.

Imagine that a specific round has to hit the mantlet at a 67^ angle to deflect into the roof. 66^ is non ricochet as is 68^. (Fictitious example. ) The mantlet is symmetrical in this case. Nor does it matter what elevation the Panther gun is at. In fact, it relies upon symmetry for this effect to be a vulnerability.

The imaginings that there was a broad band of mantlet impact zones that could cause the ricochet is baseless. It may have been one specific angle foe each shell.

The hypothetical you describe would still be a band rather than a single point.

Further, I would postulate that the ogive of the shell would be critical to this effect. As would L/D ratio, angle of incidence, velocity, alloy, energy, and any compound angle.

These speculations seem rather pointless when we already know that the way the shot trap is modeled in the game is physically different than reality.

It was a rare phenomenom. It was taught as a ray of hope to outgunned allied troops. It existed but was more rare than Luke's ability to destroy the deathstar with a single shot. In fact, that should be analyzed by the statistically mimded amongst us as a comparison.

The only evidence that it was as rare as it is in the game is the game itself. It's a circular argument.

I will say again: there is no rational explanation for over 1000 rounds ricocheting off a surface inches away from the deck plate with only one of them hitting the deck plate.

[Vanir Ausf B]

BTW, all this talk about it being "rare" is missing the point. No one is disputing that it was rare, just not that rare. To put it into perspective, if the incidence of ricochet penetrations in the game were increased by a factor of 100 it would still be more than twice as rare as it was in the CMx1 games.

[Offshoot]

At the very least, the way the Panther mantlet is modeled in the game will greatly decrease the granularity of results vs. what would be seen with a rounded object.

Only if what we see in-game is also what a shell sees when it strikes. As Steve has since mentioned that "rounded surfaces (even as far back as CMx1) have special attributes to influence outcomes due to the unique nature of rounded surfaces" my guess is that this is not necessarily always the case. It is possible that a different model to the polygon one is used to do the calculations.

[sfhand]

BTW, all this talk about it being "rare" is missing the point. No one is disputing that it was rare, just not that rare. To put it into perspective, if the incidence of ricochet penetrations in the game were increased by a factor of 100 it would still be more than twice as rare as it was in the CMx1 games.

Haven't you already said no one knows how rare it was? If no one knows how rare it was how do you know CMx1 got it right?

AFAIK, CMx1 didn't use a physics engine to model the shot trap and CMx2 does. This is leading back to the "design for effect" vs "design by what ever the physics engine model is called" discussion. For me, your tests have shown the shot trap hit is possible, which we know matches reality. What we don't know is the probability of a shot trap hit which is what you seem to want to change.

In essence, you seem to be advocating for a change to the game without providing any evidence that the change would make the situation more realistic. Compare this to the CMFI Italian MG's where someone presented videos of the guns firing and reloading, after which the firing and reloading characteristics of those MG's were changed.

[c3k]

The hypothetical you describe would still be a band rather than a single point.

VaB,

I am actually sympathetic to your argument. However, calling a hypothetical 1^ angle of incidence requirement a "band" is stretching it a bit.

My point wasn't whether or not it was a band or a singularity, rather that the extent of the band is unknown. It is also unknown whether the band size is a constant for different rounds, or for the same round at different velocities.

Panther crews were confident enough in their vehicles to face Allied armor.

Allied armor, for the most part, considered the Panther invulnerable to the fornt.

Many Panthers were destroyed. Many through side and rear hits, many through self destruction for various reasons. Some through frontal penetrations. (Stop it!). A very low subset of that suffered from the ricochet effect.

How many thousands of rounds had to be fired at a Panther to get one ricochet into the hull roof?

(Likewise, if you told me the odds of two bullets hitting one another in a battle, I'd show you the various Gettysburg displays showing just that. Should that be modeled in a Civil War tactical game? Would it make a difference one way or the other? Not the same: I understand. (A Panther ricochet by an otherwise underpowered gun can be quite the game changer.))

Until there is hard data on real life ballistics, we're all just guessing. (JasonC's maths nothwithstanding.)

Ken

[c3k]

These speculations seem rather pointless when we already know that the way the shot trap is modeled in the game is physically different than reality.

We don't actually know this.

Why do you think the game differs from reality?

Ken

[Battlefront.com]

Since it hasn't been said in a while, CMx1 is not to be used as a measuring stick. That was "design for effect" when it came to something like this.

Like C3K is saying, I'm actually sympathetic to the argument that CM doesn't have it quite right yet. Otherwise I'd not be participating in this thread. But I feel I'm being asked to take a leap without any indications how far down it is to the bottom, or how far across the landing is. So until I've got better information to go on, I'm staying firmly planted on the ground.

I've already asked Charles to look into this and see if there's an obvious problem. He says there isn't. That's all I can do without a decent, sound argument for change and the circumstances/parameters for that change.

To restate one possible explanation for what were seeing...

It is possible that for 1943 through early to mid 1944 the shot trap was only a theoretical problem or perhaps a very unlikely one given the range of weapons that were firing at Panthers in that timeframe. But then in mid 1944 Panthers started coming into contact with something that changed the dynamic. So much so that by the end of 1944 the Germans had rushed a solution into production to help reduce the chances of this problem happening.

When you think of the changes in weaponry three guns come to mind:

1. Sherman 76

2. Firefly 17pdr

3. Soviet 85mm

You obviously can't test for #3 yet, but the other two can be tested now. So I recommend trying some more tests with everything the same as before except the firing vehicles being swapped out with Sherman 76s and Fireflies. If the results are exactly the same as before, or they are simply cutting through the frontal armor, then that will disprove my theory (though range could still very much be a factor). But if they start bouncing rounds into the deck... that tells us something very different.

Steve

[Freyberg]

Did Allied gunners aim for this point? If so, is this modeled in the game?

[Vanir Ausf B]

Haven't you already said no one knows how rare it was? If no one knows how rare it was how do you know CMx1 got it right?

AFAIK, CMx1 didn't use a physics engine to model the shot trap and CMx2 does. This is leading back to the "design for effect" vs "design by what ever the physics engine model is called" discussion. For me, your tests have shown the shot trap hit is possible, which we know matches reality. What we don't know is the probability of a shot trap hit which is what you seem to want to change.

In essence, you seem to be advocating for a change to the game without providing any evidence that the change would make the situation more realistic. Compare this to the CMFI Italian MG's where someone presented videos of the guns firing and reloading, after which the firing and reloading characteristics of those MG's were changed.

First of all I never claimed CMx1 got it right. I was simply using it to illustrate that there are varying levels or rarity, and how vast the differences with CMx2 really are.

Second, while it would be fair to say there is no proof that a change would be more realistic, to claim there is no evidence to that effect is not. Unless you don't consider two US tests that produced shot trap ricochets to be evidence?

[Vanir Ausf B]

Only if what we see in-game is also what a shell sees when it strikes. As Steve has since mentioned that "rounded surfaces (even as far back as CMx1) have special attributes to influence outcomes due to the unique nature of rounded surfaces" my guess is that this is not necessarily always the case. It is possible that a different model to the polygon one is used to do the calculations.

While this was certainly true in CMx1 it has always been my understanding that in CMx2 the physical vehicle model is used for ballistics modeling. I am not aware of any exceptions to this, but am open to correction if it has been stated that there are.

[Vanir Ausf B]

VaB,

I am actually sympathetic to your argument. However, calling a hypothetical 1^ angle of incidence requirement a "band" is stretching it a bit.

My point wasn't whether or not it was a band or a singularity, rather that the extent of the band is unknown. It is also unknown whether the band size is a constant for different rounds, or for the same round at different velocities.

Choose any angle or range of angles, no matter how narrow or wide, and it will be the same across the face of the mantlet on the x axis, assuming the projectile approaches from straight-on. That is what I mean by a band. If there is a significant lateral offset then that is different because compound angles come into play. But that is not a favorable scenario for ricochets into the deck.

How many thousands of rounds had to be fired at a Panther to get one ricochet into the hull roof?

We know of one US test that required 8. Another required around 80 or so, IIRC. Neither needed anything approaching thousands. That this is suggestive rather than definitive is given, but it is very suggestive.

Until there is hard data on real life ballistics, we're all just guessing. (JasonC's maths nothwithstanding.)

There are educated guesses and wild-ass guesses. Let's not lump them together.

[Vanir Ausf B]

We don't actually know this.

Why do you think the game differs from reality?

If the geometry of the Panther mantlet is WYSIWYG then it is a certainty, for the obvious reason that the real Panther mantlet was a round object rather than 8 flat polygons joined together.

[Vanir Ausf B]

Since it hasn't been said in a while, CMx1 is not to be used as a measuring stick. That was "design for effect" when it came to something like this.

Like C3K is saying, I'm actually sympathetic to the argument that CM doesn't have it quite right yet. Otherwise I'd not be participating in this thread. But I feel I'm being asked to take a leap without any indications how far down it is to the bottom, or how far across the landing is. So until I've got better information to go on, I'm staying firmly planted on the ground.

I've already asked Charles to look into this and see if there's an obvious problem. He says there isn't. That's all I can do without a decent, sound argument for change and the circumstances/parameters for that change.

Thanks for looking into it Steve. :cool:

It is possible that for 1943 through early to mid 1944 the shot trap was only a theoretical problem or perhaps a very unlikely one given the range of weapons that were firing at Panthers in that timeframe. But then in mid 1944 Panthers started coming into contact with something that changed the dynamic. So much so that by the end of 1944 the Germans had rushed a solution into production to help reduce the chances of this problem happening.

There is no reason that I am aware of to believe that ricochet penetrations were a phenomenon associated with any particular gun or round. As Amizaur pointed out in his excellent post a few pages back, any AP round that fails to penetrate is either going to disintegrate or ricochet. There are no other options. The greater the slope or curvature of the object struck, the greater the chance of ricochet vs. breakup.

The fact that the Germans didn't do anything about the shot trap until fall of '44 is more likely a reflection of the fact that the Panther did not comprise a large portion of the German armored inventory in 1943 and early 44 and was a new vehicle with relatively little combat time.

In any case, given the huge time sink this testing is I am not inclined to jump into a new round of it unless I have a reason to believe that the results may make a difference in the game, and at this point that seems unlikely.

[Vanir Ausf B]

Did Allied gunners aim for this point?

Anecdotally, yes.

If so, is this modeled in the game?

No

[c3k]

8 faces of death: or, is the panther mantlet modeled in steps, not as a curve.

I understand the question posed by this. (If you look closely, there are no curves anywhere in the game. Everything is rendered with straight line segments. The Panther mantlet is shown this way as well. It has 8 (if I counted right) segments. (Or, thinking of it as a part of a wine or beer barrel, 8 staves.)) The solution, as discussed upstream, would be to trace the arc of EVERY upward deflecting round fired from the same location and see if they follow 8 paths, or individual paths.

If they fall into 8 paths (actually 4, due to the 1/2 which would deflect upwards), then the game models the mantlet as shown, rather than as it was. If that is true, then the downward deflection model may be broken.

VaB has done a TREMENDOUS amount of work on this. I, for one, applaud you for it. Sincerely.

Ken