Panther Shot Trap Still Not Trapping

[Skwabie]

We need a Combat Mission Game Database Viewer.

Enough of this countless testing. It is tiring, time consuming, and at the end of day still unconvincing to the average joe.

Enough of the repeated request on gun vs armor penetration tables. Even there is one, it still is an indirect reflection of how the game mechanism works.

Yes, a database VIEWER, NOT editor. So we can see what values are assigned to each unit. But without the ability to modify anything.

[Vanir Ausf B]

BOOYAH

Told ya so.

But did anyone believe me? No one ever believes me...

I have to wrap up the test, then I'll post final results. But this should change the discussion, I think.

[sfhand]

It is interesting that this ricochet would, in fact, be blocked by the mantlet "chin". The statement about hitting below the transverse centerline notwithstanding, I wonder if the shot trap only works if the shell hits the bottom 4" or so of the mantlet.

Edit: it also took out a chunk of the turret on its way through the hull roof.

These were my thoughts as well when I saw the picture.

[Vanir Ausf B]

Cromwell VII vs. Panther D at 100 meters. Panthers are partial hull down behind 1 meter berm.

Total non-penetrating hits ON THE MANTET ONLY: 1214

Forward Top Hull Hit: spalling: 2

RICOCHET INTO: Forward Top Hull: partial penetration: 1

That last number is the only one that really counts. Because it now seems almost certain that those hits on the forward top hull that we have been counting as ricochets despite having the same hit text as direct hits actually are direct hits.

So, we have been working off the idea that in the previous test I got 12 ricochet penetrations out of 1096 mantlet hits, or 1.1%. It now seems the actual number of ricochet hits was zero. If we add the 1096 hits to the 1214 in this test we get 1 ricochet penetration out of 2310 total mantlet hits.

That is 0.04 percent.

[c3k]

Congratulations! You've spotted the unicorn.

- Chin models: oops, I must've been thinking about a different CM game.

- Lower 1/4 of the mantlet vs. lower 1/2 of mantlet: where did I hear that before?

- Change weapon and elevation: did someone say that the ricochet effect may only be seen at very specific ranges/engagement characteristics?

Now that you've found what works, I'd be hesitant to include the non-working tests as part of a baseline.

Ken

[Vanir Ausf B]

Unless someone can give a coherent reason why the "non-working" test shouldn't work I don't think it should be jettisoned just because it has so far produced less happy results. In terms of real world combat situations the 800 meter test represents a far more likely scenario than the 100 meter.

As for running more tests with random combinations of gun and elevation changes just to see what happens, no thanks The goal is to see if the shot trap is working properly, not to give me carpel tunnel syndrome. There is no reason to suppose that minor shifts in elevation will dramatically change results, and large changes will obscure the shot trap area. As for guns, based on what we know about characteristics that affect ricochet likelihood (or think we know given that it is all theoretical), I'm not aware of an AP round in CMBN that would be a markedly better ricochet candidate than the 75mm I've been using. However, if you have access to a certain unreleased game, uncapped Soviet 122mm AP might be a good choice. Although, doing these tests with tanks as slow firing as the IS-2 would be agony.

EDIT: On second thought, 122mm would probably go right through the mantlet unless the range was very long (2000+ meters).

[Doug Williams]

Thanks for finding out that it actually does happen, though very rarely. It's this kind of attention to detail that warms a grog's heart. :-)

A 0.4% of this happening isn't going to stop me from buying Panthers or change how I use them in the game, however.

[Sgt Joch]

EDIT: On second thought, 122mm would probably go right through the mantlet unless the range was very long (2000+ meters).

nice thing about the Russian "heavies", they slice right through German armour, no need to worry about the Shot Trap.

[Amizaur]

Let me repost the picture of the Panther ricochet:

The actual height of the hit was, I believe, about the red line.

Now if the shell hit a little bit lower, it would ricochet into the front turret plate and then maybe into the roof, but that would be second ricochet and the shell would retain maybe 10% of initial energy. So I believe the blue line is the lowest possible height of hit that is likely to ricochet into the hull roof.

If the shell hit above the orange line, it would probably spend most of it's energy on deformation of the mantlet armor, then bounce back, missing the roof. The red line is what I believe highest where ricochets into the hull roof are likely.

Now if the shell hit somwhere in the middle, between the orange line and blue line, chances for ricochet into the hull roof are in my opinion quite high - maybe as much as 50%. The shell HAS TO ricochet somwhere, and it would be most likely in the direction of the hull top armor - because why not? The shell would probably retain much of it's energy, maybe enough to defeat the top armor even hitting it sideways.

IIRC the shell that penetrated the hull of Bovington Tiger possibly defeated the top hull going sideways. Look at the shape of the patch: http://www.tiger-tank.com/secure/journal36.htm

The shell penetrated 25mm of hull's top after three ricochets (!) - but each one deflected the shell's trajectory only by small angle, so it retained most of it's energy.

The Panther has only 17mm of hull top armor.

If the chances for such ricochet were in order of 0.1% then I doubt the Germans would bother modifying the mantlet in G model. They lost quite a few Panthers in this way on the East.

[JasonC]

BG - "that's a fairly subjective assessment, Jason."

I deny it - anywhere I see a p value over 0.05 I assume it could easily come about from random chance. Yes there might be cases in which I would want an even lower threshold, but 7% is high enough that anyone claiming it as statistically significant would have a lot of explaining to do. In fact any journal would reject a paper whose core claim involved pretending that something that had a 7% chance of happening randomly could not have happened by chance.

Is it significant at the .01 level or at the .001 level are additional questions, and sometimes one might want to require that. But if it isn't significant at the .05 level it flat out isn't significant. And that is what the hypothesis test I did, said - if you did have a 2.5% chance per shot and you saw a sample of 8 shots with 1 success in it, then no you couldn't take it as proof that your 2.5% estimate was too low. Your 2.5% estimate could be perfectly accurate and that could still happen. But if your hypothesis was that the prior chance was only 0.1%, then you can take that 1 in 8 sample as evidence that your estimate is almost certainly too low. You are as likely to win the daily lottery as see that outcome, if that were the right prior probability.

[Sgt Joch]

Now if the shell hit somwhere in the middle, between the orange line and blue line, chances for ricochet into the hull roof are in my opinion quite high - maybe as much as 50%. The shell HAS TO ricochet somwhere, and it would be most likely in the direction of the hull top armor - because why not? The shell would probably retain much of it's energy, maybe enough to defeat the top armor even hitting it sideways.

Good analysis, but that is only part of the equation. Not all shells which hit in the "zone" will ricochet. That depends on many factors: angle of trajectory, energy/velocity, shell construction, etc. Some shells will penetrate partially/fully, some will break up, some will ricochet.

I suspect the % of shells hitting the "zone" that would ricochet into the hull top is substantially less than 50%.

That may also partly explain why the Shot Trap problem seemed to be more prevalent at certain ranges. Those ranges probably corresponded to the shells coming down more frequently at a certain angle/energy state which increased the likelihood of a ricochet.

[JasonC]

"why the Shot Trap problem seemed to be more prevalent at certain ranges"

Cool, that suggests you have lots of data on the subject. What is it and where? What is the sample of shot trap shots and their associated firing ranges, behind the above statement?

[Sgt Joch]

"why the Shot Trap problem seemed to be more prevalent at certain ranges"

Cool, that suggests you have lots of data on the subject. What is it and where? What is the sample of shot trap shots and their associated firing ranges, behind the above statement?

no, I guess you will have to go back to counting pixels on a screenshot. Much more scientific.

[BletchleyGeek]

Cromwell VII vs. Panther D at 100 meters. Panthers are partial hull down behind 1 meter berm.

Total non-penetrating hits ON THE MANTET ONLY: 1214

Forward Top Hull Hit: spalling: 2

RICOCHET INTO: Forward Top Hull: partial penetration: 1

That last number is the only one that really counts. Because it now seems almost certain that those hits on the forward top hull that we have been counting as ricochets despite having the same hit text as direct hits actually are direct hits.

So, we have been working off the idea that in the previous test I got 12 ricochet penetrations out of 1096 mantlet hits, or 1.1%. It now seems the actual number of ricochet hits was zero. If we add the 1096 hits to the 1214 in this test we get 1 ricochet penetration out of 2310 total mantlet hits.

That is 0.04 percent.

Congratulations Vanir

[c3k]

I'm glad that VaB has found that the game does, indeed, replicate this effect.

JasonC is on a quest to show that it should happen...more often.

I don't see how you can even start to question whether the game models the frequency correctly.

Given: No one knows what conditions in real life created this effect.

Given: No one has tested the conditions in the game which create this effect.

Ergo, you cannot compare an unknown to an unknown and declare that one of the unknowns is wrong.

Now, obviously VaB has (through persistence, diligence, hard work, etc.,) found ONE condition which works. ("Work" defined as a specific weapon, range, and aspect ratio which creates an in-game ricochet/roof penetration.) What about using the SAME weapon, range, and aspect ratio, but with a DIFFERENT ammunition? Should it work? Shrug.

Now, let's start moving the aspect ratio: Slightly askew, rather than nose-on. Should the compound angle increase or decrease the odds of the rooftop penetration? (Remember, this is not an intuitive problem. If it were, do you think the Germans would've designed the mantlet that way in the first place?)

Next, let's change the range. And the angles.

Oh, we haven't mentioned different elevations of target/shooter.

Now let's change the weapon...

In short, VaB's work has shown a data point. (And some null points.) Great. Sincerely.

But to what do we compare it to? The Isigny test? Okay, tell me the muzzle velocity of the round fired, the range, the aspect ratio, the relative heights, the armor hardness, etc. All of these are variables. "We shot it from 800m." Really? Not 832m? Nor 787m?

Does that make a difference? I don't know. Perhaps that 800m ricochet was an extreme outlier.

There is no information presented which describes the real life vulnerability. Only anecdotes. So far, the game physics ALLOW this behavior. It's just not where you thought it should be. Okay, defend why it should be the way you want it to be.

Ken

[akd]

As has already been pointed out, if it were the way it is, it is highly unlikely that Germans would have recognized this as a problem and even more unlikely that the Allies would have recognized it as a vulnerability. Vanir has probably already achieved more shell to Panther mantlet interactions than occured during the course of the war.

[JasonC]

c3k - the can't know anything fallacy, and all nonsense. We know that minor changes in the range, velocity, and angle will have no effect on the correct prior probability of the event. They are the sample that the probability is meant to blend in the first place.

[Wodin]

I love this forum... You don't get threads like these in many other places.

[mjkerner]

I love this forum... You don't get threads like these in many other places.

I was thinking the exact thing.

[c3k]

c3k - the can't know anything fallacy, and all nonsense. We know that minor changes in the range, velocity, and angle will have no effect on the correct prior probability of the event. They are the sample that the probability is meant to blend in the first place.

LOL! You can't scare me with those words!

[Vanir Ausf B]

I don't see how you can even start to question whether the game models the frequency correctly.

Given: No one knows what conditions in real life created this effect.

Given: No one has tested the conditions in the game which create this effect.

Ergo, you cannot compare an unknown to an unknown and declare that one of the unknowns is wrong.

As Jason pointed out it would be a fallacy to suggest that because we don't know everything we know nothing, with the follow-on fallacy that therefore all possibilities are equally likely. That we cannot nail down the real life probability to a single certain number does not mean we cannot make a reasonable estimation of ranges and rightly view with suspicion results that lie greatly outside of that range. BletchleyGeek, Jason and myself have all done this, and although our numbers are not the same due to differences in assumptions and methodologies even the most pessimistic are larger than the test results by orders of magnitude.

There is no reason to suppose that jiggering the range, angle or gun would increase the number of positive results. It would be more likely to lower it. I am not aware of any commonly used AP round in the western allies' arsenal during the CMBN timeframe that has a length to diameter ratio more favorable than the 75mm I have been using. Increasing the targets' elevation relative to the shooter will cause the top edge of the glacis plate to obscure the lower mantlet. Decreasing the target's elevation exposes the deck plate to direct hits, which makes any concern about ricochets irrelevant. Slewing the target to the side just increases the deflection angle along the horizontal plane which would make ricochets onto the deck less likely.

At 800 meters the angle of decent for the shell is about 1°, give or take a few tenths, so that is not a factor. Range does affect velocity which will affect residual energy after a ricochet, so the 100m test may be marginally more favorable for that reason only. But it is concerning that even the most favorable test scenario we can devise only produced one unicorn out of over 1200 attempts.

[Vanir Ausf B]

LOL! You can't scare me with those words!

Your endless requests for more testing scare me

[akd]

Why did the Germans continue to demand zimmerit?

Jumping back a bit here to address a possible misconception, but the Russians captured huge numbers of German magnetic mines (Haft-Hohlladung) and were able to issue them basically as standard equipment.

[John Kettler]

akd,

News to me re Russians armed with captured German magnetic antitank mines. Where did you read about mass captures of such German antitank munitions? I've not only read of but seen footage of Russian troops using Panzerfausts, but have seen neither stills nor footage of Russians armed with ex-German hollow charge magnetic antitank mines. I always understood the Germans were being proactive in developing Zimmerit because they saw the potential havoc such weapons could wreak on their armor and developed a counter.

Regards,

John Kettler

Re

[Sgt Joch]

At 800 meters the angle of decent for the shell is about 1°, give or take a few tenths, so that is not all.

I was wondering about that, do you have a ballistics trajectory chart for 75 mm ammo? can't seem to find one on the web.