M10s beating Tiger I at 800 meters.

[Redwolf]

Do you have a source for that? Because that is far different than what I just posted.

Mumble need to check my math mumble later probably incorrect mumble.

In any case, whichever two single distances you choose BFC could have picked different values, then there could be scatter modeled but not show up in your test.

[Guest]

I don't claim to be an expert on the subject, but based on my rudimentary understanding of shatter gap I don't think I am seeing evidence of it in the game.

Well, I don't claim to be an expert on the subject either, but it's there in black and white. I don't see anything that would indicate it's not being used. If a beta tester wants to take it up and see if it needs to be tweaked or something, definitely be my guest.

There is a much higher proportion of hits that do damage to the "Front Turret" than there is to the mantlet, so we know there is an actual difference in the armor thickness. It cannot be explained away as simply a labeling issue.

Not so much a labeling issue as a "whole ton of variables" issue, which is neither simple nor explaining away. I do think your tests deserve a look. I don't think you can or should try to draw much hard evidence from hit location labels. You're talking about a huge event, lots of energy being thrown around, lots of bits and pieces touched or damaged, quite a lot of which is tracked in CM. You can't take *one* pretty broad hit location from that set, marry it up with a frankly also fairly ambiguously externalized damage model, and draw real conclusions about what's happening.

So... yes. Labeling picks one variable from a whole heap of them. It's meant (in my opinion) to be more flavor information than precise, perfect, put-it-in-a-book data. It's not an "issue", per se, unless you're trying to do that.

[Vanir Ausf B]

Well, I don't claim to be an expert on the subject either, but it's there in black and white. I don't see anything that would indicate it's not being used. If a beta tester wants to take it up and see if it needs to be tweaked or something, definitely be my guest.

I'm not sure what you mean by it's there in black and white. Are you referring to the manual (or would that be brown and dark brown)? From what I understand, if shatter gap were in effect there would be few if any penetrations at 300 meters as the shells would completely fail, since at that range it falls into the "shatter fail" portion of the diagram I posted. In testing most hits penetrate, albeit usually partially, which is more in-line with what I would expect in that situation with "normal" shells that were not flawed.

If anyone can come up with another specific test that would demonstrate shatter gap in effect I'm all ears.

EDIT: I just thought of a way to do a control group test. I'll be back...

So... yes. Labeling picks one variable from a whole heap of them. It's meant (in my opinion) to be more flavor information than precise, perfect, put-it-in-a-book data. It's not an "issue", per se, unless you're trying to do that.

Location labeling may be flavor information, but penetrations are not. At least, I don't think the crews of my Tigers would call it flavoring But I'm glad to hear you think it deserves a look.

[Vanir Ausf B]

Ok, for a control group I tested the Tigers against panzer IV Hs (late). The German 75L48 should have almost the same penetration vs. RHA at 600 meters as does the US 3" at 300 meters.

Penetration

KwK 40 L48: 123 @ 500m, 116 @ 750m, so about 120 at 600m

3" 76mm: 121 @ 250m, 115 @ 300m, so should be about the same.

The obvious difference is that the German rounds are not expected to suffer shatter gap.

Tiger I mid vs Panzer IV H late @ 600m and 10° lateral offset (roughly)

34 hits

No damage: 0

Spalling: 1

Partial penetration: 24

Penetration: 10

By comparison here is the 3" @ 300m result from earlier:

No damage: 0

Spalling: 5

Partial Penetration: 21

Penetration: 7

Not much difference. Again, I see no evidence of shatter gap in effect with regards to US 3" APCBC. Maybe the tests are set up wrong, or maybe I've misunderstood the source material or maybe the source material is wrong. Or maybe there's a bug.

[Guest]

I'm not sure what you mean by it's there in black and white. Are you referring to the manual (or would that be brown and dark brown)?

No... the code.

From what I understand, if shatter gap were in effect there would be few if any penetrations at 300 meters as the shells would completely fail, since at that range it falls into the "shatter fail" portion of the diagram I posted. In testing most hits penetrate, albeit usually partially, which is more in-line with what I would expect in that situation with "normal" shells that were not flawed.

Per JasonC's earlier comment:

"It doesn't mean deterministic failure to penetrate starting at 450 yards. It does mean failures become possible in that range window and the most common outcome by the end of it."

Failures are definitely possible - they're definitely modeled, and as mentioned I don't see any obvious bugs that would prevent that modeling from coming into play. If you want to start looking at whether they meet specific percentages of outcomes you're going to need much larger sample sets.

EDIT: I just thought of a way to do a control group test. I'll be back...

It's a good test. Too small as yet, of course. I disagree that there's not much difference between the two sets, but the numbers are too small to make a call one way or the other really.

Location labeling may be flavor information, but penetrations are not. At least, I don't think the crews of my Tigers would call it flavoring But I'm glad to hear you think it deserves a look.

I do try to look at reasonable, well-thought-through bug reports, especially with data to back them up. Penetration location info is just as susceptible as the other stuff, though. Penetrations necessarily do their thing with LOTS of bits and pieces. We can look at things more deeply on our end, though.

[Vanir Ausf B]

Per JasonC's earlier comment:

"It doesn't mean deterministic failure to penetrate starting at 450 yards. It does mean failures become possible in that range window and the most common outcome by the end of it."

I don't know where Jason got his numbers, but the source I used says differently. Again, I point you towards the diagram I posted. In some circumstance we would indeed expect to see a mixture of penetrations and failures, but in others we would expect to see almost all failures. The 300m test is squarely in all failure territory. Which means...

Failures are definitely possible - they're definitely modeled, and as mentioned I don't see any obvious bugs that would prevent that modeling from coming into play. If you want to start looking at whether they meet specific percentages of outcomes you're going to need much larger sample sets.

Larger sample sets will show us nothing we don't already know, because the expected failure rate in the 300m test is near 100% and our results are vastly different from that. A billion more samples will not change that.

[Vanir Ausf B]

I went ahead and expanded the sample size just for the hell of it. Totals now are:

Tiger I mid vs M10 TD @ 300m and 10° lateral offset (roughly)

Hits: 115

No damage: 0

Spalling: 16 (14%)

Partial Penetration: 73 (63%)

Penetration: 26 (23%)

Tiger I mid vs Panzer IV H late @ 600m and 10° lateral offset (roughly)

Hits: 109

No damage: 0

Spalling: 5 (5%)

Partial penetration: 79 (72%)

Penetration: 25 (23%)

Again, the same general pattern. Full penetrations are identical, but there does seem to be a small but probably statistically significant difference in the proportion of spalling to partial penetrations. I don't know if this is our elusive shatter gap showing up in a very subtle way or if it's due to some other factor perhaps arising from the tests not being exactly apples-to-apples.

I doubt I'll do any more with this since it's not going to make any difference anyways.

[Fizou]

Thanks for doing these tests and hunting bugs Vanir. Such dedication to CM

[Sgt Joch]

I don't know where Jason got his numbers, but the source I used says differently. Again, I point you towards the diagram I posted. In some circumstance we would indeed expect to see a mixture of penetrations and failures, but in others we would expect to see almost all failures. The 300m test is squarely in all failure territory. Which means...

You seem to be basing your info on Rexford's book. Rexford has done some interesting research, but not everyone agrees with his conclusions. Everyone agrees there was a "shatter gap", but not everyone agrees it can be reduced to a simple formula.

For example, Rexford's theory of "shatter gap" against Tiger/Panther armor is based on a very small number of tests:

http://www.battlefront.com/community/showthread.php?t=990

You also have this classic exchange between JasonC and Rexford:

http://www.battlefront.com/community/showthread.php?t=898

I tend to agree with JasonC that actual combat shows better result for 76 mm ammo than Rexford is willing to admit.

There is also this interesting article which summarizes the 1944 live firing tests:

http://worldoftanks.com/en/news/21/chieftains-hatch-us-guns-vs-german-armour-part-1/

[Guest]

Larger sample sets will show us nothing we don't already know, because the expected failure rate in the 300m test is near 100% and our results are vastly different from that. A billion more samples will not change that.

What I would hope to see from more samples would be evidence of shatter modeling - that there isn't a bug in the model that's preventing shatter from occurring at all. Like I said, I don't see anything that would indicate this, but tests is tests. If all you want to see is a 100% failure rate for ALL of your shot setups then... well, that's not what the data says.

Rexford's conclusion was well-researched, but arguable, based on limited samples, and certainly NOT universal in applicability. We use all of the sources you guys have access to (and some that might be harder to get, too!). If there were some universally-accepted conclusion that 3" rounds of a particular type should shatter 100% of the time at particular ranges it'd be in the game. As it is we have to attempt to model reality, which is a lot more nuanced than taking just one source as gospel... I bet Charles wishes there were just one universal source for all this info.

[Redwolf]

You seem to be basing your info on Rexford's book. Rexford has done some interesting research, but not everyone agrees with his conclusions. Everyone agrees there was a "shatter gap", but not everyone agrees it can be reduced to a simple formula.

For example, Rexford's theory of "shatter gap" against Tiger/Panther armor is based on a very small number of tests:

http://www.battlefront.com/community/showthread.php?t=990

You also have this classic exchange between JasonC and Rexford:

http://www.battlefront.com/community/showthread.php?t=898

I tend to agree with JasonC that actual combat shows better result for 76 mm ammo than Rexford is willing to admit.

There is also this interesting article which summarizes the 1944 live firing tests:

http://worldoftanks.com/en/news/21/chieftains-hatch-us-guns-vs-german-armour-part-1/

So what data do you use to model the scatter gap?

If you model the gap it has to be somewhere in the ranges.

[JasonC]

Vanir - my estimate of where shatter happens is not based on a mathematical ratio, but on after action reporting by US tankers and TD crews. Many of them are reports vs the Panther turret front, not the Tiger, but both have similar ~100mm resistence. The M10 crews report no difficulty whatever penetrating the Panther turret front at 400 yards or less, with numerous examples of successful penetrations in Normandy and in the Lorraine, all before they got any APCR to speak of. Terrain made initial ranges that low in Normandy, and fog did so in much of the Lorraine fighting.

At 800 yards, I would expect shatter to be the normal outcome, in line with crew reports from Anzio and longer range fights in Normandy and the later ETO etc. I would not expect it to be the deterministic outcome, simply because shell failure is a far from deterministic process and could vary with a lot of the details.

[c3k]

VaB,

Good job. I'm curious about the damage level sustained on the various partial penetrations. If there is a significantly lesser level of damage w/3" hits, perhaps that is the shatter gap? (Of course, true "shatter gap" would not even have a partial penetration.)

Thanks for following up on these tests.

Ken

[Sgt Joch]

its always hard to know how to interpret the results.

Based on Vanir's tests, M10 firing on a Tiger at 300 meters gets 23% full penetration and 63% partial penetration in game.

In the may 44 Shoeburyness test, 76 mm AP M62 ammo achieved full penetration of a 100 mm steel plate at a 30 degree angle at 500 meters every time with no evidence of failure, although it looks like only two rounds were fired.

so, is the game representing 76 mm AP M62 ammo fairly? Hard to say.

[ridethe415]

so, is the game representing 76 mm AP M62 ammo fairly? Hard to say.

Well I just lost a Tiger from an M10 at ~837M (frontal hit) in The Forest of Wild Beasts. So something is being represented.

[Vanir Ausf B]

A quick update, as that is all I have time for. I did go back and do quite a bit more testing and I think I may have found the elusive higgs bo-, er, I mean shatter gap. I moved the Tiger test back to 500 meters and again used the Pz IV as a control. The results showed a very significant difference in favor of the Pz IV. I then tested at 800 meters and saw an even bigger difference, larger than I can explain by any factor I am aware of other than shatter gap.

I then was getting a little excited and started testing against the Panther mantlet at 500m, which should be a slam dunk shatter gap scenario. I'm not done yet so the sample sizes are too small to be definitive but so far am see virtually identical results between the 2 tests, which is not at all what I expected after the Tiger tests and is a little disappointing. I'll finish up the testing tomorrow and post the results.

[Redwolf]

Well I just lost a Tiger from an M10 at ~837M (frontal hit) in The Forest of Wild Beasts. So something is being represented.

Study with sample size of 1?

[ridethe415]

Study with sample size of 1?

If it were real life, a sample size of one would be enough for the crew

[JasonC]

Vanir - on AAR evidence I think 300 meters is too close for shatter gap to be a serious issue yet. If you are seeing significantly more failures at 500, that sounds about right, and lots of them at 800, again what I would expect.

[Sgt Joch]

I have been running my own series of informal tests. Not as scientific as Vanir though. Basically a series of 4 v 4 fights, various US/CW AFVs vs Panthers, front to front, all regulars, 600 meter range, in CMFI and CMBN.

The tests never last long since the Panthers generally wipe out whatever is placed before them in short order. They remind a lot of M1 Abrams in CMSF.

Even when the Panthers are ordered not to fire, it is very difficult to kill them. Even 17 pdr APCBC shells from CW fireflys generally tend to just bounce off from the front lower hull.

Given the general lack of hard data, it is difficult to conclude there is anything wrong.

[Redwolf]

The tests never last long since the Panthers generally wipe out whatever is placed before them in short order.

Remove main gun ammo in the editor?

[womble]

Remove main gun ammo in the editor?

Make them fanatic and give them short covered arcs. They'll squirm a bit under fire but probably won't shoot back.

[Sgt Joch]

Even with regulars/normal motivation, they will not fire if you give them short cover arcs/hide order.

Ran a few more tests, july 44, sherman fireflys/US 76(w) shermans at 200/400 meters. Both can kill a panther from the front.

U.S. 76 mm APC M62 will penetrate the lower hull Panther armour at 400 meters, although most shots bounce off. The 17 pounder APCBC shell achieves more reliable penetration, although again many shots bounce off. It looks more or less in line with the 1944 test results.

[JasonC]

Sgt Joch - the lower hull hit reports are interesting, but what about the main event - the front turret hits? That is where most of the kills should actually come from.

[Vanir Ausf B]

I do intend to finish up my testing soon. I've been busy with other things. In light of the lack of shatter gap against the Panther mantlet I'm going to expand the testing to include the front turret armor as well.

EDIT: Note that most US 76mm hits on the mantlet do fail to penetrate at 500 meters, but so far this appears to be due to the rounded shape of the armor rather than shatter gap.