Armor Penetration Curves

[fnord_disc]

Did you mix up the graphs or do the 406 mm shells of the Große Currywurst really have better penetration than the 420 mm shells?

That would make the 420 mm shells entirely pointless outside of drug references, especially considering they increase reload from 29 to 32 seconds.

 

Does the 10 m/s loss of speed really make so much more of a difference than the 190 kg increase in mass?

 

Well, the 420 shell retains more penetration. Check the ranges 12km+.

 

Speed loss is probably more or less offset by the mass increase, but a bigger diameter also reduces penetration.

 

In my opinion their penetration is more or less the same and it comes down to damage/reload and maybe overmatch if someone finds a 29mm armor plate.

 

 

I can tell from my perspective, that I've surely not expected such curves and they may not matter that much in game. 

 

As we can see the russian 152 mm AP shell is vastly superior over the German 150 mm in terms of penetration, yet I struggle to citadel cruisers with Budjonny in the very same situations where I deal multiple citadels in one salvo in the Königsberg. It's not even close, though it should be. I can't offer any explantion either.

 

Maybe I made a mistake, who knows? You could check using your Königsberg and make a training room, but I don't have that ship anymore.

 

Technically the Königsberg should have pretty good penetration, but the krupp value is so low.

[Namolis]

Fixed plots for Dunkerque, Scharnhorst and Bayern because of wrong drag.

 

As expected based on the previous plot, Bayern has very bad drag and penetration quickly goes to the dogs.

So, according to Lesta, the Germans already had a shell with 0.3093 drag coeff. and then goes ahead and "upgrades" it to a larger version with drag coeff. = 0.4092. Because, you know, its tech tree competitors already have 50% more guns, so clearly its bigger gun needed worse pen too. That, right there, is some world class logic for you.

 

(yes, for anyone wondering; that number is shape only. It doesn't scale with size).

 

Also, Nassau at 0.5337.. I found a paper showing that a modern 130 mm shell has a Cd of ~0.23 in normal configuration and ~0.63 when fitted with a "recovery plug" (essentially modifying the tip to intentionally sabotage its ballistics as much as possible. They want to prevent the shell from burying too hard into the landing area so it can be recovered intact for study). Nassau is closer to the latter. Clearly it had its shell designed by traitors and saboteurs.

 

 

Anyway, thanks for clearing it up for me

[Captain_Hook_]

I asses the calculation was done with F-Formula. In the case of german projectiles(ww2) the formula underestimates performance by about 10-15%.

Otherwise excellent work  :thumb up

[Mucker]

 

Maybe I made a mistake, who knows? You could check using your Königsberg and make a training room, but I don't have that ship anymore.

 

Technically the Königsberg should have pretty good penetration, but the krupp value is so low.

 

I have no reason to think the numbers are wrong, I just wanted to say that one should not judge a guns capabilities purely on the penetration numbers.

The way a Königsberg annihilates criusers at sub 10 km ranges is a sight to be seen, an I could not reproduce it with Budjonny at all, although the guns should deliver in spades. But that's probably just me...

 

 

[MakiseKurisuEU]

 

Well, the 420 shell retains more penetration. Check the ranges 12km+.

 

Speed loss is probably more or less offset by the mass increase, but a bigger diameter also reduces penetration.

 

In my opinion their penetration is more or less the same and it comes down to damage/reload and maybe overmatch if someone finds a 29mm armor plate.

 

 

Which happens to be Montana's upper deck.

 

btw the air drag of Yamato is 0.292, NoCal/Iowa/Montana's SHS is 0.352 and German 406 and 420 are both 0.299 (0.2994 to be exact). I don't know if German guns data is still valid because it's an old data from Chinese forum, at the time when info about German Battleships started to leak out.

[fnord_disc]

 

Which happens to be Montana's upper deck.

 

btw the air drag of Yamato is 0.292, NoCal/Iowa/Montana's SHS is 0.352 and German 406 and 420 are both 0.299 (0.2994 to be exact). I don't know if German guns data is still valid because it's an old data from Chinese forum, at the time when info about German Battleships started to leak out.

 

Those are the values I'm using, yes.

[Unintentional_submarine]

 

Which happens to be Montana's upper deck.

 

 

I'm assuming you mean the oddly armoured central weather deck on Montana? 28mm. 406mm AP overmatches it, so the 420mm won't overmatch anything a 406mm won't either.

The 420mm really baffles me. I see the 420mm being liked more, but the values really indicate that the 406mm should be better.

 

 

I have no reason to think the numbers are wrong, I just wanted to say that one should not judge a guns capabilities purely on the penetration numbers.

The way a Königsberg annihilates criusers at sub 10 km ranges is a sight to be seen, an I could not reproduce it with Budjonny at all, although the guns should deliver in spades. But that's probably just me...

 

 

 

It might have been. My experience with the Budyonny and Shchors tells me that the Soviet 152mm is significantly superior to the German 150mm. There isn't really any contest. The Soviet 152mm is just the perfect package for the game. Every strength it had in real life is amplified and it's weaknesses are reduced. If there is a gun I would place right behind Zao's 203mm for tier power, it would be the Soviet 152mm at T6 (thank goodness you only get access to 9 of them).

[Mucker]

I guess It's just not meant for me to have success with the russian cruiser line. Couldn't make Kirow work, couldn't make Budjonny work. Scrapped the tree afterwards. But let's get back to topic:

 

May I request a plot for Myogi and Tenryu/Kuma please? 

[fnord_disc]

I guess It's just not meant for me to have success with the russian cruiser line. Couldn't make Kirow work, couldn't make Budjonny work. Scrapped the tree afterwards. But let's get back to topic:

 

May I request a plot for Myogi and Tenryu/Kuma please? 

 

I'm not too fond of generating curves for these low-tier ships because nobody plays them for long and they're not really balanced either. I also don't really want to aid in any seal-clubbing endeavors.

 

Myogi stock is the Kongo stock guns, so lighter shell, much less krupp and terrible drag (but not as bad as Ishizuchi) -> moon arcs, bad penetration.

 

Kuma uses the 140mm CCS, drag is fine so relatively flat arcs but abysmal krupp -> bad penetration. This also affects Nagato's secondaries, which fire AP, so I will generate Kuma's curves for that.

[MakiseKurisuEU]

 

I'm assuming you mean the oddly armoured central weather deck on Montana? 28mm. 406mm AP overmatches it, so the 420mm won't overmatch anything a 406mm won't either.

The 420mm really baffles me. I see the 420mm being liked more, but the values really indicate that the 406mm should be better.

 

 

Yes the weather deck. According to armor model on GM3D site that deck is only 29mm which is fully overmatched by 420mm guns. However armor model on that site is quite outdated as the data is from patch 5.2. I'm looking forward to seeing the newly modelled Montana armor in public test 5.11.

[fnord_disc]

I'm running out of real estate on the first page, so I'll have to put links to the later pages there.

 

Also, it's definitely cool how this thread is just not getting stickied.

[Ictogan]

Could you please make a plunging fire graph for 127mm USN DD guns?

[fnord_disc]

[inaccurate]

[Unintentional_submarine]

They achieve plunging fire that close? You are sure you are using normalisation after auto-bounce calculation right?

 

But even so, USN 127mm AP should be able to plunge through Yamato's 57mm weatherdeck at around 13km. Or Amagi in general as soon as plunging is achieved. That should be easily testable.

[fnord_disc]

They achieve plunging fire that close? You are sure you are using normalisation after auto-bounce calculation right?

 

But even so, USN 127mm AP should be able to plunge through Yamato's 57mm weatherdeck at around 13km. Or Amagi in general as soon as plunging is achieved. That should be easily testable.

 

I fixed the script and this is with normalization after autobounce.

 

I should add that I tuned my ballistics model on the belt penetration data published by WG and to a lesser degree on experimental data. That doesn't mean that my ballistic model is the same as Lesta's, only that it performs equivalently with respect to belt penetration. It wasn't tuned for deck penetration and while I feel confident enough in the results to publish these curves, I don't think they're as accurate as the belt armor curves.

 

I would definitely be interested in hearing from someone who tests them to know whether they're usable or whether I should refine my ballistics before making more deck curves.

 

The problem with belt vs deck is that deck penetration is very sensitive to impact angle whereas belt penetration is much less sensitive to errors in impact angle. Both deck and belt penetration are sensitive to errors in the velocities and ultimately the drag model, but belt armor penetration suffers much less from angle errors, and I did not optimize this error source during development.

 

tl;dr I'm a lot less sure of the deck penetration results and testing would be nice.

[Woofbark]

Can you do deck pen graphs for Roon and for Cleveland since I have those ships and might test them in training room?

[fnord_disc]

[inaccurate]

 

That's Cleveland, ~13km. Roon never penetrates deck armor unless it overmatches.

[Unintentional_submarine]

 

I fixed the script and this is with normalization after autobounce.

 

I should add that I tuned my ballistics model on the belt penetration data published by WG and to a lesser degree on experimental data. That doesn't mean that my ballistic model is the same as Lesta's, only that it performs equivalently with respect to belt penetration. It wasn't tuned for deck penetration and while I feel confident enough in the results to publish these curves, I don't think they're as accurate as the belt armor curves.

 

I would definitely be interested in hearing from someone who tests them to know whether they're usable or whether I should refine my ballistics before making more deck curves.

 

The problem with belt vs deck is that deck penetration is very sensitive to impact angle whereas belt penetration is much less sensitive to errors in impact angle. Both deck and belt penetration are sensitive to errors in the velocities and ultimately the drag model, but belt armor penetration suffers much less from angle errors, and I did not optimize this error source during development.

 

tl;dr I'm a lot less sure of the deck penetration results and testing would be nice.

 

Yeah, I don't know why I didn't factor in the angle of attack, especially when I have even made inquiries about that in the past. It was a derp moment. Obviously Yamato's 57mm weatherdeck will be much thicker than that as the shell will impact at around 40 degrees. So in reaily USN 127mm AP shouldn't be able to penetrate the Yamato deck (the bow and stern are only 32mm so possible there).

Amagi's protection is uniformly 32mm across it's deck profile (not counting the turrets and superstructure of course), so it ought to be the best test subject.

[fnord_disc]

 

Yeah, I don't know why I didn't factor in the angle of attack, especially when I have even made inquiries about that in the past. It was a derp moment. Obviously Yamato's 57mm weatherdeck will be much thicker than that as the shell will impact at around 40 degrees. So in reaily USN 127mm AP shouldn't be able to penetrate the Yamato deck (the bow and stern are only 32mm so possible there).

Amagi's protection is uniformly 32mm across it's deck profile (not counting the turrets and superstructure of course), so it ought to be the best test subject.

 

Sorry, I was unclear. Your thinking was correct - these are effective values including the impact angle. For a strict 90° angle, you can simply consult the first page. I did not duplicate these curves.

 

The deck curves I am generating are effective penetration, so they do predict a penetration Yamato's weather deck by 12.7cm USN AP.

 

Like I said: please test them and don't take my word for it.

[Unintentional_submarine]

 

Sorry, I was unclear. Your thinking was correct - these are effective values including the impact angle. For a strict 90° angle, you can simply consult the first page. I did not duplicate these curves.

 

The deck curves I am generating are effective penetration, so they do predict a penetration Yamato's weather deck by 12.7cm USN AP.

 

Like I said: please test them and don't take my word for it.

 

All right.

 

So I tested some things.

I went into Training Room with Cleveland against 3 Yamatos. Cleveland's AP pen should be better, so I picked that one first. I sat at close to max range (about 14.3km) and started out dialing in on proper target areas, then I locked the guns and went into Free Cam mode over the Yamato I was firing at. Then I began firing, taking note of the hits on the deck and the corresponding results in the hit markers.

 

After sinking all three Yamatos I can say that there were only about 4-5 hits in total (out of thousands) that might have been penetrations of even the 32mm plating (obscure areas that are hard to see were hit). All the rest on the deck were ricochets. Not even non-pens. All the observable penetration markers on the ship (those blacked holes), were all on the sides, none on the deck. So about the time I began on the third Yamato, I decided to look at the incoming shells from the sides (looking at the stern of the Yamato from directly behind). The angle of fall doesn't appear to support plunging fire, it looked like 25 degrees of fall at most.

 

So I went into a new round with my Sims (with my AFT Cleveland captain for 15.5km range) against four Amagis. And the results are not yet finished. But so far, not a single penetration of the deck at 12.3km. However, unlike Cleveland I don't see as many ricochets, but a hell of a lot of non-pens (unfortunately the dispersion sends a lot of shells into the sides and the superstructure leading to a lot of pens, but they are thankfully very obvious). Also looking at the angle of fall, the shells appear to fall at more than 30 degrees at 12.3km, a lot steeper than Cleveland's. So plunging fire is achieved, but the penetration isn't up to the task at 12.3km.

 

Will move further out for more tests. The reason for the short range is the position of the Amagis on the map. I want broadside shots to be able to hit the rear deck around the superfiring rear turret as it is the most open and the most free from superstructure.

 

[EDIT] HAH... score

I missed an Amagi hiding on the other side of the spawn. Perfect scenario. 15.1km

And the results were shockingly different. The plunging fire very much managed to penetrate Amagi's deck. The penetration ratio of the hits was around 75%, which is impressively high (meaning 75% of all hits hit the deck rather than the sides or the turrets). Admittedly I was aiming for the best area for this (between third and fourth turrets), but it was still impressive. Shots that hit the sides failed to pen obviously. Angle of fall look to be around 42-45 degrees.

Will move somewhat closer to see when the non-pens take over again.

 

[EDIT2]

Huh, well that was quick. at 14.1km, there haven't been a single penetration of the deck. All non-pens (all ricochets are against the side of the ship or the turrets).

Moved away slowly (moving in steps of around 100m at a time) and penetrations against the deck began very sharply at 14.8km. Suddenly it went from all non-pens to all pens. There wasn't a sort of overlap, which surprised me greatly.

 

With these results against 32mm plating it doesn't seem likely that Yamato's 57mm deck can be penetrated at all.

[fnord_disc]

When I designed my ballistic model, I played around with it until the velocities, and therefore the belt penetration, fit the values I was seeing in the game. But that doesn't mean the impact angles are the same in my model and in Lesta's model. My personal impression was always that Lesta's model produces somewhat flatter arcs than mine, but I never ran tests. If that is the case, then all deck penetrations would be much weaker in Lesta's model since the angle is sharper and that does seem to be what you tested for Cleveland.

 

There is a very obvious explanation for the flatter arcs. I only simulate drag in the x direction (between ships), and if Lesta is using some kind of linear approximation for the y drag (height, there is no analytical solution though) then it would counteract the gravity and produce flatter arcs at minimal effect on terminal velocities. Basically, velocities wouldn't change much, but the arcs would be flatter.

 

I don't know a lot about linear approximations for y drag though, so I culled the y drag completely because its basically irrelevant for belt armor.

 

I guess that decision is coming back to haunt me now.

 

What I can do is make a numeric simulation with very small timesteps. Hopefully the ranges are not so enormous as to render this numerically unstable. Can you get me 2-3 definite deck penetrations of (ship) on (target) at (range)?

 

I should be able to make a reasonably accurate numeric simulation for decks, then.

 

edit:

Once you have the test data for me, I will remove the two plots I generated and work on a better model for deck pen.

[Unintentional_submarine]

Can you get me 2-3 definite deck penetrations of (ship) on (target) at (range)?

 

I should be able to make a reasonably accurate numeric simulation for decks, then.

 

With Cleveland? I doubt it. The angle just appears to be too shallow. Well unless she overmatches of course, but that hardly matters.

 

The USN 5 inch however had a very interesting sharp transition at 14.8km from only non-pens to only pens. Against cruisers it should be possible to get closer in of course, but cruisers are smaller, and they are considerably more cluttered, making the penetration of other stuff more likely.

I admit that my 5 inch test was much shorter than the Cleveland test, but it was just so... easy I suppose. It was so binary.

[fnord_disc]

 With Cleveland? I doubt it. The angle just appears to be too shallow. Well unless she overmatches of course, but that hardly matters.

 

No, any ship. Light shells with bad drag have the best fall angles so things like USN 12.7, Yorck, etc would work.

 

You did USN 12.7 and the tipping point is 14.8. Was that a transition of non-pen -> pen or bounce -> pen?

 

If non-pen -> pen, on Amagi's deck, right? How thick is that?

 

I'll do the simulation later today.

[Unintentional_submarine]

 

No, any ship. Light shells with bad drag have the best fall angles so things like USN 12.7, Yorck, etc would work.

 

You did USN 12.7 and the tipping point is 14.8. Was that a transition of non-pen -> pen or bounce -> pen?

 

If non-pen -> pen, on Amagi's deck, right? How thick is that?

 

I'll do the simulation later today.

 

The transition was non-pen to pen. Ricochets on the deck stopped way before. At 12.3km they were already non-pens.

I didn't think to move closer to find the transition between ricochet and non-pen.

 

Amagi's deck is uniformly 32mm with extensive open spaces behind the superstructure. Which is why she is such a great test subject. It is very easy to see when shells hit the deck, or something else, so the chances of missing a penetration or the reverse, is rather lower than on other ships with non-uniform deck plating and more clutter.

[Namolis]

 

All right.

 

So I tested some things.

I went into Training Room with Cleveland against 3 Yamatos. Cleveland's AP pen should be better, so I picked that one first. I sat at close to max range (about 14.3km) and started out dialing in on proper target areas, then I locked the guns and went into Free Cam mode over the Yamato I was firing at. Then I began firing, taking note of the hits on the deck and the corresponding results in the hit markers.

 

After sinking all three Yamatos I can say that there were only about 4-5 hits in total (out of thousands) that might have been penetrations of even the 32mm plating (obscure areas that are hard to see were hit). All the rest on the deck were ricochets. Not even non-pens. All the observable penetration markers on the ship (those blacked holes), were all on the sides, none on the deck. So about the time I began on the third Yamato, I decided to look at the incoming shells from the sides (looking at the stern of the Yamato from directly behind). The angle of fall doesn't appear to support plunging fire, it looked like 25 degrees of fall at most.

The autoricochet angle for american CAs is not 60 degrees off the normal, but 67,5 unless that has changed since 0.5.6.1. I'm not sure if the minimum angle where autoricochet is possible is still 45 degrees for USN or if that too has been shifted by 7,5 degrees, but when the angle is between "certain" and "certain not" (which will happen quite often), there is rng whether autoricochet happens or not. Afaik, this rng is rolled with odds being relative to how far off the certain angle you are (I don't actually know the if the probability change is linear or some other function though, I've just always assumed it's linear). So your results sound very consistent with all of this.

 Huh, well that was quick. at 14.1km, there haven't been a single penetration of the deck. All non-pens (all ricochets are against the side of the ship or the turrets).

Moved away slowly (moving in steps of around 100m at a time) and penetrations against the deck began very sharply at 14.8km. Suddenly it went from all non-pens to all pens. There wasn't a sort of overlap, which surprised me greatly.

Remember; we know that there is no rng when it comes to penetration, so a sharp line should be expected on that account. Your results may also mean that there is no rng when it comes to muzzle velocity (perhaps not surprising; basically you get exactly what it says on the can) AND that final velocities are not worked out individually for each shell, but is worked out only once for each dispersion ellipse based on the exact point you aimed for - even if shells are distributed potentially quite far apart inside an ellipse following different trajectories and (visually at least) impact at different times (all of which could naively suggest different impact velocities). It would certainly make sense to do it that way as it saves some calculations.

 

When i tested Bayern, I did have one point where i got a smaller (but nonzero) number of penetrations vs non-penetrations. That may simply be due to bad statistics, me making some kind of mistake, or due the fact that BBs are larger than DDs, causing guns to be spread out more, with one set being close enough to penetrate.