fnord_disc

No, that's not how it works. I tried explaining it in the other topic too, but I guess I'll do it again. It doesn't matter that they're sister ships. This has nothing to do with history at all, only with programming. The values on the wiki are exported from the ingame armor model using the script that also generates values for the API. There is no human being looking at the armor model and writing out the values that seem most sensible to him. It's completely automated. But for the script to know where it has to look, every armor layer in the game has to be allocated to a group, such as citadel, belt, casemates, superstructure, torpedo belt... If you look at GM3D, you can see that in the labels assigned to the armor plates, but GM3D sometimes fails at exporting the correct labels, so don't trust them too much. The script can easily get confused if the way the game groups armor layers is different between Scharnhorst and Gneisenau, and this would easily happen if a different 3D artist did Gneisenau's armor scheme. So even if the two ships have exactly the same armor, the script might give completely different values if the armor groups are funky. And even if the script gives the same values, it doesn't mean the armor is actually the same, or has the same thickness. The values displayed in the port and on the wiki are "official" in the sense that they put there by Wargaming and based on the ingame values, but the process by which they are put there is full of error sources and never checked by a human being. The numbers are meaningless for any kind of reliable discussion. How about this? I will bet you 5€, payable via PayPal, that Gneisenau has 350mm maximum belt armor in the game. Winner is determined by checking the datamined armor model on GM3D whenever it is put up. edit: If you want to see one of the many examples where the wiki is wrong already, look at Wyoming's armor scheme http://gamemodels3d.com/worldofwarships/vehicles/pasb004 If you click through the armor layers, you can see that the only area on the ship that has 305mm are the turret faces. The belt has 279mm and 292mm is the conning tower. Yet the wiki claims that Wyoming has 305mm of citadel armor, even though both the datamined armor model says 279mm and the historical blueprints. Wyoming did not, historically, have 305mm of belt armor. The armor values that UI and the wiki tell you are meaningless.

I'm saying that there is no way to know until the armor model has been mined. The values displayed in the port and given on the wiki are meaningless. It might be worse. It might be the same. It might be better! The numbers mean nothing.

It depends on how WG defines armor areas for the UI and the API. It is literally impossible to know what the armor model looks like or how it performs from the values displayed in port and on the wiki. Depending on which areas WG defines as belt armor, citadel armor or casemates, the UI will average adjacent values. Depending on which areas the UI groups into one, it sometimes average belt armor tapering from 350 down to 170 as 260, but it will not average a 40mm citadel roof because it isn't a continuation of the belt armor but a separate armor layer. You cannot say anything at all about the armor performance of a ship based on the port values. They are meaningless. It might be that the ingame armor is not historical. It might be historical. You can't know unless you datamine the armor model from the game. That Gneisenau and Scharnhorst have different displayer armor thicknesses also means nothing. They have different hulls, so the UI and the API might group completely different armor layers into one to calculate whatever they display. It means nothing. edit: You need to understand that the values in the UI and on the wiki/the API are generated by a script and not by a human being. The script just spits out a couple of values from the sections it is made to analyze, but that doesn't actually say anything about the armor model. For example, if the 170mm under the waterline and below the 350mm belt still count as belt armor as far as the script thinks about it, it will just average these areas thinking the entire belt is 260mm even though the 170mm area is tiny. Then it looks at other parts of the citadel and sees the 40mm armored deck above the citadel and says, all right, citadel has between 40-260mm. Or this might not what the script does at all! Maybe the belt armor is only 270mm thick and it also takes the 250mm casemates to get to 260mm. Maybe the entire belt is 260, screw history. Or maybe the armor model actually looks completely different and it's just glitching out! The port and the wiki show you a bunch of numbers, but they don't really tell you anything at all about the armor of the ship in the game.

When are people like you going to understand that those values are meaningless? Armor model or go home.

Are you going to spam every thread Thrawn just because you're unable to understand that the UI shows average armor values, not maximum?

Until I see the armor model, all this is just hot air.

German engineering is too strong for WoWs. It has to be nerfed or there would be no German ships lower than T8 or T9.

No, those values are for the unupgraded range. If you're getting that off shipcomrade or the API, then the dispersion is wrong.

Yes, ß=ss, ö=oe, ü=ue... is considered correct orthography if the system does not allow Umlaute.

I'm sure you know this already, but your site only displays stock ships correctly. For example, upgrading fire control should also increse dispersion (because the range increases), and whenever a hull upgrade changes the shells, the artillery substats are not changed to the new shells. Hipper HE shells for example. AA DPS values, distributions and names are also not updated. Thank you for your site, though.

Can you explain to me how you're getting the server to use the n+1 (the higher) trajectory? Understanding it will help me with other things related to the game's ballistics. PM if you don't want to make it public.

I wasn't even talking about the recruitment, but whatever. Nor was I saying WG is a collection of saints. But whatever. Some people just refuse to see the world in shades of grey.

I think I have an idea how he's doing that, though I don't know if it's really a bug or whether he's modified his game client to send spurious requests to the server. To be honest, I'm leaning towards the latter.

The issue is not as clear-cut as a lot of people in Europe like to think. For one thing, the European staff in Paris get much high wages (because Europe) than their equivalent colleagues in Russia.

It would be so great to have a 12k money event for 1 guy on EU. People totally wouldn't screw up the lower tiers leveling their multi-accounts and definitely wouldn't try to TK people they suspect of doing the same at the beginning of the match. Definitely not. I'm sure it would all work out perfectly and without envy and all.

I don't understand the point of the ship, then. Hipper is already the second-worst T8 cruiser after New Orleans going off the statistics. She's not terrible, but not good either. What's the point in making the premium even worse?

Holy crap, Prinz Eugen. WG better make the ship nice.

The reason why naval tactics are screwed up in this game isn't because of the wrong dispersion pattern. But I've had this discussion too often by now and I'm not gonna take the bait again.

The 42 behave like the American 406 in weight and damage. So the more novel gun is actually the light, higher-rof 406.

A line centered around armor, secondaries and HE. Nice.

It's just a courtesy. I write it after almost any match that I wasn't insulted in, even if it was a bad game. It just means "thank you" to me.

I know how experiments work. But making meaningless experiments produces meaningless results. It isn't complex... Also I trust nothing this game tells me, otherwise I wouldn't be reverse-engineering it. That's normal. *sigh* You are assuming that I don't know why experiments are important. I do. I am asking you what information, what truly useful information, your experiments have given you. This is a completely bizarre argument. What are you trying to tell me? Let me explain to you why your experiments are useless then. You are assuming that the game stores a to-hit percentage somewhere in the game code and when you're in, for example, a Kongo and you shoot at a Minekaze, the game looks this up and rolls and dice and says, hurrah! it's a hit. This makes no sense. This is how games 20 years ago did their calculations. If this were what the game did, why can I hit targets that I'm not locked on to? Why can I hit friendlies? Why can I shoot blind into a smoke screen and hit someone? Does the game try to guess who I might be shooting at to determine your hit% chance? Why do I hit islands that are in the way? Why does the game have hitboxes if there are no trajectories, according to you? You can hit specific parts on the enemy ship if you aim correctly. How can there not be trajectories?!? What you did was, I assume, record hit% over range for ship A against ship B. 4 Variables. You are not going to get any kind of formula from this that lets you predict to-hit rates. The best you can hope for is an extrapolation for a different range, but only for the same shooter and the same target, not for different ships. So to get truly new information, you would have to do more and more experiments. What's the point of experiments if it doesn't predict anything? The goal of experiments is not to gather endless amounts of information. The goal is find a description of the system that describes it analytically based on the experimental data. It's staggeringly obvious if you just observe the ingame mechanics that the game calculates each shell individually and gives each shell a trajectory into some dispersion ellipsis, but from just 4 variables you will not reverse-engineer the dispersion ellipsis. It's completely hopeless. If you want to understand how accuracy works, you should do the following experiment. I would do it because it's actually really interesting, but I don't have the time for it. Find a friend who joins your training room as observer Fire a lot of salvos at some distant point, but always the same point Your friend puts his camera directly above the target point and looks down so that you get a 2D map He screenshots every salvo impact You make a 2D frequency map from the shell impacts and from there you solve for the multivariate normal distribution This would give you a general description of accuracy for this range and ship and from there and some data-mining you can backtrack to the general formula for all ships. This still doesn't give you the chance to hit because, like I said, it depends on the size and shape of the target, but you would be pretty close and the rest is simple and explained in my last post.

German ships have hardly been underpowered so far.

I wouldn't consider it debunked. What Darth_Glorious said is still possible, i.e. that Dunkerque's line goes lower at 8-10km than the American/Japanese/German lines. That would make her much more accurate at ranges around 8-13km (but not at max range) compared to other battleships. It's not possible for me to judge that without more French battleships. This comparison is only valid for American/Japanese/German accuracy and to a certain extent also for British/French at strictly max range.

Yeah but this is just a visual thing. If the turrets are a long way apart, the dispersion will look big while the shells are in flight, but they'll converge into the same dispersion ellipsis at the target as any other turret configuration with the same dispersion/sigma. I suppose if the turrets are very far apart compared to the range, their dispersion ellipsis will start to drift apart, but that's only noticeable at really short ranges. Just don't make dispersion more mystical than it is. All battleships have pretty much the same horizontal dispersion at shorter ranges.