pcasey
Junior Member
Posts: 58
|
Post by pcasey on Jun 28, 2019 8:53:31 GMT -6
As we all know, the game gives very limited feedback on the damage potential of shells or the damage caused by shells. There's good in-game information about the penetration capabilities of shells, but not so much on the damage potential of shells.
In an effort to get to the bottom of this, I've been doing some not very rigorous experiments in my most recent games, comparing the performance of a couple of really quite extreme designs to see if I could deduce what was going on.
I've built SMS Archer with 8 20" guns.
I've build SMS Odin with 12 18" guns.
I've build SMS Thor with 12 16" guns.
I've build SMS Unusual with 12 12" guns.
I've build all of them in the early 1940s when gun penetration values have far outstripped armor to the point there at medium combat ranges, every one of those guns is going to penetrate a typical AI dreadnought (the AI really like armor between 9" and 12").
And then I've played with them ... enough to have an impression of their performance, but not enough to have a statistically valid sampling of their performance e.g. what follows is subjective, not based on a lot of math.
Subjectively, it seems to take just as many 20" penetrating hits to sink a BB than it does 12" penetrating hits.
Naturally the 20" guns have better range so they can score hits at ranges the 12" guns cannot.
Looking only at hits from within the engagement envelope though, it looks like it takes the same number of 20" shells to sink a BB that it does 12" shells.
Given that a 20" shell is going to have something like 5X the mass of a 12" shell (20/12^3) I'd reasonably expect something like 1/5 the hits required to sink a target. Even assuming for some amount of overkill (the structure right where the shell struck isn't any more destroyed on a 20" vs a 12" shell), I'd still expect there to be a really noticeable ratio between the damage potential of a huge shell like a 20" and a small shell like a 12", and in game I'm just not.
Possible theories:
1) The game engine doesn't model damage potential in shells, just penetration. The superior performance of larger guns is modeled via their superior ability to penetrate armor at greater ranges, but not via the superior damage caused by such penetrating hits.
2) There is, in fact, a really material damage ratio, but the dataset I looked at is somewhere between a statistical outlier and an unlucky accident, so what I'm seeing isn't representative of the underlying math.
3) There is a damage ratio, but its a lot less extreme than the 5X I postulated, and as such isn't going to produce the expected skew in the in game data.
Has anybody else gotten to the bottom of the damage model?
I'm really liking this game, but also really trying to work out the underlying engine mechanics. For me this game is 100% an optimization challenge, and I feel like I'm optimizing a system w/o information on how it works, which is turning a fun optimization challenge into a science project to try to determine the underlying parameters of the system.
|
|
|
|
Post by dorn on Jun 28, 2019 9:02:37 GMT -6
I do not play too much RTW2, but in RTW1 even the difference between 14" and 16" shell was enourmus.
If I remember well William present here coeficient of increased damage by increasing shell caliber. I think it was about 2/3.
|
|
|
|
Post by williammiller on Jun 28, 2019 9:06:39 GMT -6
Different shell masses do not cause linear damage on naval targets, either in real life or in the game. Yes, shells do more damage the larger they are in the game, but it is not a linear relationship.
|
|
|
|
Post by entropyavatar on Jun 28, 2019 9:10:56 GMT -6
I do not play too much RTW2, but in RTW1 even the difference between 14" and 16" shell was enourmus. If I remember well William present here coeficient of increased damage by increasing shell caliber. I think it was about 2/3. Yeah it was supposed to be roughly proportional to (shell mass)^(2/3). Though I gather there were a bunch of other factors. |
|
pcasey
Junior Member
Posts: 58
|
Post by pcasey on Jun 28, 2019 9:12:06 GMT -6
Different shell masses do not cause linear damage on naval targets, either in real life or in the game. Yes, shells do more damage the larger they are in the game, but it is not a linear relationship. I believe its non linear, but we're talking about shells with a 5X mass difference. Even if you assume that "true" damage is a function of the square root of the mass differential we should still expect the bigger shell to do like twice the damage. If you assume its the cube root of the mass differential you're back looking at like a 1.66X damage ratio. I'm just not seeing even that much difference in performance of even the (really pessimistic) cube root model. |
|
|
|
Post by dorn on Jun 28, 2019 9:27:09 GMT -6
Different shell masses do not cause linear damage on naval targets, either in real life or in the game. Yes, shells do more damage the larger they are in the game, but it is not a linear relationship. I believe its non linear, but we're talking about shells with a 5X mass difference. Even if you assume that "true" damage is a function of the square root of the mass differential we should still expect the bigger shell to do like twice the damage. If you assume its the cube root of the mass differential you're back looking at like a 1.66X damage ratio. I'm just not seeing even that much difference in performance of even the (really pessimistic) cube root model. Did you study logs? I have just fleet excercise. 2 ships of same class. One ship got 21 heavy hits, her structure dmg was about 1/4, floating dmg about 1/3. The second ship got only 11 hits but she was sunk. Reading log it was clear as where hits were. |
|
|
|
Post by rodentnavy on Jun 28, 2019 9:30:40 GMT -6
Different shell masses do not cause linear damage on naval targets, either in real life or in the game. Yes, shells do more damage the larger they are in the game, but it is not a linear relationship. I believe its non linear, but we're talking about shells with a 5X mass difference. Even if you assume that "true" damage is a function of the square root of the mass differential we should still expect the bigger shell to do like twice the damage. If you assume its the cube root of the mass differential you're back looking at like a 1.66X damage ratio. I'm just not seeing even that much difference in performance of even the (really pessimistic) cube root model. When you say the same number of shells to sink a battleship are you looking through the logs and weeding out the hits marked ## or not? The thing is that the number of shells ploughed into a sinking battleship is likely to remain about the same given how quickly you call the dogs off and manage to get them redirected. One of the issues with shelling and torpedoing sinking vessels is that opening up flooded compartments to the sea is not going to change that much. Thus this alone will tend to skew the result unless you are eliminating shots that land after a ship is definitively sinking and even then a lot of the hits before that may well be redundant. |
|
pcasey
Junior Member
Posts: 58
|
Post by pcasey on Jun 28, 2019 9:30:52 GMT -6
I believe its non linear, but we're talking about shells with a 5X mass difference. Even if you assume that "true" damage is a function of the square root of the mass differential we should still expect the bigger shell to do like twice the damage. If you assume its the cube root of the mass differential you're back looking at like a 1.66X damage ratio. I'm just not seeing even that much difference in performance of even the (really pessimistic) cube root model. Did you study logs? I have just fleet excercise. 2 ships of same class. One ship got 21 heavy hits, her structure dmg was about 1/4, floating dmg about 1/3. The second ship got only 11 hits but she was sunk. Reading log it was clear as where hits were. I did, but, again, I'm not claiming that I did a full statistical workup. That being said, my tests didn't show the same results as yours ... took more or less the same number of 20" hits to sink a ship as it did 12" hits. Could have been a fluke or statistical anomaly of course. |
|
Warspite
Full Member
Sky of blue/And sea of green
Posts: 230
|
Post by Warspite on Jun 28, 2019 10:09:24 GMT -6
This 2800 ton AV took 66 hits from 14 and 16 inch guns before sinking. I've no idea why but she took far more punishment than a lot of BBs I have sunk in my games. I have attached the battle log if anyone is interested. Battle Log.rar (2.12 KB)  |
|
|
|
Post by mycophobia on Jun 28, 2019 10:16:40 GMT -6
Did you study logs? I have just fleet excercise. 2 ships of same class. One ship got 21 heavy hits, her structure dmg was about 1/4, floating dmg about 1/3. The second ship got only 11 hits but she was sunk. Reading log it was clear as where hits were. I did, but, again, I'm not claiming that I did a full statistical workup. That being said, my tests didn't show the same results as yours ... took more or less the same number of 20" hits to sink a ship as it did 12" hits. Could have been a fluke or statistical anomaly of course. Bear in mind that actual damage inflicted depends hugely on *Where* you hit, assuming you can penetrate the armor regardless, hit location is the single most important factor that determines damage. As in addition to immediate damage and flood point the hit cause, secondary effect like fire, progressive flooding, crew damge etc will further contribute to actual damage inflicted, and thus, shell required to sink. Based on my own experience, even in rtw 2 16in is marked improvement over 12in guns in terms of effectiveness in sinking ships, but that's based on actual combat situation where penetration likely played a significant role. The game does model different damage value based on shells, but I don't think it should be an important design consideration vs penetration. To actually test this, you basically must test ship that are hit at the same location by different shell, and compare damage point before any additional flooding kicks in. Imo that's just far too many factors, and in this case, optimizing based on theoretical additional damage is just not worth it when compared to other factors. |
|
|
|
Post by Capsized on Jun 28, 2019 10:19:30 GMT -6
This 2800 ton AV took 66 hits from 14 and 16 inch guns before sinking. She was sunk after 5 hits. All the hits with ## in front are AFTER the "Ship sinking" message and are more or less useless. |
|
|
|
Post by mycophobia on Jun 28, 2019 10:22:19 GMT -6
This 2800 ton AV took 66 hits from 14 and 16 inch guns before sinking. I've no idea why but she took far more punishment than a lot of BBs I have sunk in my games. I have attached the battle log if anyone is interested. Looks like she took a lot of repeated hits in a fairly short interval of time ( <20min), most of which being HE. Flooding and sinking is a progressive process, shes likely doomed by the time 6-7th shell hit her, but didn't actually take on enough water to actually sink yet. The rest of the hits just speed up the process. ^Edit: As Capized points out, most subsequent hits does nothing since the ship is already gone. How did I miss that x.x Also, I am not sure how exactly the game models this, but if a ship is hit by HE, even if penetrated her outer armor, it may not reliably damage her interior space as well as an AP hit(provided the AP didn't overpen, though even with overpen it might still cause more flooding if it basically went through the ship below waterline). So in this case, HE will do a good chunk of damage against most light targets, and will not be almost useless when over penetrating superstructure or both ends of the ship. But I suspect an AP that did go in and detonate in machinery will still be much more devastating than HE that explode on hull(which is probably still very effective nonetheless). |
|
|
|
Post by Fredrik W on Jun 28, 2019 10:35:32 GMT -6
We have a damage simulator where we have tortured ships through 1000s hits ad 1000s of sinkings just to make sure the damage model is sound. I am confident in the results, and there is a huge difference in the damage done by a 20 in shell compared to a 12 in shell. Though location and luck are large factors in the effect of individual hits. Hits in naval warfare is like real estate, the three things that matter most are location, location and location!
|
|
|
|
Post by entropyavatar on Jun 28, 2019 11:36:10 GMT -6
Wow, that damage simulator sounds awesome. So you must have a pretty good idea how decisions like narrow or inclined belt play out in various scenarios.
|
|
|
|
Post by christian on Jun 28, 2019 12:38:57 GMT -6
We have a damage simulator where we have tortured ships through 1000s hits ad 1000s of sinkings just to make sure the damage model is sound. I am confident in the results, and there is a huge difference in the damage done by a 20 in shell compared to a 12 in shell. Though location and luck are large factors in the effect of individual hits. Hits in naval warfare is like real estate, the three things that matter most are location, location and location! as cool as that sounds it is still INCREDIBLY annoying we have absolutely 0 way to guage what kind of damage/explosive filler a gun will do upon penetration it could be some totally random number which makes no sense like what you find in the gundata file where the numbers mean nothing unless you know the code what does a reload of 25 mean I HAVE NO IDEA even this would help compare guns damage to eachother because as of currently we have NO WAY AT ALL to gauge the damage potential is it worth to mount 6 inch guns on destroyers due to supperior damage ?? how much does sap shells do compared to he shells and compared to ap shells can we get something which tells us the penetration of he sap and ap shells the damage from he sap and ap shells the damage dosent have to be in readable numbers it could be just like the gundata file also knowing a bit more about sap and he (maybe include a selector in the immunity zone calculator where you can select diffrent shells he ap sap) alot of these things is something you would hope admirals know (there is a reason for penetration tests) and in real life navies knew how much filler they put in and its effect due to tests this goes for both sap and ap he less so but the explosive filler should be visible in some way even if its in numbers which cannot be properly read (again like the gundata file) oh yeah knowing a bit about torpedo damage both airborne and shipborne would be nice same goes for bombs penetration for bombs would be nice and explosive value for both torpedo and bombs while the game is extremely good its also extremely hard to see what you are actually trying to protect against i saw a 1000lb sap bomb go through 5.5 inches of armor ? what do i do know how much pen do other bombs have how much does that bomb have and how much damage do sap bombs do compared to ap bombs or gp bombs SO MANY QUESTIONS |
|
