Naval Gun Penetration

[JWE]

Enough! This ends now!

This is my GD thread and I will not stand for it's being hijacked! If ya'll must chat, do it pm or sidebar, but keep it off my frikkin lawn! The very next one of these will cause this thread to be locked, and I will be very very very pissed.

[JWE]

ORIGINAL: Nikademus

ORIGINAL: JWE
We do consider all ranges. Thing is we cannot modify the coded algorithm, and so must do our best to "fit" things.

And therin lies the rub. Its not simply a matter of getting the most "correct" data into the OOB. its how the algorithm works that data into the routine that generates the armor vs. pen. Sometimes its easier to manipultae the OOB data to get the desired effect vs. trying to rewrite the algorithm. I did this sort of thing in my old Nikmod series. The OOB data was not "correct" in places per paper stats....but the effect it generated improved results (for the most part. [:D] )

Trust Nik to make the explanation simple. Wish I had his gift for relevant brevity. [;)]

Can't expose the code, but can do what Nik did, and may be able to do a graph of generated results. It will be normalized, so at least it will be uniform for whatever basis people choose, without violating either the letter or the spirit of the restrictions. Offsets will be problematic, but legitimate modders can depend on assistance.

[JWE]

ORIGINAL: JuanG
When doing my own asessment of the guns, I also tried to fit the values within the same 'bracket' as stock, as I felt moving away from this might break gameplay. In retrospect this was a bad move, and I intend to redo alot of the data for AltWNT so that we no longer have BBs with an IZ that extends to 11,000yards. I will also try to account for the effect of explosive in HE rounds, though I'm a little worried this might break AA fire. I may try to scrap the 'weight' system of effect altogeather and do a composite 'kinetic + chemical energy' system to see how it works.

There are some other factors though, that without knowing how the game uses the penetration value, we cannot resonably account for;
-Formula for deck penetration at long ranges; how is the game calculating this from the penetration value? Is it even doing this at all? How can it account for the difference in a HV low elevation gun and a LV high elevation gun??
-Effect of armour thickness to shell diameter ratio (T/D), and its effects on penetration; does the game estimate shell size from effect? Is this considered at all? Probably not.

Juan, you are a man after my own heart! Someday, we just have to meet, munch on boliche, drink tinto, listen to fado, and just chat. I get to the Med every now and then, and can burn a day and fly over. Woof.

Your thoughts are identically like mine, or maybe mine are identically like yours. Either way, you "got it". You have my email, I believe. I would enjoy, and look forward to, hearing from you.

Ciao. John

[CJ Martin]

Fair enough, I believe I've made my point.

I don't want to see this thread locked as there is worthwhile discussion going on.

-CJ

[JWE]

ORIGINAL: JuanG
Penetration values in game are unforunately a hazy area ... so I believe that one has to not only look at the raw data ... but also fit that to provide believeable results at the ranges which dominate surface combat in WitP - I personally consider this to be the 5,000 to 15,000yard range band.

Yes, but the algorithm doesn’t work quite that way. However, I’m plotting actual pen curves over the algorithm and calculating the sweet spot to be within those general parameters; more a function of range than a specific min/max, but you get the idea.

When doing my own asessment of the guns, I also tried to fit the values within the same 'bracket' as stock, as I felt moving away from this might break gameplay.

It most definitely will.

I may try to scrap the 'weight' system of effect altogether and do a composite 'kinetic + chemical energy' system to see how it works.

It won’t. Not unless you have exactly the same absolute value spread between your min and max, as for the weight parameter.

Formula for deck penetration at long ranges; how is the game calculating this from the penetration value? Is it even doing this at all? How can it account for the difference in a HV low elevation gun and a LV high elevation gun??

It does, actually. It calculates trajectory v range and there is an angle of incidence factor.

Effect of armour thickness to shell diameter ratio (T/D), and its effects on penetration; does the game estimate shell size from effect? Is this considered at all? Probably not.

Well, the algorithm doesn’t consider that at all, per se, but it’s all part of the calculations that we do to determine just what is the appropriate penetration basis, at what appropriate range. You are looking at third order effects. The game code is rather limited, but frankly pretty damn good in general, and can be accommodated reasonably well. Obviously, I look at second and third order effects, but once acquired, they must fall within a simple curve fit.

[JWE]

So ... having said all that, I think the best penetration values, for any particular gun, should be set at a basis of penetration at "82% of nominal max range" calculated by Nathan's program, modified by SoftShell and FlatShell.

It's relational, just like the program.

Will be doing this for Da Babes mod, and perhaps can port it over to the Official scenarios - time will tell. It won't change many things, just tweak some things up (or down), but it will provide a mathemetical basis to justify the values to the "this-isn't-historical" Nazis out there.

[edit] will certainly provide plots, etc., and explanations.

[mariandavid]

I for one are more than satisfied with the system that was developed. There is sometimes a tendency to believe that the figures cranked out by Okun and others are always and irretrievably definitive, but they only form a useful guide. In truth penetration turned out to be seriously impinged by relative aspect of target to line of impact (not declination), round and armour relative quality and sea-state. Effect was endlessly varied, just as examples, by everything from size of burster (USN 16" poor) to nature of impacted material (RN SAP rounds particularly affected) . I therefore feel that any attempt to claim more plausibility is doomed, while evaluations based on single nation characteristics are prone to error. 
 
My only concern (hopefully a trivial one, but influenced by what I have sort of noticed in my own games) is that the formula generates a higher hit rate than was actually achieved. On reading the latest book on the war in the Mediterranean (where actual damage is far better documented than in the Pacific) it is truly astonishing how very, very few hits were achieved by well-trained ships firing over considerable periods. 

[JWE]

ORIGINAL: mariandavid
I for one are more than satisfied with the system that was developed. There is sometimes a tendency to believe that the figures cranked out by Okun and others are always and irretrievably definitive, but they only form a useful guide. In truth penetration turned out to be seriously impinged by relative aspect of target to line of impact (not declination), round and armour relative quality and sea-state. Effect was endlessly varied, just as examples, by everything from size of burster (USN 16" poor) to nature of impacted material (RN SAP rounds particularly affected) . I therefore feel that any attempt to claim more plausibility is doomed, while evaluations based on single nation characteristics are prone to error. 

Oh, heck yeah! There's really 2 parts to the game, the editor numbers, and the code. Everyony gets them confused and thinks they are the same thing.

Editor numbers are fun, but those are the ones everybody "demands" must be acording to their version historical. They are pretty, however, and if the point is to show people a "pretty" OOB, the editor will let you do that. But, the code don't care about pretty, so a historically precise set of editor fields (according to opinion) would probably bugger the game; but what the hey, most historico Nazis don't actually play, so they can make all the "pretty" OOBs they want - they won't work, but they will be 'pretty".
Need to know how the code works in order to efficiently "prettify" the field values.

My only concern (hopefully a trivial one, but influenced by what I have sort of noticed in my own games) is that the formula generates a higher hit rate than was actually achieved. On reading the latest book on the war in the Mediterranean (where actual damage is far better documented than in the Pacific) it is truly astonishing how very, very few hits were achieved by well-trained ships firing over considerable periods. 

You are confusing one thing with another. % hit is a totally different algorithm from % penetration. What Juan & I & others are doing is lining up the ducks on the penetration side of things. % hit is another topic entirely.

[witpqs]

ORIGINAL: mariandavid

... sea-state.

And as you guys are pointing out, when you otherwise have everything calculated (at least for statistical likelihood), Mother Nature comes along and says:

"I'm changing the impact angle, in all three planes. "

[:D]

[Buck Beach]

ORIGINAL: JWE

Oh, heck yeah! There's really 2 parts to the game, the editor numbers, and the code. Everyony gets them confused and thinks they are the same thing.

Editor numbers are fun, but those are the ones everybody "demands" must be acording to their version historical. They are pretty, however, and if the point is to show people a "pretty" OOB, the editor will let you do that. But, the code don't care about pretty, so a historically precise set of editor fields (according to opinion) would probably bugger the game; but what the hey, most historico Nazis don't actually play, so they can make all the "pretty" OOBs they want - they won't work, but they will be 'pretty".
Need to know how the code works in order to efficiently "prettify" the field values.

This is rather disappointing to know that any of our tweaks out here (to erroneous data entries, missed OOB, etc) that we care to make with the editor have no value? Hope I am misinterpreting your statement above.

[mariandavid]

You are confusing one thing with another. % hit is a totally different algorithm from % penetration. What Juan & I & others are doing is lining up the ducks on the penetration side of things. % hit is another topic entirely.

Apologies - I incorrectly believed that you were dealing with both! In any case deserves its own files!

[JWE]

ORIGINAL: Buck Beach
This is rather disappointing to know that any of our tweaks out here (to erroneous data entries, missed OOB, etc) that we care to make with the editor have no value? Hope I am misinterpreting your statement above.

They have a great deal of value, Buck. The tweaks just have to stay within the general boundaries of the paradigm. Just about everything we've done to date has been well worth doing. The caution is against being arbitrary and just picking a number. The number might be historically precise, but may not "fit" within the playing field.

For example, we spent a lot of time with the tonnage model, and making the values work within the code. If someone were to come along and say "I think it's better to have it in gross tons"; well, it's a legitimate value in some cases, and it would be "pretty", and match what the internet says, and make some people more comfortable, but it would break the paradigm, and cause all sorts of problems. It wouldn't "fit".

[bklooste]

ORIGINAL: Buck Beach



This is rather disappointing to know that any of our tweaks out here (to erroneous data entries, missed OOB, etc) that we care to make with the editor have no value? Hope I am misinterpreting your statement above.

No he is not the algorithm is fixed i believe from early WITP days , play test in the development would have shown certain values too high or low so the values were probably adjusted.

eg the penetration model may have worked well for ships up to cruiser size but BBs were not penetrating at ranges where they should and hence BBs would have had their pen values increased . This is a lot easier than retuning a number of complex formulas.

[JWE]

Ok, think we found a good way to find a workable penetration number. These fit best into the game’s calculations and give the closest approximation to actual penetration over the widest set of ranges. Fits well into Juan’s 5k to 15k.

It’s not exactly penetration at the same specific range, rather it’s actually the penetration at a given % of max range for each gun: about 16-17-18-19% of max effective range, 16 for small shells/small guns, 19 for big shells/big guns.

If ya don’t want to work that hard, a simplified model is to use pen at 6500 yds for BBs, 5500 yds for CAs, 4500 yds for CLs, and 3500 yds for DDs (or 6000, 5000, 4000, 3000, to make life even simpler). Those 500 yds don’t make very much difference, but I wouldn’t move too far outside that spread.

Individual gun/pen stats all ought to come from the same place so that the all the data is internally self consistent – Nathan’s prog is very good for this. Not surprisingly, the calculated values are in the same ballpark of those Gary Grigsby devised back 8 or so years ago (GG seemed to know which end of the gun the bullet came out of). Some go up or down about 5-10% or so. There’s a few errors that need fixing, but not that many.

Basically this will put all the little ducks in a nice row, rationally related to one another, and internally self consistent. It will make things a lot easier to mod because the basis is known and determinable, and is the same for everything. Change is a simple matter of applying a % offset (up or down).

Did this for maybe 125 different guns so far (worked every time). Will stick it into Da Babes. If we’re lucky, it might get included into the main scenarios in patch-2.

[JuanG]

Sounds like a good approach. I used elevation myself, by percentage of range sounds like a better idea to be honest. Just requires a little more work to get the numbers out...

Look forward to seeing them in your scenario - any chance you could post them in a spreadsheet or something on here? Just curious how theyre affecting certain key weapons.

Regarding your replies to my earlier points, I have to admit it sounds like I've underestimated the ballistics model from what you say - its a shame we cant get a look under the hood, but I can understand why. I'll just have to settle for being both jealous and thankful at once! [;)]

Thanks for the effort,
Juan

[JWE]

ORIGINAL: JuanG
Sounds like a good approach. I used elevation myself, by percentage of range sounds like a better idea to be honest. Just requires a little more work to get the numbers out...

Look forward to seeing them in your scenario - any chance you could post them in a spreadsheet or something on here? Just curious how theyre affecting certain key weapons.

Regarding your replies to my earlier points, I have to admit it sounds like I've underestimated the ballistics model from what you say - its a shame we cant get a look under the hood, but I can understand why. I'll just have to settle for being both jealous and thankful at once! [;)]

Thanks for the effort,
Juan

Yeah, we did it that way because the game first calculates pen as a function of a range fraction. So we plotted the game curve against actual penetration curves, shifted the game curve up or down to get the best line fit, and selected pen at the first crossover. After normalization, we noticed that the selected pen values were all within a tidy little group of range fractions. It just happened to work out that way, and I was quite surprised and pleased at the result. Made life easier.

Sure thing, can do a spreadsheet and send it up. It will have gun ID, shell wt, range, and pen values.

btw, if you want to keep it really, really simple, and only do a simple batch file for running Nathan's prog, it is probably ok to run everything at 5k yds. BBs will be maybe 10% more powerful, DDs will be maybe 15% less powerful, CAs and CLs probably won't change more than a couple %, either way.

[edit] reminds me, spreadsheet will have 4 pen columns (at least for Allied guns): pen at 5k yds (uniform) against Japanese VH, pen at 5k yds (uniform) against std Class-A, pen at % range against Japanese VH, and pen at % range against std Class-A. This should give a very nice range of options.

Ciao.

[JWE]

Ok, here's a taste; the Brits. There's some values in blue. They are interpolations. God only knows why the original OOB folks made so many 4in/45s and 4in/40s. Got some good numbers for the nominal tubes, but as to the rest, I'm thinkin who cares what went on in 1924. Once the war got going, Uk built what it could, and your ships shot what they got. Lot easier to reline a tube than retool a factory, so the numbers in blue are right out of my butt, but I would be really pleased if somebody could provide specific and directed data for these little pukes.

Didn't do anything below 4in. It's a who cares, so far as the game is concerned. Ya'll really, really don't need to know (or care) about 3in penetration, I'll see what I can do. Anyhow, that's about it. Brits are maybe 25% of the magilla, others being run as we speak, but Brits are what most people are whining about, so thought I'd post them first.

It's an xls file inside a zip.

Ciao.

[Bradley7735]

ORIGINAL: JWE

Ya'll really, really don't need to know (or care) about 3in penetration,

That's what I keep telling my girlfriend.

in all seriousness, you guys are awesome. I hope your changes can make patch 2

[JWE]

Ok, well .. here’s the second pass, with UK and US data. Same deal, an xls file inside a zip.

Found some interesting stuff on the US BB shells. Seems they weren’t quite the hot pazotties that everyone thinks (even the new ones). They were quite reasonable, but only that.

BuOrd knew of this and had a fun project going on with respect to body steels that resulted in a major increase in hardness and therefore penetration. Found some acceptance trails minutes and build specs which indicate the earliest these guys would be available (in limited quantities) was 10/44 and most likely 2/45. These were the hot pazootie shells.

Game OOB doesn’t capture this, so I have 2 lines for each BB gun; one with the nominal shell, one with the hi-Brinnel shell (indicated as 10/44). Notably, the nominal shells don’t suck, but they sure aren’t the cat’s meow.

Working on figuring out how to do a data upgrade within the editor, without specifying a different gun in the ship/class files. Getting fancy, I know, but as soon as I figure it out (or find it can’t be done that way), I’ll be sure and post.

[JWE]

And here’s one with the Japanese on page 3. Boy, the DDs were a stone witch. Whoda thunk they didn’t have AP (or SAP or even CPBC), but just HE with an instantaneous nose-fuse, and show fairly uniform 35-40 pen from 0 to .765 * max range. Woof!!

Can’t do that because the game curve doesn’t work that way, so had to actually find a pen value. Groin!! Anyway, it’s pretty close to actuals at nominal engagement ranges; higher at shorter ranges, lower at longer ranges, but functional. The short model 12cm and 20cm guns are even worse. Gonna have to think long and hard for those.

We have obliquity and striking angle curves up the wazoo, so with a bit of extrapolation, think we can come up with very acceptable results for the small “popguns”. Only thing we are missing is good data on body weight. Relative % of body wt v shell wt is all over the map, so can’t just shine it. We may have to, but if anybody can provide precise data, that would be preferable.

For the Dutch and French, there weren’t too many native factories making replacements after 1940, so I’m using their gun parameters but assuming that shells were manufactured to Brit standards (steels, core, charge, etc.).