Weapon Accuracy for Devices

[Big B]

A good site for everything you ever wanted to know about naval weapons calculations is here: Nathan Okun

[bigred]

ORIGINAL: Buck Beach

ORIGINAL: JuanG

ORIGINAL: MateDow




Where did the constant of 10.25 come from? Did I miss that in the explanation earlier?

It comes from simplifying the constants in the equation. These are 16.1 (from the caliber basis), Pi (from the sectional density calculation), 4 (also from the sectional density calculation, a consequence of converting R^2 -> (D/2)^2 -> (R^2)/4 ) and 1/2 (from the 'divide by 2' final operation).

Equation wise;
Initial equation;


Convert R^2 -> (D^2)/4, 4 moves to top;


Rearrange constants;


Resolve constants = 10.2495 = 10.25.

Uhhh, what ever happened to "you push the little value down and the music goes round and round and it comes out here"[:D][:D]

I [&o] this thread. Easy to see why I went infantry.[:D]

[Symon]

ORIGINAL: JuanG
John, any chance you could clarify which you intended as the basis? Also, should the L and crh basis be applied before or after the divide by 2 bit? I applied them afterwards, but that was in the interest of simplification for the example.

Hola,

I use crh more to identify the "performance outliers". Some guns were very accurate, some were very inaccurate (dispersion, range, etc..). Found crh for those and there is some degree of correlation in the ratio. This being one of my more demented periods, I averaged crh for each caliber and ratioed the individual gun values to the caliber average. They are all 'close' to 1, but a few were notably bigger or smaller (looked at things outside +/- 1 and 1.5 sigma from the mean).

Turns out this, in combination with the L ratio, is a resonable way to justify the more spectacular (and notably unspectacular) performers without being arbitrary and capricious [8D]

Ciao. John

[GaryChildress]

For anyone interested I've made an MS Excel spreadsheet (readable in OpenOffice Calc as well) which does the calculation JuanG posted above. All you need to do is plug in shell weight (in pounds), bore diameter (in inches) and rate of fire (per minute) in the respective cells and the accuracy will be autumatically calculated in the cell labeled "Accuracy". The sheet is protected so the formula cannot be accidentally deleted but the password to unprotect it if you want to throw caution to the wind is "password".

Spreadsheet is attached below:

[GaryChildress]

Hmmm. Maybe I've miscalculated the formula in my spreadsheet but according to what I'm getting for the German 12.7cm/45 SK C/41 which, according to navweapons.com has a RoF of 15 rpm and a shell weight of 61.7 lbs, the accuracy is coming out to 115 [X(]

Did I calculate something wrong or is that the value I need to put in under accuracy in the DB?

EDIT: Plugging in the Bofors 40mm round I come out with 770 accuracy! [:D]

[JuanG]

ORIGINAL: Gary Childress

Hmmm. Maybe I've miscalculated the formula in my spreadsheet but according to what I'm getting for the German 12.7cm/45 SK C/41 which, according to navweapons.com has a RoF of 15 rpm and a shell weight of 61.7 lbs, the accuracy is coming out to 115 [X(]

Did I calculate something wrong or is that the value I need to put in under accuracy in the DB?

EDIT: Plugging in the Bofors 40mm round I come out with 770 accuracy! [:D]

I get 95.5 for the 12.7cm/45 C/41.

The US 5in/38 Mk12 comes in as ~85 without the L corrections.

I think the issue here is that it is intended to model the larger caliber rifles (probably 8in upwards), so the model may break down for the small fast firing weapons. This is likely because determination of 'nominal' RoF is fairly difficult for these weapons, or atleast moreso than the large rifles. Certainly, 15 which is the 'sustained' RoF for both of these weapons is not the same thing.

To account for that, this is why I use a power of 0.34 for the RoF factor in my own fomula - it gives a fairly gentle increase, which reflects the reality of RoF being of secondary importance in surface combat (firecontrol was much more important). Of course, there are instances in which RoF is important, particularly close range encounters or barrage fire against light targets, but the model cannot cover everything. On the otherhand, when dealing with AA accuracy, I used a power factor of 0.7, which reflects the much greater importance of rate of fire on AA fire compared to surface fire.

The actual formula used in the current iteration of HDM in plain form is;
Surface Accuracy = 5 + ((RoF^0.34) * 20 * Modifier)

The 'Modifier' here is a value the is a composite of the firecontrol system associated with the device, as well as the platform type (BB, CA and CL main batteries receive a bonus) and other modifiers (including my own personal tweaks to some guns up and down depending on historical performance). For most BB's with middle of the road FCS systems (say Nagato or Colorado at the beginning of the war), it works out to ~1.3.

For AA accuracy it is;
Anti-Air Accuracy = 5 + ((RoF^0.7) * 10 * Modifier)

In this instance the Modifier is fed slightly different values, but the general idea is the same. RoF in bothn of the above is simply sustained rate of fire.

I am in the process of (or will be as soon as I move back onto Naval stuff from Aircraft) revising the formula to incorporate some of John's suggestions, though it will in all likelyhood end up looking somewhat different.

[GaryChildress]

OK, I know this makes no sense whatsoever and is based on nothing remotely corresponding to reality but I toyed around with numbers and JuanG's equation for HDM and came up with this equation which creates accuracy values sort of similar to the current numbers in the version of AE I'm playing. Basically in this equation rate of fire and barrel caliber both factor into the accuracy. Granted there is nothing to take fire control systems into account.

Ideas? Thoughts? Denunciations?

[JuanG]

ORIGINAL: Gary Childress

OK, I know this makes no sense whatsoever and is based on nothing remotely corresponding to reality but I toyed around with numbers and JuanG's equation for HDM and came up with this equation which creates accuracy values sort of similar to the current numbers in the version of AE I'm playing. Basically in this equation rate of fire and barrel caliber both factor into the accuracy. Granted there is nothing to take fire control systems into account.

Ideas? Thoughts? Denunciations?

A few stray thoughts;
-Dont bother trying to fit the 40mm or other 'rapidfire' weapons into the same formula. I use mine down to ~3inches, as well as the breech loaded smaller guns (6pdrs, etc.). Basically anything with an RoF of over ~20 or so uses a different formula.
-I presume 'Cal.' in your formula is barrel length in calibers. I'm not sure about using that directly, because as mentioned a lot of short barreled weapons actually had better ballistic characteristics from an accuracy standpoint. Typically, it is the weapons that also have high muzzle velocities/energies relative to barrel length that tend to result in dispersion problems.

[el cid again]

ORIGINAL: Symon

ORIGINAL:
Listening to little clones using 2006 RHS data is noooot the greatest idea I ever heard.[8D]

Someone does not understand RHS data. It uses the stock system for naval gun accuracy. OTHER things have been modified, but not accuracy. I will confess to not seeing a problem with the base system and a good deal of statistical testing indicates it is remarkably good for such a simple algorithm. Elaborate changes to it do not appear justified.

[Symon]

ORIGINAL: el cid again
Someone does not understand RHS data. It uses the stock system for naval gun accuracy. OTHER things have been modified, but not accuracy. I will confess to not seeing a problem with the base system and a good deal of statistical testing indicates it is remarkably good for such a simple algorithm. Elaborate changes to it do not appear justified.

Over all the years, I have never once looked at RHS. I have never cared about RHS. None of the AE development team have looked at RHS. Nothing in AE comes from RHS. I know this for a fact.

RHS data is irrelevant. It does not use any of the info gathered over the years and incorporated into Michaelm's beta executables. In fact you have no clue what we have done. You are trying to sell a one-horse buggy with a tired Percheron, in a world of (well, not Porsches, but maybe) Honda Accords or an Infiniti Q.

Someone does not understand the modern world's data. Ciao. [;)]

[DOCUP]

Guys thanks for all your hard work at improving this wonderful game. I have a few questions, if I may. You may not be able to answer some of my questions due to game secrets and that's ok with me.

1. Does the formula work for land based guns also? Ie CD guns and such.

2. The accuracy number in the data base. Is that in percent or something else? IE 16 inch MK 7 has an accuracy number of 20 is that 20 hits out of a 100?

3. How does fire control radar help with accuracy?

4. Aren't some ships better gun platforms than others? Wouldn't fixed land or bigger ships have better accuracy than smaller one?

Thanks for your help in advance.

[Symon]

ORIGINAL: DOCUP

Guys thanks for all your hard work at improving this wonderful game. I have a few questions, if I may. You may not be able to answer some of my questions due to game secrets and that's ok with me.

1. Does the formula work for land based guns also? Ie CD guns and such.

It works for land based 'Naval' Guns, i.e., CD guns and the 'Naval' part of DP guns. Different formula for AA guns and the AA part of DP guns.

2. The accuracy number in the data base. Is that in percent or something else? IE 16 inch MK 7 has an accuracy number of 20 is that 20 hits out of a 100?

It is something else. It is not a %. It does have a little bit to do with the possibility of obtaining a hit, but how it works is something that's embedded in the code and we aren't allowed to say how it does it.

3. How does fire control radar help with accuracy?

It doesn't. Fire control radar was not implemented in the code for AE. Basically an irrelevant device type.

4. Aren't some ships better gun platforms than others? Wouldn't fixed land or bigger ships have better accuracy than smaller one?

Thanks for your help in advance.

Sure, ideally. But don't forget that 'accuracy' is just a convenient word used to describe a number in a data field. It is not accuracy. It is less than that and much more than that. It is simply a number, that takes many, many, many things into account, that's plugged into an abstract computer model.

Ciao. J

[DOCUP]

Thanks John for your input. Very interesting. That's disappointing about the FCR. Oh well, keep up the good work guys. Thanks for the education also. Love a free education.

[Symon]

Oh, heck. You paid for your 'free' education, so wtf ... yeah?

Ok, this isn't news because some smart people out there found out about it and posted something on it way back in the day. Don't remember who it was or when they did it, but I know they did 'cause I remember discussing it. So, since it's already out there somewhere, and since somebody else can do the same thing and get the same results, why not ... !

The algorithms that determine hits and damage are rather abstruse. But they surely depend on range. Range is an abstraction, but it does represent points along the curve that are appropriate for all guns.

To put it simply, the farther you shoot the greater your spread (less “accuracy”) and higher the trajectory. Yes the game thinks in simple trajectory terms. The range closes, and trajectories get flatter (more ‘accuracy”). Now the 8inch guns come into play. Whether they hit decks or hulls “at that range” is a function of their specific range.

In the real world, accuracy is a function of range. There is a sweet spot. It changes with every shot fired from that tube.

So at a given range, how many rounds would one expect to fall within a specific parallelogram in a “standard” time period. Now, do this for each separate one of a set of 5 different ranges. Now, do this for each separate one of 18, 16, 15, 14, 8, 6, 5, 4.7, 4, and 3 inch guns, and for each of the 5 different ranges, and you begin to see the eleranahanapant

Weird stuff, I know, but if you know what the playing field is like, you can always find a nice Mr Math that makes sense out of all the Xs and Os .

[DOCUP]

Thanks again John. That's a lot of work for all those guns. I'm not a math guy so I'll trust those who are. Thanks for making my game better.

[Alfred]

ORIGINAL: Symon

Oh, heck. You paid for your 'free' education, so wtf ... yeah?

Ok, this isn't news because some smart people out there found out about it and posted something on it way back in the day. Don't remember who it was or when they did it, but I know they did 'cause I remember discussing it. So, since it's already out there somewhere, and since somebody else can do the same thing and get the same results, why not ... !

The algorithms that determine hits and damage are rather abstruse. But they surely depend on range. Range is an abstraction, but it does represent points along the curve that are appropriate for all guns.

To put it simply, the farther you shoot the greater your spread (less “accuracy”) and higher the trajectory. Yes the game thinks in simple trajectory terms. The range closes, and trajectories get flatter (more ‘accuracy”). Now the 8inch guns come into play. Whether they hit decks or hulls “at that range” is a function of their specific range.

In the real world, accuracy is a function of range. There is a sweet spot. It changes with every shot fired from that tube.

So at a given range, how many rounds would one expect to fall within a specific parallelogram in a “standard” time period. Now, do this for each separate one of a set of 5 different ranges. Now, do this for each separate one of 18, 16, 15, 14, 8, 6, 5, 4.7, 4, and 3 inch guns, and for each of the 5 different ranges, and you begin to see the eleranahanapant

Weird stuff, I know, but if you know what the playing field is like, you can always find a nice Mr Math that makes sense out of all the Xs and Os .

Symon,

There are several threads where you discussed this issue previously. The ¨best¨ early one you discussed this was the following thread:

tm.asp?m=2268790&mpage=1&key=ballistics

from 2009. There was a subsequent, even more detailed, thread where you discussed it with your fellow gun bunny JuanG in 2011.

The 2009 thread I have linked I think explains how it fits into the game code in relatively easy to understand language for the layman. The 2011 thread is heavy with maths and comparisons with data from Okum´s site.

Alfred

Edit: Oh what the heck. Here is the 2011 thread referred to above.

tm.asp?m=2887820&mpage=1&key=ballistics&#2889693

[Symon]

ORIGINAL: Alfred
Symon,

There are several threads where you discussed this issue previously. The ¨best¨ early one you discussed this was the following thread:

tm.asp?m=2268790&mpage=1&key=ballistics

from 2009. There was a subsequent, even more detailed, thread where you discussed it with your fellow gun bunny JuanG in 2011.

The 2009 thread I have linked I think explains how it fits into the game code in relatively easy to understand language for the layman. The 2011 thread is heavy with maths and comparisons with data from Okum´s site.

Alfred

Edit: Oh what the heck. Here is the 2011 thread referred to above.

tm.asp?m=2887820&mpage=1&key=ballistics?

Thank you very much Alfred. I lose track of these things: have recollections, but ... it's not quite the same.

It's good to have things pulled into one big blivet [;)] You probably help me as much as you help the general community.

Thanks again, and Ciao. JWE

[sandman455]

Pretty hardcore computing going on in this thread. Perhaps some would be interested in hearing about a different approach?

Anyone else spend their youth playing with naval miniatures? General Quarters? Bismark? Seekrieg? Fun times. It was a good intro to surface engagements. Sufficient enough for me to conclude that while I love this game, it is very far removed from the realities of tactical engagements. In this game you can't expect much other than plenty of randomness which is actually a great attribute. Luck was a huge factor and many kudos to Gary Grigsby and 2by3 Games for integrating it into WITP's naval engagement routine.

When ship meets ship in a struggle for surface supremacy, you are really talking about a lot of variables that are extremely important to the engagement. Pick the historical fight and then think about all that was in play. Then measure those variables against what we have in our game to simulate what happen. Not a pretty picture is it. I mean who gets to sail up the straits when we duke it out in hex 82,88? It was after all the defining factor in that fight. Sadly, we can't really do much to recreate the real events to even test our success in simulating the real thing. You simply can't control the ships which is tantamount to determining the outcome of a naval engagement.

Not really complaining, we are playing a game covering the entirety of the war on a strategic level. Tactical engagements are not its forte. Yet using what we do have to control, how can we best simulate what actually took place during a surface fight? Look at what we have on the table:

1. Devices (with all the data points like quantity, penetration, accuracy, ammo, etc)
2. Ship Characteristics (again more data points like armor, speed, maneuverability, etc)
3. Crew experience
4. Captain skill ratings
5. TF commander skill ratings
6. Detection levels on the TF's
7. Weather and visibility

There's probably a few more that I just can't think of. And after everything gets assigned a value the game engine brings in the RNG and we all cross our fingers and hope that lady luck is nice to our side. So how can we improve upon this? Well one idea I struck upon was that I can minimize the role of rather insignificant variables thereby enhancing the role of other more significant ones.

Can anyone think of a surface engagement where accuracy (as it is defined in the game which is mostly ROF but as prior posts attest, some are trying to take it further) was a factor? Even a small, tiny, miniscule factor? Unless I took away some very bogus ideas from my readings, I can't think of a single one. Even the heroic efforts of the destroyer escort Samuel Roberts at Samar. Her main armament fell well below their effective ROF. Clearly it was not for lack of targets or crew motivation. Then why?

Well it comes down to how surface engagements pan out in reality. Often ROF is heavily determined by each ship's maneuvering and range to the target. Oh boy, just more variables out your reach in our game. Perhaps the most important factor effecting ROF was the simple fact that if you don't want to get hit, then you'd best move like someone is shooting at you. Yet this tactic is seriously bad news for your own accuracy and ROF. If you want to hit something, you best stop moving like you don't want to get hit. The logic is as profound as it is simple.

So what's the point? Well I kind of took the naval gunnery devices and gutted their accuracies within the database. I want to remove the data point as a significant variable between guns of similar capabilities. By leveling the playing field I've enhanced the variables that are left which I feel are far more likely to play a part in the outcome of the engagement.

My first step was to group guns up by six different classes:

1. The real big ones (over 13"/325cm)
2. The kind of big ones (10-12"/250-300cm)
3. The medium ones (8-9"/200-240cm)
4. The small ones (5.8-7"/145-180cm)
5. The tiny ones (<5.8"/145cm)
6. The Dual Purpose ones (all were 5.8" or less and must have AA capability - I also took their penetration away: 30-38mm max - since the number of AP rounds in the magazines was very limited and were mostly for scuttling duties not fighting engagements)

Six gun classes get only six different accuracy settings for every naval gun in the game:

1. 25 (a good 20-25% better than normal for these big guns in the database)
2. 30 (about the median in the database)
3. 35 (about 5-10% below median because CA's are often busy doing stuff that hampered accuracy if not ROF so much)
4. 40 (a good 15-20% below the median because CL's are just like CA's but more so)
5. 45 (this not far from the median)
6. 50 (this is 10-15% below the median and yes the jury is still out if I'll see unrealistic late war AA results)


I restrict the DP's for a variety of reasons not the least of which was that they were often mounted on poor gun platforms that were busy using all their speed and maneuverability to do things like launch torpedoes or lay down screens. Even when DP's were mounted as secondary batteries on bigger ships they would often get restricted in favor of optimizing main armament performance. The reduction had me a little concerned due to their AA capability but so far (early 1942) it doesn't seem to make much difference. Also because of the reduction in penetration given to these guns, cruisers won't find themselves outclassed by tin cans unless they are completely out numbered.

As for the heavy main armament on BC/BB's, it gets a boost because these ships were doing just the opposite of the small combatants. Even in the face of dangerous opponents they would frequently do everything possible to improve the accuracy of their main armament. ROF and even accuracy between guns in this class was meaningless in real engagements. Against smaller ships, big ships would rarely waste time with tactics like salvo chasing or smoke screens which would have dramatically reduced their ROF and accuracy. Most players seem to agree that the game has a tendency to short change the big boys in surface actions. I completely agree and hope bumping the accuracy of their guns will help resolve the issue.

The TLDR version: In my mod, I've removed the variation in accuracy of individual guns within their group, hoping to increase the significance of the other variables in the outcome. BTW, I'm not putting forth any groundbreaking ideas here. Just dig into any tactical naval gunnery game. I've yet to play one that has ever given ROF a significant role other than to ensure that weapons get to fire somewhat relative to what they could in relation to other guns. Some go further and factor in range to the target which is just another nod at it being insignificant since a big gun close to a target might be firing faster than a small gun much more distant. Clearly WITP's accuracy data can't account for such behavior so again we are hopelessly borked in recreating reality but perhaps we can at least weight the values that truly did make a big difference - even if it turns out to be plain old dumb luck.

[Symon]

ORIGINAL: sandman455
Pretty hardcore computing going on in this thread. Perhaps some would be interested in hearing about a different approach?
<snip>

Yepperino. There's a gazillion different ways to model this. But every model must fit within the abstraction constraints of the code. It's ground-up. You have a system that works for you within the game constraints, more power to ya.

[GaryChildress]

OK.

I'm going to give Plan 8-8-8 at least one more try (I basically gave up a while back). What I need at this stage is a good way to determine the "accuracy" value for ALL the different devices. I need a formula or however many formulas it takes where I can plug in certain known values and come up with numbers compatible with how the hard code for AE works for "accuracy". If there needs to be more than one formula for different device types like dual purpose guns versus battleship guns or tanks or bombs or whatever then I need ALL those formulas, otherwise this scenario of mine is just not going to work. Does anyone know of or have a list of these formulas or alternative formulas which work just as well with AE's computation engine (EDIT: and is willing to share)? All I really need are the mathematical expressions like JuanG gave earlier in this thread. "Draw a curve and plot points on the curve" just doesn't help me a whole lot. Call me stupid I guess. I'm basically looking for some concrete formulas which I can replicate in MS Excel to make life a bit easier.

Thank you very, very much for any and all help!

Gary