Let's Understand and Review: GUNNERY. (Please read OP before posting)

[Steeltrap]

EDIT of 23/06/20:

Not going to continue with this specific thread, or any future ones I had in mind. You can read my last post for more detailed explanation.

================================================================

I'm wondering if any of you like the idea of making several of these threads as it seems to me it's potentially a good way of drawing together our very wide experience, knowledge and talents and present them to the devs in formats that are relatively specific and focussed.

I do appreciate I'm asking you all to post things you very possibly have done previously, but I hope we'll end up with a comprehensive set of thoughts on each big topic area.

I'm happy to run with a few of these and monitor/update as we go. I might ask if I can get this put as a sticky, and possibly even see if I can get some better form of editorial control within this (and other) threads if the idea gains sufficient support to make it worthwhile.

============================================================

Hi all,

I thought it might be helpful to take stock of the GUNNERY MODEL as it stands, especially now that we're able to see EXACTLY the ENEMY ships' tech, guns, gunnery statistics through a battle, and all their real time gunnery modifiers.

With that in mind, could I ask you use the following HEADINGS format within posts. I realise not everyone will want to/care to follow it, but at least for those who do I'll provide some common format.

If you wish to discuss someone else's post, please make sure you continue to caption your own contribution with the "GUNNERY FACTOR" and/or quote the other post to achieve the same thing.

GUNNERY FACTOR - post with clear heading/statement of the specific gunnery model question/issue you are raising. Where the issue is a specific factor included on screen, please "USE EXACTLY THAT term" within quotation marks. Doing so will allow me to sort/update relevant contributions as we go.

EXAMPLE(S) - provide screen shots to illustrate.

Not at all seeking to teach anyone to suck eggs (despite my professional background I am surprisingly lousy with these things, LOL), but I tend to do the following:

I pause the game and use the mouse to bring up any relevant info if it's necessary to get a tooltip to display, or I want real-time gunnery factors on screen.

I sometimes make a few different images if I can't get all the info in the one (I might need more than one tooltip, for example), or am doing a "before/after" demonstration.

I use the built in snip via Left WIN + SHIFT + S as it cuts down on image size and makes me focus on what's important in the image. I also like it because if I select it immediately I can then CNTL + S to save the image and I use those save names to describe what's in it/why I'm saving it. I mention it solely as a tip for anyone interested, but I'm sure most of you are way ahead of me (and I don't like image hosting sites but perhaps I ought to get over that).

DISCUSSION - give the meat as to why you've raised the specific GUNNERY FACTOR, such as what you see as an issue with it, whether you think it's a good factor yet less than ideally implemented, or a poor one (obviously explain why) and would be better off being removed (and possibly replaced with another to achieve the intended result in a better fashion).

SUGGESTION - make a suggestion as to how to address the issue you're raising.

As mentioned in the intro, if enough people think this sort of thread might help bring feedback/experience together I'll make some others, so this is something of a trial run.

Cheers

Edited June 22, 2020 by Steeltrap
23/06/20: advising people I won't be continuing with this idea

[Steeltrap]

GUNNERY FACTOR: "Target Manoeuvre"

EXAMPLE:

DISCUSSION:

This is taken from a target travelling at less than 1 knot (hence the title of my screen snip, lol).

1 knot = 1.852km/h, or (1852/3600 = ) 0.51m/s. For a shell flight travel time of 20 seconds, say, the target is going to travel 10.2m.

I suggest it's nonsensical for it to be able to apply a penalty through "manoeuvre" that is GREATER than the bonus its slow speed gives to anything shooting at it. This is part of the broader "range rate" discussion I expect to see, which is to say it's the degree to which you alter your position compared with your enemy's prediction of where you'll be as they fire that is really the important point. It's also why "Target Speed" is equally a poor factor on its own.

I think that both this AND "Target High/Low Speed" are poor factors and ought to be replaced with something that is a better simulation of true gunnery, specifically range rate. I understand this will have a cascade effect through the model, but I don't see how the current state of affairs is defensible other than "it's easier to programme".

SUGGESTION:

Remove Target Manoeuvre and build a proper factor that looks at "range rate" (where range rate is "the rate at which the distance from the measuring equipment to the target or signal source that is being tracked is changing with respect to time").

Will probably cause a greater examination of the gunnery model in total, but I happen to think that's a good thing.

Edited June 15, 2020 by Steeltrap

[DougToss]

I really appreciate you doing the legwork on this. I like what you're doing, I like the format and I'll do my best to contribute when I get the chance. 

[Tycondero]

4 hours ago, Steeltrap said:

SUGGESTION:

Remove Target Manoeuvre and build a proper factor that looks at "range rate" (where range rate is "the rate at which the distance from the measuring equipment to the target or signal source that is being tracked is changing with respect to time").

Will probably cause a greater examination of the gunnery model in total, but I happen to think that's a good thing.

First I would like to applaud this sort of initiative. I think it can massively improve the overview of issues to the developers. 

About the suggestion made: yes, I completely agree now that you spelled it out like this. Makes no sense if a target is slowly travelling for it to just get a manouvre buff because it uses its rudder. Could they not just implement a system that updates the location of ships every second or so that determine the change compared to the previous tick/second. The delta (change) in location (along with shell travel time) would then determine the difficulty of the shot.

[Steeltrap]

GUNNERY FACTOR: "Target Manoeuvre"

19 minutes ago, Lieste said:

Recall that all assumed target speeds, ranges, headings et al are derived from sequences of range and bearing measures, from an unstable and moving platform and often in poor conditions of visibility. They generate an assumed track, using mechanical computers and experienced operators, and this assumed track, projected forward is the target for directed shooting. Changes to rates have similar effect to increased noise in the tracking, and computed track will lag behind the actual as *real* changes are detected. Spotting corrections can also be used to compensate for ambiguity or error in rangefinding measures and their derived plots, but again will lag by around two 'time of flight' (one for the shot to arrive, and one for the resulting spot to be relayed and implemented into the track).

This is exactly why I said I suspect we're going to end up looking at the broader issue of the whole simulation of "the gunnery solution", how a ship gets one and then how it recognises it's no longer acceptable and produces a new one, which can even lead to shifting from full RoF back to reduced, spotting fire.

As it is now, the system updates your numbers as soon as you fire, and then they shift in real time as you/your target manoeuvre or close/open the range.

[Steeltrap]

Topic's been here a week, nobody has put any posts related to specific issues other than myself. I was going to post 2 more, but have changed my mind.

The devs, however, ALREADY HAVE all the details. They could share them and ask for our opinions, such as "here's the table of speed v to hit adjustment, what do people think?". I was in fact compiling such a table when I concluded it was a waste of time and effort.

They've had months and months of feedback/constructive criticism and suggestions.

None of which makes the slightest difference.

Meanwhile, their answer, if the hotfix announcement is to be read as written, is simply to "make stuff up" ("dispersion value" changes to address problem of missing close targets) to produce a different result and see if people are happy with that. How does that address the genuine root causes of why it was so hard to hit certain targets? "Dispersion values", which I fear are DREADFUL BLOODY HIDDEN NUMBERS JUST AS WG LOVES DOING, may well have contributed, but what about the HUGE penalties for target speed and manoeuvre? They're irrelevant are they?

As I said elsewhere, I'm not going to bother any further because the simple fact is all evidence points to the devs not caring one way or another.

Cheers

Edited June 22, 2020 by Steeltrap

[madham82]

I'm afraid you are right. Unless they express some interest or real direction on the core issues, not worth our effort to provide more discussion. 

[Cptbarney]

31 minutes ago, madham82 said:

I'm afraid you are right. Unless they express some interest or real direction on the core issues, not worth our effort to provide more discussion. 

I hope it isn't the case, the only thing i can think of is that they will unvail this when the campaign arrives. I havent provided anything yet as im waiting for what they will say.

I understand they cann't spill the beans on everything, but some clarification would help as to stop peeps from getting too annoyed or burnt out. Im not sure what they use to calculate shell deviation and drag rates but i hope its not going down to the road of sigma and dispersion like in world of dankships.

Armour and AP mechanics need an overhaul or major improvements, gunnery needs to be changed for some aspects and shell deviation should start when the shell has left some way after the barrel and start to deviate from its initial shell grouping.

[Steeltrap]

2 hours ago, Cptbarney said:

I understand they cann't spill the beans on everything, but some clarification would help as to stop peeps from getting too annoyed or burnt out. Im not sure what they use to calculate shell deviation and drag rates but i hope its not going down to the road of sigma and dispersion like in world of dankships.

Next time you open the game, read the "news" section, namely the notes for version 7.

Part of it speaks about (I'm paraphrasing from memory) "changes to sensitivity of gunnery, so ships like DDs have a better chance of evading shots" etc.

Now the hot-fix speaks about changing dispersion so closer targets, specifically mentioning DDs, ought NOT be so difficult to hit.

So they introduce a change in v7 to make them harder to hit, now a hot-fix to make them easier?

It's almost as though the gunnery model is the problem, and rather than playing hokey pokey with factors not known to the players (an approach to design I personally despise for good reasons, but that's an entirely different matter) perhaps it might be a good thing to look into the core of the gunnery model as we've been asking for only 4-6 months.

The impression they're giving is they're throwing darts at a board and hoping to hit something a majority of players might like. How that relates to a good design, let alone "realism", escapes me. Chasing popularity is NOT part of core element design. You do that later once all your interlocking essentials are solid, reliable and passing the success criteria established for them.

That I'm prepared to write posts like this is the surest sign there is that regardless of what the devs are doing, it's high time I took a break.

Which is exactly why I'm going to.

Lastly, if you really DO want something less pulled out of, ahem, thin air and more deserving of the claims to "reality", might I suggest people spend less time posting about all sorts of WW2 ships, not to mention extreme weapons and other trivia, and CONSTANTLY press them on the basics? Assuming of course you don't all reach the same conclusion RAMJB and now I have.

Cheers

[DougToss]

I'm sorry I haven't posted in here earlier. I have posted so frequently about gunnery in so many threads, it really is as @Cptbarney said at the point of burn out. There is no point taking hours to read about naval gunnery, on top of a career as an artilleryman, and now working on defence projects, reading about how gunnery is simulated in my spare time, and condensing that into posts for it to all fall on deaf ears. 

I am happy to read firing tables and Admiralty reports, if I think that contributing that research is a useful exercise. If not, I could devote my spare time to other interests. So many of the most engaged and active contributors on this board have burnt out. That's really disappointing because @Steeltrap, @akd, @RAMJB and others have clearly put in many manhours of unpaid work to help out with a project they are clearly passionate about. That would be any developer's dream when they put out the call for testing. If used right, it saves them the cost of a researcher or a consultation with a naval historian. That's invaluable, and yet, what have the results been? 

A PDF made of the longform posts made by these forum members would be tens of pages long and cover just about any problem or complaint expressed about the game, and quite likely, any roadblock hit during development. How does fire control work? @RAMJB has posted about how it works in reality and how the existing implementation is flawed and can be improved. What are the terminal ballistics of various calibers of shell? @akd has posted about the effect on target of different rounds. 

My limitation here is that I am a professional in the business of ballistics, but not very well equipped to discuss video games. I don't know how to communicate in terms of code, or mathematics. Maybe that's the missing link, because I can readily see the problems, but can't express solutions in terms of what code would do a better job. 

Having said that, anyone can evidently compare real speeds, gunnery tables, combat records and so on to the existing game and see something needs to change, and I agree that that is a worthwhile pursuit for everybody. 

3 hours ago, Cptbarney said:

Armour and AP mechanics need an overhaul or major improvements, gunnery needs to be changed for some aspects and shell deviation should start when the shell has left some way after the barrel and start to deviate from its initial shell grouping.

I've butted heads with @Cptbarney often, but that we agree on this is a pretty clear indication that there are foundational problems here worthy of addressing. 

[Bilderberger]

I deeply despise the degree of asymmetric information we're facing. And I also think that's the core of the issue at hand. Fortunately, some of it has been lifted recently. Yet, so many questions remain - I need not tell what systems ought to be improved as many people have already contributed many good and precise arguments concerning said issues, gunnery, armor and the like. 

I think many of us have the impression that our input (as a community) is more or less ignored. And this might very well be the case - or not. Obviously, the devs cannot reveal everything they have planned. That is impossible. Still - I think there is a lack of communication. Let me give you an example.

 

Concerning gunnery and speed penalties - Many of us have made points about it, they spent much of their spare time researching and analyzing data to improve the situation. Perhaps the devs read them and decided to rework the system. Perhaps they did not. We do not know. We cannot even say whether many mechanics are final or not.

Are they really solely working on the campaign right now, something they've told us would take yet another 6 months, just weeks before the official release ought to have happened? Or are they also working on something else? Maybe they've put the data to good use and they're cooking one of the best armor and ballistic systems. Or maybe they call the current Status Quo a day and leave it at that. 

I dislike that. I really do. I do not ask for much but I think some clarification is direly needed!

As Doug already mentioned, certain members have invested incredible amounts of time and effort into this project. But will they continue doing so? Are people willing to take an investment if the shroud of uncertainty grows ever larger?

[bshaftoe]

On 6/15/2020 at 3:29 AM, Steeltrap said:

GUNNERY FACTOR: "Target Manoeuvre"

EXAMPLE:

DISCUSSION:

This is taken from a target travelling at less than 1 knot (hence the title of my screen snip, lol).

1 knot = 1.852km/h, or (1852/3600 = ) 0.51m/s. For a shell flight travel time of 20 seconds, say, the target is going to travel 10.2m.

I suggest it's nonsensical for it to be able to apply a penalty through "manoeuvre" that is GREATER than the bonus its slow speed gives to anything shooting at it. This is part of the broader "range rate" discussion I expect to see, which is to say it's the degree to which you alter your position compared with your enemy's prediction of where you'll be as they fire that is really the important point. It's also why "Target Speed" is equally a poor factor on its own.

I think that both this AND "Target High/Low Speed" are poor factors and ought to be replaced with something that is a better simulation of true gunnery, specifically range rate. I understand this will have a cascade effect through the model, but I don't see how the current state of affairs is defensible other than "it's easier to programme".

SUGGESTION:

Remove Target Manoeuvre and build a proper factor that looks at "range rate" (where range rate is "the rate at which the distance from the measuring equipment to the target or signal source that is being tracked is changing with respect to time").

Will probably cause a greater examination of the gunnery model in total, but I happen to think that's a good thing.

Wouldn't it be much easier to limit the malus of "Target Maneuver" to be a function of the current speed... or even of the current malus of "Target Slow speed"? Or would this just be a patch over a patch?

Couple of alternatives:

1. Choose what would be the default speed for hard and evading maneuvers. Let's say 25 or 30 knots. The smaller your speed is in relation to that speed, the more the"Target Maneuver" is diminished. If, so to speak, and using invented numbers, the default malus for "Maneuver" is -25%, you can multiply it by CurrSpeed/25. The lesser curr speed, the lesser the final malus. So if you're evading or turning at the absurd velocity of 1 knot... you will have a malus of -1%. If a destroyer at 38 knots is turning hard to starboard, the final malus will be 38/25 * 25%... or this is... 38% malus. I don't know if I would still keep the speed malus.

2. Simply apply somehow the BONUS you are giving your enemy because you're going slow as a limit factor of the MALUS you're giving your enemy because you're evading/turning/maneuvering. So, for example, quick and dirty thumb estimation: simply substract this factor from the maneuver malus. So... you're going slow, and this makes easier to an enemy to fire at you (it gives them a 28% bonus)... and you're "maneuvering" in a way (hard turn, zig zag, whatever) that would normally penalize your enemy's aiming substracting a 31% of his possibilities of hitting you... With this approach, 31-28 =3%. So the final maneuver malus to your enemy is a mere 3%, because you're going SO slow, that yes, you're turning, but this factor is negligible or almost negligible. Conversely: you're going ultra fast... so fast that the enemy is having a hard time hitting you (malus for speed of -10%). -10 -28 = -38% total malus, because now your speed supports your maneuvering, and not limit it.

This is, and in short: simply make the "maneuvering" malus a function of the current speed OR of the current speed bonus/malus.

Edited July 2, 2020 by bshaftoe

[HMNZS Achilles]

Thank you for opening this thread. I've only been playing around with the game since late in Alpha 5.

 

I do have some thoughts on the complex simultaneous equation that is topic of gunnery.

Agree rate of change is one of the more important factors in naval gunnery (probably two rates of change - rate of range change and rate of baring change, but any whoo) Essentially if there is a low rate of change the firing ship and the target ship are travelling roughly on same course at a very similar speed (assuming the rate of change calculation is correct) and while that is maintained the laying of guns on the point the target will be when the shells land is simpler.

High rate of change means, therefore, course and or speed of firer and target are widely different then the prediction becomes a bit harder and the consequences of not correctly calculating rate of change more.

Calculations were made by mechanical computers of varying levels of automation and plotted on a plotting table or its equivalent (at least once longer range director controlled fire came along).

To overcome the many problems noted above with calculation (e.g. accuracy of range finding, correct deduction of relative course etc) observation of fall of shot is used to empirically correct the plot. So the longer both the target and shooter (both likely playing both roles of course) the better the accuracy of the plot should get even at a high rate of change (though not as good as low).

The plot becomes less reliable when either the target or shooter changes course (and to a lesser extent speed) and the observed fall of shoot is likely to fall off target and the target will need to be reacquired and the accuracy the plot re-established.

What's the game implication of that? And the game mech seems to be doing some of this,.

A) Accuracy should build the longer both target and shooter both maintain course, and improve particularly following the fall of shot

Accuracy should degrade when either change course, the bigger the change in course (well change in rate really) the bigger the degrade and more so if the shooter violently changes course because that will likely to throw out the gyro stabilizers that are key components of the plotting table/computer and they take a while to settle down again.

C) A high speed target, even at 55 knots, that is holding a steady course (and therefore steady rate of change or range and bearing) should become a predictable target to a steady shooter. I don't think the game yet does that. Speed of target seems to be a permanent degrade on accuracy (though my testing isn't limitless yet).

 

Happy to have the developers tell me I'm teaching them how to suck eggs and they are doing this.

cheers.

 

 

Edited June 30, 2020 by HMNZS Achilles
B) got changed to a sunglasses emoji -