This needs to stop

[brucesim2003]

The AI is STILL building ships that have far too much emphasis on speed. Tied in with the target speed gunnery modifier, it results in an almost unhitable ship.

This is destroying my interest in playing the game. The 2 hits I did get were early in the game before the AI was up to speed. Now my hit chance is less than 2%, with 8x 13" mk 3. The AI is hitting me with virtually every broadside. As the battle of Jutland proved, speed is NOT armour.

This is a 1912 battle. How does a 23k ton ship get 32 knots in 1912? If you can stop the AI from doing this, put a hard cap (increasing with year) on the speed of ships. better yet, rethink the engine algorithms, because that sort of speed was simply not possible for a battleship in 1912.

The biggest problem though is the gunnery modifier. It needs adjusting, with some alacrity.

[arkhangelsk]

I hate to break it to you, but for a ~23000 ton ship, historically you only need about 90000HP for 32 knots (ref Courageous). While the game underestimate HP requirements, battlecruiser engines were getting close to that power rating by then. Tiger had 85000HP. Including forcing, she can go over 100,000SHP. It's a little early, but not maddeningly so. And while speed might not be armour, it is also willing to drop all the way to 9-inch guns for some more armor. In general terms, this is a plausible design.

The game does favor speed. On the other hand, but history does suggest that once you get above a certain speed in the 25-30 knot range, ships do get a lot harder to hit.

Also, actually look at the hit rates for each gun. 9" and 12" are favored calibers. 13" is not. That is actually not very realistic, but so many people complained the destroyers are unkillable (they WERE killable!) with little guns the devs must have given in...

[brucesim2003]

Yes, it dropped to 9" guns. All 16 of them. And has almost 17" of (I assume) belt armour. And maximum bulkheads. And he has 15% chance to hit vs my 2%.

Speed is not THAT good at protecting a ship. The Lion class (27-29kts, depending on sources) got absolutely hammered at Jutland. The large light cruisers (32kts) got riddled at 2nd Helgoland. The gunnery modifier is ridiculous.

[arkhangelsk]

Yeah, but they are still little. The 17" may well be conning tower or turret. The elevated hit chance is why the 9" is the "hero gun" of the game - it is right at an apex before the hit rates drop again, said apex put in because of some people's impatience with the secondaries. The game AI is just exploiting a mechanic put in response to the whines of human players.

https://en.wikipedia.org/wiki/Second_Battle_of_Heligoland_Bight

Quote

Casualties and losses:
1 light cruiser damaged

https://en.wikipedia.org/wiki/HMS_Courageous_(50)

Quote

Courageous fired 92 fifteen-inch shells and 180 four-inch shells during the battle,[19] and the only damage she received was from her own muzzle blast.

https://en.wikipedia.org/wiki/HMS_Glorious

Quote

One 15-inch shell hit a gun shield of SMS Pillau, but it did not affect her speed. At 8:33 the left-hand gun in Glorious's forward turret was wrecked when a shell detonated inside the gun barrel. At 9:30 the 1st Cruiser Squadron broke off their pursuit to avoid a minefield marked on their maps. The ships turned south, playing no further role in the battle.[17] Glorious required five days of repairs to fix damage caused by premature detonation and her own muzzle blast.[18] She fired 57 15-inch and 213 four-inch shells during the engagement.

Doesn't look very affected by the enemy gunfire to me.

Oh, and before you get to 30 knots, the reduction is not that severe in game.

Edited April 12, 2020 by arkhangelsk
Paragraphing

[Shaftoe]

I agree. AI ship designs are often completely unbalanced. It's not even the fact that they can go this fast, it's the fact how often AI builds unrealistically fast vessels. That's unacceptable, if this game is truly meant to be realistic.

[Shaftoe]

@Nick Thomadis

 

[Steeltrap]

[In case it's not obvious, all offered in the spirit of constructive criticism and suggestions. We're in Alpha and plenty can and will change, and I have no expectations things ought to be "just right"]

Logically it would be reasonable to suspect the most important points about target speed are::

1. ability to estimate it, and

2. potential for it to allow significant spacial separation between anticipated location at which the gunnery solution has directed you to shoot and actual location due to application of rudder.

If you're travelling at 30 knots on a known course, once I've determined both it seems to me there's not a lot of difference in the difficulty of generating just as accurate a solution for that target against that of doing the same for a ship travelling at 20 knots.

Which really brings us to the issue of the sophistication of the fire control system and how well it can gather information (to determine range, speed, course) and use it to take all the relevant factors (shell travel times based on shell velocity, flight characteristics and range, pitch and roll, and so on)  into consideration to fire shells at where the target is likely to be when the shells arrive AND land shells there. It also has to take feedback, obviously including shell splashes, to amend the solution in reasonable time.

Here's what I would prefer (in no particular order):

- overhaul the 'target ship speed' modifier. I find on its face the idea that a 47,200t ship travelling at 36 knots can apply a -90% factor to my gunnery absurd (yes, HMS Hood in the Battle of Denmark Strait mission, I'm looking at you, which is all I COULD do because I sure as shit couldn't HIT you....and the idea of ANY BB/BC travelling at that speed prior to nuke power is bollocks), UNLESS there are some other mitigating factors that make this less significant. Radar appears to add a MUCH greater "target ship size" modifier, often to very large numbers, so there is that. Even so, I find the idea a ship travelling at 36 knots is so much harder (-90%) to hit than one travelling at about 15-16 knots as highly questionable at the very least, and more likely frankly absurd.

- revisit the whole "own ship speed" modification. If I get +22% for travelling at 25 knots while giving my enemies a penalty of -30%, yet can surrender that +22% for an increased penalty of -50% at 31 knots, it OVERWHELMINGLY favours charging around at the highest speeds possible because the benefit YOU get from travelling at lower speeds is substantially less than the benefit your ENEMY gets from you now being a substantially easier target to hit.

(I realise the previous two are more or less opposite sides of a "speed penalties/bonuses" coin. I happen to think they are significantly out of order as they appear to ignore entirely the degree to which gunnery degraded FROM the ship travelling at high speed. You surrender a bonus, but at no point do you start to accrue a penalty, yet you should. Vibration alone could cause difficulties with gunnery optics, and vibration accompanied increased revolutions on the screws in pretty much every ship to varying degrees. Why is there no "ship exceeding x-speed", where 'x' would logically be a % of listed flank speed? It's not clear to me why the band of the speed that attracts the maximum bonus is as narrow as it is, either, as I'd have expected it to go lower than the 21-24 knots of a 31 knot speed ship, for example.)

- tie the "target ship manoeuvre" penalty to speed. I've raised this before, how a ship travelling at 0.5 knots can apply a -30+% penalty to your gunnery is clearly nonsensical. I'm sure it's a case of the modifier being a crude factor based on the actual rudder position of a ship, and that the implications of that have not been worked through. It needs to change.

- make the gunnery model adjustments more in tune with a real world experience. As things stand, shell splashes are irrelevant; the model is updated as soon as you fire. We KNOW more or less how this stuff worked. Why not have the system do its level best to recreate it as closely as possible, which includes the point about shell splashes etc? Your own ship should revert to "ladder aiming" IF the observed splashes show your current gunnery solution is no longer valid. It might seem a minor quibble, but I think it's significant in HOW it looks to the player. Why not have a "spotting shell splashes" interim between "ladder aiming", ultimately resulting in "locked"?

- reduce the speed with which gunnery is locked. One pair of shells lobbed at an enemy at just about any range ought NOT result in "locked" IMO. I am aware the game uses an INCREASING bonus for successive shots AT a "locked" target, which is fair enough, but I think it's too important a concept to have it hidden in the fine print of gunnery details many won't necessarily look into with such focus. I'd rework the whole presentation of ladder aiming, locked etc on the gunnery panel.

When I played the scenario of the approximation of the Denmark Strait, I was both shocked and laughing when I saw all 47,200t of Hood doing pirouettes at 36 knots, barely slowing, and applying a GARGANTUAN -90% penalty to my gunnery. Really? LOL, sorry, just NO.

The irony is that I had popped into the forum to start a thread about something that bothers me already, namely "ship handling and performance". The game is building in GROSSLY inflated ship performance, and I don't mean just the ridiculous speeds of BB/BCs that are possible, I mean the rates at which they (and in fact EVERYTHING) can CHANGE their speeds. I've posted elsewhere the chart used by the RN for speed curves of just about all their major capital ship classes of the late 20s/early 30s, and the game is presenting an utterly alien model of ship handling. I'll still write that, but this thread was sort of related.

Happy to hear other thoughts/criticisms of course.

Hope everyone's keeping safe.

Cheers

Edited April 12, 2020 by Steeltrap
Format changes, slight alterations of wording of content

[Cptbarney]

2 hours ago, Shaftoe said:

I agree. AI ship designs are often completely unbalanced. It's not even the fact that they can go this fast, it's the fact how often AI builds unrealistically fast vessels. That's unacceptable, if this game is truly meant to be realistic.

Yeah i had frenchies going bloody 30knots plus at 1910 roughly while my bc's were doing close to 29 themselves.

You can hit them, but it takes longer to do so. I think along with older ships and modern dd and cl hulls we need to fix the AI as well.

Also it seems some of the modifiers are bonkers atm, but hey this what this alpha is about finding stuff and fixing it before general release *loud breathing noises outside*

*cocks 32 gauge shotgun*

Also with smoll dev team (3-5) its gonna take awhile aswell.

Edited April 12, 2020 by Cptbarney

[madham82]

I can see a fast small ship at range being hard to hit. I do not see why a large fast ship at medium to close range is hard to hit. To me the closer the range the less significant the speed advantage becomes. The issue with accuracy in these cases is shell travel time and distance the target ship travels in that time span. The closer in that doesnt really matter until you get to a point the guns can't traverse fast enough to keep a bead on it. Maybe I'm missing some key things, but not finding any real world examples that prove otherwise. 

[lordcmdr]

Another reason to just let us design our own enemy ships 😉

[GUTB]

I like how no one points out how the AI cheats like crazy. There's no way that ship is getting to 15% accuracy at that speed at tech level. Also that hull maxes out 22.5k t, the AI just slaps what it wants on it and doesn't care about weight restrictions, possibly up to a limit. That hull is also illegal, you can't arrange those funnels that way. I know because I just tried building it -- the closest I could get it messing around with funnel options and accuracy got up to 9% or so at close range...that's with a particularly slow AI.

Also, the AI cheats like mad in damage control. No amount of money or weight can be scarified to get AI-levels of damage control cheating.  The AI just cheats. The player is forced to build ships with maximum stability and gunnery tech to get decent accuracy -- that also means not having 3 funnels generating huge amounts of smoke interference -- while the AI just cheats. Simple as that.

[Nick Thomadis]

4 minutes ago, GUTB said:

I like how no one points out how the AI cheats like crazy. There's no way that ship is getting to 15% accuracy at that speed at tech level. Also that hull maxes out 22.5k t, the AI just slaps what it wants on it and doesn't care about weight restrictions, possibly up to a limit. That hull is also illegal, you can't arrange those funnels that way. I know because I just tried building it -- the closest I could get it messing around with funnel options and accuracy got up to 9% or so at close range...that's with a particularly slow AI.

Also, the AI cheats like mad in damage control. No amount of money or weight can be scarified to get AI-levels of damage control cheating.  The AI just cheats. The player is forced to build ships with maximum stability and gunnery tech to get decent accuracy -- that also means not having 3 funnels generating huge amounts of smoke interference -- while the AI just cheats. Simple as that.

AI surely, if you want to believe me, does not cheat. I hate AI cheating myself and do not want it in any game I am part of. So please, do not say that again (10x times is enough in this forum), because it becomes personal. And when I say "Cheating" I mean some kind of bonus to the AI for the same stat. AI might be more efficient in building a better ship than an average player, but that is not a cheat. If player has issues with his accuracy, he can try the auto-design feature and see if the ship accuracy is improved. If there is any related bug, we will investigate and fix, but as we know there is none on this matter.

The only advantage/disadvantage right now is that the AI does not have a stable restriction to operational range (as the player) for Custom Battles and Naval Academy. So sometimes, AI might have some extra weight to use for gunnery/speed. Sometimes it might have full operational range and have less available weight. This is a known and temporary issue, that -we have told- we are going to address in next updates, and will be solved permanently when we introduce the campaign, so the balance must work there logically for all cases, for player and AI. We keep this "issue" because otherwise we could have unbuildable or very weak AI ships, which would be a much bigger problem.

[arkhangelsk]

Here. Have a 1912 HMS Barfleur.

https://photos.app.goo.gl/WxqqPycqbKzyowj27

It really is no effort to put the thing together, complete with 16 9 inch guns, 32.5 knot speed and plenty of armor.

The computer estimates about 85000 SHP to get to the required speed. Based on historical precedent, while it is not blatantly impossible, it is a very favorable ratio and IMO shouldn't be granted to a Dreadnought hull (this is Dreadnought I, maxing out at 27000 tons), but a battlecruiser hull.

In any case, since it is perfectly feasible for a human to throw the hull together, it is not a case of cheating by the AI.

[brucesim2003]

Well, I didn't imply (or mean to) in the OP that the AI is cheating. But the high speed is far too effective at knocking down accuracy.

[GUTB]

25 minutes ago, arkhangelsk said:

Here. Have a 1912 HMS Barfleur.

https://photos.app.goo.gl/WxqqPycqbKzyowj27

It really is no effort to put the thing together, complete with 16 9 inch guns, 32.5 knot speed and plenty of armor.

The computer estimates about 85000 SHP to get to the required speed. Based on historical precedent, while it is not blatantly impossible, it is a very favorable ratio and IMO shouldn't be granted to a Dreadnought hull (this is Dreadnought I, maxing out at 27000 tons), but a battlecruiser hull.

In any case, since it is perfectly feasible for a human to throw the hull together, it is not a case of cheating by the AI.

You're using the wrong funnels.  But I'll admit that I thought the OP was showing a BC, not a BB in which case the hull layout isn't illegal. Trying it as a dreadnaught I was able to achieve the 15% accuracy at close range....now obviously 9" guns are useless to a player vs the AI, but vs a player the rof and being stuck with non-enhanced damage control and protection / penetration it can spell trouble.

[madham82]

58 minutes ago, Nick Thomadis said:

AI surely, if you want to believe me, does not cheat. I hate AI cheating myself and do not want it in any game I am part of. So please, do not say that again (10x times is enough in this forum), because it becomes personal. And when I say "Cheating" I mean some kind of bonus to the AI for the same stat. AI might be more efficient in building a better ship than an average player, but that is not a cheat. If player has issues with his accuracy, he can try the auto-design feature and see if the ship accuracy is improved. If there is any related bug, we will investigate and fix, but as we know there is none on this matter.

The only advantage/disadvantage right now is that the AI does not have a stable restriction to operational range (as the player) for Custom Battles and Naval Academy. So sometimes, AI might have some extra weight to use for gunnery/speed. Sometimes it might have full operational range and have less available weight. This is a known and temporary issue, that -we have told- we are going to address in next updates, and will be solved permanently when we introduce the campaign, so the balance must work there logically for all cases, for player and AI. We keep this "issue" because otherwise we could have unbuildable or very weak AI ships, which would be a much bigger problem.

Should also have mentioned that Naval Academy missions are not based on same tech for player/AI. AI does get better tech in cases to provide a challenge. Many people forget that and believe the AI is cheating. 

There are some issues of AI building rules probably not working as expected. There's a screen shot in the Alpha 5 feedback thread showing an AI BB with secondary tower hanging over the stern for example. 

[Nick Thomadis]

1 minute ago, madham82 said:

Should also have mentioned that Naval Academy missions are not based on same tech for player/AI. AI does get better tech in cases to provide a challenge. Many people forget that and believe the AI is cheating. 

There are some issues of AI building rules probably not working as expected. There's a screen shot in the Alpha 5 feedback thread showing an AI BB with secondary tower hanging over the stern for example. 

Yes, true, thanks for noticing. This minor bug is known (for french BB) and will be addressed, hopefully, in the next update.

[GUTB]

1 minute ago, madham82 said:

Should also have mentioned that Naval Academy missions are not based on same tech for player/AI. AI does get better tech in cases to provide a challenge. Many people forget that and believe the AI is cheating. 

There are some issues of AI building rules probably not working as expected. There's a screen shot in the Alpha 5 feedback thread showing an AI BB with secondary tower hanging over the stern for example. 

It boggles me how someone can't tell the AI is cheating. Even if you're willing to pretend the AI has the same accuracy and penetration / detonation chance, etc, as the player, you absolutely can't fail to notice the AI's miraculous damage control. Hey where can I equip a mass teleporter bilge pump?

[madham82]

1 hour ago, GUTB said:

It boggles me how someone can't tell the AI is cheating. Even if you're willing to pretend the AI has the same accuracy and penetration / detonation chance, etc, as the player, you absolutely can't fail to notice the AI's miraculous damage control. Hey where can I equip a mass teleporter bilge pump?

You'll have to post some examples and screens of what you are seeing (start a new thread). I've never seen anything odd about AI damage control...yet. I have seen issues with the damage model in general. 

[Tycondero]

Damage control can be very effective for the player too, if he/she is willing to invest in it.

In general I think the speed accuracy penalty should be there, but I feel that it should be based on the relative speed vector of both ships. If both ships are going at the same bearing with similar (high) speed, the penalty on accuracy should be way smaller as the relative difference is much less. Ships that go extremely slow will still get an accuracy penalty for target speed, but could get an accuracy buff due to other factors, such as platform stability (not moving much) for instance.

Edited April 14, 2020 by Tycondero

[Steeltrap]

16 hours ago, GUTB said:

It boggles me how someone can't tell the AI is cheating. Even if you're willing to pretend the AI has the same accuracy and penetration / detonation chance, etc, as the player, you absolutely can't fail to notice the AI's miraculous damage control. Hey where can I equip a mass teleporter bilge pump?

To answer your derisory question, it turns out I can equip that "mass teleporter bilge pump" simply by choosing maximum bulkheads, a few component choices, and looking up the resulting damage control stats in the ship builder.

Amazing! Who knew??

FIRST thing I do on my designs is shift bulkheads to maximum because I know how disproportionately effective that is. "Broken" would be a much harsher term, and a conclusionary statement rather than one of fact, so I don't use it.

I do my best not to get personal, but in this case the issue seems to be one of your behaviour/attitude, so it's all but impossible to reply without doing so.

You DO realise you're telling the programmer of the game, Nick, someone with a well deserved reputation for doing great things with AI within games (his TW: Empire "Darth Mod" was simply a "must have" for the game to be worth playing, and a terribly indictment of what CA allowed to be released), that HE'S LYING, right?

Calling anyone a liar, let alone the lead programmer who's made a clear statement contradicting your claim, is, at least as I see it, entirely uncalled for. With no evidence provided to support your accusation, it's also entirely unjustified.

Might I suggest the following instead?

Provide what you consider to be EVIDENCE of this cheating, remembering that for it to be a cheat it MUST be something the AI can do THAT PLAYERS CANNOT.

Let's take your allegation about damage control. As it happens, it's one I and many others have commented on.

I think it's absolutely valid to suggest that the bulkhead status is vastly more prominent in damage control IN ALL WAYS AND SITUATIONS than it ought to be. I think that because it's exactly something I have commented on more than once myself. That's NOT the same as it being a cheat, however.

To be an AI cheat, you need to demonstrate that YOUR damage control abilities are not as good as the AIs when all other factors are considered.

Can DDs with Maximum Bulkheads soak "questionably vast" amounts of damage and recover? Yes, they certainly used to. Which is why, I believe, Nick made some changes in some scenarios to prevent the AI having DDs with max bulkheads.

Same goes for Transports. Were Nick to be bothered, and I'm sure he won't (I hope he doesn't, lol), he could tell you how often I've complained that Transports were ludicrous damage sponges, able to soak 30+ 5-6" shells, not to mention even 8-9" ones, get down to single digit floatation showing on the damage portrait 6 compartments flooding, yet recover to 40-60% floatation and sail away IF you didn't keep pumping round after round into it until it sank.

He's told me he altered TR durability so they're easier to sink. I've no reason to doubt him, plus I've observed it myself. It DOESN'T mean those with max bulkheads aren't questionably durable to lower calibre shells in particular (the very shells a raiding warship would use to sink them), but that's a question of fine tuning.

Regardless, if your argument the AI cheats due to its ability to control damage, I would suggest the following:

1. the damage control model as it stands is very basic. It appears, which is all I can say as I don't know, to treat ALL ships as having the same potential damage control capacities. Those are then modified by tech, such as anti-flood versions 1 and 2, auxiliary engines that affect water pumping rates, particularly bulkhead status (minimum can make any ship sink easily, far too much so IMO, while maximum makes them god-like in damage control) and so on. Thus a Transport appears to have the same potential for damage control as a battleship.

2. to continue from point 1, it's worth listing many obvious things it does NOT consider at present. Availability of trained damage control teams, pumping capacity INCLUDING the ability to counter-flood and just how much can be done, firefighting equipment, repair parties and so on. In fact, you can safely assume the model at present assumes all ships appear to have an infinite, base line of ALL these, modified by tech, as mentioned in 1.

3. Do YOU enjoy the same benefits? YES, YOU DO. TRY IT AND SEE. Do a simple "random battle" and build a ship with whatever levels of aux motors, anti-flood and so forth but set it at MINIMUM bulkheads. Now do it again, with exactly the same design but MAXIMUM bulkheads. Come back and tell us what you saw happen.

You seem to be conflating not liking the shortcomings of the damage control model with the AI having advantages YOU will not if you build to the same tech levels. It's proof of nothing other than you not liking how well ships can control damage (and I'll bet in all cases it's because they have maximum bulkheads, although 'many' can be tough, too). Again, where's the evidence of cheating?

Edited April 14, 2020 by Steeltrap

[Steeltrap]

16 hours ago, Nick Thomadis said:

AI surely, if you want to believe me, does not cheat. I hate AI cheating myself and do not want it in any game I am part of. When I say "Cheating" I mean some kind of bonus to the AI for the same stat. AI might be more efficient in building a better ship than an average player, but that is not a cheat.

(Minor edits removing some bits to distil the critical points, see below as to why)

I happen to think this is possibly the most important thing you've said, absolutely crucial in a single player game. I recommend that it appears (as a header, footer, somewhere) on EVERY EXAMPLE of information about the game made available in public. It's SO at odds with usual industry attitude, and is a real selling point for such a game.

Start making it available on EVERYTHING and perhaps people might just wonder why it's not said anywhere else. Yeah, that's probably asking too much of the average gamer, but perhaps NOT so for the well known YT game reviewers. Surely SOME of them will notice, but not if they're not told explicitly. Wouldn't it be nice if devs started getting asked about this specific point in single player games, and presumably follow up questions as to how the AI cheats and WHY?

[Also a sharp point of difference from, and implied deserved criticism of, the vast majority of game devs who take the quick and easy path to increasing difficulty, namely piling on loads of meaningful stat bonuses to AI units etc (and, yes, CA Total War, I'm looking very squarely at you and your abysmal, lazy AI for so many years)]

[madham82]

8 hours ago, Tycondero said:

Damage control can be very effective for the player too, if he/she is willing to invest in it.

In general I think the speed accuracy penalty should be there, but I feel that it should be based on the relative speed vector of both ships. If both ships are going at the same bearing with similar (high) speed, the penalty on accuracy should be way smaller as the relative difference is much less. Ships that go extremely slow will still get an accuracy penalty for target speed, but could get an accuracy buff due to other factors, such as platform stability (not moving much) for instance.

The more I think about it, relative speed really doesn't add up. The fact is the fast ship is harder to hit because of the distance it travels between shots (and time it takes the shell to reach the target), coupled with dispersion and other gun accuracy factors. To me, this means the margin of error is greater because of the fast speed of the ship. This is all assuming you have precise range, otherwise that margin of error increases significantly. But as range decreases, shell travel time drops, and generally the shells dispersion reduces speed is less of a factor. That seems to be the main issue with the game modeling now. 

Another issue I see is fast speed not having a negative impact on that ship's own accuracy. A ship moving at or near it's top speed would not be a very stable ship, especially in rough seas I would think. So if there is a bonus to accuracy for sailing at cruise speed, why would there not be a penalty to sailing at top speed? This would IMO balance the mechanic more. As you would use flank speed to escape, cruise to engage the enemy (which matches real world tactics). Right now I believe the AI will just sail at flank speed all the time. So if AI did that with his 38kt BC, then not only would we not be able to hit him, he shouldn't be able to hit anyone either. Then we can start balancing the accuracy for game play. 

[Wowzery]

Just did a video where I tried to design the USS Texas, the AI designed a 36kt battleship, with 16" main guns, and up to 17" of armor for the year of 1925.  (Also had a secondary battery of 8", 6" and 4" guns.)