Jump to content
Game-Labs Forum

>>>Core Patch 0.5 Feedback Hotfix v90<<<


Recommended Posts

  • Replies 887
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Popular Posts

Greetings Admirals, The next major patch for Ultimate Admiral: Dreadnoughts has just become available! This large update is the precursor of the first campaign version of the game. We decided to rel

Hello all, We understand your anticipation for the campaign and it is welcome. The last patch which you all play has revealed very few remaining issues that we already sort out to fix in the

Hello all, Tomorrow a new hotfix patch will become available. It will include the following: *Hotfix v90* (30/9/2021) - Refactoring of the Auto-Design code, addressing various issues that coul

Posted Images

1 hour ago, Steeltrap said:

Am yet to find a recent game where formation sea keeping is in any respect competent.

It's a total mess in War on the Sea.

Bizarre thing I don't understand is that it ought to be VERY EASY to work backwards from what is necessary to produce something that delivers it. The AI, after all, has all the necessary parameters.

I think of it this way:

- start with how navies DID their formations; WE ought to be able to set templates however we want, or use standard provided ones.

- look at standard processes for changing formation, or changing course.

- take those processes and build an invisible, moving template of "required position" for each and every ship every 2,3 or perhaps 5 seconds. I think of each ship following a trail of breadcrumbs which, if each does so perfectly, will result in a seamless alteration of course, formation, speed and combinations thereof.

- the AI ought to have the necessary performance numbers to have the ships get VERY close to those required positions effectively in real time.

There would be no bizarre veering out of line because of some less than ideal 'collision avoidance' system that presumably is placeholder because the AI would have been told where each ship is going to be in 2,3, 5 or 10 seconds from now and thus recognise there is no danger.

Again, thinking off top of head. No idea how feasible it is. Indeed it's entirely possible what I'm describing ISN'T achievable, although I find that hard to accept.

Regardless, from what I've seen NO naval game seems to have tried this OR they can't manage it. They all seem to use some pared back, basic thing that is patently unfit for purpose thus ought not be released. I just don't see how they've not cracked the issue of "start with what the results must be, design the steps to get there (template, waypoints tied to time, whatever) then build the AI to execute".

Perhaps it's just seen as too much work for not enough return, but it certainly grates in 'proper' naval wargames.

In this, WotS, and anything else I've played, the first thing I do upon entering a battle is UNDO every formation. It's a pain in the butt, yet it's the only way to get ships to do what I know they OUGHT to do.

Suppose, for example, a formation is changing course to port (and can we get the ability to input courses? I don't think any navy in the world at this time sort of pointed in some direction and said "let's go that way") ships to port of the centreline need to slow while ships to starboard need to accelerate.

Yet for some reason the AI remains seemingly entirely bamboozled by such concepts, which is why I wonder if it is because it doesn't have an effective, near real time template of positions that, if followed, will achieve the required result. Or something that achieves the same effect.

NOT saying it's unacceptable in an Alpha, of course. Anything is fine in an Alpha provided there's evidence it's changing. Or at least evidence it's intended to change.

Am simply thinking aloud as I'm writing (something I do quite often if it's not obvious).

Would love to know what the underlying design is in terms of process/logic, or at least the final intended version, but such curiosity is an occupational hazard of mine, lol.

Getting the AI to perform the way you're thinking is actually quite a feat.

Honestly, most people look at AI and game design in general without having any idea what it takes to make even the basic, discreet systems that go into getting the game to actually work. This isn't a shot at you - you're just asking a lot of questions and trying to fill in the blanks - it's an observation of the gaming community in general.

Anyway, let's look at AI as a general system, or more accurately, a collection of individual systems designed to replicate plausible human behavior. Let's even restrict it to station-keeping, target selection, and formation movement in UA:D.

The first thing you need to understand is there is no way anyone is going to design some sort of giant program with all the necessary sub-systems to manage all the things needed to get the enemy and non-player controlled ships to replicate a reasonable facsimile of human behavior. That's IBM-level Deep Blue stuff. What UA:D almost certainly has is multiple discreet systems which handle specific aspects of enemy and non-player performance. So how many are there and what do they do?

Here's a speculative and non-exhaustive list:

  1. Main battery target selection.
  2. Secondary battery target selection.
  3. Torpedo target selection.
  4. Station-keeping.
  5. Damage-induced retirement.
  6. Torpedo evasion.
  7. Collision avoidance.
  8. Screen orders.
  9. Scout orders.
  10. Battle line orders.
  11. Follow orders.
  12. Spacing sub-orders.
  13. Formation movement orders.

Each of those are separate, coded systems. Each has some impact on the others. Each has to be adjusted based upon actual experience in seeing them interact with the others. It's a lot of very time-consuming work, and probably isn't all that much of a priority at this point in time. Why? Because it's enough to know they function. If they function, the base system is sound, and can be improved at any time. For a game that is functionally pre-alpha (Steam Early Access would qualify as alpha in more traditional terms), it's vitally important that they work, but they don't have to work well. At least not just yet. That's usually late alpha or early beta stuff in traditional dev environments.

There are exceptions. If you want Nick and the team to be able to make incremental improvements to the AI as we go, record clips of bugs and AI failings, and submit those. The reason the AI auto-design system gets incremental improvements is because players relentlessly submit its problems in these patch threads, often with visual documentation.

If you want to see discrete AI systems in action, fire up UG:CW if you have it. Play any battle that has skirmishers and line infantry, then:

  1. Move infantry toward an enemy skirmisher unit.
  2. Observe how the skirmishers fall back when the infantry comes within range.
  3. Do it again, but instruct the infantry to hold fire before moving into range of the enemy skirmishers.
  4. Observe how the infantry can move much closer to the enemy skirmishers, up to a certain point.
  5. Do it again with hold fire orders. Notice how the skirmishers fall back at a specific distance every single time.

That behavior is governed by a system unique to skirmishers. They automatically retreat if an enemy unit approaches within a certain range, but the reason they seem to fall back every time they're fired upon isn't because you shot at them, it's because your unit targeted them. The "AI" uses the fact that it's been selected as a target to fall back, thus convincingly replicating a behavior that seems human. Something that simple makes it seem human, but that's a finished game with much simpler unit interactions than UA:D.

 

That's kind of a long answer, but hopefully that helps a bit.¬†ūüôā

Edited by killjoy1941
  • Like 2
Link to post
Share on other sites
On 10/8/2021 at 4:30 PM, madham82 said:

Think it goes back to a point I have made many times since getting the game. Control of the sea (i.e. the battle map) should be the only thing that counts as a victory (in campaign). That would leave your nation's intelligence service as the only way to know if something badly damaged made it back to port unless intercepted again. 

In custom/missions, it should be control of the sea first, then use a points calculation based on damage dealt/received and a bonus for ships sunk outright. Lots of ways to make this part work. 

I don't think there are any ways to make this part work. And I know a bit about it. First fact in custom missions there is no calculation for victory at sea after sinking a capital ship, you have to sing all of the small ones too. So in custom missions it should be control of the fire power superiority first.

  • Sad 1
Link to post
Share on other sites
On 10/14/2021 at 9:00 PM, Cptbarney said:

Nah its fine, i can see great potential with this game. And judging from the scope it will be tricky to get right, but if they do muh goodness.

Also hope yer keeping good as well mister steel.

Nah I can hardly see any potential in this game. First of all it uses an archaic graphical system which is a bad sign because we are in 2021 and pc graphical hardware has advanced far beyond even for low budget games like this. The concept of the gameplay is extremely archaic feeling, we cant even shoot the guns, we just click and do nothing. the fact that the game calculates everything which could be fun gameplay is a boring concept. We should be able to aim and shoot the guns ourselves and to see through a rangefinder and other equipment and steer the ships ourselves. the models are really bad and etc. there are even no cool ships, only a few German ones. So this game is just disappointing at this piont.

Edited by Elrerune The Honorbound
  • Sad 1
Link to post
Share on other sites
1 hour ago, Elrerune The Honorbound said:

Nah I can hardly see any potential in this game. First of all it uses an archaic graphical system which is a bad sign because we are in 2021 and pc graphical hardware has advanced far beyond even for low budget games like this.

It is using unity which is far from archaic, the graphics do need updating and so does sound, but that's a small issue that can wait. Gameplay is infinitely more important than Graphics ever will be.

1 hour ago, Elrerune The Honorbound said:

The concept of the gameplay is extremely archaic feeling, we cant even shoot the guns, we just click and do nothing. the fact that the game calculates everything which could be fun gameplay is a boring concept. We should be able to aim and shoot the guns ourselves and to see through a rangefinder and other equipment and steer the ships ourselves.

Fam world of warships exists for a reason, same goes for war thunder. This game as repeatedly said to be something akin to RTW, but 3D.

1 hour ago, Elrerune The Honorbound said:

the models are really bad and etc. there are even no cool ships, only a few German ones. So this game is just disappointing at this piont.

If the models were fully done up, the game would be laggy as all hell due to having to render soo many little details. Maybe once the core game mechanics are done they can focus on that tidbit as well. They probs could have a system to allow people to dress up the ships, and also a better designer in general, but that will have to wait for a bit as they do the campaign.

Theres a tonne of issues, which im sure nick is aware of but they do have a small team so i would give them a break at least. The major issue is getting trying to get a designer that isn't tied to the AI or failing that at least make the AI use templates from player made creations.

  • Like 4
Link to post
Share on other sites
2 hours ago, Elrerune The Honorbound said:

I don't think there are any ways to make this part work. And I know a bit about it. First fact in custom missions there is no calculation for victory at sea after sinking a capital ship, you have to sing all of the small ones too. So in custom missions it should be control of the fire power superiority first.

I see your point.

 

2 hours ago, Elrerune The Honorbound said:

Nah I can hardly see any potential in this game. First of all it uses an archaic graphical system which is a bad sign because we are in 2021 and pc graphical hardware has advanced far beyond even for low budget games like this. 


The graphics look bland, yes, but there is far worst cases. But I see your point here too.

 

2 hours ago, Elrerune The Honorbound said:

 The concept of the gameplay is extremely archaic feeling, we cant even shoot the guns, we just click and do nothing. the fact that the game calculates everything which could be fun gameplay is a boring concept. We should be able to aim and shoot the guns ourselves and to see through a rangefinder and other equipment and steer the ships ourselves. 

Sorry, but this is (supposed to be) a realistic RTS. In 90% of RTSs, you dont direct control the unit, because there is often no point in it. You are supposed to control several units, not one. Its a strategy type genre, not an action one. Besides, of all RTS-Genres, you had to pick realistic naval warfare, the type of combat where realistically you sit for maybe even hours aiming at the enemy with, shooting, missing, waiting for a 30 second reload, and repeat until you are lucky to get a hit after 10 salvos. 
Sorry, but this is not WoWs. 
If you want a build-your-ship action game, I agree with barney, pick waves of steel. You will be able to build your ship to surprisingly great detail and direct control it all you want.
 

 

2 hours ago, Elrerune The Honorbound said:

the models are really bad and etc.


Its not that the models themselves are bad. But GL often recycles and resizes them, and takes too long to make new ones. It is a issue of insuficient quantity, not a issue about the quality of the models themselves.

 

2 hours ago, Elrerune The Honorbound said:

there are even no cool ships, only a few German ones. So this game is just disappointing at this piont.

No, cool ships, what does that mean? No historical ships? No Super-Battleships? That is a very vague statement.

And yes, the game is disapointing at this point. But not because it isnt build-your-ship WoWs, but because it is failing to recreate the (great) 3D Rule the Waves concept it set out to.

 

  • Like 3
  • Thanks 1
Link to post
Share on other sites
6 hours ago, killjoy1941 said:

Getting the AI to perform the way you're thinking is actually quite a feat.

Honestly, most people look at AI and game design in general without having any idea what it takes to make even the basic, discreet systems that go into getting the game to actually work. This isn't a shot at you - you're just asking a lot of questions and trying to fill in the blanks - it's an observation of the gaming community in general.

Anyway, let's look at AI as a general system, or more accurately, a collection of individual systems designed to replicate plausible human behavior. Let's even restrict it to station-keeping, target selection, and formation movement in UA:D.

The first thing you need to understand is there is no way anyone is going to design some sort of giant program with all the necessary sub-systems to manage all the things needed to get the enemy and non-player controlled ships to replicate a reasonable facsimile of human behavior. That's IBM-level Deep Blue stuff. What UA:D almost certainly has is multiple discreet systems which handle specific aspects of enemy and non-player performance. So how many are there and what do they do?

Here's a speculative and non-exhaustive list:

  1. Main battery target selection.
  2. Secondary battery target selection.
  3. Torpedo target selection.
  4. Station-keeping.
  5. Damage-induced retirement.
  6. Torpedo evasion.
  7. Collision avoidance.
  8. Screen orders.
  9. Scout orders.
  10. Battle line orders.
  11. Follow orders.
  12. Spacing sub-orders.
  13. Formation movement orders.

Each of those are separate, coded systems. Each has some impact on the others. Each has to be adjusted based upon actual experience in seeing them interact with the others. It's a lot of very time-consuming work, and probably isn't all that much of a priority at this point in time. Why? Because it's enough to know they function. If they function, the base system is sound, and can be improved at any time. For a game that is functionally pre-alpha (Steam Early Access would qualify as alpha in more traditional terms), it's vitally important that they work, but they don't have to work well. At least not just yet. That's usually late alpha or early beta stuff in traditional dev environments.

There are exceptions. If you want Nick and the team to be able to make incremental improvements to the AI as we go, record clips of bugs and AI failings, and submit those. The reason the AI auto-design system gets incremental improvements is because players relentlessly submit its problems in these patch threads, often with visual documentation.

If you want to see discrete AI systems in action, fire up UG:CW if you have it. Play any battle that has skirmishers and line infantry, then:

  1. Move infantry toward an enemy skirmisher unit.
  2. Observe how the skirmishers fall back when the infantry comes within range.
  3. Do it again, but instruct the infantry to hold fire before moving into range of the enemy skirmishers.
  4. Observe how the infantry can move much closer to the enemy skirmishers, up to a certain point.
  5. Do it again with hold fire orders. Notice how the skirmishers fall back at a specific distance every single time.

That behavior is governed by a system unique to skirmishers. They automatically retreat if an enemy unit approaches within a certain range, but the reason they seem to fall back every time they're fired upon isn't because you shot at them, it's because your unit targeted them. The "AI" uses the fact that it's been selected as a target to fall back, thus convincingly replicating a behavior that seems human. Something that simple makes it seem human, but that's a finished game with much simpler unit interactions than UA:D.

 

That's kind of a long answer, but hopefully that helps a bit.¬†ūüôā

Also killjoy brings up a pretty good point here too be fair. I just want this game to do well frankly. Edging more to realism than arcadey stuff. I wonder if the devs should look at naval art or waves of steel for inspiration if they are stuck on what they could replace the current designer.

But yeah the AI needs, improvements as it tends to go mental at times or bug out. But we will need to screenshot more stuff. It does well when with one ship type, but seems to struggle with multiple still (also, if ships colliding into each other because of mis-understood commands we will need this too be visible so we understand so that we can provide better feedback in-general).

  • Like 1
Link to post
Share on other sites
1 hour ago, Stormnet said:

. Besides, of all RTS-Genres, you had to pick realistic naval warfare, the type of combat where realistically you sit for maybe even hours aiming at the enemy with, shooting, missing, waiting for a 30 second reload, and repeat until you are lucky to get a hit after 10 salvos. 
Sorry, but this is not WoWs. 

Therein lies the problem. Though I appreciate you preempting @RAMJB @akd or @Steeltrap having to be the ones to say so. 
 

I think I said on the last page, in reference to the secondary battery thing, there’s no convincing people who aren’t interested enough in naval warfare to be convinced. 

  • Like 2
  • Sad 1
Link to post
Share on other sites
2 hours ago, Stormnet said:

And yes, the game is disapointing at this point. But not because it isnt build-your-ship WoWs, but because it is failing to recreate the (great) 3D Rule the Waves concept it set out to.

I think trying to please the WoWs crowd who were wowed¬†(see what I did there?) by seeing¬†trailers and came in to testing is a big reason why. They endlessly complained that the game ‚Äúdidn‚Äôt feel right‚ÄĚ (like WoWs) or wasn‚Äôt ‚Äúbalanced‚ÄĚ (like WoWs), and in trying to please them we drifted away from the concept.

Every positive step to correct that drift, like improvements to gunnery and damage, makes them upset and around and around we go.

Fundamentally, they did not buy into the core concept, and have not been brought around to¬†it - despite ‚Äúrealistic‚ÄĚ appearing 13¬†times in the¬†description of the battle system.
 

I hate to say it, but it the root of their criticism is ‚Äúis not World of Warships‚ÄĚ, and we know that cuts against the core concept which has been reiterated again and again by @Nick Thomadis¬†and the grogs and rivet counters - 3D Rule the Waves - at what point do we start discounting their feedback to stay on track and reestablish a consistent design consistent with the philosophy for the title?

 

Edited by DougToss
  • Sad 1
Link to post
Share on other sites
1 hour ago, DougToss said:

I think trying to please the WoWs crowd who were wowed¬†(see what I did there?) by seeing¬†trailers and came in to testing is a big reason why. They endlessly complained that the game ‚Äúdidn‚Äôt feel right‚ÄĚ (like WoWs) or wasn‚Äôt ‚Äúbalanced‚ÄĚ (like WoWs), and in trying to please them we drifted away from the concept.

Every positive step to correct that drift, like improvements to gunnery and damage, makes them upset and around and around we go.

Fundamentally, they did not buy into the core concept, and have not been brought around to¬†it - despite ‚Äúrealistic‚ÄĚ appearing 13¬†times in the¬†description of the battle system.
 

I hate to say it, but it the root of their criticism is ‚Äúis not World of Warships‚ÄĚ, and we know that cuts against the core concept which has been reiterated again and again by @Nick Thomadis¬†and the grogs and rivet counters - 3D Rule the Waves - at what point do we start discounting their feedback to stay on track and reestablish a consistent design consistent with the philosophy for the title?

 

Devs are pretty impervious to the kind of influence you describe. Nick and his team have a very specific vision for what this game is going to be, and they're not deviating from that as far as I can determine. They have their own plan for their own game, and it clearly doesn't involve being a MOBA naval game with historical dressing.

You're much better off simply waiting for the campaign to see what UA:D is really going to be.¬†ūüôā

  • Like 1
Link to post
Share on other sites

For the record I think he's just joking around since the Australian stereotype (well-founded) is that they always say "cunt" even when it's not meant insultingly.

Link to post
Share on other sites
1 hour ago, Speglord said:

For the record I think he's just joking around since the Australian stereotype (well-founded) is that they always say "cunt" even when it's not meant insultingly.

Well, I'm not australian, and I didn't know that. Either way, I'd find it insulting in all scenarios.

Link to post
Share on other sites
2 hours ago, Speglord said:

For the record I think he's just joking around since the Australian stereotype (well-founded) is that they always say "cunt" even when it's not meant insultingly.

 

1 hour ago, Commander Reed said:

Well, I'm not australian, and I didn't know that. Either way, I'd find it insulting in all scenarios.

Stop going off topic in the Feedback thread. ;)

Link to post
Share on other sites
12 hours ago, Cptbarney said:

It is using unity which is far from archaic, the graphics do need updating and so does sound, but that's a small issue that can wait. Gameplay is infinitely more important than Graphics ever will be.

Did you forget that this game is 3 years in development? Just because it uses Unity engine does not mean it is not archaic because it is utilizing barely the barebone base of what unity has to offer graphically. System requirements state DX11 but the game has NO DX11 assets and effects that Unity offers. So yes it's very archaic state. Yeah sure gameplay is important but only equally as graphics because we are in 2021 no 2005. But if you don't want to sell your game then go ahead and don't do care about graphics.

12 hours ago, Cptbarney said:

Fam world of warships exists for a reason, same goes for war thunder. This game as repeatedly said to be something akin to RTW, but 3D.

No, that is a bad and boring concept. It's basically: bad gameplay + bad graphics. But the way you praise it make it seem like you think it's better that Anno 1800. go take a look at a concept of good graphics + good gameplay. (But fair warning get yourself a decent gpu before playing as that game hes real DX12 features.)

12 hours ago, Cptbarney said:

Gameplay is infinitely more important than Graphics ever will be.

This is definitely not true as proved by many most modern and most demanding games of today. Such as modern Battlefield and Call of Duty games.

12 hours ago, Cptbarney said:

If the models were fully done up, the game would be laggy as all hell due to having to render soo many little details.

With modern CPU's and GPU's that would not be a problem at all even with high distant details. Especially not for the hardware in the recommended requirements with the current models.

CPU: Intel Core i7 3770 3.4 GHz, AMD FX-8350 4.0 GHz
GPU: Nvidia GeForce GTX 1060, AMD Radeon RX 580

12 hours ago, Cptbarney said:

Theres a tonne of issues, which im sure nick is aware of but they do have a small team so i would give them a break at least. The major issue is getting trying to get a designer that isn't tied to the AI or failing that at least make the AI use templates from player made creations.

Yes. I know. But I'm afraid after the Campaign is done the sales won't improve just thanks to gameplay. As most people who see the game will be like "oh look this is a simple looking game with super-fun gameplay, I'm gonna buy it!" And then they will see the system requirements and be like " whaat?? for what does this game need such hardware, it looks like it could run on a 660 Ti on ultra 4k!" 

Edited by Elrerune The Honorbound
Link to post
Share on other sites
11 hours ago, Stormnet said:

Its not that the models themselves are bad. But GL often recycles and resizes them, and takes too long to make new ones. It is a issue of insuficient quantity, not a issue about the quality of the models themselves.

I meant low-poly. + Historically unrealistic = Bad models.

Link to post
Share on other sites
On 10/15/2021 at 10:44 AM, Skeksis said:

Nick has said the game will not support over the top designs

I'm sorry.  I must be the only person here who's seen AI-made ships in custom battle with 4x4 18" or larger guns then.  For Christ's sake man, anything over 80,000 tons at full load is over the top.  Even the Yamatos and their nine 18.1" guns only topped out at 73k metric tons.

EDIT:  Am I the only one who thinks its utterly ridiculous that 130k metric ton hulls even exist for a game that's supposed to end in the 1940's?  That's a full third again what a modern-day Gerald R. Ford CVN displaces.

Edited by SpardaSon21
  • Thanks 3
Link to post
Share on other sites
1 hour ago, Skeksis said:

Ridiculous, fiction or otherwise, they are already here, this must be Dev‚Äôs ‚Äėvision‚Äô for the game. I think the 80 thousand+ ton¬†ships is to cover the ‚Äėwhat if‚Äô scenarios, what if the paper ships came to life, and to¬†cover the what if beyond that. Maybe alittle bit towards them being the carrier end game replacement. Definitely crafted for an end game.

 

IMO they exist to represent what would happen in a alternative history where doesn't exist any naval treaty between nations. Naval powers would continue to push the limits in ship and gun size if they could. In that context makes sense to see 20 inch guns or ships with 100k tonnage. 

However it could be a nice feature to see in the campaign a political pressure between nations to implement a fictional naval treaty after a war.  

  • Like 2
Link to post
Share on other sites

it somewhat amazes me that people are surprise/outrage that a game about building alternative ship design allows to build not historical designs...

  • Like 2
Link to post
Share on other sites
10 hours ago, SpardaSon21 said:

EDIT:  Am I the only one who thinks its utterly ridiculous that 130k metric ton hulls even exist for a game that's supposed to end in the 1940's?  That's a full third again what a modern-day Gerald R. Ford CVN displaces.

No, but such "a-historical" details (they do actually have historical precedence, look up the H-class designs) can be fixed by hyper-realists in mods once the game is launched.

We've all seen the meme-ships, but RN the devs need to focus on campaign mechanics & formations instead of wasting even more time trying to fix the AI auto-designs with anti-meme ship balancing, because as we've seen time and again, even if they do fix the meme ship problem, somebody is going to cry foul over the fixes and how they spoil the realism, and we will have waited another couple of months going back and forth on a balancing issue that realists could easily fix with mods once the core game is in place and the game has launched on steam.

It's just not worth the time and effort right now. Their time & energy is better spent ironing out the last formation issues & getting the campaign ready for launch, and we have been promised mod tools (Nick rose to prominence in the development community making the excellent darth mods for the total war series) so those issues should be easily solvable to each our own tastes once the core game releases ūüôā

Link to post
Share on other sites
2 hours ago, SiWi said:

it somewhat amazes me that people are surprise/outrage that a game about building alternative ship design allows to build not historical designs...

It’s not about the designs, it’s about the underlying systems. 

  • Like 2
Link to post
Share on other sites
  • Nick Thomadis locked, unlocked, locked and unpinned this topic
Guest
This topic is now closed to further replies.

×
×
  • Create New...