When is TWs actually going to add a major feature?

Users who are viewing this thread

StewVader

Sergeant
@stevehoos

I bet if you keep complaining that'll speed things up!

Seriously though, you whine all over these forums like a child. How many times has "real features when?" been asked across these forums over the last year? It's like an actual, unironic circlejerk around here about "something something unplayable something something no features something something skeleton game". We all feel it. We all want depth and features. But clogging the forums up with this drivel doesn't help anyone or anything.
Oh FFS dude. Post like this are necessary to remind people why they should never buy another TWs game.

I honestly believe TW deceived us. From past videos and dev diaries to all the other BS. We were sold a false bill of goods.

So much for working with the community during EA,. Or ****, so much for even developing the game. Its essentially the same as it was at launch, with more quest and a bit of ascetic polish.
 

AxiosXiphos

Sergeant Knight
M&BWBWF&SNWVC
Post like this are necessary to remind people why they should never buy another TWs game.
..... What?

I think people can make up their own mind on how they spend their money. Perhaps focus on reminding TW on what needs improving instead. The world doesn't need you to protect them from Taleworld products...
 

StewVader

Sergeant
..... What?

I think people can make up their own mind on how they spend their money. Perhaps focus on reminding TW on what needs improving instead. The world doesn't need you to protect them from Taleworld products...
Can you read? I'm not instructing anyone do anything. Just pointing out that post like this are a reminder to not trust TW again. You know, not buy their ****ty products again, because they lack quality.
 

mehmeteking

Veteran
WBVC
Agents in the sense of Agent-oriented Programming are not what you would consider AI. It is a non-deterministic system but it is not powered by Machine Learning.
Well, AI is a superset of machine learning, so yeah. But I am curious as to how it is not the Actor-model but AOP. Agents are supposed to be autonomous based on some BDI framework (or equivalent.) Is that the case in Bannerlord? Or actually the better question is, what exactly convinced you that certain parts of Bannerlord is developed by AOP?
 
In Bannerlord, AI agents are considered on the same level as the player. Time and time again TW devs have declined feature suggestions because of the impact it would have on the AI agents. Everything the player can do, the agents must be able to do as well. Or from a different perspective, the player is limited by what can be coded for the AI agents. If this does not raise all sorts of warnings that Bannerlord is a game in name only, I don't know what to tell you. A game is ALL about the players. Everything exists to create a cool experience for the players.

A good example is the new 1.5.10 "feature" of adding stances to your clan's parties. What had been asked for a very long time is to allow the player to have control over what our clan parties can do. TW response is to just allow the player to slightly influence the AI agents priorities. Allowing the player to tell its clan parties to siege castle A while we ourselves siege castle B, effectively splitting the ability of the defenders to protect themselves, would have been super easy to do code-wise (I know this for a fact, by the way) and it would have added a great vehicle of agency and tactical value for the player. Did TW do that? No, because this level of sophisticated coordination is not something that the AI agents would also be able to do.

Another good example that really shows TW philosophy is the Skill Perks. The thing is riddled with completely useless perk effects for the Player! In a freaking single-player game!!

Third example and then I'll shut up. The economy. For those that have been here since the beginning of EA, work into Bannerlord's economy consumed the first several months of development. But why waste all that effort on it though? It is a simulation that is almost completely opaque to the player. Worse, the player can barely influence it because, if options to do so existed, they would also have to be available to the AI agents, and balancing that would be a nightmare, never mind that it would make the game a better game and its economy vastly more interesting for the player.

Bannerlord is mostly a great demo of how Agent-oriented Programming can create really sophisticated (from a software perspective) simulations where you have mostly independent software agents interacting with each other to work towards a common goal. It is objectively a bad computer game. Does that mean you can't have fun with it? Absolutely not, I have had at least 20+ hours of true fun with it. The problem is I have "played" 100+ hours of it trying to have fun and I would really like to have those hours back.

Edit: The strong and negative sentiment in my words is not directed at you @SOku , apologies if it reads that way.

I have to applaud the attempt though, if they we’re able to create this program (program? Sure, I dont know) they could roll it out to a multitude of their own games or sell it to different gamecompanies.

Create a bannerlord but rebrand it so partys are battleships, towns become islands, etc.. pirate game!
Or partys become spaceships, towns planets, the battle sim a spaceship fight...

If only it worked though..
Cyberpunk jumps to mind, with its placebo AI..
 

SOku

Sergeant
TW seems to be more interested in making the simulation aspect of Bannerlord successful rather than making it a rewarding gaming experience for the players.

In Bannerlord, AI agents are considered on the same level as the player. Time and time again TW devs have declined feature suggestions because of the impact it would have on the AI agents. Everything the player can do, the agents must be able to do as well. Or from a different perspective, the player is limited by what can be coded for the AI agents. If this does not raise all sorts of warnings that Bannerlord is a game in name only, I don't know what to tell you. A game is ALL about the players. Everything exists to create a cool experience for the players.

A good example is the new 1.5.10 "feature" of adding stances to your clan's parties. What had been asked for a very long time is to allow the player to have control over what our clan parties can do. TW response is to just allow the player to slightly influence the AI agents priorities. Allowing the player to tell its clan parties to siege castle A while we ourselves siege castle B, effectively splitting the ability of the defenders to protect themselves, would have been super easy to do code-wise (I know this for a fact, by the way) and it would have added a great vehicle of agency and tactical value for the player. Did TW do that? No, because this level of sophisticated coordination is not something that the AI agents would also be able to do.

Another good example that really shows TW philosophy is the Skill Perks. The thing is riddled with completely useless perk effects for the Player! In a freaking single-player game!!

Third example and then I'll shut up. The economy. For those that have been here since the beginning of EA, work into Bannerlord's economy consumed the first several months of development. But why waste all that effort on it though? It is a simulation that is almost completely opaque to the player. Worse, the player can barely influence it because, if options to do so existed, they would also have to be available to the AI agents, and balancing that would be a nightmare, never mind that it would make the game a better game and its economy vastly more interesting for the player.

Bannerlord is mostly a great demo of how Agent-oriented Programming can create really sophisticated (from a software perspective) simulations where you have mostly independent software agents interacting with each other to work towards a common goal. It is objectively a bad computer game. Does that mean you can't have fun with it? Absolutely not, I have had at least 20+ hours of true fun with it. The problem is I have "played" 100+ hours of it trying to have fun and I would really like to have those hours back.

Edit: The strong and negative sentiment in my words is not directed at you @SOku , apologies if it reads that way.

Are you a (professional ) dev ? Did you look at the actual Bannerlord code ? Assuming you're either an enthusiast or a modder.
None of what you stated above has to do with complying with a paradigm, this a paradigm in the end, not how you design a game.
( or a sofware for that matter )
Heck, considering your first point, this is entirely possible with OOP, why wouldn't it ?
Make a base class like BaseSoldier and inherit both the players and NPCs from it, rince repeat. 100% OOP.

I suspect you get confused on how the word Agent are generally used in game industry and most popular game engines :

Agent: An in-game character or object that uses AI to interact with other objects in its environment.

Since I really don't know where did you find that Agent Oriented Programming, say both from the devs or at looking into the actual code, I'm assuming this above. But just so you know, all of your points have nothing do to with how you choose a certain paradigm in your code base.
This is the field of game design and how you connect the wires in each other to create a game, not a programming paradigm.
Additionnaly C++ C# being in itself, mostly used following OOP.
 

Dreed89

Sergeant at Arms
Right! None of the mechanics are layered. Its just one shallow mechanic after another. There is no payoff for anything the player does, no sense of accomplishment. Its a remarkable achievement in complete and total failure.
I had also thought about this made some posts in the past.

The features (shallow features) are not connected to each other. The game is not enjoyable nor interactive.

It is just a collection of features thrown here and there.

I know the reason for this, though. Development-wise, this is easier. Their code is such a mess that they are afraid to integrate stuff together. Instead, they find it safer to add separate stuff.
 

SOku

Sergeant
Well, AI is a superset of machine learning, so yeah. But I am curious as to how it is not the Actor-model but AOP. Agents are supposed to be autonomous based on some BDI framework (or equivalent.) Is that the case in Bannerlord? Or actually the better question is, what exactly convinced you that certain parts of Bannerlord is developed by AOP?
AI in the context of game dev or in a broader sense video games is just NPCs powered by algorithms, behavior trees etc, shallow or not.
This has nothing to do with Machine Learning ( for now .. )

edit : and I would add, for a Strategy game context, opponents of course, everything that interacts with the environment. Since it's not a Character per se.
 

guiskj

Squire
Well, AI is a superset of machine learning, so yeah. But I am curious as to how it is not the Actor-model but AOP. Agents are supposed to be autonomous based on some BDI framework (or equivalent.) Is that the case in Bannerlord? Or actually the better question is, what exactly convinced you that certain parts of Bannerlord is developed by AOP?
In all honesty, it could be Actor model rather than true AOP (which can be considered as the ultimate extension of the Actor model), or something completely different. I have a high degree of certainty that it is asynchronous-based in order to handle this many actors/agents at once all acting kind of independently. I worked with AOP back in college a some time ago and how the "AI" in Bannerlord behaves certainly feels like that.

This, plus how in all the C# code I have seen from Bannerlord references these specifically as Agents, formed my assumed connection that they are using some form of AOP. Otherwise why not reference them in-code as Actors? We programmers like our nomenclature well established, after all.

...And back in college the international AOP conference that my Master's teacher went to was in Turkey... (this is anecdotal only, not argumentative)

I think my points stand regardless of the technological framework that brought Bannerlord to life, though. Agreed?
 

mehmeteking

Veteran
WBVC
AI in the context of game dev or in a broader sense video games is just NPCs powered by algorithms, behavior trees etc, shallow or not.
This has nothing to do with Machine Learning ( for now .. )

edit : and I would add, for a Strategy game context, opponents of course, everything that interacts with the environment. Since it's not a Character per se.
I don't know who brought up machine learning or why, actually. My question literally has nothing to do with it.
ML is just a subset of AI where "intelligent behaviour" is achieved through learning algorithms. AOP has nothing to do with ML (unless one decides to have learning agents.) Strategy games more often use techniques like CBR, which is also AI. (I developed one such AI during Master's studies.)

I just want to know why the OP considers it AOP, because as I said AOP is a very specific model where agents are autonomous, communicate through generalised protocols (some ACL), having their own beliefs, desires, intentions, etc. I have no idea if AOP is used in Bannerlord, could very well be the case but I always thought the AI follows rather simple scripts (or behaviour trees, etc). The question still remains though, where does this claim come from because I am genuinely curious :smile:
---------
In all honesty, it could be Actor model rather than true AOP (which can be considered as the ultimate extension of the Actor model), or something completely different. I have a high degree of certainty that it is asynchronous-based in order to handle this many actors/agents at once all acting kind of independently. I worked with AOP back in college a some time ago and how the "AI" in Bannerlord behaves certainly feels like that.

This, plus how in all the C# code I have seen from Bannerlord references these specifically as Agents, formed my assumed connection that they are using some form of AOP. Otherwise why not reference them in-code as Actors? We programmers like our nomenclature well established, after all.

...And back in college the international AOP conference that my Master's teacher went to was in Turkey... (this is anecdotal only, not argumentative)

I think my points stand regardless of the technological framework that brought Bannerlord to life, though. Agreed?
Yes, of course. Thanks for the answer. I was wondering because I actually did plan at some point to create a game with AOP and therefore you can imagine my curiosity; I was about to analyse the hell out of BL :smile: So it may still be AOP, we don't know for certain.

I am not saying this to protect Armagan; but what you study especially in Master's or even what you specialise in during UG can have an effect on how you approach problems in the industry. In my case it certainly did. There is always some bias towards certain solutions where you feel more comfortable. This may also effect just how you name things, so perhaps Armagan named them "agents" because he wanted to turn them into truly autonomous agents at some point; or just liked the name. If he used AOP, I would certainly call it overkill for what it achieves but not necessarily hold it against Armagan. Does that make sense, or am I too lenient? :smile:

Edit: BTW, yes; AOP is an extension (or a special case) of the Actor Model, and that's how I learnt to love the actors in the first place :smile:
 
Last edited:

MadVader

Duke
M&BWB
@guiskj You are wrong about the motivation for taking control from the player, this is not some kind of weird tech/UX experiment. The Taleworlds intent is to make the game simpler and more accessible for wider audience and sometimes they achieve this by simplifying features and removing complex player actions, because casuals may be confused by complex things.
It's definitely a wrong way to go about making the game accessible. You can keep the depth and complexity, but still make the game casual-friendly if the complex features are not necessary to play the game, but add to the experience. Taleworlds are stupid and don't care about their core fans.
Edit: Just look at this TW dev thinking about the battle size slider and it will become clear why they do stupid things.
 
Last edited:

guiskj

Squire
Are you a (professional ) dev ? Did you look at the actual Bannerlord code ? Assuming you're either an enthusiast or a modder.
I suspect you get confused on how the word Agent are generally used in game industry and most popular game engines :
Is this really necessary? I was cordial to you, why not return the favour?

...
None of what you stated above has to do with complying with a paradigm, this a paradigm in the end, not how you design a game.
( or a sofware for that matter )
Heck, considering your first point, this is entirely possible with OOP, why wouldn't it ?
Make a base class like BaseSoldier and inherit both the players and NPCs from it, rince repeat. 100% OOP.

...

But just so you know, all of your points have nothing do to with how you choose a certain paradigm in your code base.
This is the field of game design and how you connect the wires in each other to create a game, not a programming paradigm.
Additionnaly C++ C# being in itself, mostly used following OOP.
OOP is, in its very nature, a synchronous paradigm. You can write asynchronous code in OOP languages, but as the complexity of your asynchronous code grows it becomes a losing battle. I would challenge the notion that simple inheritance could manage the complexity of orchestrating two thousand soldiers in a 3D environment. (If you allow me to target a single aspect of your post instead of tackling it as a whole... Not so nice, is it?)

I mainly had a single point, that TW seems more interested in creating a simulation environment than a game. Then I provided some, in my eyes at least, supporting evidence:
  • Lack of good game design
  • Focus on automation
  • Opaque features
  • etc
TW games are unique from a technical standpoint, and to me it seems like furthering that technology (independent AI agents doing things) is more important to them than creating a good game. Whether it is AOP, Actor-model or even OOP is unimportant to the argument and I shouldn't have even brought it up as it detracts from the conversation.
 
Last edited:

SOku

Sergeant
In all honesty, it could be Actor model rather than true AOP (which can be considered as the ultimate extension of the Actor model), or something completely different. I have a high degree of certainty that it is asynchronous-based in order to handle this many actors/agents at once all acting kind of independently. I worked with AOP back in college a some time ago and how the "AI" in Bannerlord behaves certainly feels like that.

This, plus how in all the C# code I have seen from Bannerlord references these specifically as Agents, formed my assumed connection that they are using some form of AOP. Otherwise why not reference them in-code as Actors? We programmers like our nomenclature well established, after all.

...And back in college the international AOP conference that my Master's teacher went to was in Turkey... (this is anecdotal only, not argumentative)

I think my points stand regardless of the technological framework that brought Bannerlord to life, though. Agreed?
Did you read my message ?
So afterall, nothing against you, you're making a bold assumption just because you've seen the word Agent ( like I said a term widely used in the industry ).
The fact that people acknowledge your message without any critical thinking just because either you made a long post, made a point ( bold assumption ) or two, makes me sad.
Before writing my message I asked you if you were a dev or if you've seen a large amount of their code base.

Your points don't stand, because all of these can be achieved with OOP, forget about what you feel.
Thus making me wonder why do you stick with AOP if your only valid point is because you've seen the word Agent on their codebase somewhere.

Say I name my function FunctionToFunction, does it make my codebase functional programming ? Short answer no.
Debating about how they stick into a paradigm rather than how healthy and efficient is their codebase is beyond non-sense, reminds me those long debates on how language A is better than language C ... the only ones who care are the ones who spend a long amount of their time arguing on a tip of an iceberg. Me included at this point.


PS : I could stand corrected if one of TW's devs could tell us what they use, I doubt they will, that doesn't make your message valid tho.

EDIT :

Is this really necessary? I was cordial to you, why not return the favour?

Could you point out how am I not cordial in this message ? Those questions have not to be read as " You know nothing I do " but rather " Am I talking to someone experienced or knowledgeable enough or just a Wikipedia reader ? ", here me out, I have no problems at all to be corrected by actual professionals or enthusiasts, Internet being itself, everyone can act that they know everything just by reading short bits of google results and spill it on forums.
My questions are simple enough to target who am I discussing to.
Me ? I personally didn't read Bannerlord code by myself, or even experienced modding it, maybe you did that's the sole purpose of these questions.
So I got the habit to ask you these, if you don't see me being cordial on those specific sentences, I guess the problem isn't in my side.
 
Last edited:

guiskj

Squire
I'm not giving TW another cent. I'm reserving my Bannerlord cash to support modders that actually care about game design. TW tricked gamers into basically funding Armagan's academic desire to keep exploring Agent-oriented Programming multi-agent software systems. If you look at Bannerlord through these lenses it starts making sense why they want everything in the game to be automated and reduce player agency even though this, well you know, is a single-player game.
TW seems to be more interested in making the simulation aspect of Bannerlord successful rather than making it a rewarding gaming experience for the players.

In Bannerlord, AI agents are considered on the same level as the player. Time and time again TW devs have declined feature suggestions because of the impact it would have on the AI agents. Everything the player can do, the agents must be able to do as well. Or from a different perspective, the player is limited by what can be coded for the AI agents. If this does not raise all sorts of warnings that Bannerlord is a game in name only, I don't know what to tell you. A game is ALL about the players. Everything exists to create a cool experience for the players.

A good example is the new 1.5.10 "feature" of adding stances to your clan's parties. What had been asked for a very long time is to allow the player to have control over what our clan parties can do. TW response is to just allow the player to slightly influence the AI agents priorities. Allowing the player to tell its clan parties to siege castle A while we ourselves siege castle B, effectively splitting the ability of the defenders to protect themselves, would have been super easy to do code-wise (I know this for a fact, by the way) and it would have added a great vehicle of agency and tactical value for the player. Did TW do that? No, because this level of sophisticated coordination is not something that the AI agents would also be able to do.

Another good example that really shows TW philosophy is the Skill Perks. The thing is riddled with completely useless perk effects for the Player! In a freaking single-player game!!

Third example and then I'll shut up. The economy. For those that have been here since the beginning of EA, work into Bannerlord's economy consumed the first several months of development. But why waste all that effort on it though? It is a simulation that is almost completely opaque to the player. Worse, the player can barely influence it because, if options to do so existed, they would also have to be available to the AI agents, and balancing that would be a nightmare, never mind that it would make the game a better game and its economy vastly more interesting for the player.

Bannerlord is mostly a great demo of how Agent-oriented Programming can create really sophisticated (from a software perspective) simulations where you have mostly independent software agents interacting with each other to work towards a common goal. It is objectively a bad computer game. Does that mean you can't have fun with it? Absolutely not, I have had at least 20+ hours of true fun with it. The problem is I have "played" 100+ hours of it trying to have fun and I would really like to have those hours back.

Edit: The strong and negative sentiment in my words is not directed at you @SOku , apologies if it reads that way.
@SOku I have quoted myself and replaced the one mention of AOP to a more generic term. Do you see that both posts still stand without loosing anything? Please stop focusing on it.
 

guiskj

Squire
@guiskj You are wrong about the motivation for taking control from the player, this is not some kind of weird tech/UX experiment. The Taleworlds intent is to make the game simpler and more accessible for wider audience and sometimes they achieve this by simplifying features and removing complex player actions, because casuals may be confused by complex things.
It's definitely a wrong way to go about making the game accessible. You can keep the depth and complexity, but still make the game casual-friendly if the complex features are not necessary to play the game, but add to the experience. Taleworlds are stupid and don't care about their core fans.
Edit: Just look at this TW dev thinking about the battle size slider and it will become clear why they do stupid things.
I disagree (well, obviously given my last few posts haha). If the intent is to make things simpler and more accessible why all the complexity?

I get the complex combat due to the multiplayer facet of the game, but why make the sandbox/campaign things so complex? Why the simulated economy or making sure that almost every action the player can take the AI can do so as well under their own non-deterministic attitudes?

From a player perspective, Bannerlord is a very simple game. Why make all this software development work for yourself unless you are actually interested in the technology itself?
 

SOku

Sergeant
I specifically asked you why all of this had to with a paradigm, and rather with a open-question calling you to elaborate your thoughts.
Hence you answered me with that long message.
You wouldn't if I didn't ask at the first place. Color me surprise when you're going the " don't focus " route.
I fully agree with you about game design and philosophy. Let's not forget this whole debate started because of something else.
 

guiskj

Squire
...
EDIT :
Could you point out how am I not cordial in this message ? Those questions have not to be read as " You know nothing I do " but rather " Am I talking to someone experienced or knowledgeable enough or just a Wikipedia reader ? ", here me out, I have no problems at all to be corrected by actual professionals or enthusiasts, Internet being itself, everyone can act that they know everything just by reading short bits of google results and spill it on forums.
My questions are simple enough to target who am I discussing to.
Me ? I personally didn't read Bannerlord code by myself, or even experienced modding it, maybe you did that's the sole purpose of these questions.
So I got the habit to ask you these, if you don't see me being cordial on those specific sentences, I guess the problem isn't in my side.
If you did not mean offense then I am certainly not going to keep chewing on it. Let's chalk it up to inference reading and miscommunication on our parts.
I am a professional software developer and I have modded Bannerlord, though not much. I had grand plans for it but gave up after becoming disillusioned with Bannerlord.
 

MadVader

Duke
M&BWB
I disagree (well, obviously given my last few posts haha). If the intent is to make things simpler and more accessible why all the complexity?

I get the complex combat due to the multiplayer facet of the game, but why make the sandbox/campaign things so complex? Why the simulated economy or making sure that almost every action the player can take the AI can do so as well under their own non-deterministic attitudes?
The combat is not complex compared to Warband mods. It lacks sophistication.

The economy is the fault of a particular developer who likes complex models. He did this in Warband too, but it was less visible to the player. It's over-engineering - when engineers go wild and are not stopped by management. It's an exception to their overall design philosophy.

The AI taking equivalent action is the preoccupation of another developer (and a fan's hero) who got a lot of complaints about the AI lords having it too easy respawning and marching on you with full complement of troops, so he worked to make their behavior more complex. When he asked permissions from management to work on more complex features, they were turned down as a rule.

From a player perspective, Bannerlord is a very simple game. Why make all this software development work for yourself unless you are actually interested in the technology itself?
No it's not. You still have much more to do and learn than in your average shooter. You are grasping at straws here.
Consider that you might be wrong and not every idea you come up is brilliant.
 

guiskj

Squire
The combat is not complex compared to Warband mods. It lacks sophistication.

The economy is the fault of a particular developer who likes complex models. He did this in Warband too, but it was less visible to the player. It's over-engineering - when engineers go wild and are not stopped by management. It's an exception to their overall design philosophy.

The AI taking equivalent action is the preoccupation of another developer (and a fan's hero) who got a lot of complaints about the AI lords having it too easy respawning and marching on you with full complement of troops, so he worked to make their behavior more complex. When he asked permissions from management to work on more complex features, they were turned down as a rule.


No it's not. You still have much more to do and learn than in your average shooter. You are grasping at straws here.
Consider that you might be wrong and not every idea you come up is brilliant.
Let's tackle the obvious attack first:
I am not providing an idea, simply my opinion. Every opinion is brilliant to its owner. And honestly, why the attack?

Saying the combat is not complex because modded Warband was more sophisticated is the wrong counter-argument here. My argument about Bannerlord combat was meant in more of a generic way. If the goal of TW is to dumb things down to casual players as you proposed, I think these types of players would not have cared one bit if combat was much more straight-forward. And to my point, why would TW go through the trouble of designing the combat and then training the AI on it if the demographic is indeed what you propose? Honest question here.

And on this vain, shooters can be quite complex. Even though I suck at those types of games, I can understand it takes certain skill to master. A broad diversity of competency at any skill is a good demonstration of its complexity.

Edit: Also I don'y buy that the whole economy and lord behaviour were things that were implemented on a developer's whim. We all joke about TW being bad at project management, but that would be complete anarchy.
 
Top Bottom