Almost two months into EA. Satisfied?

[somethinglightthat]

The bar is set higher than ever for game development today.

Looking at the patch history I can see a clear pattern of hard work and progress that only comes from a decent team. There's a lot of things that are built on foundations in game and software development, some things must come before others. The order in which they are fixing things suggests to me that they are on the right track here, despite the "new content now" demand. Delivering new content on top of things that need to be fixed will not result in better development, imo. In hundreds of hours of running a game that puts stress on cpu/gpu to near maximum levels I've only seen a couple of crashes. For an EA product, that's good.

On the negative side, perks that weren't globally complex should have worked at launch. I think enough people have said this for it to be obvious. The state of attribute and perks for companions is also in this category. Siege battles should have been polished more, especially when they are promotion material that draw people to the product. Having to auto resolve sieges for bad maps / AI is what I'm on about.

Things seem to be getting stable in the crash and bug department, and I'd look for the focus to shift to fleshing things out after that. The shell of this game is good, and it does need more content, which puts it in its rightful place, early access.

 

[Kentucky 『 ＨＥＩＧＵＩ 』 James]

For an EA product, that's good.

Did you pay an early access price for it to match the lack of features?

This is what I can never wrap my head around. You're paying for the game as is, not what they say it'll be like in a few months. There isn't anything legally binding them to make the game good or even finish it. Yet people keep taking "early access" at face value and insisting that the game will one day match the expectations that made them buy the game for that price. It's incomprehensible.

 

[eddiemccandless]

Did you pay an early access price for it to match the lack of features?

This is what I can never wrap my head around. You're paying for the game as is, not what they say it'll be like in a few months. There isn't anything legally binding them to make the game good or even finish it. Yet people keep taking "early access" at face value and insisting that the game will one day match the expectations that made them buy the game for that price. It's incomprehensible.

At the end of the day it is a matter of trust. I bought the game because I think that I paid way less for Warband than it was worth, and because I think that they have done a good job with final deliveries in the past. Which makes me think that they will come around eventually this time as well. If am wrong and they don't, well I guess that sucks, but I definitely had 100 bucks worth of enjoyment from Warband alone, and I paid less than that overall. Had this been a new game from another company I would probably have waited before buying it.

 

[TwistedCoil]

I guess Taleworlds sells consoles as well then.

Not sure how that changes anything, but if you want another example: Ubisoft

 

[Bob Gnarly]

Did you pay an early access price for it to match the lack of features?

This is what I can never wrap my head around. You're paying for the game as is, not what they say it'll be like in a few months. There isn't anything legally binding them to make the game good or even finish it. Yet people keep taking "early access" at face value and insisting that the game will one day match the expectations that made them buy the game for that price. It's incomprehensible.

Then don't buy EA?

I'm a long time M&B fan and I know that the game will be completed eventually and then modders will keep the game fresh for years to come after release. I've already got my monies worth out of the EA build alone, everything after this is just a bonus.

 

[Retsu]

I think that one of the biggest issues with bannerlord is that no one knows what kind of game is supposed to be, is it supposed to be a medieval simulator? it takes so many things from different genres some from Grand strategy games, other from RPG, some people call it a Sandbox,etc different players have different expectations, you can easily see how many people complain about the RPG elements being mediocre especially the roleplaying ones, others complaining about the sandbox elements not allowing different impactful and meaningful ways of playing besides constant battles, others want depth logistic, intrigue diplomatic systems. and of course the people that are being very vocal about needing more content because they found the game to lack replayability will not be happy with more fetch quests because more fetchs quest does not address the issue.

My dissapointment does not come from these design decisions though, as I think all these issues could be fixed with mods or more development time before i played the game, the reason of my dissapointment is the AI, i dont like criticizing devs or to single out a team and i really wish to be wrong because honestly, it would make me sad if all the hard work that went to to make bannerlord as moddable as it is go to waste because of the AI, especially considering how much it could limit the game after a few years.

Just to be clear, i am not expecting deepmind or super advanced AI that can mimic human behaviour or anything like that, just an AI that actually supports mods like a goal oriented one instead of FSM, its been done before with games and it can be done again.

 

[somethinglightthat]

Did you pay an early access price for it to match the lack of features?

This is what I can never wrap my head around. You're paying for the game as is, not what they say it'll be like in a few months. There isn't anything legally binding them to make the game good or even finish it. Yet people keep taking "early access" at face value and insisting that the game will one day match the expectations that made them buy the game for that price. It's incomprehensible.

As an early access title, yes, this is how it works, you're in the development phase, some things won't work, others won't be implemented yet, and the final product may not be what you thought. If that's not acceptable you can wait for a release and then if you don't like that you can not buy it.

Price? Hmm, let's see. I paid about $40 after a discount through an online retailer. I've spent 464 hours so far in this game. That's 8 pennies an hour for my entertainment so far. If I never touch the game again that's fine.

 

[stevehoos]

I have thought about this some more. Nope.

 

[somethinglightthat]

the reason of my dissapointment is the AI

Could you give an example of a game that has AI similar to what you would want to see in Bannerlord?

Could you give a name of an AI style that you'd like to see used?

It's my understanding that most games are Finite State Machine or based on some tweaking of FSM.

 

[Retsu]

Could you give an example of a game that has AI similar to what you would want to see in Bannerlord?

Could you give a name of an AI style that you'd like to see used?

It's my understanding that most games are Finite State Machine or based on some tweaking of FSM.

Most AI that have been developed and used on RTS, GSG and 4X games have moved on from FSM since Jeff Orkin Created GOAP in 2005, CA particularly used it for Total War Empires back in 2009 and most of their next title AI have been inspired on GOAP architecture in some form while bringing innovation like using the MCTS algorithm for Rome II and this was more than half-decade ago, i have no idea what you are talking about here.

 

[somethinglightthat]

Most AI that have been developed and used on RTS, GSG and 4X games have moved on from FSM since Jeff Orkin Created GOAP in 2005, CA particularly used it for Total War Empires back in 2009 and most of their next title AI have been inspired on GOAP architecture in some form while bringing innovation like using the MCTS algorithm for Rome II and this was more than half-decade ago, i have no idea what you are talking about here.

Goal-Oriented Action Planning (GOAP)

From: http://alumni.media.mit.edu/~jorkin/GOAP_draft_AIWisdom2_2003.pdf

"A GOAP system does not replace the need for a finite-state machine (FSM) [Fu03], but greatly simplifies the required FSM. A plan is a sequence of actions, where each action represents a state transition. By separating the state transition logic from the states themselves, the underlying FSM can be much simpler. "

What I'm talking about is that GOAP is largely a planned sequence of FSM. I mean yes, it's different but I thought maybe you were talking about some completely different system than something I've seen. I don't know that it is so different that they could not work this into the game from here.

If you have something that would open my eyes to a completely different approach I would like you to post it. I'm open to correction on my understanding.

 

[Retsu]

Goal-Oriented Action Planning (GOAP)

From: http://alumni.media.mit.edu/~jorkin/GOAP_draft_AIWisdom2_2003.pdf

"A GOAP system does not replace the need for a finite-state machine (FSM) [Fu03], but greatly simplifies the required FSM. A plan is a sequence of actions, where each action represents a state transition. By separating the state transition logic from the states themselves, the underlying FSM can be much simpler. "

What I'm talking about is that GOAP is largely a planned sequence of FSM. I mean yes, it's different but I thought maybe you were talking about some completely different system than something I've seen. I don't know that it is so different that they could not work this into the game from here.

If you have something that would open my eyes to a completely different approach I would like you to post it. I'm open to correction on my understanding.

Its seem that you do not understand the difference between FSM, GOAP, etc they are systems, of course every AI regardless of architecture will be using the same stuff like libraries, agents, states ,etc the difference is how they are DESIGNED TO use them, which means that at its core they are still using the same fundamentals of logic which is why some FSM principles are present ,English is not my first lenguage so its really difficult for me to explain it any further and no if you designed a game AI to use FSM is impossible to implement GOAP over it you have to start over and write the AI again. Also that paper is from 2003, currently GOAP is seen as a principle used in the way the architecture of the AI is designed, they works backwards from its goals in terms of logic.

 

[somethinglightthat]

Its seem that you do not understand the difference between FSM, GOAP, etc they are systems, of course every AI regardless of architecture will be using the same stuff like libraries, agents, states ,etc the difference is how they are DESIGNED TO use them, of course at its core they are still using the same fundamentals of logic ,English is not my first lenguage so its really difficult for me to explain it any further and no if you designed a game AI to use FSM is impossible to implement GOAP over it you have to start over and write the AI again.

I see. There were a lot of problems with the release of Total War Rome II, Joe Vargas (aka Angry Joe) complained about AI problems, and the creative director Mike Simpson would later go on to state, in a second public announcement about new and upcoming fixes, about asking for further player input while also "hoping we can fundamentally treat our releases differently in the future." I'm not sure if any of the performance issues had to do with the AI or some other problem, but they certainly had enough problems with the release to apologize. That MCTS algorithm looks computationally expensive, thanks for mentioning that, I can look at it later.

I do not understand the differences well enough, as you point out. That's why I'm asking. If they have to rewrite from scratch then I doubt it's going to happen. Thanks for the answers.

 

[Retsu]

I see. There were a lot of problems with the release of Total War Rome II, Joe Vargas (aka Angry Joe) complained about AI problems, and the creative director Mike Simpson would later go on to state, in a second public announcement about new and upcoming fixes, about asking for further player input while also "hoping we can fundamentally treat our releases differently in the future." I'm not sure if any of the performance issues had to do with the AI or some other problem, but they certainly had enough problems with the release to apologize. That MCTS algorithm looks computationally expensive, thanks for mentioning that, I can look at it later.

I do not understand the differences well enough, as you point out. That's why I'm asking. If they have to rewrite from scratch then I doubt it's going to happen. Thanks for the answers.

Its ok as i said on my first post, i really wish to be wrong and their AI design is actually modern and ok instead of just tweaked FSM, i am just really afraid thats the reason we see loops like several lords with a low amount of troops sit outside of enemy siege campents for several days or when caravans get stuck outside of Tyal, its just scream to me hardcoded values and FSM doing Fleeing>heading somewhere>seeing danger>repeat

 

[somethinglightthat]

It is really obvious that the AI is stepping through very predictable decision making steps for its actions. These lords probably can't form an army due to the required influence and yet are still instructed to defend the siege, yet won't attack because of values that rightfully tell them they would get uselessly killed, or the order to form an army isn't in the code. The caravans getting stuck I haven't seen. If you sit in a town where enemy caravans try to enter they approach > retreat > approach in a loop without entering the town.

If I'm not mistaken AI programming is the hardest part? There are many parts of the game that work well, but I think I get what you're saying. I'm not sure specifically what you would like the AI to do instead of what it does now. Even in a basic loop there should be ways to determine how many times its run and recognize stuck conditions and take some other action, right?

 

[Virthuss]

So far, extremely disappointed by the lack of flexibility of the multiplayer.

The single player mode is funny but could be improved as so far, it's literally warband with fixes and better graphics...

 

[stevehoos]

So far, extremely disappointed by the lack of flexibility of the multiplayer.

The single player mode is funny but could be improved as so far, it's literally warband with fixes and better graphics...

I think personally Warband is more immersive despite some of the limitations. They have a long way to go here.

 

[Retsu]

It is really obvious that the AI is stepping through very predictable decision making steps for its actions. These lords probably can't form an army due to the required influence and yet are still instructed to defend the siege, yet won't attack because of values that rightfully tell them they would get uselessly killed, or the order to form an army isn't in the code. The caravans getting stuck I haven't seen. If you sit in a town where enemy caravans try to enter they approach > retreat > approach in a loop without entering the town.

If I'm not mistaken AI programming is the hardest part? There are many parts of the game that work well, but I think I get what you're saying. I'm not sure specifically what you would like the AI to do instead of what it does now. Even in a basic loop there should be ways to determine how many times its run and recognize stuck conditions and take some other action, right?

My main issue with the AI is that while these situations can be fixed with more tweaks to FSM as long as is just FSM is going to still be predictable and will require much more tweaking compared to an AI that work backwards from its goal, i suggest you to read Kipsta post on this thread https://forums.taleworlds.com/index.php?threads/hows-the-ai.423677/ as i think it will be able to solve most of the question you are asking.

 

[Virthuss]

J

I think personally Warband is more immersive despite some of the limitations. They have a long way to go here.

Just saying that some very cool features ( advanced diplomacy, creating patrolls, way more advanced economical system to have different ways to fight a war, population control, etc... ) that I was expecting are not there, or at least not yet.
The combat system and the new physic is fine, but I have the sad impression that the best of this game will once again be found on the mods...

 

[babelfisch]

Did you pay an early access price for it to match the lack of features?

This is what I can never wrap my head around. You're paying for the game as is, not what they say it'll be like in a few months. There isn't anything legally binding them to make the game good or even finish it. Yet people keep taking "early access" at face value and insisting that the game will one day match the expectations that made them buy the game for that price. It's incomprehensible.

I believe that's why Taleworlds told people to not buy the game if they have doubts.
What's incomprehensible is that some people somehow still managed to convince themselves into buying it and now act all entitled and betrayed.