TALEWORLDS, LISTEN TO ME. WATCH MY EYES, DON'T LOOK AWAY.

Users who are viewing this thread

Workplace troubleshooting conversations can take days rather than minutes.
Not quite. On the contrary, decisions are taking around 1-2 minutes for new features - while they shouldn't be decided in such a short timeframe. Yes. They are not even get discussed on internal-chat. It's getting decided on calls - on spot. You can't possibly think all the good and bad effects of the feature in 2 minutes and give decisions. But this is exactly what's happening at the moment.
For bugs, probably they are not even "deciding" anything but simply changing the order of their sprint/backlog. And since siege is working in some scenes - they might not be spending all resources in there.
I bet only 1 guy is checking this problem at the moment. Because for AI - they don't have a huge team anyway and they also have other problems like stupid AI on several occasions - getting stuck on some scenes because of pathfinding etc. And it's not easy to hire new people that are comfortable with C++ for the game industry in Turkey. This especially hard if your codebase isn't related to any known pattern but simply an ad-hoc implementation over the years
Then there's budgetary pressures. I imagine that is why we're in EA in the first place.
That's what I thought too but it seems like this is not the case. Release date decision for EA is taken in Summer - it wasn't a last-minute call after seeing that COVID goes wild and people stuck at home. But that overlapping events certainly benefitted TW.
To ensure that there is reliable income to buy a small design team enough time to work through time consuming issues like this.
This is not a design issue. This is old codebase spaghetti issue. Design is clear. "Don't mess up sieges. Check Warband in doubt. Improve that." And you can't buy your way out as I previously explained on top. And TW's engine code is only touched and known by a handful of people in the company - and nearly all of them are Veteran Taleworlds employees ( i.e. Cem )
I'm not saying that you can't have poorly written or performing code in your legacy codebase. That's perfectly understandable. But if this mess is causing you so much trouble that even your promised core functionality fails for a year that you can't fix it in one month - it's a red flag. Like, big ass red flag, like in the main menu.
And since talking about flags, you can hide away and say "Banners aren't priority cuz they are cosmetic" but you can't say "Sieges aren't priority cuz they are cosmetic" so you have to fix that. It's not maybe but question but yes or not question for them. And answer is yes. Yes they have to fix that.
I mean, this thread has what, a hundred or 2 views,
Taleworlds Devs or managers are not giving a damn about Forum. I think earlier you guys realize this is better. They are relying on Dejan to bring suggestions - Arda and others to bring issues. They won't engage discussions with playerbase - and half of the company don't have enough "knowledge" about the game itself to even keep up with discussion anyway. The reason why mexxico is loved by many people is that he knows the game. And he is also playing it or at least focusing his part intensively. For him, it's not a simple Jira task.
 
Not quite. On the contrary, decisions are taking around 1-2 minutes for new features - while they shouldn't be decided in such a short timeframe. Yes. They are not even get discussed on internal-chat. It's getting decided on calls - on spot. You can't possibly think all the good and bad effects of the feature in 2 minutes and give decisions. But this is exactly what's happening at the moment.
For bugs, probably they are not even "deciding" anything but simply changing the order of their sprint/backlog. And since siege is working in some scenes - they might not be spending all resources in there.
I bet only 1 guy is checking this problem at the moment. Because for AI - they don't have a huge team anyway and they also have other problems like stupid AI on several occasions - getting stuck on some scenes because of pathfinding etc. And it's not easy to hire new people that are comfortable with C++ for the game industry in Turkey. This especially hard if your codebase isn't related to any known pattern but simply an ad-hoc implementation over the years

That's what I thought too but it seems like this is not the case. Release date decision for EA is taken in Summer - it wasn't a last-minute call after seeing that COVID goes wild and people stuck at home. But that overlapping events certainly benefitted TW.

This is not a design issue. This is old codebase spaghetti issue. Design is clear. "Don't mess up sieges. Check Warband in doubt. Improve that." And you can't buy your way out as I previously explained on top. And TW's engine code is only touched and known by a handful of people in the company - and nearly all of them are Veteran Taleworlds employees ( i.e. Cem )
I'm not saying that you can't have poorly written or performing code in your legacy codebase. That's perfectly understandable. But if this mess is causing you so much trouble that even your promised core functionality fails for a year that you can't fix it in one month - it's a red flag. Like, big ass red flag, like in the main menu.
And since talking about flags, you can hide away and say "Banners aren't priority cuz they are cosmetic" but you can't say "Sieges aren't priority cuz they are cosmetic" so you have to fix that. It's not maybe but question but yes or not question for them. And answer is yes. Yes they have to fix that.

Taleworlds Devs or managers are not giving a damn about Forum. I think earlier you guys realize this is better. They are relying on Dejan to bring suggestions - Arda and others to bring issues. They won't engage discussions with playerbase - and half of the company don't have enough "knowledge" about the game itself to even keep up with discussion anyway. The reason why mexxico is loved by many people is that he knows the game. And he is also playing it or at least focusing his part intensively. For him, it's not a simple Jira task.
+1 for your comment. You are spot on in my opinion.
 
+1 for your comment. You are spot on in my opinion.
Good to see that your opinion is changing now. Because you were trying to insult me 4 months ago :smile:
" I learned long ago, never to wrestle with a pig. You get dirty, and besides, the pig likes it." Don't waste too much energy with pointless discussions and the rabble Callum. You got bigger fish to fry.
And even in that post, I mentioned about Siege issue.
4 months later, we are still having the same issue.
And with this pace, we will have some issue in next 4 months too.
Warband Sieges >> Bannerlord Sieges. Why? It's not just stupid AI. This clusterfork also hits hard on performance in sieges as well. Which turns funny bugs into rage-material bugs for players.
 
Not quite. On the contrary, decisions are taking around 1-2 minutes for new features - while they shouldn't be decided in such a short timeframe. Yes. They are not even get discussed on internal-chat. It's getting decided on calls - on spot. You can't possibly think all the good and bad effects of the feature in 2 minutes and give decisions. But this is exactly what's happening at the moment.
Mind if I ask how you know this? The way you worded it makes it sound like you have some kind of insider knowledge into how TW in particular operates.
Or are you just extrapolating how game dev studios generally operate and applying it to TW?
 
Siege tower AI has been raised a million times.

It's actually not easy to fix. Anyway it's pointless to raise it again. They have heard about it constantly from the community.

I feel your pain but..... yeah, ease off with the capital letters. Just makes you seem like a bit of a........
 
Good to see that your opinion is changing now. Because you were trying to insult me 4 months ago :smile:

And even in that post, I mentioned about Siege issue.
4 months later, we are still having the same issue.
And with this pace, we will have some issue in next 4 months too.
Warband Sieges >> Bannerlord Sieges. Why? It's not just stupid AI. This clusterfork also hits hard on performance in sieges as well. Which turns funny bugs into rage-material bugs for players.
You are totally right man. Thanks for pointing that out. I am just fed up with their sloooow pace of development, which you explained the why above. I know the developers are hard at work, its the leadership that sucks balls. I really really hope the best developers at TW leave, form their own company, and create the game we all deserve.
 
Siege tower AI has been raised a million times.

It's actually not easy to fix. Anyway it's pointless to raise it again. They have heard about it constantly from the community.

I feel your pain but..... yeah, ease off with the capital letters. Just makes you seem like a bit of a........
The guys was venting but he has a point anyway, easy fix or not it should be top priority because sieges are a core feature and basically 1/3 of mid-late game.
 
1/3?

How long do your sieges take in battle? Mine take like 60 seconds.
not in duration (which should be much longer too) but in importance and reappearance during gameplay, mid-late game you are mostly set gear and income wise so you start to focus more on army x army field battles and sieges to capture or defend territory and this important part is barely functional since the begining of early access.

Attacker and defender AI are pretty bad, they have a hard time using the siege engines (the siege towers being the worst offender), defender AI can't defend for **** and you can win attacking while heavily outnumbered because the defender just ignore enemies in their path to go to the breaches and once there barely defend at all just holding position, same with the gates where they just make a shield wall in-front of it and stay static while the attackers swarm them (heard this part was improved in 1.5.8 beta, trully hope so)
 
Last edited:
We literally fixed lot of this things in RBM (which proves that they just need dedicated AI guy). We have no issue if TW just copy pastes it.
 
He can't. He is not in that team. He can at best mention about this in design discussions but that's all. But that would be pointless because I bet the entire office is already aware of the fact that siege AI sucks. If they are not aware of it - let's drop the game immediately because it basically means they have no idea what they are doing.
If they know it - and if it's not resolved for 1 year, there are two options:
- It's just they are not "prioritizing" it properly. Taleworlds dev team is literally not giving a damn about community and issues at the moment ( some devs aside ) Which is 90% true actually. But they do check bugs and resolve them.
- Some funky **** is going on in engine side which takes a lot of time to resolve. And they are understaffed about AI team.

My bet would be on the second one - since the engine itself is extremely old and by looking at current "modern" mess on the gameplay side, it's safe to assume that they have really really really horrible codebase on C++ side.
I mean, if you can mess up fresh C# code in 4 years, I can't even imagine what you can do to C++ codebase in 12 years. And looking at their pace about how they resolve C# related gameplay issues - you can also assume that this mess is happening on somewhere really deep.

They have to fix it though. I bet console peasants can't even see the first frame in Sieges with their low computation power if they port it like this. And for PC, for release, they have to fix this anyway. So I'm pretty sure they will fix it soon. When soon? That I don't know.
I was not talking about hypothetical solutions but things we either already did with a mod and they helped or things that will almost certainly help but we cannot mod them because they require either access to their engine or at least to map (you cannot mod original maps right now, if this was enabled it would literally allow modders to fix lot of the AI related issue).
 
We have no issue if TW just copy pastes it.
So you are saying the issue is on engine side or not? If you fixed with mod -then it's not on engine side - but then later you are saying `we can't mod them because it requires access`. What you did isn't fixing the issue - it's applying bandaid to a deep cut.
Without knowing how the engine works - you are proposing hypothetical solutions for engine side. You don't know if they are able to open those endpoints directly or not. You didn't see the code. Also, it could be that they know siege code is a placeholder and they can simply decide throwing out the entire code to trash and rewrite that with different engine endpoints. Their game, their engine, their code.
After seeing that Big Mac sauce tastes different, you can't mix up two burger sauces and tell McDonald's "Hey we found the Big Mac sauce, you can take it if you want".
 
So you are saying the issue is on engine side or not? If you fixed with mod -then it's not on engine side - but then later you are saying `we can't mod them because it requires access`. What you did isn't fixing the issue - it's applying bandaid to a deep cut.
Without knowing how the engine works - you are proposing hypothetical solutions for engine side. You don't know if they are able to open those endpoints directly or not. You didn't see the code. Also, it could be that they know siege code is a placeholder and they can simply decide throwing out the entire code to trash and rewrite that with different engine endpoints. Their game, their engine, their code.
After seeing that Big Mac sauce tastes different, you can't mix up two burger sauces and tell McDonald's "Hey we found the Big Mac sauce, you can take it if you want".
what? the guy was just saying there are multiple problems, the ones on the engine side they can't access to fix, the ones "public" to modders he already fixed in his mod (which is amazing btw)
 
So you are saying the issue is on engine side or not? If you fixed with mod -then it's not on engine side - but then later you are saying `we can't mod them because it requires access`. What you did isn't fixing the issue - it's applying bandaid to a deep cut.
Without knowing how the engine works - you are proposing hypothetical solutions for engine side. You don't know if they are able to open those endpoints directly or not. You didn't see the code. Also, it could be that they know siege code is a placeholder and they can simply decide throwing out the entire code to trash and rewrite that with different engine endpoints. Their game, their engine, their code.
After seeing that Big Mac sauce tastes different, you can't mix up two burger sauces and tell McDonald's "Hey we found the Big Mac sauce, you can take it if you want".
We were able to fix regular ladder almost on all maps. We were also able to make units counterattack and surround attacker when gate and walls are breached (these are fixable in the currently available code). However we discovered issue with towers that simply require new queue points which can be only added map wise (currently there is only one queue point for each tower, that serves 3 ladders at the time which is causing lot of mess to the AI). We were able to reduce soldiers climbing down by reducing their "melee agro range" in sieges, but preferrable solution would be hardcoding the ladders to go only up. As for AI improperly spawning and standing in wrong places on siege maps, this is editable in the map, only problem is that maps are not editable by modders (you can actually make brand new map, including castle with current modding tools, you just cannot change already existing maps from TW, thats how we know its matter of manual editing). Same aplies to AI improperly reacting to rocks or big trees in field combat maps, it is actually possible to make AI react it would just require editing.
 
Is there any reason why they haven't allowed modders to edit vanilla maps? does the tools simply don't have this function yet or something?
 
Back
Top Bottom