2.8.1 (2190) Beta AI Feedback Megathread

[MrFreake_PDX]

Just as a note for everyone, this thread is winding down and will be locked over the next week or so. Anyone who has anything else useful to add should do so now before that happens. Thanks to everyone for your feedback, this thread has been extremely useful and we've got dozens of actionable points that we will be taking a look at.

As you're all likely aware at this point, there will be big changes coming in the next update, as such collecting feedback from 2.8.1 is already limited in its usefulness. However, that being said, to everyone who's taken the time to do testing and followed the rules of this thread, from everyone on the Stellaris Team, I'd like to say THANK YOU!

We will likely be doing more of these in the future, so stay tuned.

 

[glee8e]

More details required, especially since I can see you're not using debug_ai, and obviously haven't read the OP.

Yes I did read the OP. Not using debug_ai is indeed my fault but I think the situation is crystal clear. The friendly AI fleet is taking a massive detour around the world to get pass the starbase with some hyperlane inhibitor and that AI fleet to a starbase. The friendly AI fleet is strong enough to take on both the starbase and hostile fleet together.

More information required. A one-line response isn't useful feedback here.

I do attached a savefile, no? Is that savefile broken or when it loads it is not building agriculture districts?

 

[HugsAndSnuggles]

More information required. A one-line response isn't useful feedback here.

Want more? Sure: simply enable console for ironman games couple days prior to my post and notify everyone of the change.

 

[Angelo217]

Ok, I will give that a try and let the game run without scaling difficulty. Maybe that's truly the reason why the AI feels so "borked".

I would like to report back that without scaling difficulty the AI feels and acts more "competitive". It's also building much earlier megastructures. However, certain things which I have pointed out are imho still valid, like for with or without scaling difficulty the AI could build starbases with shipyards a bit more, or that the AI still doesn't terraform planets to an Ecumenopolis if the requirements are met or sometimes empires, especially machine empires are often lacking of enough unity income.

 

[Tartan Spartan]

Hi,

A final comment and some save files on the previously reported issues of unemployment - this appears very very prevalent in my most recent play through and really undermines the performance of some AI empires.

Please see the confederation of Achemar in the attached save files - which continues to build City districts until the planet is full urbanised - despite major unemployment and resource shortages in most categories.

3 saves attached
2278.10.23_GKhan.sav - the base, cities are built until planet full
2281.03.03_district_test.sav - post intervention - AI builds sensible buildings after economy is stabilised
2288.03.02_districts_still_cities.sav - ai reverts to building city districts only

The behaviour is very wierd - and I struggle to understand the cause, I have been unable change through editing the districts file - even with huge weights for rural districts.

 

[Mímisbrunnr]

@MrFreake_PDX

Just as a note for everyone, this thread is winding down and will be locked over the next week or so. Anyone who has anything else useful to add should do so now before that happens. Thanks to everyone for your feedback, this thread has been extremely useful and we've got dozens of actionable points that we will be taking a look at.

As you're all likely aware at this point, there will be big changes coming in the next update, as such collecting feedback from 2.8.1 is already limited in its usefulness. However, that being said, to everyone who's taken the time to do testing and followed the rules of this thread, from everyone on the Stellaris Team, I'd like to say THANK YOU!

We will likely be doing more of these in the future, so stay tuned.

To manage expectations, is there any reasonable likelihood [no promises at this point, just a probability] of getting a 2.8.2 update with some fixes [be it with AI-related points raised here or general bugfixes], even if its only small? I'm asking as I'm getting to the point where I'd be soon ready to start a new campaign, but as campaigns take a long time I'd rather wait a bit if a hypothetical 2.8.2 is not that far off so my campaign can benefit from its improvements from the beginning. If there is minimal to no chance of getting 2.8.2 and instead we're likely to have several months until the next fixes I may start up a campaign in the meantime.

If you're able to offer any insight on this matter, your advice would be greatly appreciated.

 

[MrFreake_PDX]

@MrFreake_PDX

To manage expectations, is there any reasonable likelihood [no promises at this point, just a probability] of getting a 2.8.2 update with some fixes [be it with AI-related points raised here or general bugfixes], even if its only small? I'm asking as I'm getting to the point where I'd be soon ready to start a new campaign, but as campaigns take a long time I'd rather wait a bit if a hypothetical 2.8.2 is not that far off so my campaign can benefit from its improvements from the beginning. If there is minimal to no chance of getting 2.8.2 and instead we're likely to have several months until the next fixes I may start up a campaign in the meantime.

If you're able to offer any insight on this matter, your advice would be greatly appreciated.

I'm not sure anybody knows for certain at this stage.

 

[Mímisbrunnr]

@MrFreake_PDX

I'm not sure anybody knows for certain at this stage.

1A.) Well, that's definitely helpful for answering my question - I know not to assume that a patch is coming soon, so I might as well start in the meantime. Thanks for that information.

1B.) Admittedly, I'll confess that this is also a bit disappointing, as there is currently a major bug (although not related to the AI in the topic) that is causing breaches of galactic law to do nothing as of 2.8 and 2.8.1, which means sanctions also do nothing now and the high-end resolutions from Federations (a paid DLC feature) that are meant to be put many of your enemies in breach aren't working in the way they were meant to either. I had hoped that you'd be able to implement a fix for that alongside a few of the AI improvements suggested by the user data posted in this thread so that such a key feature from a major DLC wouldn't be left broken for potentially months, or even at a minimum just a minor bugfix without the AI updates. So I hoped there would be at least concrete plans to address bugfixes. I beseech you and/or PDX to try to at least restore the galactic law breaches to functioning as soon as possible so your Federations DLC resolutions work as intended.

I'll say no more on that subject here there, as that's more properly belonging to the other threads. Let me change subjects now...

2.) To try to bring my post more on the subject of AI, may I make a general suggestion - the AI has long lacked the ability to replace districts (its far beyond a 2.8 issue, existing since 2.2), which also generally makes it unable to build ecumenopoli. The new district change with industrial districts should help this to an extent, but if the AI ever builds a single agriculture/generator/mining district on a world it will be forever unable to make an ecumenopolis unless you grant it the ability to replace districts. Is it possible to add logic to the AI to have it realize that it should replace a district if its very close to meaning the ecumenopolis conditions? Giving how much the building and district system is changing, I wouldn't recommend changing this until the new system is implemented, but I think its worth considering.


Thank you for your time.

 

[MrFreake_PDX]

Thanks to everyone who provided feedback for the beta! This thread has now been locked.