Capitalism 2 and Capitalism Lab Forum

Hello,


I've had this issue back in 6.3x without the banking DLC and while it is less prevalent, there are still three key issues with AI subsidiaries set up by me :

- their aggression is very, very low, even when given billions in cash and sufficient products to cover the entire food industry they expand *very* slow, if at all. AI competitor aggressiveness does not seem to make a difference, my subsidiaries are always very timid.
- they seem to completely ignore or fail to process certain businesses, banking and insurance for example, at the very best they will buy/start researching relevant tech but will never go beyond building redundant RD centers
- the AI often complains about not having precisely 15m of cash to expand with yet another RD center, even when they have a billion in cash. As a programmer this makes me believe the AI is checking against the wrong value in certain cases


Attached you will find my save on the latest version + all DLCs + RealWorld mod (which is confirmed to work with this release). You can see that i have set up three subsidiaries:

- Royal Food: food producing sub that has a wide range of products enabled, this is my only sub that is independently doing what is is supposed to do, although its far, far too slow to expand regardless of funds given
- Royal Homes: meant to be a real estate company, they will build a single office building every few years but do nothing beyond that, the apartment buildings they have were all built by me in an attempt to make them stop losing money at least
- Royal Bank: this is a complete failure, the AI did build a single RD center (while complaining about that missing 15m again), then bought and started researching relevant techs but never cared to do anything else, didn't set up a single bank or insurance HQ. I did manually build them a number of RD centers (as they always complained about not having 15m to build yet another), and the Insurance HQ you see built was also done by me in an attempt to kickstart the AI from wherever it got stuck, but it completely ignores the firm.

I've also played around with all of their settings including their preference for profit, allowing a wider range of products (which is difficult when they are supposed to be specialized), etc, but they don't seem to work. If you for Example allow Royal Bank to expand in retail it *will* start building stores, albeit very slowly and they will still ignore their intended purpose.

https://filebin.net/p5tbd12qo9csj07l/PL ... t=9qhj85uo

Thanks

I have requested the programmer to look into your save game and here are his findings:

- Royal Food: food producing sub that has a wide range of products enabled, this is my only sub that is independently doing what is is supposed to do, although its far, far too slow to expand regardless of funds given

The CEO of this subsidiary is conservative in business expansion. You can see this on the Personality page of the Person report. Please see the attached screenshot.

- Royal Homes: meant to be a real estate company, they will build a single office building every few years but do nothing beyond that, the apartment buildings they have were all built by me in an attempt to make them stop losing money at least

The AI did not build any apartments because there is an oversupply of apartment in every city. You may check the Supply and Demand Index graph on "City Graphs 1" on Cities report.

- the AI often complains about not having precisely 15m of cash to expand with yet another RD center, even when they have a billion in cash. As a programmer this makes me believe the AI is checking against the wrong value in certain cases

If you see that message again, please send me your save game and I will ask the programmer to investigate it.

- Royal Bank: this is a complete failure, the AI did build a single RD center (while complaining about that missing 15m again), then bought and started researching relevant techs but never cared to do anything else, didn't set up a single bank or insurance HQ. I did manually build them a number of RD centers (as they always complained about not having 15m to build yet another), and the Insurance HQ you see built was also done by me in an attempt to kickstart the AI from wherever it got stuck, but it completely ignores the firm.

The CEO of this subsidiary is also Conservative in business expansion. You can see this on the Personality page of the Person report.

A conservative CEO will not expand until he/she sees that the corporate cash level is high enough.

The programmer will add messages about insufficient cash for bank and insurance expansion on the Corporate Detail screen in the upcoming version 6.4.06.

In addition, I did a test by starting a new game. The CEO of my subsidiary has a Moderate aggressiveness. And the cash level of the subsidiary is higher than that in your save game.

The AI was able to build a bank in less than one year.

I provide the save game for you to test - please download the BankSub.RAR attached in this post. Then load the save game and let it run for a few months and you will see it building a bank.

Thank you, I haven't considered the Character property when hiring the CEO, I'll try my subs with more aggressive people, although the bank had over a billion cash at some point and it was still idle.

However, I've started two additional subs (one for telecom/internet/media, the other for furniture/tools/generic ikea) in the same save shortly after I uploaded it and they are both completely idle, even when having 700m in cash, and against all my attempts to build firms for them they completely ignore everything. The absolute best they will do is buy tech but nothing beyond that, regardless of available funds or how profitable the market is (pretty much every market is a win currently). To me, these scenarios look a lot more like the AI failing to successfully evaluate some decision and getting perpetually stuck on it than low aggressiveness, I can spam them with billions and they will do nothing with it.

Is it possible to add a small debug-like window somewhere on the subsidiary interface that lists you all major decisions the sub AI has made? Similar to a rolling console window, continuously listing significant evaluations and their conclusions by the AI. That would help massively to understand whether my AI is just waiting for additional 10m cash to their 1 billion already to start building, or it had a stroke a long time ago and cash won't make a difference.

This is my current save with the idle furniture sub, I hired an aggressive CEO two years ago but it is still idle and never complains about anything:

(everything in History was done by me)


https://filebin.net/9jbw60ql0r89f2qr/PL ... t=f8mwtnrx

Would it be possible for you to provide us with a save game with a single subsidiary company company that apparently failed to execute the plan?

To check your save game, the programmer has to run the game and look into the programming code and it could be a fairly time consuming process. As you know, we have limited resources and there are many tasks scheduled as well as improvements requested by users (some requests have been pending for a long time, due to the large number of tasks.)

In order to test your game efficiently, the programmer needs a save game from you in the simplest form that shows only the failing subsidiary company with limited or no business activities of the main player company.

Thanks for your understanding.

Also, the programmer did tell me that the AI code for running subsidiary companies is the same AI code that runs rival AI companies. So you are not put into any disadvantages with subsidiary companies. But the AI code is complex and it is difficult and very time consuming to perfect it. We can only try to improve those with obvious problems.

Another thing is that we have discussed a long time ago while doing beta-test on the Subsidiary DLC that it is not our game design goal to make the subsidiary company AI as powerful as possible, as it will take the fun out from the game. Imagine that you could just set up a few subsidiary companies and you can win the game, doing very little business decisions. This is not the design goal of the game. Hope you could understand that technically achieving excellence on subsidiary AI may look noble as a concept, but in practice, the player will end up just abandon the game if the AI is too good and the game becomes devoid of fun. In the world of game design, we really have to consider all these. Hope you could understand.

No worries David, my field of expertise is not very far from what you were talking about so I understand.

As for a save with minimal clutter, if I sold/deleted everything I own apart from the sub in a dedicated save, waited a year being idle then saved that, would that do? I no longer have the sub on the screenshot as I had to absorb it to prevent going bankrupt (my bank imploded), but I will do that when I notice a new sub is just idling again. Or you would need a fresh save where I can reproduce this issue?

_white_ wrote: ↑Wed May 06, 2020 5:46 pm As for a save with minimal clutter, if I sold/deleted everything I own apart from the sub in a dedicated save, waited a year being idle then saved that, would that do?

Yep, that would do. Thanks.