NEW Poll: AI needs Improvement

[David]

Produce the most profitable product first and then move on to the others and not just focus on 2 products and quit expanding

I started a new game with 324% difficulty and no DLCs. One of the AI companies was already producing 4 products by early 1994 - see the attached screenshot.

So it seems that the suggested improvement is already working in the game.

By the way, the AI was programmed to have different preferences and capabilities. Like in the real world, not every single company will pursue aggressive expansion strategies. Some of them do and some of them don't. This leave room in the market for the player to explore and expand into, which is part of the fun playing the game.

[nejxis]

In some cases that happens, but most of the time they either stop at two or three products and move into a different product class or just stop there and accumulate cash.

We're talking about them expanding throughout various years, in my example below it is a company about 40 years old.

In the case below they have 100% dominance in two apparel products and didn't try to expand more into the apparel sector and I am using real world mod, so there is a lot of products to expand in.

[David]

Would you please send me your save game file? I would like to take a look at it.

[nejxis]

this is a different game I started after that game as it autosaved on the same file.

In this game I just bought up a bunch of shares and let the AI do almost everything without me interfering in any operations

Some companies expanded pretty good in the same product type like body care products, some others diversified a lot and have like 6-7 types of product classes instead of focusing on 1 or 2, but at least they kept expanding.

But if you check the file, most of them stick to few of the products, even though they are 30+ years old and making a profit - for example the worst case scenario is in the drugs sector, the biggest company (Pfizer in this case) has a cash hoard of 23bil and still didn't expand in all drugs..

[David]

My opinion is that it is in line with the real world situation where not all companies will strive to diversify. Some companies are just content with being focused on a specific market.

I think such a mix actually makes the game more interesting than having all the AI pursuing diversification endlessly.

[David]

I started a new game and let the AI run for 30 years. Many of the AI corporations with a diversification strategy were able to produce products spanning more than one product class.

I prefer to use the standard data set instead of a MOD as it is a better way to assess the performance of the AI.

The game settings may have something to do with the AI's aggressiveness and performance as well.

I've attached the save game here in case you would be interested in taking a look at it. It was saved using the latest version 5.8.03, which you may download from https://www.enlight.com/forum/viewtopic.php?f=7&t=6704

[David]

Start caring more about operational costs and choose the right factory size in comparison with the revenue

If you have a save game without using a mod that can illustrate this issue, please send it to me and let me know specifically which factory has the issue, I will then take a look at it.

[David]

We have improved the algorithm for AI choosing factory sizes. Please download the latest post-release beta version 6.4.01 from: http://www.capitalism2.com/forum/viewto ... =10&t=7467