Dear Paradox Forum, I never thought it would happen to my country...

  • We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
If I didn't know any better, I'd think only Sweden and Scandinavia have an AI that will build up logging camps. And I've been buying up wood across the globe for years. So, if I was under an embargo from half the planet, yeah, it would make sense that maybe nobody else is increasing wood production because they don't care about my needs. But the countries I've listed are either neighbors of me (so no convoys required to trade) or in my market, and they should be willing to spam logging camps to sell to me.

I think I should bug report this, but I'm not sure how to frame the discussion.

I don't think "The AI sucks" is an acceptable bug description under Paradox guidelines.

More seriously though, even if it were framed as "The AI doesn't take return-on-investment/profitability/productivity in its decision process when deciding what to build" that wouldn't qualify as a bug report because I'm not sure it was ever programmed that way.

I guess it picks randomly from a list of goods with a deficit but checks for positive profitability? Maybe? Who even knows?
 
  • 2Like
Reactions:
I've been playing a new game as Germany, and I'm starving for wood again. But this time, I wasn't being peaceful, and I've been annexing a bunch of stuff.

And I noticed something about the logging industries of various countries.
[...]
Make a bug report and toss me the link, I'll take a look at it when I get back from my vacation tomorrow.
 
  • 3
  • 1
Reactions:
More seriously though, even if it were framed as "The AI doesn't take return-on-investment/profitability/productivity in its decision process when deciding what to build" that wouldn't qualify as a bug report because I'm not sure it was ever programmed that way.
from what i gather the Ai is looking (at the tooltip) for profitability (which isn't always good) and to increase it's military so it never builds what it needs
 
  • 1
Reactions:
Make a bug report and toss me the link, I'll take a look at it when I get back from my vacation tomorrow.

Will do.

from what i gather the Ai is looking (at the tooltip) for profitability (which isn't always good) and to increase it's military so it never builds what it needs

See, even if it looked at the tooltips, even if those tooltips aren't always accurate, it would still see that building logging camps would be a solid choice. If I saw a few extra logging camps, I'd be like "AI just can't budget for construction properly, but at least it knows it needs to build these things." This situation seems different. If Scandinavia wasn't building logging camps, I'd think the AI just didn't know how. But since it is, there's something else goofy happening.
 
Will do.



See, even if it looked at the tooltips, even if those tooltips aren't always accurate, it would still see that building logging camps would be a solid choice. If I saw a few extra logging camps, I'd be like "AI just can't budget for construction properly, but at least it knows it needs to build these things." This situation seems different. If Scandinavia wasn't building logging camps, I'd think the AI just didn't know how. But since it is, there's something else goofy happening.

The AI doesn't currently consider dual production buildings correctly in many cases, I think they are fixing some issues in 1.1 around that, but we'll see the results (it doesn't consider the price of explosives when building a chemical plant, the price of military ships when building shipyards, etc.). They are fixing an issue with the PM calculation, the AI weight for those is missing so many things that the only circumstance that will allow AI to change to Hardwood production is if the state has a hardwood bonus, and since there are going to be states without that, the default PM is going to be to not produce hardwood, so the AI won't consider the value of hardwood at all for building a logging camp.

Also the AI economic strategy heavily favors certain building types (1.5-2x multiplier), and I believe the resource one is the least likely, so even when a logging camp would be a good idea, they say nah, let's build a barely profitable factory because I'm on industrialization mode. The AI has no weighting in its building selection to consider that increasing its wood production would lower the input cost of wood in its factories and increase their profitability. There are strong weightings to get out of an absolute shortage, but there are weightings that prevent it from creating the shortage in the first place. There is a weak weighting for if a good is really high price which needs tuned up, and there are some favored goods weights that could probably be used if they turned on/off when appropriate instead of always being on.

The AI weights also makes it shun upgrading buildings, so it ends up spamming random level 1 factories everywhere. There are also fairly hefty disincentives to it building in unincorporated states, which is often where the wood/rubber/etc. end up being in, although the weights that result in building level 1 factories are still greater than the unincorporated aversion.

I've adjusted quite a few of the weights in a mod, and the result is fairly decent most of the time for in game AI, it's nowhere near the hyper industrialization of custom scripted AI, but it also isn't going to cut down game speed. (It may actually be faster than unmodded late game, since the AI is less likely to fragment pops as much with level 1 factories in every state in the world.) The AI still gets into debt crises and breaks sometimes (especially the UK), but its less frequent and many countries that you'd never see develop will get to at least a decent place. (They still suck at reforming their laws, but that's kind of a historical reality vs. gameplay thing, I think some countries need some more events to have a better chance to make a historic shift, but that will come over time as the game continues to be developed.)
 
  • 1
  • 1Like
Reactions:
Also the AI economic strategy heavily favors certain building types (1.5-2x multiplier), and I believe the resource one is the least likely,

This is odd to me.

Not that I don't believe you, but I usually see decent amounts of iron and coal available in other countries. I wonder if this is due to just having more iron and coal mines at start, or if something else is going on.

To be clear, I can't run a decent sized economy off iron and coal imports, but unlike wood, there's more available. I wonder if this is a combination of dual production bugs (as you point out elsewhere) combined with other issues.
 
This is odd to me.

Not that I don't believe you, but I usually see decent amounts of iron and coal available in other countries. I wonder if this is due to just having more iron and coal mines at start, or if something else is going on.

To be clear, I can't run a decent sized economy off iron and coal imports, but unlike wood, there's more available. I wonder if this is a combination of dual production bugs (as you point out elsewhere) combined with other issues.

I checked the files again, so if the AI goal is on agricultural expansion or plantation economy, it's going to weight those buildings at 2.0 and manufacturing at 0.5, looks like resource buildings are untouched at 1.0 so it will likely build mostly farms/plantations.

Resource expansion goal gives it 1.5 to rubber, oil, mining, and logging.

Industrial expansion is 1.25 to manufacturing, rubber, oil, mining, and logging, so they should be even from a multiplier perspective. Been poking around in console for a bit, even under resource expansion 1.5* multiplier, odds are some random factory is going to be weighted higher, so when it's on par with factory weight, it doesn't have much chance unless the factory outputs fall under 75% price and the resources are over 125% price. Note that there is a bonus to industrial/military goods, and since iron and coal are categorized industrial, but wood isn't, this gives them another advantage in weight.

The production methods chosen to compare weights for building a new building is either the existing one in a state or what the default one will be for constructing a new building, so in somewhere like the France or Russia, since all the logging camps start with hardwood production, any new camps will have hardwood production and use its values and price thresholds for comparisons, while every other markets will use no hardwood. So this means a country like Russia is unlikely to build iron or coal mines unless it becomes a critical shortage because they start with the lowest mining PM, but they will sometimes build logging camps (unless they have a shortage of tools at some point and decide to change a logging camp down to use no tools, at which point the country's default will be no tools logging, and it will never have a chance against other buildings until it decides to change that PM back to sawmills.)

The reason France does so well with building relative to other countries, is that it starts with good PM in all mines and lumber camps, has enough production of every main good that it isn't going to have a price swing that will glitch out the profitability calculations or accidentally run into a shortage, and has some pretty good trade routes at game start that are likely to remain in effect.

And I think the other thing with mines is that once they reach condensing engine they are fairly profitable and can somewhat compete against random industry while logging camps with sawmills aren't going to be weighted favorably until they get electric saw mills, but then the cost of electricity or a shortage there could block the AI at that point, and techs are also unlocking higher PM for factories at that point pushing them even higher once they switch to those.
 
  • 7
Reactions: