Victoria 3 - Dev Diary #9 - National Markets

[lachek]

That’d be true if base price is close to equilibrium price, but it can’t be so for all markets and all technology levels. If the base price is enough to keep pre-Bessemer steel mill barely profitable, it should generate huge margin for an EAF mill, so EAF will be able to cope with constantly insufficient demand and low prices and still pay decent wages. Or am I understanding the system wrong?

No, you got it right! And yes, pre-Bessemer Steel Mills are barely profitable (prime candidate for subsidizing if you want to be an early Steel-producing power) while late-game Steel Mills can be very profitable indeed. However, those late-game Steel Mills also require expensive input goods as well as more qualified and higher paid employees, and by that point any decent economy would have seen average wages inflate quite a bit compared to 1836. Since buildings compete with each other for workers what is considered a "decent wage" changes over time, and the high potential productivity of late-game Steel Mills can only be realized with access to cheap advanced input goods.

 

[lachek]

To clarify: situation where "machine tools"(tm) good produced at let's say 10 units, demand is 20 (assuming it is below shortage treshold), the world can operate as if 20 units are actually produced, for as long as buyers can afford high price?I'm especially interested in the case of state, since state can have a very deep pockets. But, are 10 or 20 units actually available, and if 20, whom pockets the difference, for those 10 units that were not produced, but we're bought from somewhere?

In V2, there was a curious dynamic, where you had to create machine tools factories, or your chance of rapid industrialization was zero. Market simply would have supply to actually build those factories.

In V3, that wouldn't be the issue. As long as I'm willing to she'll out extra cash, I can industrialize Russia in 1936, and for a few years the world will just create machine tools out of thin air, for me, right?

Then, question, where will that money go? Or does not closed system also means that money can be payed out into thin air for goods provided from it?

If you have zero access to Tools, the shortage mechanic will ensure you cannot industrialize.

If you have some access to Tools but your demand for it is somewhat higher than that, you will be paying a lot more to industrialize than your more capable competitors. The premium you're paying for Tools will go to your Tool-producing industries and the Pops who own them. If your Tool supply come from imports, it means greater profits for Tool-producers in the exporting countries.

 

[lachek]

I'm afraid I don't quite understand why such things couldn't be modeled in the old system. Could you give some examples?

One fundamental example is that when supply is insufficient, you have to distribute it between prospective purchasers by some rational mechanism. So when you have a Steel Mill and 100 Pops who all want to buy Coal, who gets to purchase what fraction of the available supply?

Option 1: Determine some priority sequence, e.g. buildings always buy first, then Pops buy in order of descending Wealth. This means the Steel Mills might buy up all available Coal, leaving Pops with excess money but no heating. Pops start dying, and the player has to decide to wait until the price of Coal increases due to low supply such that their Steel Mills have failed so Pops get access to the supply, or close/destroy their Steel Mills temporarily until they can secure their Coal supply. Or, if Pops buy first, the Steel Mills get no Coal and close down, thereby causing Pops to lose their jobs so they can't afford to buy Coal, leading to a cycle of different agents in the system failing and/or downward spirals. Furthermore, with sequential purchasing that depends on the outcome of the previous purchase, we cannot multithread the logic. Hard priority buy sequences like this feel artificial and cause frustrations where players ideally want to control the distribution top-down with sliders, which of course is both fiddly and nonsensical from an immersion perspective.

Option 2: Have an overall "trade manager" split supply proportionally between purchasers based on desired purchasing amount. This is the rational decision since it can both be multithreaded and doesn't lead to systemic cyclical breakdowns. However, when combined with a Goods Substitution feature, such a trade manager would have to balance two parameters (availability and price) across a number of different permutations to make a single rational purchase decision that maximizes the yield for potentially tens of thousands of agents. There may be some cool math that can be applied here but none that we can instrumentalize such that it can reliably execute several times per second.

This is already a computationally difficult problem. Now insert local price conditions based on Infrastructure and Market Access. How much of the market's allotment of Coal compared to Oil should a Farmer in Michigan get to buy when Michigan only has 75% Market Access but Coal is produced locally?

With an open economy where stockpiles are abstracted into the price conditions, we can have all purchases resolve on their own thread and tally up into a total combined price. If the price is too high for some of those purchasers to sustain it in the long run, their own economic conditions will deteriorate and demand will decrease. This means we can approach an equilibrium over time by each economic agent doing their thing, rather than trying to compute an optimal distribution between all agents given a multitude of variables that all influence each other.