So I built him a bunch of railways yet he doesnt want to subsidize them ruining his own state's eco. How to nod the AI into the right direction?
- 1
-
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.
I don't really think you can.
The AI is shit at managing infrastructure, and frankly, the whole infrastructure system is an overengineered and underthought piece of dung.
Maybe the AI really doesn't have any money to spend on subsidies? Try bankrolling it or reducing subject payments. However, this is a very broad answer to a specific problem, and unlikely to help. There's no explicit option to subsidize your subjects' industries, even infrastructure, you just have to live with that.
The AI is shit at managing infrastructure, and frankly, the whole infrastructure system is an overengineered and underthought piece of dung.
Maybe the AI really doesn't have any money to spend on subsidies? Try bankrolling it or reducing subject payments. However, this is a very broad answer to a specific problem, and unlikely to help. There's no explicit option to subsidize your subjects' industries, even infrastructure, you just have to live with that.
- 7
Not only at managing infrastructure.
A similar thing happneed to me but regarding convoys. Playing as Dutch East Indies. All was okay while Netherlands was in the British market but as soon as the Dutch had their own market again, overseas market access dropped and the Dutch refused to build more ports.
- 3
What is the demand of transport in Zulia?
It seems it is not hiring because it is not profitable, it could be due to low demand or high input costs.
Subsidizing is an option but not the best one.
At 52% market access…
Last edited:
Has the state enough willing workers to be hired? You can import steam engines to the Venezuela market to decrease input cost.
- 1
How does profitability and transport demand come into question when you are in infrastructure deficit? It shouldnt. If it does, thats a bug. The issue is present and needs to be solved. How do we humans solve it? Very easily: build railroads and subsidize them. Why cant the AI follow this logic? Lets pretend we wouldnt abuse the AI to build them 1000 levels of railroads, lets not go there.
It has, you can see has 1.6m pops, that should be more than enough to hire for 10 levels of railroads with subsidies.
- 1
I do not subsidize railroads anymore, I increase demand for transport above supply and that's enough for my railroad companies to be profitable as I keep input prices below base price. I only subsidize if I need more infra than the provided by the railroads sized for state demand for transport as I said, but that's very rare.
AI follows my logic, subsidizing is very dangerous. As you can see there are already 10 levels of railroads but none is filled. If they were subsidized the AI would be losing money.
My guess is that the AI has not switched to transport PM in buildings in the state because there are more peasants than jobs available and that makes demand for transport very small. It does not justify to employ anyone, specially at those engine prices due to the 52% access to the market.
PS: if this is the case, what you want for the AI to use railroads is to build enough buildings to employ all peasants so the AI starts to automate, aka, using transport.
PS2: start to build everything that has local demand so the buildings will be profitable no matter the access to the market (unless using advanced PMs that require outside goods). Try to build low infrastructure buildings first.
PS3: start to share knowledge with them, they might not have all the transport PM yet.
I am guessing the problem here is the 52% access to your market. By subsidizing the railroads the infrastructure problem would be solved.
Look it from the AI point of view. By cutting access to your market their POPs are consuming mainly the local products and the government has not to pay subsidies to railroads.
I think AI is very smart here! (irony on)
Last edited:
- 2
- 2
1.6m pops did not mean that there were enough willing jobseekers.
Check the Buildings tab of the state and it should show how many people want to change their current job
From the first screenshot, it looks like there are 0 unemployment and 0 peasants pops in the state.
In this case, there could be a high number of vacant job positions in other buildings with better wages.
Maybe if you own several of them then you can downsize these buildings to decrase demand for the workforce
Last edited:
- 1
You are right, there are no peasants nor unemployment but 204k jobseekers.
If the AI is not subsidizing railroads, the only way to employ them is to make them very profitable so they can pay higher wages and get those jobseekers.
Because we do not how much transport is demanded in the state, we can’t know if there is a problem with techs or the AI script to automate.
Guys, that is exactly why you MUST subsidize. Its paramount to fix this extreme infrastructure deficit. Any sensible human would do it in an instant. What are we even arguing about guys?
- 2
We are trying to understand why the AI is not making use of the railroad building.
Let's look at the game files. On the defines we see that the game follows the common sense logic and will give a very high value to subsidize if the building provides 100% of the infrastructure. However, in Zulia this is not the case due to ports providing infrastructure.
So this may be an oversight or there is something else going on.
I was trying to apply what I do in the game, and because I do not subsidize, I asked you how much demand is there for transport and what techs Venezuela has.
If we look at the AI weights to choose a PM we have the following:
Automation seems to be mainly about profitability, as I expected. Thus, this makes the AI not automate because transport is very expensive and this may be making transport demand -> 0.
It is something that as a Player I do not have a problem with as I switch to automate to increase transport demand to increase railroad occupation, even if that hits profitability of the building that automates.
The AI logic seems to not work as you have denounced.
Last edited:
- 2
AI needs to be fixed big time.
Another example from my Dutch East Indies run: Price of sulphur already -21 %. What did the Swede and Dutch private investors build in my country? 6 sulphur mines with a profitability of only 8.0! Why didn't they invest in the 16.0 or 22.0 ones, that produce really needed goods?
Another example from my Dutch East Indies run: Price of sulphur already -21 %. What did the Swede and Dutch private investors build in my country? 6 sulphur mines with a profitability of only 8.0! Why didn't they invest in the 16.0 or 22.0 ones, that produce really needed goods?
profitability is not shown in the game, I guess you are describing productivity.
AI takes into account profitability but also availability of labour...
Clearly the AI could improve its performance, I give you that.
# Below data is used to calculate how much weight the AI places on a production building
PRODUCTION_BUILDING_RANDOM_FACTOR = 10.0 # The higher this is, the more random AI construction will be
PRODUCTION_BUILDING_STATE_RANDOM_FACTOR = 1.0 # The higher this is, the more random AI state selection for building types will be
PRODUCTION_BUILDING_BASE_VALUE = 1000 # If no ai_value is scripted in the building
PRODUCTION_BUILDING_NO_AVAILABLE_WORKFORCE_FACTOR = 0.25 # Total weight is multiplied by this if the state lacks enough workforce to staff the building
PRODUCTION_BUILDING_GOODS_PROFIT_FACTOR = 5000 # Per pound of predicted building profit (divided by employee count per level)
PRODUCTION_BUILDING_GOODS_DEFICIT_FACTOR = 10000 # Per pound of predicted building deficit (divided by employee count per level and multiplies a negative value)
PRODUCTION_BUILDING_GOODS_DEFICIT_SUBSIDIZE_FACTOR = 5000 # Overrides PRODUCTION_BUILDING_GOODS_DEFICIT_FACTOR for subsidized buildings
PRODUCTION_BUILDING_PRODUCED_VALUE_FACTOR = 0.25 # Per pound of produced goods in base value
PRODUCTION_BUILDING_INCORPORATED_INFRASTRUCTURE_USAGE_FACTOR = 0.05 # Total weight is divided by ( 1 + this * building infrastructure usage ) in incorporated states
PRODUCTION_BUILDING_UNINCORPORATED_INFRASTRUCTURE_USAGE_FACTOR = 1.0 # Total weight is divided by ( 1 + this * building infrastructure usage ) in unincorporated states
PRODUCTION_BUILDING_INPUT_NO_LOCAL_PRODUCTION_FACTOR = -0.5 # Add this value to a building for each unit of money worth of goods consumed that are not produced in the state but *are* produced by buildings in the market
PRODUCTION_BUILDING_OUTPUT_NO_LOCAL_CONSUMPTION_FACTOR = -0.5 # Add this value to a building for each unit of money worth of goods produced that are not consumed in the state but *are* consumed by buildings in the market
PRODUCTION_BUILDING_OUTPUT_HIGH_PRICE_THRESHOLD = 1.25 # Lower threshold multiplier of base price for applying 'high price' modifier below
PRODUCTION_BUILDING_OUTPUT_HIGH_PRICE_FACTOR = 0.25 # Add this value to a building for each unit of money worth of goods above high price threshold that this building produces
PRODUCTION_BUILDING_OUTPUT_HIGH_PRICE_WANTS_HIGH_SUPPLY_FACTOR = 0.75 # Add this value to a building for each unit of money worth of goods above high price threshold that this building produces and which we want a high supply of
PRODUCTION_BUILDING_OUTPUT_HIGH_PRICE_LOCAL_CONSUMPTION_MULTIPLIER = 2.0 # All factors for high prices are multiplied by this value if any buildings in the state consume the good in question
PRODUCTION_BUILDING_OUTPUT_LOW_PRICE_THRESHOLD = 0.75 # Upper threshold multiplier of base price for applying 'low price' modifier below
PRODUCTION_BUILDING_OUTPUT_LOW_PRICE_FACTOR = -0.5 # Add this value to a building for each unit of money worth of goods below low price threshold that this building produces
PRODUCTION_BUILDING_SUBSIDIZE_PRICE_FACTOR_MULT = 1.25 # Multiply the effects of price factors on buildings that the AI will consider subsidizing
PRODUCTION_BUILDING_OUTPUT_NEW_GOODS_FACTOR = 100.0 # Add this value to a building for each unit of money worth of goods that currently isn't available in our market/state but the AI wants to prioritize
PRODUCTION_BUILDING_OUTPUT_NEW_GOODS_STATE_INCORPORATED_POPULATION_THRESHOLD = 250000 # When the AI prioritizes 'new goods' in a state rather than a market (because of local goods) it only cares about incorporated states with at least this much population
PRODUCTION_BUILDING_OUTPUT_NEW_GOODS_STATE_UNINCORPORATED_POPULATION_THRESHOLD = 1000000 # When the AI prioritizes 'new goods' in a state rather than a market (because of local goods) it only cares about unincorporated states with at least this much population
PRODUCTION_BUILDING_OUTPUT_WANTED_INDUSTRIAL_GOODS_FACTOR = 0.25 # Add this value to a building for each unit of money worth of industrial goods that aren't overproduced in the country and are used by buildings
PRODUCTION_BUILDING_OUTPUT_WANTED_MILITARY_GOODS_FACTOR = 0.25 # Add this value to a building for each unit of money worth of military goods that aren't overproduced in the country and are used by buildings
PRODUCTION_BUILDING_OUTPUT_NEEDED_INDUSTRIAL_GOODS_FACTOR = 0.5 # Add this value to a building for each unit of money worth of industrial goods that are underproduced in the country and are used by buildings
PRODUCTION_BUILDING_OUTPUT_NEEDED_MILITARY_GOODS_FACTOR = 0.5 # Add this value to a building for each unit of money worth of military goods that are underproduced in the country and are used by buildings
PRODUCTION_BUILDING_FAVORED_GOODS_FACTOR = 0.5 # Add this value to a building for each unit of money worth of building goods output that the AI favors
PRODUCTION_BUILDING_DISFAVORED_GOODS_FACTOR = -0.25 # Add this value to a building for each unit of money worth of building goods output that the AI disfavors
PRODUCTION_BUILDING_MISSING_QUALIFICATIONS_MULT = 0.50 # Multiply score by this if there's not enough qualifications to staff the building (disabled for now)
PRODUCTION_BUILDING_DESIRED_INFRASTRUCTURE_SURPLUS = 2 # If infrastructure 'surplus' in state is less than this, AI wants to improve it
PRODUCTION_BUILDING_FREE_INFRASTRUCTURE_TARGET_WHEN_LACKING_WORKFORCE = 5 # When lacking workforce, the AI will stop building in the state once available infrastructure is this or less
PRODUCTION_BUILDING_EXCESSIVE_INFRASTRUCTURE_SURPLUS = 10 # If infra 'surplus' in state is at least this, AI doesn't want to improve it
PRODUCTION_BUILDING_REDUCE_SHORTAGE_MULT = 5 # Multiply score by this if we're reducing the state region's shortages
PRODUCTION_BUILDING_INCREASE_SHORTAGE_MULT = 0.01 # Multiply score by this if we're increasing the state region's shortages
PRODUCTION_BUILDING_OTHER_BUILDING_TYPES_UNDER_CONSTRUCTION_DIV = 0.5 # Divide score by (1 + this * number of other building types queued in the state)
PRODUCTION_BUILDING_LONG_CONSTRUCTION_TIME_THRESHOLD = 40
PRODUCTION_BUILDING_LONG_CONSTRUCTION_TIME_MULT = 0.5 # Multiply score by this if construction time would take more than PRODUCTION_BUILDING_LONG_CONSTRUCTION_TIME_THRESHOLD weeks at full construction industry usage (ignores MAX_WEEKLY_CONSTRUCTION_PROGRESS)
PRODUCTION_BUILDING_VERY_LONG_CONSTRUCTION_TIME_THRESHOLD = 60
PRODUCTION_BUILDING_VERY_LONG_CONSTRUCTION_TIME_MULT = 0.25 # Multiply score by this if construction time would take more than PRODUCTION_BUILDING_VERY_LONG_CONSTRUCTION_TIME_THRESHOLD weeks at full construction industry usage (ignores MAX_WEEKLY_CONSTRUCTION_PROGRESS)
PRODUCTION_BUILDING_COMPANY_MULT = 2.0 # Multiply score by this if we have at least one established company for the building type
PRODUCTION_BUILDING_FOREIGN_INVESTMENT_HAS_RECENT_NATIONALIZATION_MULT = 0.1 # Multiply score by this if the construction is a foreign investment into a country that has recently nationalized buildings held by investor country
PRODUCTION_BUILDING_STATE_RANDOM_FACTOR = 1.0 # The higher this is, the more random AI state selection for building types will be
PRODUCTION_BUILDING_BASE_VALUE = 1000 # If no ai_value is scripted in the building
PRODUCTION_BUILDING_NO_AVAILABLE_WORKFORCE_FACTOR = 0.25 # Total weight is multiplied by this if the state lacks enough workforce to staff the building
PRODUCTION_BUILDING_GOODS_PROFIT_FACTOR = 5000 # Per pound of predicted building profit (divided by employee count per level)
PRODUCTION_BUILDING_GOODS_DEFICIT_FACTOR = 10000 # Per pound of predicted building deficit (divided by employee count per level and multiplies a negative value)
PRODUCTION_BUILDING_GOODS_DEFICIT_SUBSIDIZE_FACTOR = 5000 # Overrides PRODUCTION_BUILDING_GOODS_DEFICIT_FACTOR for subsidized buildings
PRODUCTION_BUILDING_PRODUCED_VALUE_FACTOR = 0.25 # Per pound of produced goods in base value
PRODUCTION_BUILDING_INCORPORATED_INFRASTRUCTURE_USAGE_FACTOR = 0.05 # Total weight is divided by ( 1 + this * building infrastructure usage ) in incorporated states
PRODUCTION_BUILDING_UNINCORPORATED_INFRASTRUCTURE_USAGE_FACTOR = 1.0 # Total weight is divided by ( 1 + this * building infrastructure usage ) in unincorporated states
PRODUCTION_BUILDING_INPUT_NO_LOCAL_PRODUCTION_FACTOR = -0.5 # Add this value to a building for each unit of money worth of goods consumed that are not produced in the state but *are* produced by buildings in the market
PRODUCTION_BUILDING_OUTPUT_NO_LOCAL_CONSUMPTION_FACTOR = -0.5 # Add this value to a building for each unit of money worth of goods produced that are not consumed in the state but *are* consumed by buildings in the market
PRODUCTION_BUILDING_OUTPUT_HIGH_PRICE_THRESHOLD = 1.25 # Lower threshold multiplier of base price for applying 'high price' modifier below
PRODUCTION_BUILDING_OUTPUT_HIGH_PRICE_FACTOR = 0.25 # Add this value to a building for each unit of money worth of goods above high price threshold that this building produces
PRODUCTION_BUILDING_OUTPUT_HIGH_PRICE_WANTS_HIGH_SUPPLY_FACTOR = 0.75 # Add this value to a building for each unit of money worth of goods above high price threshold that this building produces and which we want a high supply of
PRODUCTION_BUILDING_OUTPUT_HIGH_PRICE_LOCAL_CONSUMPTION_MULTIPLIER = 2.0 # All factors for high prices are multiplied by this value if any buildings in the state consume the good in question
PRODUCTION_BUILDING_OUTPUT_LOW_PRICE_THRESHOLD = 0.75 # Upper threshold multiplier of base price for applying 'low price' modifier below
PRODUCTION_BUILDING_OUTPUT_LOW_PRICE_FACTOR = -0.5 # Add this value to a building for each unit of money worth of goods below low price threshold that this building produces
PRODUCTION_BUILDING_SUBSIDIZE_PRICE_FACTOR_MULT = 1.25 # Multiply the effects of price factors on buildings that the AI will consider subsidizing
PRODUCTION_BUILDING_OUTPUT_NEW_GOODS_FACTOR = 100.0 # Add this value to a building for each unit of money worth of goods that currently isn't available in our market/state but the AI wants to prioritize
PRODUCTION_BUILDING_OUTPUT_NEW_GOODS_STATE_INCORPORATED_POPULATION_THRESHOLD = 250000 # When the AI prioritizes 'new goods' in a state rather than a market (because of local goods) it only cares about incorporated states with at least this much population
PRODUCTION_BUILDING_OUTPUT_NEW_GOODS_STATE_UNINCORPORATED_POPULATION_THRESHOLD = 1000000 # When the AI prioritizes 'new goods' in a state rather than a market (because of local goods) it only cares about unincorporated states with at least this much population
PRODUCTION_BUILDING_OUTPUT_WANTED_INDUSTRIAL_GOODS_FACTOR = 0.25 # Add this value to a building for each unit of money worth of industrial goods that aren't overproduced in the country and are used by buildings
PRODUCTION_BUILDING_OUTPUT_WANTED_MILITARY_GOODS_FACTOR = 0.25 # Add this value to a building for each unit of money worth of military goods that aren't overproduced in the country and are used by buildings
PRODUCTION_BUILDING_OUTPUT_NEEDED_INDUSTRIAL_GOODS_FACTOR = 0.5 # Add this value to a building for each unit of money worth of industrial goods that are underproduced in the country and are used by buildings
PRODUCTION_BUILDING_OUTPUT_NEEDED_MILITARY_GOODS_FACTOR = 0.5 # Add this value to a building for each unit of money worth of military goods that are underproduced in the country and are used by buildings
PRODUCTION_BUILDING_FAVORED_GOODS_FACTOR = 0.5 # Add this value to a building for each unit of money worth of building goods output that the AI favors
PRODUCTION_BUILDING_DISFAVORED_GOODS_FACTOR = -0.25 # Add this value to a building for each unit of money worth of building goods output that the AI disfavors
PRODUCTION_BUILDING_MISSING_QUALIFICATIONS_MULT = 0.50 # Multiply score by this if there's not enough qualifications to staff the building (disabled for now)
PRODUCTION_BUILDING_DESIRED_INFRASTRUCTURE_SURPLUS = 2 # If infrastructure 'surplus' in state is less than this, AI wants to improve it
PRODUCTION_BUILDING_FREE_INFRASTRUCTURE_TARGET_WHEN_LACKING_WORKFORCE = 5 # When lacking workforce, the AI will stop building in the state once available infrastructure is this or less
PRODUCTION_BUILDING_EXCESSIVE_INFRASTRUCTURE_SURPLUS = 10 # If infra 'surplus' in state is at least this, AI doesn't want to improve it
PRODUCTION_BUILDING_REDUCE_SHORTAGE_MULT = 5 # Multiply score by this if we're reducing the state region's shortages
PRODUCTION_BUILDING_INCREASE_SHORTAGE_MULT = 0.01 # Multiply score by this if we're increasing the state region's shortages
PRODUCTION_BUILDING_OTHER_BUILDING_TYPES_UNDER_CONSTRUCTION_DIV = 0.5 # Divide score by (1 + this * number of other building types queued in the state)
PRODUCTION_BUILDING_LONG_CONSTRUCTION_TIME_THRESHOLD = 40
PRODUCTION_BUILDING_LONG_CONSTRUCTION_TIME_MULT = 0.5 # Multiply score by this if construction time would take more than PRODUCTION_BUILDING_LONG_CONSTRUCTION_TIME_THRESHOLD weeks at full construction industry usage (ignores MAX_WEEKLY_CONSTRUCTION_PROGRESS)
PRODUCTION_BUILDING_VERY_LONG_CONSTRUCTION_TIME_THRESHOLD = 60
PRODUCTION_BUILDING_VERY_LONG_CONSTRUCTION_TIME_MULT = 0.25 # Multiply score by this if construction time would take more than PRODUCTION_BUILDING_VERY_LONG_CONSTRUCTION_TIME_THRESHOLD weeks at full construction industry usage (ignores MAX_WEEKLY_CONSTRUCTION_PROGRESS)
PRODUCTION_BUILDING_COMPANY_MULT = 2.0 # Multiply score by this if we have at least one established company for the building type
PRODUCTION_BUILDING_FOREIGN_INVESTMENT_HAS_RECENT_NATIONALIZATION_MULT = 0.1 # Multiply score by this if the construction is a foreign investment into a country that has recently nationalized buildings held by investor country
Last edited:
this is a problem that is caused by inadvertent design
devs dont want railroads to be profitable and makes them cost a ton for their output at all tiers, Ai plans and builds like your private sector queue does but with military and government buildings unlocked, and since railroads are BY DESIGN only profitable if you have a shortage while being your only infrastructure booster...
AI does not consider building, subsidizing, or generally laboring them up since it sees doing so as a waste of money it could dump into autarking that state to offset the infrastructure price malices as that creates more profitable buildings anyways
devs dont want railroads to be profitable and makes them cost a ton for their output at all tiers, Ai plans and builds like your private sector queue does but with military and government buildings unlocked, and since railroads are BY DESIGN only profitable if you have a shortage while being your only infrastructure booster...
AI does not consider building, subsidizing, or generally laboring them up since it sees doing so as a waste of money it could dump into autarking that state to offset the infrastructure price malices as that creates more profitable buildings anyways
Nothing to worry about. My qualm is about profitability being such an important metric in the game but not being shown anywhere.
I understand productivity is the most important metric for GDP as wages are 100% going to the GDP calculation and you want to focus on the productivity number to be as high as possible.
However, profitability or the share of surplus that goes to the owners of the building after deducing wages is very important now in the game as it determines hiring, investment, and many other factors.
The only way a player can see profitability level at a glance is in the savings bar of the building. If the bar is green or gold, the building is profitable, if red, the building isn’t enough profitable, even if the productivity is on the green.
Last edited:
- 2
It amazes me that after 2,5 years they still haven't made the very obvious and easy solution of either:
- Make all buildings' employment percentage dynamically adjust their infrastructure usage since why is a factory with 50 people using as much infra as one with 5000 people?
- Remove railway employment dynamically adjust infra supply since if 50 people use as much as 5000, it only makes sense 50 people supply as much as 5000.
Since when infra crashes factories fire people, it'd auto-correct itself dynamically even with the incompetent AI.
- Make all buildings' employment percentage dynamically adjust their infrastructure usage since why is a factory with 50 people using as much infra as one with 5000 people?
- Remove railway employment dynamically adjust infra supply since if 50 people use as much as 5000, it only makes sense 50 people supply as much as 5000.
Since when infra crashes factories fire people, it'd auto-correct itself dynamically even with the incompetent AI.
- 1
I think this is already the case and employment level dictates infra usage and production.
For sure production. Usage with auto downgrade buildings level when not at full employment, it has the same effect as dynamic infra.
- 1
I guess build an engines factory in his state, and potentially even a steel mill and iron an coal there if he has it
The engines are expensive for this railway. But then again, at 52% market access that will happen if the province does not produce engines locally. It's a large premium to buy it from the Dutch market, no premium to buy it from within the state. But then the engines factory will need steel, which would be pricey if it needs to come from outside the state too, and the same goes for coal and iron.
however, if strang labor compettion is involved, then railways will usually lay of no matter what, and the AI might indeed still not fund its railways anyway even if engines and coal are rather cheap. Base railway profitabillety is annoying for that purpose.
Now if you can produce cars, and if a sufficient quantity of the people there will buy cars, then you can also have added infrastructure trough car consumption. Car consumption is my ticket usually to not have to give a damn about subsidizing railways, but it takes some high SOL as to make the people buy enough cars that you need 0 railways, something like average 30 sol tends to do it.
The engines are expensive for this railway. But then again, at 52% market access that will happen if the province does not produce engines locally. It's a large premium to buy it from the Dutch market, no premium to buy it from within the state. But then the engines factory will need steel, which would be pricey if it needs to come from outside the state too, and the same goes for coal and iron.
however, if strang labor compettion is involved, then railways will usually lay of no matter what, and the AI might indeed still not fund its railways anyway even if engines and coal are rather cheap. Base railway profitabillety is annoying for that purpose.
Now if you can produce cars, and if a sufficient quantity of the people there will buy cars, then you can also have added infrastructure trough car consumption. Car consumption is my ticket usually to not have to give a damn about subsidizing railways, but it takes some high SOL as to make the people buy enough cars that you need 0 railways, something like average 30 sol tends to do it.