Understanding of industry_increase_every

[Václav]

Recently I started new game. As often with

industry_increase_every = 5000

Originally all cities have about 10150 citizens
Currently they have about 18250 citizens

So, logically new industry should be already placed - but still there are only original factories that were available in the beginning of game.

Where am I wrong? What is causing that no new factories were built?

[Giziar]

If I understand correctly the industry_increase_every = x means a new industry chain will appear at a city if the city reaches a population x*2^n.
In your particular case that means at around 5k, 10k, 20k, 40k etc.
Since your cities started with a population just above 10k one would expect to see industry chains start appearing when the cities get to (or near) 20k

[DrSuperGood]

x*2^n.

Frankly that is a rather stupid model which is easily abused. This means that it is best to spam cities close together (so they do not get very big) to maximize population growth and minimize average population per industry boost. Also I am not sure industry growth is all that smart (I could swear a tiny eco-power station counts the same as an enormous multi-chain manufactory industry like concrete or ACC Stones.

So to maximize industry.
1. Get a decent passenger and mail network setup (you need a lot of destinations to assure >70% pickup).
2. Zone an area with efficient transport (underground station coverage, good bus coverage, whatever you want).
3. Cover area with 10-20+ cities.
4. Wait a few years as you will get 20 new industries (with another 20 soon after).

[prissi]

This system was far better than a linear model. It will spawn much less industrie in later play.

This goes together with city growth, so achieve a linear grow in size the growth of large cities is deliberately made faster. Hence industy growth must be a bigger intervalls.

But you are invited to suggest a better model which still supports initial play and does nopt spawn too much later.

[DrSuperGood]

This goes together with city growth, so achieve a linear grow in size the growth of large cities is deliberately made faster. Hence industy growth must be a bigger intervalls.

As far as I could tell glancing at the code there are 3 city growth rates (there is not much documentation on exact mechanics out side of source code).
1. Small town.
2. Large town.
3. City.

These define the maximum possible growth rate of the city with a City of 20,000 or 200,000 (which I achieved on some servers) growing at the same rate.

Although it does say "growth per person" or something of the kind I did not actually see city population being factored into the calculation. As such multiple small towns will give you considerably more growth than 1 single large city even if the large city has many times more people in it than all the small towns combined.

Obviously this is only important for when you can build new cities as otherwise you are limited to what you have.

A slightly better model could perhaps calculate global growth and distribute it among cities. This would obviously be based on map size and current population etc. This would mean that if you have 100 small towns growing or 2 huge cities you still would get the same overall growth for the same total population of the world. Additionally industries could be constructed based on an ever growing population requirement for the next industry (kind of like hero level exp requirements in most RPG games) to prevent excessive industry build up on a global scale. However that is more something for another topic to look at.

[Václav]

1. Get a decent passenger and mail network setup (you need a lot of destinations to assure >70% pickup).

Intercity rail network where trains go from two or more smaller cities into one large - and then large cities are connected between each other - also with trains or with airports

2. Zone an area with efficient transport (underground station coverage, good bus coverage, whatever you want).

I often build streets in cities as a circuits - and bus/trolleybus/tram lines have circuit routes too. But trams break this plan sometimes.

3. Cover area with 10-20+ cities.

I play mostly with 31 cities - because I play on large maps (like 2000 * 1316). It causes that start is a little more difficult, but later money are made very fast.

[Spenk009]

DrSuperGood, I like the idea of having a global growth setting. This could take into account a normal, realistic and steady increase in population volumes and distribute the new population based on successful trips and industry success. Also if editable for scenarios, mass population increases could be simulated for challenges and the like.

How difficult would implementing trip success and overall population factoring into the global growth value be?

Steps for towns to grow could change to being:

1. calculate how much global growth is availabe: (trip_success_% + modifier) * log(global_population) * global_growth = growth in citizens for map
2. decide by successful trips to and from, compared to global trips, how much of the global growth the town can take
3. Grow based on value given

Notes: "trip_success_%" is how many trips were made of all theoretically wished for in the month. "modifier" is a simple helper to ensure at least some growth happens even if there is no focus on pax/mail. "log(global_population) would be the n of 10^n of the world population, so that large maps get more growth than little maps. "global_growth" is the simuconf.tab editable value for scenarios and general adjustment dependant on style of play, maybe even taken from a table of different values for different time periods.

Having levelled industries doesn't seem quite too far sighted, as most industries that are implemented are needed in any "world". The thought of a specific building being built for different industry sizes is very compelling, because it would be great to see e.g. smalled far off islands having smaller infrastructure in their industries. But that would probably be part of the land value / zoning process (which was in the works in Experimental), requiring painting and balancing.

[DrSuperGood]

I play mostly with 31 cities - because I play on large maps (like 2000 * 1316). It causes that start is a little more difficult, but later money are made very fast.

For that size of map you want at least 200-500 cities. Yes I am talking abuse, not trying to play reasonably (which the current mechanics discourage). The found city tool is in the utility section. Pak64 especially allows you to found a lot of cities unless you forcefully make them cost a lot. Pak128 has less money available unless you abuse aircraft so it is slightly harder.

I am not sure if experimental uses the same mechanics but there this kind of abuse can be seen where players have literally covered a 7000*5000 map with several hundred cities. The result is absolutely crazy industry.

DrSuperGood, I like the idea of having a global growth setting. This could take into account a normal, realistic and steady increase in population volumes and distribute the new population based on successful trips and industry success. Also if editable for scenarios, mass population increases could be simulated for challenges and the like.

Yes and since it is not coupled to city volume it means 10 cities in a huge map can end up with >1 million people each or 100 cities with 100 thousand etc. The problem is fairly distributing growth among cities (probably using weighting) which may be tricky to set up. The main advantage is you can define "maximum" growth, unlike the current model where number of cities defines that (which players can change).

[prissi]

Simutrans is a game that can be easily abused. However, then it is not fun playing anymore. I highly doubt that there is any mechanism that cannot be cheated. (Apart from transporting as much as you can is the goal. I would not consider that cheating.)) Anyway, freedom of choice: If you like to "cheat" then simutrans will allow this too.

Remember, SImutrans was single player only for most of its existence. ANd many people still use it as a large model railway and switch industry growth off.

[Spenk009]

Yes and since it is not coupled to city volume it means 10 cities in a huge map can end up with >1 million people each or 100 cities with 100 thousand etc. The problem is fairly distributing growth among cities (probably using weighting) which may be tricky to set up. The main advantage is you can define "maximum" growth, unlike the current model where number of cities defines that (which players can change).

To avoid such huge volumes, we have the factor of global growth that is coupled to the amount of citizens present as to not grow too much or too little depending on how many people actually exist. I'll try putting the makeshift formula into numbers:

Code Select Expand

      (0.30 + 0.40)         *            6           *    1500       = 6300
(trip_success_% + modifier) * log(global_population) * global_growth = growth in citizens for map

Here we have 30% of all trips requested being made, a modifier of 0.4 to not have population descrease while there is no pax transport, 10^6 population on the entire map ( 1 million) and the global growth value set to 1500. The resulting increase in population for the entire map is 6300 citizens per month.

The individual cities will then take from those 6300 free people according to their share of trips made. The scenarios for different coverages and resultant growth is logical. A town with 10% of all trips made globally will get 10% of those 6300.

Scalability is automatic, as a higher global population will increase population increase more slowly as values of 10^n are achieved more slowly.

My own problem with this is that using the log value with such a high numeric importance in the equation can be problematic for balancing.

EDIT: to answer the quote, depending on coverage and success this is true. But the map generation can stay the same, it's very good. Distribution of growth happens when the city compares its trips (which are already measured) to all trips made in the month and takes its share of growth (potential for places to rot and see no change exists, maybe some minimum growth could be included to alleviate that). Maximum growth per city could be encorporated, but I agree that sometimes having a limit is nice in gameplay.

prissi, maps of other players often shock me. I tend to play Exp-Britain trying to stay realistic: http://imgur.com/a/f59XW.

[Václav]

For that size of map you want at least 200-500 cities. Yes I am talking abuse, not trying to play reasonably (which the current mechanics
discourage).

200-500 cities? You are crazy.

Playing Pak128 reasonably may be hard in the beginning - but it is possible. And without many cheating addons. I played pak128 for some (quite long) time and so I know that it is possible.

And now I play Pak96.comic (Expansion) and there are else economic conditions - that allow reasonable playing.