The International Simutrans Forum


Author Topic: How important is industrial density?  (Read 470 times)

0 Members and 1 Guest are viewing this topic.

Offline Matthew

  • *
  • Posts: 99
  • Languages: EN, some ZH, DE & SQ
How important is industrial density?
« on: April 26, 2018, 11:29:33 PM »
A few days ago I began playing the Expanded nightly builds, after several years of playing Experimental 11.35. There are too many excellent new features to list in full, from passenger classes to new species of horse, but I'm grateful to all those (especially James Petts) who've worked on them and I've been impressed by how well many of them have been incorporated into the in-game help system (where I think zook2 lent a hand).

However, there is one area that has left me puzzled, and where it might be helpful to have an explanatory note in the FAQs. It seems that industries have been rebalanced (according to one of Jando's posts, they have one-eighth of their previous demand) and that shipment demands of business-to-consumer industries (B2C = end consumers such as bookshops and bakeries) are far lower than I expected. Based on playing the Railroad Tycoon series and Simutrans over a few decades, I expected that shipment demands would be of the order of multiple freight wagons per game-month, but in fact they often start in the order of a few crates/barrels per game-month. Even when the passenger network has built up enough to allow visitor numbers in the high hundreds, they still only seem to seem to demand a couple of dozen wagons-worth.

Now, this may be a quirk of either the era I am playing in (1840s; 1890s), one particular map, or (most likely) deficiencies in my playstyle. There are so many possible variables. And I am by no means complaining about the situation. On the contrary, I am sure that a great deal of thought has gone into setting these levels, and part of the beauty of Simustrans is that players must create a strategy that fits the industrial landscape they find, not simply build model canals/railways/airlines. And I can see from Github that James has added a constant 25% chance of creating new B2C industries in order to remedy the problem a little.

However, the effect of the balance improvements seems to be that in order to overcome the high fixed costs of 18th and 19th century transport infrastructure, it is important to have a high density of industry. I note that James Petts recommended to the maintainer of the Canterbury-Whitstable online game that the density be set at 5 industries per town. Generating a handful of test maps suggests that you can have 1,000 industries on a 1024x1024 map, or 2,000 industries on a 2048x2048 map (= about 15 industries per town), and it only adds tens of seconds to generation times. There appeared to be four effects of such a higher industrial density.

Firstly, it made the map look much more like the British Isles, because large areas of the map were completely covered by towns and farms (like southeast England and western Ireland IRL) while other areas had empty spaces (=heathland?) between villages (as in the Scottish Highlands). This might not be desirable for Pak128.Sweden-Ex, but it looked great in Pak128.Britain-Ex. I know that some people have complained about 50 grain farms supplying one bakery, and I share their frustration,

Secondly, many of the cattle farms and sheep farms were either missing or failing to display their fields (they only a couple of squares of fields, leading to an output of about 4 head, which was lower than I'm used to).

Thirdly, it meant that there was huge economic potential if all these industries could be linked up, which seems to be the desired situation. Pretty much every village had a market, for example, and medium-sized towns (30,000+ people) usually had more than one. Although I've not playtested the maps, it would seem likely that a profitable network could be established early on and that players would do well to emulate that most Victorian of consists, the Mixed Goods train.

Fourthly, creating those networks was going to be rather expensive, because you would have to destroy farmland in order to build canals and railways. That is not necessariky an undesirable outcome. Land purchase costs made up about 20% of UK railway construction costs in the 1840s. But British railway construction also meant spending a great deal of money on lawyers and Parliamentary agents in order to get the necessary Act of Parliament passed, which really about land purchase through 'eminent domain' (the company formation aspect could be, and almost invariably was, avoided through the use of scrip). The current 5,000¢ cost of purchasing farmland might need to be reconsidered though.

So I wonder whether it might be worth mentioning in the FAQ that it is easier to play with a higher industrial density. It might also be useful for James to ponder what kind of industrial density you regard as optimal as you carry out cost rebalancing.

Offline wlindley us

  • Devotee
  • *
  • Posts: 953
    • Hacking for fun and profit since 1977
  • Languages: EN, DE
Re: How important is industrial density?
« Reply #1 on: April 27, 2018, 02:53:57 PM »
Three cheers for Pak128.Britain-Ex progress.  Most of my games lately have been 1820-1850, the changeover from canals to railways. 

The farm situation might be helped if farms automatically built bridleways (the lowest capacity roadway) from their central building to their boundary, if fields were not built on at least one quadrant, -- or some way of accessing their production without 20,000¢ of bulldozing.

As for consumption, one textile mill probably needs to supply more than one or two clothes shops; the early ironworks needs lower employment levels to avoid perpetual shortages in all but the largest cities... Perhaps similar minor pak adjustments could make a huge difference?

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18030
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: How important is industrial density?
« Reply #2 on: April 28, 2018, 12:00:27 AM »
This is a very interesting experiment: thank you. I have not fully explored the upper limit of the industries per town ratio. However, I note that a 1024x1024 map equates to an area only 128km x 128km, a rather small area: Simutrans-Extended is designed to be able to have maps much larger than that (the current Bridgewater-Brunel game on the server is the scale size of Great Britain at 125m/tile), so I am not sure how feasible that so many industries would be at that sort of size.

Incidentally, as to the rationale for industry production rates: these are scaled by the short time scale (i.e., that shown in minutes and hours in the bottom left hand corner of the screen), rather than the long time scale (that measured in months and years). The idea of having two time scales is to prevent players having to wait real life years for game years to pass. The ratio between the two scales is 6.4 game hours per game month. Thus, the amount of production and demand that an industry will have in a game month is the same amount of production and demand that the real life equivalent would have (if I have found the correct data and interpreted it correctly, or otherwise extrapolated/interpolated correctly where data are not available) in 6.4 hours.

As to consumer industries, this is a known issue, and is scheduled to be dealt with when town growth is reworked, after the current projects relating to convoy/vehicle management are dealt with, which themselves have been delayed somewhat by a large number of bug reports of late. Industries automatically building connecting roads is also planned, but this is an even lower priority than the bug fixing, vehicle management and town growth, and will be quite a lot of work. If anyone would like to help to contribute to the coding, this might help to expedite matters considerably.