The International Simutrans Forum

 

Author Topic: Simutrans Experimental pre-1900, balancing.  (Read 4760 times)

0 Members and 1 Guest are viewing this topic.

Offline Dutchman on Rails

  • *
  • Posts: 239
Simutrans Experimental pre-1900, balancing.
« on: April 12, 2010, 07:42:08 AM »
Following several people speaking out having problems with the balance of Simutrans Experimental when playing a pre-1900 game, I've promised James Petts to write a few comments here on how I play that. I normally start my test games in 1750, and though I normally find events to report, after some tinkering I found some settings in the .tab files as well as some management methods that work for me in creating a challenging and often slow starting game, but not one that is impossible to win.

A small disclaimer. This is a playing style and my playing style, but not the playing style. It works for me, and I suppose it will turn a small but steady profit if replicated. But different configuration settings and playing styles will turn different results, and there are hundreds of ways to play, and any playing style that works for any player is a good one.

First of all, the settings. Two settings have turned out to be vital for my games:

1) Bits_per_Month = 20 (default = 18). When first playing in Simutrans-Ex 7.0, it turned out all of my freight convoys were making a loss. Cross referencing with earlier Simutrans-ex games played with Pak.German showed that the Bits_per_Month setting was the problem. With this setting at 20, most freight vehicles should turn a profit, and one should find at least one profitable vehicle type for each kind of freight. A word of caution though, not all vehicles will be profitable for all kinds of freight. As an example, the goods stage wagon will only be profitable on the piece goods with the highest transport prices.

2) Maintenance_Buildings = 100-500 (default IIRC 1800). This one is pretty straightforward and needs little explanation. The horses and carts make little profit, so your cheapest stops shouldn't cost dozens or hundreds of pounds to maintain. If they do, you make a loss bigtime. In previous versions, I set the Maintenance_buildings to 100, because there was no cheap freight stop. With the arrival of the mailbox, I feel in future games 400-500 will do nicely.

Regarding the management, two important differences need to be noted from Simutrans Standard.

First of all, in Simutrans standard, passenger travel is the money maker. Even between townhall only hamlets, connect them all and you'll have hundreds of passengers to move that pay at every intermediate stop and generally spin big money.

In Experimental in the 1750's it simply doesn't work. Passengers want to travel, but as long as you can't deliver anything better than slow, shaking, noisy coaches on bad roads, almost everyone prefers to stay close to their villages. And they pay when they arrive at the destination or where they need to change, scattering your profits and generally lowering them. I find I may be able to earn some money using omnibuses in the larger villages and towns, and sometimes to a nearby village or bigger lines to a metropolis, but generally my source of income is freight, then mail, and maybe if I'm very lucky some passengers.

Second, in Simutrans standard, the preferred management style by many players I follow is to build for the future. It costs many more times to rearrange a track than to build it right for high capacity in the first place.

In Experimental in the 1750's. my starting company normally simply can not sustain such farsighted planning. I'm busy enough merely surviving the first decades and finding opportunities for growth and more profit, as there's precious little to begin with per line. That means building where the business opportunities are, to start making money quickly and cheaply, and make my income work for me to get more profit. Meanwhile there are a lot of limitations in my infrastructure I simply have to grudgingly accept until either they become intolerable (for instance because of later engines) or, much less likely, until I'm swimming in money and can afford to spend it for the benefit of future generations. :)

I hope this may help managers of early companies a bit. The 1700's and 1800's are not the easiest years to make money, but there are businesses and people out there that need their cargo and mail transported, and even sometimes want to travel. For me, it's a challenge I particularly like.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Devotee
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Simutrans Experimental pre-1900, balancing.
« Reply #1 on: April 12, 2010, 11:18:02 PM »
Dutchman,

thank you for your detailed explanation and feedback - it is very interesting to read how a Simutrans-Experimental game plays out in practice. It certainly was the intention to put the player in the realistic position of not having the luxury of the funds to build the network that will eventually be required right at the very beginning - just as in real life!

The situation with the buildings I am working on - a more sophisticated solution is required than just reducing building maintenance, since that might  make the game too easy in the later stages. The support for this is already partly enabled in the code (for stations, not depots), but not used in any pakset yet. I am hoping to have a way of differentiating depots in the next release of Simutrans-Experimental, and make use of that, and the code for stations, in the next release of Pak128.Britain-Ex to deal better with the buildings issue.

As to passengers, what you describe is broadly the effect desired, but it may need further tweaking - it sounds as if passengers are a bit too hard at present. In the early years, it should be possible to turn a small profit with a small number of passenger travellers (a few stagecoaches going between large and medium sized towns regularly), but that profit only being sustainable on the basis of low overheads. If the passenger numbers are too low even for this, I might have to adjust either the overall passenger factor or ease the journey time tolerances a little. Until the railways, freight should be a far greater source of revenue, however.

As to the bits per month setting - did you find that this made time pass too slowly? How long does it take you, in real time, to get through fifty years of game time, for example? What is the furthest that you have got in any of your games started in the 1750s?

It would be very interesting to see your saved games so that I can see some of this play in action - it would be most helpful when thinking about how to implement new features of adjust the balance of existing features.

Thank you again for you useful analysis - and I am very glad that you are enjoying the challenges unique to Simutrans-Experimental!

Offline Dutchman on Rails

  • *
  • Posts: 239
Re: Simutrans Experimental pre-1900, balancing.
« Reply #2 on: April 13, 2010, 04:13:34 PM »
Hi James,

You're most welcome about the detailed explanation and feedback. Some reactions to your questions and comments:

Re: Not having the luxury of funds to build the network right first try: It's good to understand that I also set starting_money to between 10000 and 25000, depending on what I find (in my latest game it's 15000). So I have a few thousand to start with and set up an income of (if lucky) a few thousand per year, then I'm on my own.

Re: Building maintenance. Basically the arithmetic works as follows: In my settings a road convoy will earn somewhere between 2 and 10/month. So with a maintenance cost of some 30/month for two stops and a road depot, you need between 3 and 15 convoys to break even. This is good to note for your development, you can not have a maintenance cost of your necessary infrastructure of more than a few dozen in the 1700's.

Re: Passengers: I think again the peculiarities of my settings put you offtrack. I set passenger_factor to 1 (lowest possible!) and the average town inhabitants in the starting screen to 10, so I only have the smallest possible hamlets and a very reluctant population to travel. My results may be coloured by this, but I like it that way. Sidenote, with Simutrans standard it's still quite possible to send passenger levels through the roof even with these settings, though Pak.German has more helpful coaches than Pak.Britain.

Re: Bits per month: In the beginning, I use the turbo (+w) button a lot when saving a few years for new projects. In later stages, I have much more to mind so I can be quite busy, but still use the turbo for brief periods.

Edit: Also, this setting was, as you will have seen, done because at first it caused loss-making individual convoys into profit-making ones. I do not know why, I do not know if this is still the case, but it works for me.

Re: How long do my games last: One game with Pak128.Britain-Ex lasted until the 1780's before breaking. Games with Simutrans-Ex 6.6 and Pak.German have lasted about as long so far, the latest going from 1815 to well into 1848.

So far the games did not break because of the slow speed, but because of other reported events breaking my pleasure in playing. For the previous one(s), a massive game breaker was that the railway signals didn't work properly in Simutrans-Ex 7.1, and before that (with 7.0), the convoys moving at light speed plus the arrival of 7.1 caused a game to be stopped in the 1760's.

One problem I still have is with the industries, In my latest game, I already have 6 or 7 chains in 1756 with a population of barely 3-4000 souls - and I started with none. I would be happy if I could control that somehow in the configuration.

Re: Save games: I think a replay might help more. I have to experiment with screenshots (my current storage only keeps them for some days), but when done I could see if I can leave some save games there as well.

I hope this helps again.
« Last Edit: April 15, 2010, 04:52:13 PM by Dutchman on Rails »

Offline TygerFish

  • *
  • Posts: 107
Re: Simutrans Experimental pre-1900, balancing.
« Reply #3 on: December 05, 2010, 08:36:59 PM »
Dutchman: do you have any new findings to share since these?  Has version 8.2 made any changes in your experiences?

Offline Dutchman on Rails

  • *
  • Posts: 239
Re: Simutrans Experimental pre-1900, balancing.
« Reply #4 on: December 06, 2010, 07:14:26 AM »
Hi Tygerfish,

The gameplay experience is pretty much the same, though also of my own making as said. Since these comments, I have also been experimenting with the industry balance in one game (the beginning story minus the (expired) screenshots can be found in the screenshots section here, the last part at http://www.tt-forums.net/viewtopic.php?f=22&t=48765). The purpose of these modifications are to have lots of industry chain endpoints (about 1 per 1000 inhabitants), but low production in them (reduced from 200+ to 40-). Because the productive industries remain the same, this means one factory serves many shops, and the idea is that this should create the kind of mixed freight criss-cross traffic you find in real life.

This game broke on the high running costs of locomotives in the 1840s-70s, and the simultaneous release of Simutrans 8.2, which promised a bugfix to the runaway industry expansion (so that I don't have to workaround by using the Pak128.Britain standard industries and placing them semi-manually). A bugfix for that issue apparently is now scheduled for Simutrans 9.0. After that game broke and a brief intermezzo with a 1930s game, I got distracted by other games and only just returned to my experiments.

At the moment I'm preparing a new experiment with the industry balance. Results were promising before, so I'm curious to see how things will go. If it goes anywhere, I will once again report to http://www.tt-forums.net/viewtopic.php?f=22&t=48765, so if you're interested, you may want to set an alert on that topic.

That said, obviously it's difficult to compare my experiences with a Simutrans-Ex game with default settings. When I'm playing now, I do so with a Pak128.Britain-Ex with a few Pak128.Britain standard files, with edited and recompiled boats, road vehicles, trains (all with modifications on running costs), industries (production) and Nelson's collumn (passengers and mail factor), and with both simuconf.tab files, cityrules.tab and forests.tab edited. For me it makes an interesting game, but I'm afraid it says precious little about the game balance...
« Last Edit: December 06, 2010, 07:59:07 AM by Dutchman on Rails »

Offline prissi

  • Developer
  • Administrator
  • *
  • Posts: 9584
  • Languages: De,EN,JP
Re: Simutrans Experimental pre-1900, balancing.
« Reply #5 on: December 06, 2010, 09:34:01 AM »
I wonder, if the bits_per_month issue is fixed. The income in 5 min real time should stay the same, not matter if one game month or 10 have passed.

Offline TygerFish

  • *
  • Posts: 107
Re: Simutrans Experimental pre-1900, balancing.
« Reply #6 on: December 06, 2010, 03:14:19 PM »
I wonder, if the bits_per_month issue is fixed. The income in 5 min real time should stay the same, not matter if one game month or 10 have passed.
I saw that James increased the default bits_per_month value from 18 to 19 in pak128.Britain-Ex 0.7, and I've seen someone trying out 21 as well.  I assume that the setting is still a work in progress...

At the moment I'm preparing a new experiment with the industry balance. Results were promising before, so I'm curious to see how things will go. If it goes anywhere, I will once again report to http://www.tt-forums.net/viewtopic.php?f=22&t=48765, so if you're interested, you may want to set an alert on that topic.
I subscribed to the thread, thanks.

I've been trying out your lower maintenance costs and I think it's probably a positive step that I'll keep in most of my future games.  Track and canals cost money to maintain, whereas public roads and oceans are free, so it heightens the (real-life) incentive to make use of those things.

Regarding the early-times transportation distance, I saw that James increased the default travel time tolerances: each category is now a much wider range.  I may play around with doing that even more as well.  I think that will help the early-times problems while still giving an incentive to keep providing increased travel times: SOME people will always be willing to travel for a long time, but many more people will be willing to travel long distances when it takes less time.

Started looking at it here a while back: http://forum.simutrans.com/index.php?topic=5967.msg57632#msg57632 and I'm still mulling over ideas for how journey distance/time could be set to change over time in a meaningful fashion, maybe based on speed bonuses.