Indecision

[jameskuyper]

Simutrans seems to be a wonderful game, exactly the kind of thing I love playing - but for one key problem - too many choices. I'm suffering from decision paralysis. I'll use the "Book Empire" scenario as an example to frame my questions.

Which way type should I use? Should I use different way types for different steps between a producer and a consumer? In Book Empire, all of the main routes should obviously be either roads or railroads (you could use boats for part of the connection between Leipzig and Berlin, but that's about it). However,  which parts should be road and which should be railroad?  Road seems the best choice for City factories, but otherwise it's not very clear. Lets assume I choose to use railroads, which seem to be central to this game, and provide the widest range of options.

Which engine type should I use? There's about 15 to choose from. The game allows me, in some cases, to have multiple engines in a single convoy, but I'm going to ignore that complication for now.

Which car type should I use? There can be as many as 3 different car types for any given good category. I've come to the tentative conclusion that it will simplify planning to carry only one freight good category in any given convoy, which simplifies the decision, but not enough.

How many cars should I put in each convoy?

Which rail type should I use?  There's three different speeds. I suppose I should use the cheapest one for those places like station terminals where the train will necessarily be moving slow. Elsewhere, should I just always use the slowest track that is faster than fastest convoy that will be using that stretch of track?

I'm very good at mathematics, and have a fair understanding of microeconomics and bookkeeping. Therefore, I have  a pretty good idea, after all those decisions have been made,  how to determine how profitable it will be to deliver a specified amount of goods each month from the producer to a consumer.  But it would take be a long time to apply that analysis to every single combination to determine the best one. Could anyone give me some tips for simplifying the decision making?

[sdog]

A general, non-specific answer.

My suggestion, use a iterative optimization approach, with a random starting point. Start with a random system, and adjust it to your demands when it's planed too small or go bankrupt when it's too large. Use your economic understanding to analyse it post-hoc. Transfer your education into an informed gut feeling for it.

It really gets interesting when you have a large network, since you get a complex system with non-linear behaviour. You'd be lost calculating things then anyways. But the educated gut feeling you developed before will perhaps help you. Since our brain is rather good at understanding such systems just enough to get somewhere with a trial and error approach. (As a side-effect you might gain something for your real-life decission making competence from that training.)


more specifically:
it is often sensible to start with road networks and gather several routes with same cargo type. As soon as you have a high transport volume, you can replace the common part of that network with rail. First start with point to point rail lines between the hub and road distribution. Later you can change to pure rail networks. That is when you have an idea of the cost structure and more importantly the signaling.


Many pak-sets are designed for the "use timeline" option being enabled. Without timeline, all vehicles are always available, an overwhelming
number. It is recommendable to start at first at some later time, eg 1950, where good trucks are available and you do not have to rely on rail.

[Ters]

I think this is the first time I've seen someone ask for fewer choices, at least when it comes to ways and vehicles.

The type of locomotive to use depends (or should depend) on the type of cargo. If the cargo is heavy, you need a strong locomotive. If the cargo has a high speed bonus, you need a fast locomotive. Fast locomotives, and (ideally) strong locomotives, should both be too expensive for cargo that isn't heavy or in need of speed.

How many cars you put in a train depends on how many cars you need to outweigh the cost of the locomotive, but the more cars you add, the slower the train will run and the less money it will make. Previously, I had a simple formula I used for finding out a good combination of locomotives and cars, but the revenue calculation in the game has been changed to pay attention to the actual speed a train can reach (at least speed limits and too heavy trains), so I'm back to intuition and trial-and-error.

Ways in turn depend on the vehicles used. A high speed train requires high speed tracks to reach it full potential, or it will operate at a huge loss due to lack of speed bonus. In general, I have the same rail or road type for the entire stretch, except though cities (where the game ensures that there are city roads) and at stations where different quality track meet.

[Sarlock]

It will depend somewhat on whether most of your revenue is generated from passengers/mail or from general freight.  Passengers and mail are more sensitive to speed and therefore you will want a higher speed train and higher speed tracks with just enough cars on it to keep the speed near its maximum potential speed.  But only pursue this option if you can run your trains near full most of the time, a half empty or less train will lose a lot of money, fast.  Passengers and mail tend to be a focus in later game stages when you have larger cities and established, wide networks.  Freight, on the other hand, is more suited to higher power locomotives and top speed isn't as much of an issue.  You can load down a powerful locomotive with extra freight cars and make a higher profit margin this way (although at a certain point the train will be moving too slow and you will move past the "sweet spot" for maximum profits from a freight run).  Add more locomotives if you require.  You can match your rail type to the speed that your freight trains are moving to reduce maintenance costs.  In earlier years, pre-1970's, I generally run freight trains with 2-3 locomotives to provide enough net pulling power.
The best answer is just trial and error.  Keep trying different combinations and take your time getting familiar with the game.  It'll take a lot of playing to really figure it out... which gives the game a lot of replayability

[Dwachs]

Somebody is playing with scenarios

I would start with road connections and the cheapest vehicles. Then upgrade if necessary.

[jameskuyper]

In Book Empire, all of the main routes should obviously be either roads or railroads ...

Or trams - I forgot that option because, when I created a tram depot, the only thing it had available was a Tram Team (Pferdebahn). I'd forgotten that they need to be electrified to show the Electric vehicles, which are the only Tram vehicles that can carry anything after 1932.
When I remembered that, and took a look, I wondered why the Tram Team was still there in 1945. I checked to see whether there was anything that could be attached to it. I was surprised to find that the Bennhardt T-26 (electric) (=Lowa) could be attached to it. That seemed pointless, since the T-26 has twice the speed, and can carry passengers. Connecting it to the Tram Team only slows it down. However, I did some experimenting and found that if the T-26 engine was pulling more than 9 fully-loaded Bennhardt T-26 cars, its speed drops below 25km/hr, so attaching a Tram Team would speed it up.
Is the ability to attach the Tram Team to the Bennhardt T-26 an accident? I would think so, except for the fact that it does not otherwise make any sense for the Tram Team to wait until 1960 to retire, when the corresponding wagon retires in 1932.

[jameskuyper]

The vague qualitative advice you all have given me seems quite reasonable, and in fact I already knew it. The problem is converting that advice into discrete choices based upon quantitative judgments.  How big of a speed bonus is needed to justify using a fast train? How fast does a train have to be, to qualify for use with cargo that has a large speed bonus? I know how to determine those things analytically, I'm just daunted by the large variety of options that need to be analyzed. I'm not very good at learning such things by trial-and-error, but I guess that's the only alternative. What I'd really like to develop is some rules of thumb, so I can discard large number of options without even bothering to do the analysis. I guess I'll reach that point only by getting more experience with the game.

Speaking of experience, I just tried out a roads-only solution to Book Empire. I attached a Serf/H-Trans Book Trailer (Buchanhaenger) to a Serj/H-Trans Goods Van (Buchtransporter), and assigned it to a route that stopped at the Chemical Plant, with instructions to wait until it was 100% full of Printer's Ink. It went there, stopped just long enough to turn around without loading anything, and then continued on to the Printing Works. I tried it a second time with the Trailer removed, and that time it stopped and waited until it could load up on Ink. Can someone explain that behavior? Printers Ink is a boxed good, just like Books, and both the Van and the Trailer are marked as being able to carry boxed goods.  The Trailer's name implies that it should be used for Books, but there's nothing it the *.dat file to suggest that it's more restricted than the Van in what it can carry. Even if it was, the Van, at least, should have still been able to load up with Ink.

[Ters]

The goods list shows how much you earn by transporting a given cargo at a given speed at the present time. From the depot, you can get the capacities and running costs of the vehicles. It is then almost just to start calculating, based on the theoretical maximum speeds.

But that's just a rough estimate. You also need to calculate how much the infrastructure will cost, and how fast a convoy can go based on the road, the power and the weight. It is possible to calculate it all, the game does after all do it. Most of it is straight forward, except the speed, where the game's pseudo physics come into play. I haven't seen the formula for that, and it might not be complex, but it's based on what looks good rather than what's realistic.

[prissi]

The speed a train can achieve (and only that and the track speed limtis count for income) is actually indicated in the depot dialogue. So the information is there.

And the choices are a very important part of the game aspect. Most games are somewhat or fully deterministic but complex enough so the simple rules do not spoil the long time fun.

[Combuijs]

Speaking of experience, I just tried out a roads-only solution to Book Empire. I attached a Serf/H-Trans Book Trailer (Buchanhaenger) to a Serj/H-Trans Goods Van (Buchtransporter), and assigned it to a route that stopped at the Chemical Plant, with instructions to wait until it was 100% full of Printer's Ink. It went there, stopped just long enough to turn around without loading anything, and then continued on to the Printing Works. I tried it a second time with the Trailer removed, and that time it stopped and waited until it could load up on Ink. Can someone explain that behavior? Printers Ink is a boxed good, just like Books, and both the Van and the Trailer are marked as being able to carry boxed goods.  The Trailer's name implies that it should be used for Books, but there's nothing it the *.dat file to suggest that it's more restricted than the Van in what it can carry. Even if it was, the Van, at least, should have still been able to load up with Ink.

Actually,  ... eh ... no. Your first solution should already work, certainly if it did work later on without trailer. You must have done something wrong the first time. Are you sure you did wait for 100%?

[el_slapper]

IMHO, a full pre-analysis is not possible.

It's obviously impossible for passenger transport. The system is so dynamic that what is true in March will be false in april. And that's realistic. Your actions upon the system will change it, therefore no estimate done before the building will be accurate. You might even have a small line working perfectly, and, all of a sudden, with a connection to a bigger network, the line gets massively overwhelmed, & needs a complete remake.

It's not THAT impossible for industries. There is probably a way to mathematically optimize income on lines, especially simple ones like coal to electric station. Yet, I'm using rules of thumbs : quick estimate of the travel time(thanks to my experience; not even needed on short lines), quick estimate of the capacity needed, choice of loco with help of the tool : if I'm far from max speed, then I'm not powerful enough. This tells me what speed I'll have, & decides what kind of rail I need.

Though it could be fun(and complex with new tarrif rules) to implement a line optimizator. Though wiht last version, industry output is influenced by passenger & mail, and the system might get unstable - and therefore with no definite otpimum.

Philosophically speaking, I'd say it's also a matter of strategy. Micro-optimization May improve an existing system, but the first(and amongst the most interesting) decisions to make are strategic : industry first? Bus first? All-electric? Mixed? I tend to go for all-electric for passenger trains, & no electric for industrial trains. I tend to go industry first. But that's choices, & different ones may be useful also. Once you've decided to link Coal station A to Coal Mine B, then you can start micro-optimization. But if it is a strategic blunder, whatever your micro-optimizations, you're screwed.

[jameskuyper]

Yes, I'm sure it was set to 100%. At least, I am sure as I can be, given that a) I didn't save the game and b) I can no longer reproduce the problem.