News:

Simutrans Chat Room
Where cool people of Simutrans can meet up.

Efficient subway ?

Started by Iluvalar, November 21, 2015, 05:55:17 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Iluvalar

Hello everyone.
This is my first try at a game start with a subway. It's on pak128-iluV10 so I know the vehicles and road are balanced. It's the design that seem to fail me.





The line 1 go anti-clock wise into the right-most lane (Treadle-gaswork...) and down back in the 2nd lane.
The line 2 go anti-clock wise in up the 2nd lane and down the 3rd...
etc..
There is one last line that go across in the middle.


Obviously this is not good enough..


Again, I'm pretty sure that I have enough traffic to use the 60km/h road in pakilu. The only problem is the design of my network... did anyone managed to get a profitable subway ? How is it designed ?

DrSuperGood

Due to costs of tunnels and revenue, pak128 does not support subways. Subways will almost always operate at a loss, even in very big cities. Pak64 supports subways and they can even be profitable.

The secret with subways when using payment mode 0 (recommended) is to make them a single "inefficient" loop as long as possible and modulate throughput with number of convoys and wait for full time. The longer they are the fuller they get. The fuller they get, the more profit they make between stops. A good subway can easily make several hundred million a month in pak64. In pak128 you really need a big city and to use faster trains, even then they may struggle to break even due to the high cost of tunnels and sub 100% utilization. If you use payment model 2 then it is impossible for any subway to make profit ever and they will run at huge losses which have to be recouped from long distance passenger haulage.

In pak128 for starting I recommend ignoring subways and instead exploiting extended stop area using disjoined stop components. This saves you losing money on a commuter service, provides functions for people to walk around the city to their destinations and gathers all passengers together so you can ship them between cities efficiently and economically. Some server admins may view this as an exploit as it provides free growth and solves commuter line problems but personally it is more realistic as the stop areas are far too small anyway. I would very much like to see this becoming a supported feature, with appropriate limits and costs to balance it (unlike currently where there seems to be no limit).

Iluvalar

So no one will have a good design for me since there was no way to make it with totaldistance=2 until my pak ?


Well I'm glad I changed this :P . I do only 4k profit from a 200k investement. But this design start to work. This time I figured that by making this 2-1-2-1 depot pattern I can fit a lot of little trains and they have no way to jam. And they go bakc and forth instead of doing a loop. The east west lines dont do any profit right now. If I had to restart this I'd build them on the second basement level to see if i can make the run straight as well.

DrSuperGood

Quote
So no one will have a good design for me since there was no way to make it with totaldistance=2 until my pak ?
Few people use that option because it is unrealistic and encourages point-to-point lines. It is unrealistic in that you do not get paid for all the effort you take to deliver something, so going around a lakes or mountains earn you as much as if it was a perfectly flat plain. It also makes efficient exchanges hard to make since you only ever want to ship constructively.

Specifically in the case of transfer lines that pass the target point by a significant amount to a major hub (logistics centre) you end up losing a lot of money due to negative pay from the overshoot and high running costs. Such hubs are common with passengers and mail since they might service the nearby 5 or so cities in a circle and link to another hub by high-speed rail or by plane (which mostly have very bad operating margin due to running costs).

Your underground is not working well? there are overcrowded stops.

I usually use huge rail loops of 12 long stations for coverage and efficiency. Often 16 such stations in a single line with constant trains looping around. Number of trains per month regulated to assure good service (no overcrowding) and high utilization (no wasted capacity). One of the stations is a central hub which can connect 2 such subway systems together and forward on passengers to even bigger hubs via a high speed point-to-point rail link. Such high speed point-to-point links some times share a common backbone with other such links, but some times have dedicated links due to the volume of traffic involved (hard to make high-throughput terminals in pak64 due to lack of elevated ways).

In a pak64 server game earlier this year I had large hubs with a throughput of several hundred thousand per month (12 long trains constantly coming and leaving). All passenger areas serviced were green with 100% pickup rate of available passengers. Only crowded stations were central hubs (was not aware of mail asymmetry) and the central passenger exchange (mostly due to other players out of my control). Cities ended up covering a considerable area of the map.

gauthier

After having tried to make a profitable subway network in pak128, I think there are two major problems:
_ Revenue depends too much on speed, but subway trains are slow so they can't make enough money. When I make regional trains (faster and with fewer stops) go under cities, they still make big money.
_ Distance between stops in a subway network is short, thus trains spend a lot of time stopped. -> You have to take this into account when balancing vehicles (this is obviously not the case in pak128's current balancing).

In your game, you are trying to serve all the city (or the district, I don't know what is on the surface) with subway. Try to use subway only on most crowded lines (for instance a subway serving several hubs of the city).

Tunnels'maintenance cost is a problem. A way of balancing tunnels closer to using subways would be a maintenance cost closer to normal ways but a very high building cost to prevent players from using subways when it's not necessary.

Anyway, subways are still the most efficient way of transporting people through a city in term of capacity, buses and trams are farly outrun.

DrSuperGood

Quote
Tunnels'maintenance cost is a problem. A way of balancing tunnels closer to using subways would be a maintenance cost closer to normal ways but a very high building cost to prevent players from using subways when it's not necessary.
Except the build UI is so clunky that it does not really support that. I find that I need to rebuild ~20% of ways I make because of various reasons (off-by-1, does not work, etc). This means paying for the way and then paying for the way again to have it removed.

Quote
Anyway, subways are still the most efficient way of transporting people through a city in term of capacity, buses and trams are farly outrun.
Most efficient is to have them walk... Especially when you are already losing money due to various reasons.

The problem is the stop pickup range is so small that it becomes very problematic to service a city. In Experimental you have a much bigger pickup range for passengers meaning you can have reasonably spaced stops. In standard you need to create massive spread arrays of stops to cover cities.

Ters

Quote from: DrSuperGood on November 22, 2015, 02:52:22 PM
The problem is the stop pickup range is so small that it becomes very problematic to service a city. In Experimental you have a much bigger pickup range for passengers meaning you can have reasonably spaced stops. In standard you need to create massive spread arrays of stops to cover cities.

I find that the coverage size is fine compared to the street grid, with a stop for every two-three city blocks, but not when comparing with the vehicles' size and acceleration.

Iluvalar

Well obviously, I'm in a subway only game since yesterday. Things are going well... I converted the east-west axis (line A to E)  into that 2-1-2-1-2 pattern I made up, they are more profitable this way. I made a line 0 and a line 5 and recently I updated the line 3 and line B into 120 km/h tracks. All this from the profit of the starting network, so i'm definitively in a working subway game.



@DrSuperGood,The optimal network to travel for 1 corner to the next with paytotaldistance=0 imply a boeing 747 and a tour around the world. I'm not a die hard fan of realism but I dont think it's better. 

My pak encourage the merging of the road with the speed bonus. You can lose money on a small city line in order to merge in into a faster and longer highway trough a hub. That's actualy encouraged. However, you still deal with a totaldistance=2 system, so you can't stretch your luck more then a certain point.

@Gauthier, effectively that's how I balanced it, tunnels, elevated way and tunnel are twice as expensive but only cost 5% more maintenance. It's enough to make players prefer the normal ways but not punish anyone who need a bridge.

@Ters, I don't know for the other paks, but looking at how many small stop I had to make in this game, I don't think it fullfill any gameplay purpose, I'll increase the station coverage on my next version so see how it goes. Anyway, I'm also overcrowed so I wont save much on stop prices. I will just have less of boring similar stop to manage...

isidoro

Quote from: DrSuperGood on November 22, 2015, 12:59:22 AM
Few people use that option [totaldistance=2] because it is unrealistic and encourages point-to-point lines. It is unrealistic in that you do not get paid for all the effort you take to deliver something, so going around a lakes or mountains earn you as much as if it was a perfectly flat plain.
[...]

But you can build fake detours just to make more money if you don't use it...
That was the main reason to add that option:

http://forum.simutrans.com/index.php?topic=678.msg5559;topicseen#msg5559

DrSuperGood

QuoteBut you can build fake detours just to make more money if you don't use it...
Yes, but do you really need more money? Outside of the start you end up with so much cash in both pak64 and pak128 that you do not need to haul longer distances. Additionally it increase route length to make money, something which you can do with other means.

Specifically you can refuse to take the logical next door coal mine to coal power station route and instead opt for one on the other side of the map over 2,000 tiles away. You get paid for all that distance even though you could just stick to the next door coal mine only 10 or 20 tiles away. The problem is not really the payment model but industries making illogical links and paying for goods that are transported inefficiently. Even passengers will take impossibly expensive journeys all over the map such as being bounced around in a concord jet for several months.

The cargo bouncing exploit is cheaper than normal haulage because you can increase traffic on a way a lot, hence it becomes increasingly profitable to haul like this. For this reason any respectable server administrator should liquidate companies caught using it.

Ters

Quote from: Iluvalar on November 22, 2015, 09:50:00 PM
@Ters, I don't know for the other paks, but looking at how many small stop I had to make in this game, I don't think it fullfill any gameplay purpose

You've created subways stations with room for trains with about four carriages spaced out every few kilometers. (1 tile = 1 km, from what I've heard. Experimental has reduced that to 250 m, I think, which still means 1 km between stations.) Although 6 to 8 carriages might be the norm for stations today, London Underground has stations only a few hundred meters apart. The same have Berlin and New York. And in addition, they have buses, with stops even more closely spaced.

Not that I'm a fan of building urban transport myself, but even a medium sized city will realistically have a hundred stops.

Vladki

What about using trams in underground? They have full speedbonus at lower speed.

Ters

Quote from: Vladki on November 23, 2015, 07:57:36 AM
What about using trams in underground? They have full speedbonus at lower speed.

That's what I've usually done in pak64, since they look more subwayish than the trains, at least until the relatively recent addition of a non-ICE EMU.

DrSuperGood

Quote
That's what I've usually done in pak64, since they look more subwayish than the trains, at least until the relatively recent addition of a non-ICE EMU.
That and the fact that 120 km/h tunnel used to obsolete (not available late game) and was more expensive than 80km/h tunnel + tram track.

Quote
You've created subways stations with room for trains with about four carriages spaced out every few kilometers. (1 tile = 1 km, from what I've heard. Experimental has reduced that to 250 m, I think, which still means 1 km between stations.) Although 6 to 8 carriages might be the norm for stations today, London Underground has stations only a few hundred meters apart. The same have Berlin and New York. And in addition, they have buses, with stops even more closely spaced.
Although distance wise 1 tile is measured as 1 km this clearly is not possible because a decent 6 unit passenger train (the sort that stops in a 100m long station odd) then is 3 tiles so 3km long (comparable to some of the longest freight trains in the world). Road is also 1 km wide with a double wide track being 2 km wide. Being 100m or even 10m would be more reasonable.

Ters

I thought you knew there were different scales for vehicles and the world. This is necessary in order to simulate entire countries without making the vehicles too small to see (without zooming in so that you don't see anything else). But this duality in scale, both in time and space, is of course also the problem.

Iluvalar

#15
Quote from: Ters on November 23, 2015, 06:44:49 AM
You've created subways stations with room for trains with about four carriages spaced out every few kilometers. (1 tile = 1 km, from what I've heard. Experimental has reduced that to 250 m, I think, which still means 1 km between stations.) Although 6 to 8 carriages might be the norm for stations today, London Underground has stations only a few hundred meters apart. The same have Berlin and New York. And in addition, they have buses, with stops even more closely spaced.

Not that I'm a fan of building urban transport myself, but even a medium sized city will realistically have a hundred stops.
That's your story. In my game, the trains are 1km wide by two house worth long and they take about 4 hours to travel their own length. I have a 6km^2 stations spaced by about 6 houses from each other, in both directions.

Simutrans have some scaling problems. It struggle to make the game look good for the eyes or look good for the realism.

For the first time in my pak, i'll use realistic metric to decide of the reach of my stations. 2 approachs : a) newyork have 8.5m population with 469 stations, my little city there have 129'000k people. It should need around 7 stations. Each stations should have a reach of 9. b) In my home town, where there was in fact a bus every 2 corners there was also a 22 house long avenue between the main road deserved by bus. That would make each stop reach 11 tiles. That would make sens because my stops are full as they are right now, so even if I could space them more, I would also tend to spam them on the way of my buses to increase the total storage.

So I think I'll try a passenger game with 10 tile reach stops

Ters

Quote from: Iluvalar on November 23, 2015, 07:43:42 PM
Simutrans have some scaling problems. It struggle to make the game look good for the eyes or look good for the realism.

It's not just you struggling. If Simutrans was scaled realistically, maps would be about 100000x100000 tiles, and it would take trains an hour or more (real time and game time) to travel between two big cities.

Quote from: Iluvalar on November 23, 2015, 07:43:42 PM
newyork have 8.5m population with 469 stations, my little city there have 129'000k people.

That's another thing. Population may not fit either scale.

tubanonymous

Interesting. I actually find the scaling to work quite well, in regards to population at least. Efficient subways are possible in simutrans, they just have to be built more like real subway systems.

1. The face of the map should be so built up that only a subway can navigate the landscape without excessive bridges/underpasses or deleting of public buildings (this on its own doesnt guarantee anything, but this is how subways work--they support an already robust network, providing an *overall* advantage on the map, not just to their own lines)

2. The services should be long. Someone earlier mentioned NYC has 400-something odd subway stations. This is true, but the 6 train, for example, has I think 38 stops on it. MTA in NYC has 22 separate services, with special express services on some of those normal services. This brings us to the next point

3. Systems should share as many stations as geographically convenient.

4. Population density needs to be extremely high over a large area. To use NYC as an example again, there are about 175 city blocks that make up the length of what most people think of as Manhattan. Imagine how big of a city that is in simutrans. If you were to build a city this size in the game, it could support the previously mentioned 6 train which spans 38 stops. We know this is true because it is easily done in ny911's New York scenario

5. You need to be well networked. Just as in real life, your networks need to be linked to airports, bus stops, and normal train stops to have the ridership to support their cost.


If we are talking about using subways as early game or aesthetic tools, then maybe changes should be made. But as it is, I think simutrans reflects nicely the purpose and risks and etc involved in a subway. When you consider that 1km of subway could easily cost 250 million dollars, or that right now, NYC's new "T" line is costing 1.7billion per kilometer, we should be thankful :)

Seems to me simutrans needs to have another discussion about realism and fun and how big the grey area between is

Ters

Quote from: tubanonymous on December 03, 2015, 10:08:31 PM
Interesting. I actually find the scaling to work quite well, in regards to population at least. Efficient subways are possible in simutrans, they just have to be built more like real subway systems.

With what pak set(s) have you experienced this? This usually makes a difference between what works and what doesn't.

Quote from: tubanonymous on December 03, 2015, 10:08:31 PM
Seems to me simutrans needs to have another discussion about realism and fun and how big the grey area between is

It's not so much a discussion of fun versus realism, but of one player's idea of fun versus another player's idea of fun. And such discussions pop up at least once per year. Some like to play on the big scale and would like to build only one station per town, others like to micromanage every single building or even tree to create a world that looks pixel perfect. On another, in principle independent, axis, some players want to build a big complex network and transport everything possible, profitable or not, while others are more focused on a money, with transportation more like a means to an end.

tubanonymous

Quote from: Ters on December 04, 2015, 06:37:47 AM
With what pak set(s) have you experienced this? This usually makes a difference between what works and what doesn't.

It's not so much a discussion of fun versus realism, but of one player's idea of fun versus another player's idea of fun. And such discussions pop up at least once per year. Some like to play on the big scale and would like to build only one station per town, others like to micromanage every single building or even tree to create a world that looks pixel perfect. On another, in principle independent, axis, some players want to build a big complex network and transport everything possible, profitable or not, while others are more focused on a money, with transportation more like a means to an end.

--I've played for about 2 years on whatever versions of pak128 there have been. In the beginning, I also played 128britain, but couldn't yet make enough money to avoid bankruptcy because I was just getting use to the game, so it is somewhat irrelevant. I've also dabbled in experimental, but the visual interface for the weight of vehicles vs. weight restrictions on tracks was really hard for me to navigate (not that it isn't an interesting concept). I found the same to be true for time tables, but I haven't played experimental for maybe a year--so once again, maybe irrelevant. I am currently playing as Exxon Mobile in an online game, and I am set to hit $2billion in the next 6-7 years. 

--I agree with everything you say about playing styles. That is why simutrans is awesome to me--it matches my style of play. I like to build things as realistically as possible, which basically means networks that are not only large, but micromanaged as well. Small primary industries may not be profitable on their own, but secondary and tertiary industries which need them in their supply chains can be profitable enough to support their services.I play simutrans with standard settings, or no timeline

Ters

Having the timeline off might be crucial here. Whether something makes a lot of money or drains a lot of money depends very much on the timeline, at least with pak64. I don't know how the prices are with no timeline, but the situation will certainly be different from at least some points of the timeline.

DrSuperGood

I think speed bonus is disabled when timeline is off so many of the "late game" convoys become over powered as they are designed for the much higher speed bonuses needed late game.