Transport networks always get overwhelmed

[cosmicray]

I have had a consistent problem with my long-running OpenTTD games - the growth in cities (and their density) eventually overwhelms every possible effort to accommodate them. Even if I have intercontinental airports with 400-passenger planes and massive maglev stations, waiting passenger counts skyrocket. This also occurs with freight as industries produce more and more goods. Within the cities, even there I have trouble. Even with trams, I can't get people between stations in the city effectively.

I have experienced this problem both with and without cargodist. I'm attaching a ZIP with two savegames that illustrate my problem.

First, in Sunbridge Transport, cargodist is enabled for everything, though I haven't paid attention to anything but passengers for awhile. I've tried a hub-and-spoke system -- it helped for a bit, but not for long. I've tried direct flights and trains between cities, that didn't help. If you turn on the cargo flow legend for passengers, the issues are immediately apparent, especially around Lafinghill City. It takes a massive number of trams and even a few planes between Gredingham Airport and Lafinghill City Airport. And then notice that the stations near the Lafinghill City Airport that have trams are completely overwhelmed by passengers and there simply isn't enough roadway to do anything about it. I've played tricks with multiple tram stations and such, and it just doesn't get much better. That was a city I founded, but cities that were there in the game (such as Pradwood and Sunbridge) don't fare much better, though they are smaller. Even the intercontinental airports aren't handling enough traffic.

Then, take a look at Western Water & Rail Co. Unlike Sunbridge Transport, in this game I used the "cheat" to give myself millions at the start, in an effort to see if I could build things from the beginning to avoid these problems. Look at Shorhill. It actually uses maglev *within* the city to handle transport, despite the high cost in terms of consumed tiles of this. Even with that, and with a very high number of trains, it still can't keep up with demand. Stations like Shorhill West have thousands waiting. There is somewhat less red in the cargo flow legend in this game, but there is still a lot of yellow, and that despite me raising the limits of boats well beyond the default and really abusing the heck out of the fact that multiple boats can occupy a single tile (particularly in Noddingworth, which I designed as a water city.) It still took an awful lot of effort to keep up with growth here, as you can tell from the exceptionally complex rail networks. I laid out a grid of primary arteries before the cities grew, and they are getting overwhelmed too.

So, what am I missing? It seems like I must be missing some key piece of strategy or something. Some important wisdom on how to keep things frmo getting overwhelmed? Any ideas?

[V453000 :)]

I have looked at your savegames and enjoyed them a lot. I have not seen anyone put this good effort into a map in years. You could improve things here and there technically as I will mention later, but you do really well simply by building stuff, and a lot of it.

OpenTTD basically cannot be won. It is like tetris, it gets faster and faster until you simply cannot keep up. The original industries just explode up to 2295units of cargo per month, which is really untransportable by normal means. Passengers are more controllable as you can control the pace at which towns grow, but once you grow them, same issue applies.

There are many solutions, one of them is to simply build even bigger, even better, even more efficient. Which obviously helps you only to some degree. You can find a lot of helpful designs and tips at openttdcoop.org

Some train sets have very high passenger capacities, which compensates for the huge productions of passengers/mail. NUTS Unrealistic Train Set is one of them, I do not remember how other train sets are with passenger capacities anymore.

Most of NewGRF industry sets "sanitize" the productions. FIRS simply caps production of industries generally somewhere like 512-1k units per month. YETI makes the production only get as high, as your network can handle - you have to actually build up the production, the industries are your pets!

Last word: You are doing really, really great, but huge is relative!

[planetmaker]

There's not much to add to what V453000 said.

I have experienced this problem both with and without cargodist. I'm attaching a ZIP with two savegames that illustrate my problem.

Just for your future reference: Zipping OpenTTD savegames is somewhat pointless as OpenTTD uses a more efficient compression algorithm for its savegames than zip Thus you could have made it easier for people by simply attaching both savegames separately but uncompressed. See https://wiki.openttd.org/Savegame_format (zlib is what zip uses)

[audigex]

(Note: I'm about a million revisions out of date in my knowledge, some/all of this may have changed)

This is partially a problem with Cargodist - it isn't quite intelligent enough to handle people who aren't playing in a "Keep adding more trains" way.

It's not necessarily a bug, it's working as intended, but the nature of cargodist means that passengers stay on your network for much longer, rather than the old method of just going to the next stop (or until you tell them to get off). Think about it this way: passengers used to just get on a train, then get off at the next stop. Then we got the ability to add "pick up only" and "transfer" orders, which allowed us to send passengers across the network. Cargodist is similar but it decides for us.

Unfortuantely, when it chooses a destination for your passengers, it doesn't really work out whether
1) The route has enough capacity for extra passengers or
2) The route already has a backlog

Off the top of my head, I believe it does roughly work out the route capacity, but allows a certain number of extra passengers for "grown" - this is ok, apart from the fact that if you don't add/upgrade vehicles, the capacity doesn't grow with the demand, and a huge backlog grows. That's fine for "super-profitable" networks, like OpenTTDCoop likes to do, but not so much for us "realistic" guys, who want to keep our village branch lines worked by a couple of 2 car DMU's, not 50x 15 car ICE3's.

Ideally, Cargodist would have a "don't overwhelm routes" option which, when selected, would stop routes getting a long term backlog (ie if they get a backlog, the pax numbers drop a little until it goes away): you know you can upgrade when the trains are full, not when the station has 2,000 passengers. In a super ideal world, there would be a network map or list of routes from the "Forward planning department", showing routes which are near capacity and can probably be expanded to be even more profitable.

I can't remember if it's in trunk, but the "reduce passenger numbers" patch used to help with this a lot, especially if you increase the value later in the game. It's a bit of a blanket solution, though, and makes setting up new routes/areas a bit annoying, as they have very low pax numbers if you've had to reduce the numbers for the rest of your network.

Overall then, it's not a bug: it's working properly in terms of making the game harder as you play, but there's a "nice to have" in there which is really a feature request.

[romazoon]

i haven t checked the savegame yet, but given the description i think you should try to use some city growth scripts, this will gives you much better control on the growth of each town/city (by actually transporting enough or not enough to that town to make it grow) and so it will allow you to plan forward or decide wich town will go big and needs big capacity of transport. this will stop the "ever" growing situation of towns.
exemple of script that works well for that : NAI, RCG, City Growth Controller

then for industries, i agree with other comments, something like FIRS might be interesting too for the same reason, you will be the one deciding if industries should produce a lot or remain a small or medium producer (via supplies).

Now on the note of Cargodist, i would recommand high settings of effect of distance on demands(personaly i use around 200% or more for large maps), this way most of "the people are trying to reach a nearby connected station" instead of "most people try to reach the other side of the network", this might create a need to increase capacity of "locals links", but it will make in return Intercity service "doable" as not half the town will try to reach that other city, but more something like a quarter or even less will try to reach distant station (it becomes more close to reality, usually most of transport are made "within" the same city (think metro,tramway and bus), and just a little % of transported people are going "farther" via Intercity trains or planes)

[andythenorth]

It seems like I must be missing some key piece of strategy or something.

I'm only repeating V45300 (but shorter), but anyway: OpenTTD isn't winnable. There is no winning strategy, so you're not missing anything

[romazoon]

i opened your cargodist game, i opened the Lafinghill City Train Station windows, and sorted plannified (not waiting) passenger per Origin-destination-via, now i click on the passengers originated from "this station" and see that not a single passenger produced at that station try to go to another (tram) station in Lafinghill City ! This Is the reason why your links have not enough capacity, people have too far destination to reach.

I setted the effect of distance on effect on demand to 200% and gave a few weeks for cargodist to update links, opened the very same line to check now where passenger are trying to go, and at first glance i can see that 1/3 of the passenger produced in that station are now trying to go to another Lafinghill station. --> this reduce a lot the need for huge capacity between cities.

[cosmicray]

Thanks for all the helpful replies and encouragement! I will definintely play with the cargodist effect of distance setting and see where that gets me. It could make for a completely different type of gameplay, which would be great.

Thanks also for the suggestions for NewGRFs and scripts. I actually have fun with city growth. I don't really play to "win", but I enjoy OpenTTD sort of like I enjoyed SimCity - an open-ended game I can return to have fun with whenever I want. I know there is no "win" condition, especially in single-player, but what's been happening is that after a certain point, the fun was drained because all I'm doing is trying to keep up with unhappy citizens.

Appreciate it all!

cosmicray

[leifbk]

As you say, the game is totally open-ended. FWIW, I'm usually just ignoring passenger transport altogether, set city growth to zero, and concentrate on cargo transport instead.

[Baldy's Boss]

As you say, the game is totally open-ended. FWIW, I'm usually just ignoring passenger transport altogether, set city growth to zero, and concentrate on cargo transport instead.

While I always find the largest city,build my headquarters,and start with a passenger train from there...

[pesgores]

I think the solution is pretty simple: make trains be able to transport more cargo. Especially passenger carriages, the cap of 40 per carriage is ridiculously low. It's the sole reason why one can't keep up with increasing passenger numbers. Were it at least 80 and I'd enjoy a more satisfying sense of completion and job well done... Seriously, it's that easy.

[Baldy's Boss]

I think the solution is pretty simple: make trains be able to transport more cargo. Especially passenger carriages, the cap of 40 per carriage is ridiculously low. It's the sole reason why one can't keep up with increasing passenger numbers. Were it at least 80 and I'd enjoy a more satisfying sense of completion and job well done... Seriously, it's that easy.

Is the 68-passenger bi-level car in NARS only usable with the engines mentioned as having "cab control" in its specs?

[pesgores]

Is the 68-passenger bi-level car in NARS only usable with the engines mentioned as having "cab control" in its specs?

I don't use GRFs but I went and tested that one out. The bi-level cart can only be attached to some types of locomotive, some are stated in the specs and some aren't, moreover, most modern engines can't use it.

But let's keep to base game and the logic behind the abstract number that is 40 passengers per carriage. In my country, there is a regular long distance train (Intercidades) whose carriages have seating capacity for 82 and they're reasonably comfortable. I'm not asking for suburban-like capacities of at least 150, like the train I ride every day, because I don't think the logistics of a commuter system are supposed to be represented in the core game. 80 seems about right and would do extremely well for players who struggle with constant and ever-growing passenger overflow, ie: everybody.

[Alberth]

For base game, restrict to 256x256 map size.

Edit: That's where the original trains were designed against.

[Eddi]

But let's keep to base game and the logic behind the abstract number that is 40 passengers per carriage. In my country, there is a regular long distance train (Intercidades) whose carriages have seating capacity for 82 and they're reasonably comfortable.

that's a logic flaw. the carriage you're talking about would also be about twice as long as the wagons in this game, so a train of same length would have the same capacity.

[andythenorth]

I don't use GRFs but I went and tested that one out.

Iron Horse newgrf offers 75 pax coaches from 1985 (and smaller ones - 40 and 55 - before then). Iron Horse also offers metro trains with up to 400 pax in a two-coach unit (metro needs the Termite track newgrf to be also in the game).

Not that I'm advertising or anything. "Other newgrfs are available" etc.

[pesgores]

that's a logic flaw. the carriage you're talking about would also be about twice as long as the wagons in this game, so a train of same length would have the same capacity.

I don't see why it has to be twice as long, it's a regular-size carriage. Below is the train I was talking about, which in game terms would be 5 cars long and capable of holding (as of now) 160 people or 320 (real-life equivalent and according to my proposal).

[Kevo00]

I've got to say, much as I love UKRS, it's insistence of a capacity of 35 per carraige even on modern DMUs and EMUs seems a little bizarre to me.

[Eddi]

I don't see why it has to be twice as long, it's a regular-size carriage.

well, yes. but real passenger cars are much longer than engines or freight wagons, while in this game they all have the same length.

[img]

you can't really use an image like this as an example, because the lenses used for this kind of photo usually really screw with depth perception, making things look much shorter than they actually are.

[PikkaBird]

I don't use GRFs but I went and tested that one out. The bi-level cart can only be attached to some types of locomotive, some are stated in the specs and some aren't, moreover, most modern engines can't use it.

You must have tested an old version then - the current version of NARS has no restrictions on what vehicles can be used together.

But let's keep to base game and the logic behind the abstract number that is 40 passengers per carriage.

It is what it is. When discussing OpenTTD mechanics we have to use the defaults as a baseline, rather than "realism", because at least everyone can agree on what the defaults are - and for historical reasons are going to stay. If you want a different gameplay experience, you will have to use NewGRFs.

Personally, I'd rather deal with this issue in NewGRFs by reducing the passenger production of towns, rather than trying to rebalance every vehicle type. But other NewGRF authors have other ideas - as Andy pointed out, there are train sets out there with much higher capacities than the defaults if that's what you want.