Hacker News Reader

It's funny you mention wanting to solve transportation issues, because that's exactly the reason why I bounced so hard off Cities Skylines. At a certain point, it just turns into a "play with the roads" game. No matter how many forms of transportation you utilized, it just turned into bottleneck fix after bottleneck fix. When you tried to implement serious logical solutions (ie dedicated bus lines), you were met with more chaos. But add several roundabouts in succession and suddenly traffic fixed.

I do agree with the depth, something Cities Skylines was really lacking. Most of the logic of the game was quite shallow, and the people really don't live anywhere near as "deep" lives as in SimCity (2013). I'm hoping this is something they address in the sequel, but seeing as the community has rallied so strongly around the idea of creative designing vs functional planning, I'm not holding my breath.

I really agree with this. I've gotten into Cities: Skylines numerous times over the years and have bounced off of it each time because it was so hard to scale a city past a certain point due to transport.

I'm up for a challenge and I think transport is actually a fantastic gameplay element for the player to scale, but as it is the only way to "solve" transport feels like cheesing whatever algorithm is under the hood, rather than designing a transport system that makes any real kind of sense.

I have some beef with agent-based simulations in this regard. They're fascinating in the abstract, but IMO too many game devs are obsessed with it to their own detriment. Heavily agent-based sims tend to scale really poorly (an issue I suspect is true with SimCity 2013 and Cities: Skylines) and imperfections in modeling the agents causes extreme divergences at-scale that break the illusion of realism badly.

At this point I'd honestly much rather - especially with large-scale sims where you build cities of millions rather than a town of dozens - have non-agent-based statistical simulations. I want to build cities that I can imagine seeing IRL, not some metagame-game-mechanioc-defeating monstrosity that's optimized against the game's pathfinding.

I've played with Cities: Skylines off and on many times over the years and often do give up in frustration with endless traffic problems, but I think it's down to two main things: one, traffic management is very much "endgame" type content, as you can build a nice self-sufficient little zen garden community without really worrying about it much. But to get big cities it must be dealt with. And two, once you've already built a city, it's way harder to implement good traffic flow. This is painfully evident all the time in the real world.

Now I use a mod to unlock all the map tiles (actually, the entire map) at game start, and I build specifically around my envisioned future transportation needs. It's still a challenge, and I highly recommend the traffic management mod mentioned elsewhere in these threads, but it is not an insurmountable one.

There's really two types of "City Games" - city painters and traffic simulators, Cities Skylines is basically the second with a bit of the first, whereas SimCity 4 became the first with a bit of the second.

There is also the puzzle kind like Urbek, which has a nice SimCity feel but you need to make sure that building type X is (or is not) within Y squares over several levels of development. And more along that line there are a bunch like Farlanders that aren't really builders at all but pure puzzle games that happen to use buildings as graphics.

https://www.gog.com/en/game/urbek_city_builder_prologue

https://www.gog.com/en/game/farlanders_prologue

The Pharaoh/Caesar games also had an interesting take on the city builder but it’s a bit orthogonal.

its been years since I played but one of my favorite mods for skylines is "traffic manager: president edition" which let you do fine grained control for lanes and various traffic situations. I know its more 'fiddling' with traffic but it let me solve bottlenecks and issues that just wasn't possible in the base game.

You might enjoy Workers & Resources: Soviet Republic.

It's really rough right now, and frustratingly manual in places, but it does a good job of simulating resources at least. If you play in Realism Mode then vehicles and resources are dispatched and consumed realistically. (People too, but they teleport home at the end of their job and they don't appear to take transfers for public transit)

Workers & Resources: Soviet Republic: https://store.steampowered.com/app/784150/Workers__Resources...

OpenTTD (Transport Tycoon Deluxe being the original), is my preferred traffic simulation game.

Once you can control all routes and trains, you can never go back! You get used to the control, the timetables, the organization that's possible.

OpenTTD sadly does not simulate passengers as unique agents. Transportation demand can be very unidirectional and not make logical sense.

Have you not tried the CargoDist button?

It was added maybe 5 years ago, but as an Open source project, advertising for this feature is pretty bad.

It turns passengers/mail into agents who will perform 2 or 3 or more hops to reach their destination.

Seconding this, if you dropped OpenTTD because of the simplistic passenger/cargo routing (i.e. take any cargo to any place and pay is only based on how far you took it) then definitely give the CargoDist option a try.

Every passenger (or mail or manufactured good) has a unique destination in your system that you need to get it to. They are roughly balanced based on demand (two stations in large cities will generate a lot of traffic between them, but a small town will only generate a small amount of traffic to and from a large city). It makes the transport routing game way more interesting and more like the routing of passengers in more traditional Sim City type games.

Beyond just that, the ability for passengers to "route" themselves (Take airplane from CityA -> CityB, then rail from CityB -> CityC, then bus from CityC to CityD... and back) is the bigger, more important effect of the CargoDist feature.

That means you can form "Hub and Spokes" for passenger traffic, and the individual passengers are smart enough to figure out how to route themselves to their ultimate destinations.

------------

Its the most "realistic feeling" traffic mod I've found of any of these games, honestly. Even Cities Skylines seems rather basic in comparison.

-----------

I play on Passengers/Mail as CargoDist, and then all other traffic (Coal, Wood, Cargo, Oil, etc. etc.) as default / original rules.

I also set the "symmetry" of CargoDist to 80%. That is, 20% of traffic is one-way, which seems to be correct in my opinion? I think it defaults to 100%, but it doesn't make sense for _ALL_ traffic to be bidirectional (ie: for all agents to make a return trip each time).

EDIT: I do think that a "big mod" to OpenTTD under a new name would be beneficial to the project. Something that redefines OpenTTD's sizes (each square is supposed to be a square km, but it doesn't "feel" like it. Roads aren't a square km wide after all...). I think that if each square were 10m x 10m nominal, and then for all other parts of the game to be redesigned off of that (scaling up docks, rail depos, cities, etc. etc. as appropriate to this new size), would be really all that OpenTTD needs. Plus CargoDist Passengers/Mail by default.

OpenTTD sadly does not simulate passengers as unique agents. Transportation demand can be very unidirectional and not make logical sense.

Personally this isn't really a negative. I've found that I only care about "agents" when the scale of the game focuses on caring about individual units. Once you get beyond that, I don't think the simulation necessarily is more interesting just because it uses unique agents. And conversely, I've never thought while playing SC4 or similar game that "This game would be better if it simulated all citizens as unique units".

A number of games simulates unique agents: Tropico, Rollercoaster Tycoon, Theme Hospital, Two Point Hospital, Ceasar 3 / Cleopatra etc. etc.

Its a different game entirely compared to games like Sim City or City Skylines. I've noticed that these "individual agents" games are very "unstable". They're only designed for a limited number of agents before everything goes to crap.

As such, "individual agent" games tend to be scenarios that you play for a short time, rather than trying to play for months on the same map (like Sim City or City Skylines or OpenTTD can be).

---------

Tropico Boom/Busts grow too variable and uncontrollable. You'll eventually have an island-wide boom/bust collapse that probably triggers the instant-loss endgame (aka: Soviets and/or USA invades).

Ceasar 3 / Cleopatra are reasonably stable, but definitely a smaller map and designed for scenario play instead.

Openttd is a great logistics simulator, but it's not really a traffic simulator. Traffic by it's nature requires a lot of independent agents deciding how to use a shared resource, and the resulting conflicts, like Braess' paradox[0]. As far as I've seen, the only game that does well at this is Cities Skylines. OpenTTD on the other hand is my favorite part of factorio.

[0]: https://en.wikipedia.org/wiki/Braess%27s_paradox?wprov=sfla1

I've personally found it pretty easy to get around. In particular, just build long stretches of roads without any intersections. Just "s-curve" like a snake, especially around industrial centers (all traffic originates in industry, and don't seem to "return" to industry in Cities Skylines).

Maximize the distance of these singular roads, and build out your road system like a computer-science definition of a tree. Its not very realistic, but it solves more or less all the issues of Cities Skylines. (Or perhaps it is realistic, and matches the behavior of neighborhood planners in the real world).

In contrast, OpenTTD encourages hub/spoke model and recycling of centers. I think its sizes are off but the overall gameplay loop is correct here.

--------

OpenTTD's "traffic" is pretty dumb, but its infinite. As long as you've got high station ratings, the station will expand and traffic will increase. It doesn't matter what design you have, the traffic increases until you can no longer support it.

So OpenTTD is somewhat self balancing. The stronger a player you are, the more traffic you generate. The weaker a player you are, the less traffic you generate.

> all traffic originates in industry, and don't seem to "return" to industry in Cities Skylines

It absolutely does come back? I've had major traffic headaches from returning trucks. If I don't build rail for trucks, my cities are usually screwed over by that.

> Maximize the distance of these singular roads, and build out your road system like a computer-science definition of a tree. Its not very realistic, but it solves more or less all the issues of Cities Skylines. (Or perhaps it is realistic, and matches the behavior of neighborhood planners in the real world).

Trees create massive chokepoints when traveling within the city, so I don't think its likely to work in a large city (though if you squint, road hierarchy is somewhat tree-like, I suppose, though with more intra-branch connections). I suspect you simply haven't gotten cities big enough to actually have traffic, especially if you think most traffic is trucks. It does take a bit for personal car traffic to be heavy enough to be a problem.

The unique thing about car traffic vs station demand is that the overall speed depends on the quantity of cars on the road, with maximal throughput occurring before a fully occupied road. That's how you get weird effects like Braess paradox.

OpenTTD is much better at simulating public transit systems, Cities Skylines is better at simulating traffic.

> It absolutely does come back? I've had major traffic headaches from returning trucks. If I don't build rail for trucks, my cities are usually screwed over by that.

Don't the trucks go into the Cargo station, and then disappear?

In OpenTTD, if you have a truck -> rail station, the truck drops items off, then have to return to Coal/Forests/whatever to pick up items again.

In Cities: Skylines, the truck hops onto the train, and then vanishes. Its been a while since I've played Skylines, maybe I'm mis-remembering?

> Trees create massive chokepoints when traveling within the city, so I don't think its likely to work in a large city (though if you squint, road hierarchy is somewhat tree-like, I suppose, though with more intra-branch connections). I suspect you simply haven't gotten cities big enough to actually have traffic, especially if you think most traffic is trucks. It does take a bit for personal car traffic to be heavy enough to be a problem.

Or maybe I'm just abusing the behavior of cargo stations too much.

The cargo-station "teleports" traffic somewhere else, and seems to magically erase return traffic. Use / abuse of cargo-stations seems key in making cities grow larger.

----------

EDIT: Ah, I misremembered. Seems like the return traffic does need to be accounted for. In any case, the industry -> Cargo Train terminal is really all you need to design against. With a proper industry block (or S-road / snake-like road), all industry goes to cargo train, and all of them return to industry, with no intersections potentially wrecking them.

> (though if you squint, road hierarchy is somewhat tree-like, I suppose, though with more intra-branch connections)

Intra-branch connections are your mistake. Don't do that. If you're purely a tree-shape, there's no Braess paradox because each neighborhood (or housing unit) only has _one_ path to the rest of the city.

The minute you have 2, 3, or 4 different paths to reach the same house, is the minute that paradox starts to bite you in the ass. So just avoid it completely with a 100% tree-like branching behavior. One path means you have absolute control and force the little cars to go down the exact paths you want.

-------------

Yes, roads and intersections will saturate. At that point, stop building in that neighborhood and build another branch of your tree.

I don't think I ever found a reason to do a non-tree design in the game, aside from simulating realism. But if you're just designing neighborhoods to minimize traffic and maximize population, a pure tree is the obvious approach.

IIRC, there's no residential-to-residential traffic. Its all residential-to-commercial, and then commercial-to-residential. That's why a pure tree works. Or industry-to-commercial, and then commercial-to-industry.

> Cities Skylines was great, but I am looking for something a bit more in depth.

Incidentally, Cities Skylines 2 was announced just earlier today. There aren't many details yet, but perhaps that will scratch your itch.

That is exciting. Weird that the trailer is so light on detail or gameplay, but it says coming this year, so more details will come soon I imagine.

SC4 + NAM is what you are looking for https://www.sc4nam.com/

It has been in development for 20 years.

> the transportation flexibility of Cities Skylines

It's worth mentioning that Colossal Order (the studio behind C:S) cut their teeth on transit simulators, namely the Cities in Motion series.

Did you play Sim City 3000?

Did you play Sim City 2000?