Back around when my family got its first PC in the 90s (a 166Mhz Pentium beast), I also bought what were my first three PC games: SimTower, SimIsland, and SimCity 2000, in a neat little multi-pack.
They were all developed by Maxis, the studio responsible for the first Sim game, the original SimCity, back in 1989. The story goes that developer Will Wright was working on a war game called Raid on Bungeling Bay when he realised that designing the level maps was the best bit; he co-founded Maxis, which would go on to develop games simulating everything from ants to the Earth itself.
These games would pick up on specific structures and systems – farms, evolution, skyscrapers – and build what we call “sandboxes”. Inside the game there weren’t any objectives, beyond what the player sets themselves; but the edges of the game, and the resolution of the simulation within it, were clearly defined. There was only ever so much reality per pixel.
In SimCity 2000, for example, the player can lower the crime rate by building more police stations, or by passing local government ordinances like setting up a Neighbourhood Watch or increasing the budget for anti-drug education schemes in school. They’re the crudest of crime prevention strategies, with unrealistically straightforward links between cause-and-effect, but they’re also the easiest to simulate. It wasn’t until SimCity 4 that other factors like the education system were taken into account, but even that meant an over-simplification for the sake of the sandbox. Lack of education isn’t the only reason people commit crime; and white collar crime, as in life, was overlooked in favour of muggings and burglary.
A couple of weeks ago, EA announced it would be closing the Maxis studio in Emeryville, California, which it’s owned since 1997, and was absorbing most of its staff. (Not that there was much left of the company from its heyday; Wright left in 2009, after the release of Spore, to run an entertainment think tank.) Games developer and theorist Ian Bogost has a thoughtful eulogy for Maxis over at the Atlantic, in which he argues that the “abstraction” of the Sim games successfully helped “wrest us from the temptation of personal identification”:
“In these games, players experience a model of some aspect of the world, in a role that forces them to see that model in a different light, and in a context that's bigger than their individual actions.
“The best games model the systems in our world – or the ones of imagination – by means of systems running in software. Just as photography offers a way of seeing aspects of the world we often look past, game design becomes an exercise in operating that world, of manipulating the weird mechanisms that turn its gears when we’re not looking. The amplifying effect of natural disaster and global unrest on oil futures. The relationship between serving size consistency and profitability in an ice cream parlor. The relative unlikelihood of global influenza pandemic absent a perfect storm of rapid, transcontinental transmission.”
The Maxis of the mid-90s, when I bought that triple-pack of games, had a very particular aesthetic that reflected this psuedo-educational side. There was a black coffee, short-sleeved shirt and tie tone, shared with reference apps like Encarta. And, as a geek, I loved, and pretended to understand, the short essays from each game’s lead designer, explaining which social theorist or scientist they’d been reading when drawing up their first drafts. James Lovelock even wrote the intro to the SimEarth manual.
Is that kind of thing necessary for a good city builder, though? I ask this because I’ve been plugged into Cities: Skylines for the last week, and it’s a game which a lot of people are saying is the game that SimCity (the 2013 version, not the 1989 one) should have been. And yeah, it’s excellent, in many ways. But it’s primarily excellent at being a Maxis-style city builder.
C:S comes from Colossal Order, a Finnish studio whose only two previous titles are Cities in Motion 1 & 2 – games which are very much spiritual successors to Chris Sawyer’s beloved 1994 classic Transport Tycoon. And it’s possible that C:S wouldn’t exist (or at least, wouldn’t have easily sold more than 500,000 copies in less than a week after release) if the last SimCity hadn’t been such a dud.
That edition was deliberately kneecapped before release by EA: among other issues, it was a single player game that required an always-on internet connection, and for too long after release there weren’t enough servers for everyone to actually connect to. And there’s been something of a pent-up demand for a “real” sequel to 2003’s excellent SimCity 4. C:S delivers that by borrowing some of the better ideas from SimCity (2013), and adding some of its own.
Each game begins the same way: a blank square of land, connected to the rest of the world via a highway offramp. You may be a mayor, but there are no elections, and your control over the planning system is near-total. Your job is to grow your settlement from a dirt road with farms and bungalows, up into a city of a million citizens, with an airport, mining and farming sectors, cargo port and subway system.
Or maybe not; it’s a sandbox, after all, and you might prefer something else. Just balance the budget and plan new suburbs and districts as and when your citizens demand them, and you’ll be fine.
For some, playing this kind of game is a meditative experience. Others will find it deathly dull.
Short of a complete disaster, it’s hard not to make money, and the game is relatively unforgiving if something goes wrong. And, as long as you’re comfortable working with road shapes other than grids, it’s surprisingly easy to create beautiful skylines – this is a good-looking game.
But perhaps the most major thing that C:S borrows from SimCity (2013) is agent-based modelling. The residents of City: Skylines aren’t abstracted into vector shading on a map overlay, as in earlier SimCity titles – they’re all individual characters, with their own daily routines, education levels, families, wants, desires.
Click on a car and you’ll see the name of the driver, where they live and where they work; you can see exactly which vehicles are responsible for carrying wood from a specific forest to a specific factory, and which stores in your commercial district receive van deliveries of goods from that same factory. Every dog has an owner, and every person waiting at a bus stop is waiting to complete their own unique journey.
This is a huge change to how city builders work. Traffic is a tangible thing in this game, and you will know when you have a traffic jam – there’s going to be a big line of cars and lorries and vans and scooters backed-up halfway across the map. Traffic flows like water, and dams in the same way. (Oh, and I haven’t even mentioned how fun it is to build hydroelectric dams and realise you’ve flooded your city’s downtown. Sometimes, with poop. The dynamic water physics are great.)
A significant part of the joy of the game is constructing ever-more-efficient traffic intersections. Traffic is the hypnotic core of this game.
It’s unfair to attribute C:S’s quality and success entirely to being a SimCity that doesn’t suck, of course. This game is published by Paradox Interactive, whose own games – like superb medieval dynastic struggle strategy game Crusader Kings 2 – not only allow modding, but openly invite it. C:S is no different.
This makes it easy to forgive the obvious problems with the basic game, and there are a few. If you build a one-way road, you have to demolish it and rebuild it if you want to change its direction (except there’s now a mod that fixes that); you have to manually demolish abandoned or burned-down buildings (except there’s now an auto-bulldoze mod that fixes that); the city architecture can look a little bland at first, but there are thousands more you can download from other players if you want to add some variety. Only a week after release, C:S is already a significantly improved game thanks to its fans. It’s certainly the best city simulation game currently available.
Consider its approach to corpses. Since the game simulates the lives of every individual citizen, it has to chart out their deaths too. The system for dealing with trash is the same as that for the dead, but instead of landfill sites and incinerators sending out garbage trucks, it’s cemeteries and crematoria with fleets of hearses. (One crucial distinction: incinerators generate power and pollution, crematoria don’t.)
Each hearse carries ten bodies; if they get stuck in traffic, the bodies pile up throughout the city. The deaths come in waves, too, that correlate with the waves of immigration that accompanied the construction of each new neighbourhood in earlier in-game years. Bulges in your population pyramid have knock-on effects on the city’s economy.
When a game is capable of this level of simulation, it’s hard not to notice the things that it purposefully doesn’t simulate with the same care. Crime, for example, is still as crudely-calculated as it’s been in every other game – as a combination of education, policing and general happiness, measured systemically. Even though every individual citizen has an education level, you can’t identify criminals from among the citizenry, just the houses that have been burgled by the abstraction. Each simulator’s sandbox has an edge that defines its ambition, and this is just one example.
While C:S might very well be the best Maxis-style city builder there is, it seems worth asking whether this is going to be the last, or one of the last, of its kind. The models that we use to understand things like society exist because trying to account for every single interaction between complicated things like, say, humans, is really hard. Better to invent a model by taking into account a few known things, and tweaking it to make it fit. Maybe it’s right, maybe it isn’t, but it’s an acceptable process considering our epistemological limitations.
Yet give everyone inside a video game a quantifiable life, and you invert the modelling process. You don’t need an equation to figure out where the traffic’s bad: you can already see everyone backed-up on the highway turnpike, because you’ve doubled your city’s office space without, for example, doubling the public transportation budget.
This is reflected in the assumptions that have always existed in SimCity titles, and by extension C:S. Despite being a Finnish production, it is a resolutely American manifestation of a city, where there’s always plenty of room to build. Government regulation is lax, and devolved to a surprising degree into the hands of you, the mayor. (You can even choose to legalise marijuana, if you want.)
The default tax rate is a laughably low nine per cent, and your city will empty if you approach even half of what would be necessary to fund a modern European welfare state. There is no history in these games: no street patterns laid down centuries ago that make public transportation and walking necessary defaults; no need to think about architectural or cultural heritage when demolishing something. Everyone earns the same wage, and if they don’t like it, tough shit. There aren’t any unions.
Does a city builder have to keep going along with those assumptions? Why always playing in the same sandbox? Maybe the true spirit of 90s Maxis isn’t found in perfecting this one archetype of a city, but instead to start thinking about more fundamental relationships between people within cities – whether it should be as possible to build a communist utopia as a libertarian dystopia with the same tools, for example. Maybe personal narratives aren’t the opposite of the simulation genre; maybe they’re the thing that will improve it.