Tuesday, September 14, 2010

Economies and World Building

I just got into a spirited discussion about economies in world-building (either for a game or just as a setting). The discussion revolved around what factors are important - food, energy, building materials, etc, etc. I quickly derailed it into a pages-long argument for transportation being the only thing that really matters.

This post is pretty much a technical description of why I say that, and what my experiences have been. If you're not into world-building, skip this.

When I'm world building, regardless of technological level, the first thing I think of is transport. How do goods and information get from one place to another?

Some of the people arguing against my position take the position that, at least in high-tech environments, transportation is cheap enough to disregard. However, regardless of price, transportation is what enables the growth of economies. All economic centers are also transportation hubs. This is a symbiotic rather than a causal relationship: a stronger economy promotes more import and export, and more import and export promotes a stronger economy. This is not simply a function of infrastructure, either: it's a fundamental matter of how humans think and plan.

Having built many worlds with this philosophy, I can say that it allows you to quickly and fluidly generate a dense and real-seeming set of nations and cities, regardless of technological level or what economic rules you have underlying the simulation. This is because almost everything about a city and nation is heavily influenced by how many transport routes they have running to/through them, and how open/closed/taxed those routes are.

Things like war, disease, and the spread of culture all happen along transport routes, all of which are marked by a size denoting their max load. Size is not length: it's the amount that can be transported in a time unit, which includes not just the actual road/river/spacelane in question, but also the docks on the tips. Safety can be assumed to be a function of size. Things like strife and war can also be fairly reasonably simulated using nothing but transport lines to determine how nations feel about each other, as long as those transport lines include a rating not just of size, but also of taxation/closedness.

Example: it's ancient times, farming's just been invented, and we use a random seeding system to create tiny little villages. After that, everything is determined by the trade routes. Rivers and coasts are easy to transport along since you only have to build some docks, and roads can be built if you take an economic hit for the initial investment. Not a certain amount of money, but a decrease in your "economic rank".

Speaking from experience trying to simulate a China-like environment, what we find is pretty realistic. Cities rise up on rivers and deltas and coastlines. A network of roads springs up as well, but inland cities never get very far. Not because they lack resources, but because cities on rivers and oceans have a strong head-start by then and dominate the route-building game.

In late game, cities will pop up on land routes simply because the land route is there. This, too, is realistic. Add in a random assortment of disasters that can damage trade routes, and you get a pretty reasonable feeling of organic complexity.

Adding some political acumen to the game works flawlessly, and you can see the borders form and wars start... based almost entirely on the idea of jockeying for transport hubs and routes. A second layer of restricting transport along certain routes or to certain countries further extends this political spectrum and, at this point, it becomes close to realistic. Adding in a dash of mutating culture and you have something that can pass muster with ease.

Of course, this can all be done manually, as well. I find most people manually building worlds end up thinking in terms of transportation, anyway.

A high-tech example: you have FTL drive and are all about building colonies on planets. You rank various planets with a basic "economic value". It's not even important whether the value is food-based, ore-based, or scientific-oddity-based.

The cost transporting world-to-world within a star system is low and flat, the cost for transport between star systems is per light year. All transport must be along defined "routes", but defining a route is cheap, consisting of building a landing dock.

With these simple rules and a completely random star/planet layout, you will quickly find a dense web of trade routes that ends up almost identical to the one in our primitive version. The biggest difference is that we're now working in 3D rather than 2D. The difference is subtle at the moment, but if we add in the "upper layers" of the simulation to get political, the 3D world makes for some very odd political maneuvers.

If we add in wormholes for very cheap long-range transport, wormhole star systems become the most valuable, and almost instantly become massive centers of commerce: just like you'd expect, when you "cheat", you win.

Let's turn around and say that you don't need to define routes: any route is fine. This is actually unrealistic, since established trade routes are more than just a landing bay at each end. An established route is important for insuring company longevity, so companies would naturally start to form specific trade routes even in the absence of any physical need to. It's the nature of trade.

But ignoring that, even when treating every shipment with its own independent path-finding, we still end up with trade routes as they follow the same best path every time. The differences are actually pretty minor, although they grow more extreme as the simulation continues if you continue to disregard things like policing routes, taxing/banning trade, and so on.

Another factor that can be fun to include would be information transport routes. This is especially fun if the information transport system is fundamentally different from the physical transport system, ending up with a complicated "two-map" set of trade routes. Either way, it's a lot of fun if we presume that economies are two-stage: initially, goods are what matters, and the physical trade routes are stressed. As the economy grows past a certain point, goods plateau or even begin to decrease in importance and information creation/transport rapidly grows more and more important (exponentially). Info transport and creation can also be the primary source of technological and cultural innovation.

ANYWAY, this post is all about my pet peeve, so it was a bit dull. The basic lesson is: if you ignore transport in your world-building, you are ignoring a major source of depth.

What's your experience?


Darius Kazemi said...

For a great sci fi novel that is 100% worldbuilding around a novel mode of transportation (and indeed describes how a sudden change in transport mode completely changed a society), check out The Stars My Destination by Alfred Bester.

Craig Perko said...

I agree. Actually, most scifi has strong tints of this, including such unlikely candidates as the Honor Harrington series and Star Wars novels.

Claes Mogren said...

I'm interested to see how Civ 5 will play out now when roads cost upkeep and are generated automatically, not something you build. In earlier versions you could just cover everything with roads/railroads.

Anyway, thanks for an interesting post, and I'm curious about that world generator. Any chance of a post about how it is built/could be built?

Craig Perko said...

For me, the problem with Civ games is that you play a single nation from start to finish. That means you have to end up in a world where you civilization never loses, never gets hit with any serious misfortune.

Given that most of the world was shaped by misfortune, it seems like a bad thing to make a failure condition. :P

You know, I meant to post videos of the China Sim, and I never got around to it.

The sim itself is just a proof of concept, so it's not very playable and has some seriously misimplemented features. So it would be best to post a video, rather than an app.

I'll post on the blog if/when the video goes up.

Craig Perko said...

Addendum: the post will contain a lot of technical details about how to actually implement this kind of generation. But the truth is that it's REALLY straight forward. Maybe it's just because I've made so many of them, but these are not complicated rules.

Bill said...

Nice articles. I'd like to see a video or something of your China Sim if its an implementation of some of these ideas.

Matthew Rundle said...

So, what I wonder is, is this specifically useful for generating economies? To what extent does transport get you stuff like local colour, and quests?

Craig Perko said...

It is useful for describing traffic patterns, which can be used to extrapolate cultural drift and so on. However, for actually creating cultures and quests, it's useful only as a guideline. As far as I can tell.

Fundamentally, it's just about changing the way the 2D play happens. Normally it happens in a "capture useful locations" method, where you build cities, mines, and soldiers to take advantage of good resources and bottlenecks. Using a transport system takes the focus off capturing locations, replacing it with a map that doesn't think in terms of locations, but in terms of threads.

Andrew said...

The problem with your argument is summarized bu one statement, "Having built many worlds with this philosophy."
The problem is that you or i have never built a world. It is like you saying, because i am good at FPS, i should be a good SEAL.
I will agree that transport is important part of business but one as a subsection of supply chain management. Supply chain has more to do with inventory levels, than how you move your product.

Craig Perko said...

(This is dragged in from the pre-post conversation.)

Again, you're coming at this from the point of view of your major, rather than from the point of view of what we're actually talking about. It has NOTHING TO DO WITH SUPPLY. Or money. Or the real world. Or anything like that.

Having built many worlds with it, I can say that I have built many worlds with it. It's like saying I'm good at FPS, so I'm good at FPS. I'm not trying to start a company with these philosophies, I'm trying to build a game world.

This algorithm has nothing to do with supply chains, nothing to do with inventory management, nothing to do with worlds that rely on those things instead of these things.

If we want to talk about how to actually simulate a real-world or nearly realistic economy, then we're on another topic entirely. All I'm after is how to build a worldwide set of economies that are convincing enough to make a game out of and provide solid hooks to build cultures and wars and plots and quests.

And this does the job far better and easier than trying to muck about with stockpiles.

Andrew said...

Before commenting on supply chain management you should look up what it is about.
If you want to talk about game worlds only that is fine, but you are not. You are mixing real world facts with game logic.
If you really meant "worldwide set of economies that are convincing enough to make a game out," then you should take a good hard look at supply chain management before making you next world. People having been working on this idea of supply chain for longer than the two of us have been alive.

Craig Perko said...

The same is true of trade routes, obviously.

Here's the problem: managing supply chains is detail work. When you're talking about a world, or even a nation, there are literally billions of supply chains involved. Sure, you can focus on only the "important" supply chains, such as government stockpile of grains, but then you end up with a spreadsheet world that revolves around only those things.

The in-world population can manage supply chains. If a supply chain breaks, whoever is in-world responsible - not a player - figures out a new supply chain. Similarly, Steve Jobs doesn't go out and hire a new guy when an Apple store worker quits: there's guys in charge of that.

In fact, you can abstract nearly all of the chain management at this level out of the game entirely, instead of simulating it. You are still left with a bit of it, just enough to add depth without being enough to turn your game into a spreadsheet.

Now, you can make a good game based on managing supply chains. In fact, that's what most constructive games are built around. And, as should be painfully obvious given that you've played at least one of them, I have made games in the past that did heavily incorporate supply chains.

With that said, I find trade routes to provide more interesting play with less fiddly micromanagement, and they also provide strong in-world hooks for culture and war. Supply chain games always feel kind of forced and spreadsheety to me: it takes the game out of 2D (or 3D) space and sticks it in a spreadsheet.

Think about train games, which are all about supply chains. The rail tracks exist specifically to move the game from a world of cities and hills and rivers to a world of arrival times, max loads, and regular shipments of ore. A spreadsheet.