Why do economic models in computer games suck?

May 31, 2007 3:46 AM

Why do the economic models in computer games always suck?

I want to see markets responding to supply and demand, shortages, depressions, slumps, black markets to start growing when taxation gets too high or corruption is endemic...

The only time things like depressions / slumps happen is by scripted events, not the natural outcome of a model. (eg. Civ IV Rhyse mod)

In so many games I've played, eventually you pass the point at which money is an issue, your bank balance gets to 9999, and money doesn't matter anymore.

Or everything is based on some kind of utopian socialist economy in which the little farmers and miners work for free, and all the money goes into the player's treasury. (games like Civ City and other building / resource games come to mind here)

Is it that an accurate model would take too much computer power? That designers feel it is an unnecessary distraction from the main game focus?

Any links to relevant articles / sites would be much appreciated, as well as your own thoughts / opinions as to why more interesting / deep economic models are non-existent in modern computer games. Or refutations / counterexamples, if you have any.

Thanks in advance for any insight you can share.

posted by Meatbomb to computers & internet (28 comments total)

I want to see markets responding to supply and demand, shortages, depressions, slumps, black markets to start growing when taxation gets too high or corruption is endemic...

Those are all the emergent phenomena of millions of individual decisions made in a marketplace. Yes, that is far too much for a computer game to simulate.
posted by hoverboards don't work on water at 4:10 AM on May 31, 2007

Eve Online is said to have the most realistic market simulation of any computer game. It has stocks, market booms and slumps, shortages and monopolies. The only thing I haven't seen so far is a recession.
posted by Coolcan2 at 4:16 AM on May 31, 2007

In addition to the emergent phenominon thing, it probably has to do with gameplay and some amount of required repeatability -- if the game just had a set of rules and a set of AI actors, there's no guarantee that it would ever reach certain milestones or plotpoints. You need some kind of triggered scripted actions in order to ensure that it follows a narrative of some kind, otherwise gameplay will be incoherent and spotty.
posted by Rhomboid at 4:23 AM on May 31, 2007

Yes, on second thoughts, an MMORPG might have a realistic market, because the millions of individual decisions that form an economy would be made by real players rather than the computer. But then this would be a real market, not a simulated one.
posted by hoverboards don't work on water at 4:30 AM on May 31, 2007 [1 favorite]

The technical reason economies don't work [millions of individuals is not required] is that NPC vendors have fixed prices,and the supply of loot outstrips demand by a large margin. Many games have no consumption at all, items don't decay, so over time the amount of resources in the world grows unbounded.

The philosophical reason the games work like this is that as a game its supposed to be fun to play, not work. If the economy were realistic casual gamers would all be peasants. Nobody wants to be a peasant: people want to be rich and famous and powerful heroes. In order to make it fun for everyone, including the stupid people, the game designers have to make the game easy to play.

Some games have tried to deal with this problem: Shadowbane for example had huge money sinks in armor wear and castle maintenance that kept the guilds fighting against each other wildy for dominance of the best loot spawns. More importantly, Shadowbane did something drastic on occasion: they reset the servers. If the balance got all wacked out, they'd start over or merge two servers to make it interesting again. However Shadowbane had a large number of mature hard-core gamers who didnt get upset when their characters were deleted, games like WoW have to cater to millions of noobs, so you get the wacked economy.
posted by Osmanthus at 4:39 AM on May 31, 2007

Yeah, MMORPGs like Second Life have very well-documented and functional economies that work because they're made of real people, not AI.

The reason no game has done this in AI is probably because not only would it be a pig to program and resource-intensive, but who would seriously be interested? I don't think anyone really wants to play the Game of Economics unless they're making real money from it, as they are in Second Life et al.
posted by reklaw at 4:39 AM on May 31, 2007

I think the latest SimCity was supposed to have a fairly decent economic model. Unfortunately, the problem is that it made it not so much fun to play.

Generally, a lot of things get abstracted in games, and generally, they are the parts which are not very fun. For example, in most computer war games (Hearts of Iron, etc.), you don't have to worry about supply, or supply is not made realistic.
posted by Comrade_robot at 4:46 AM on May 31, 2007

Those are all the emergent phenomena of millions of individual decisions made in a marketplace. Yes, that is far too much for a computer game to simulate.

When you say 'too much to simulate', you mean 'too much to replicate'. It can't be that hard to simulate economic phenomena — games do a good-enough job at simulating the weather, which is a chaotic system with a far higher number of interactions. You could simulate the stock market in the short run to a sufficiently accurate standard for a computer game with a random walk, which is trivial to implement. Sure, long-run stock markets are harder to accurately model, but some combination of a random walk with upward drift, a term which correlates with the business cycle, and the occasional 'Great Depression' thrown in to spice things up, should be 'real' enough to convince most people. As for real variables like output, why not construct a really rough estimate of the initial values of things like investment and consumption in the game, keep track of technological change over time (presumably this comes from the plot of the game, or the player's progress) and then use a couple of really simple models (say, the Life Cycle Hypothesis for consumption/investment and the intertemporal labour-leisure substitution effect for labour supply, neither of which can be that difficult to simulate) and hey presto! you have a decent model of the business cycle, at least according to RBC theorists. Your data doesn't have to be a perfectly-constructed aggregation of all the little characters' individual consumption and investment levels. Games work by making reasonable approximations: when they model any object, they don't consider every single atom in it, they make efficient approximations which are sufficiently accurate to allow the player to suspend disbelief. I think simulating an economy to a high enough standard for gamers to suspend disbelief is probably not that difficult.

I apologise in advance for the abysmal quality of the Wikipedia links: they should be seen as pointers to help seek out more accurate definitions, rather than accurate in themselves.
posted by Aloysius Bear at 5:45 AM on May 31, 2007 [1 favorite]

Seconding EVE Online. Its economy, spread out over dozens of "regions" is complex enough to interest financial wizards on both sides of the atlantic

You can even download xml history of any item and analyze it to your heart's content.

(it's also a brutal, near infinite, engrossing and wonderful game)
posted by Danaid at 5:47 AM on May 31, 2007

The other way to look at it is that in a game, you're almost always a member of the privileged elite. Depressions and recessions don't really affect you. For example, in Medieval Total War II, your cities get hit by the Black Death. For you, this means you have to wait a few turns to resume recruiting Teutonic Knights.
posted by Nahum Tate at 5:52 AM on May 31, 2007

Superpower II by DreamCatcher Games has a semblance of what you are talking about, though it is not terribly smooth in its execution and can be buggy.
posted by Pollomacho at 6:11 AM on May 31, 2007

Somebody, not me, could make the argument that, if you're looking for a game with a thorough economic models, you ought to go to a trading floor or a casino. And somebody already made the point that many multiplayer games have developed their own economies. (But has anybody mentioned the economic-sim genre? Oh, wait, somebody has.)

So, instead, I'm going to suggest that everything in videogames is oversimplified (both because they're simulations and because they're games), and that you notice it more with economics because it's an area of expertise or interest or whatnot for you.

While I've never coached a professional football team, I'm pretty confident that it's significantly more difficult than Madden (or Winning Eleven) makes it look. The same can be said for racecar driving, interstellar travel, military strategy, pet ownership and economics.
posted by box at 6:13 AM on May 31, 2007 [4 favorites]

I agree with box, and with your initial thought that maybe it's not of sufficient interest to warrant development. To extend and generalize box's point from particular game skills to indices, how many games that rely on a measure of health involve realistic assessments of disease incidence (e.g., your warrior is stricken with cancer, or has sepsis set in while in recovery), aging, and the like? Or have a realistic model of how experience is accrued and translated into skill? Maybe some do, but I'm guessing that most rely on completely inadequate models.
posted by Clyde Mnestra at 6:29 AM on May 31, 2007 [1 favorite]

I think that Aloysius Bear is probably right. Coming up with some stochastic models for the economic trends is probably not too hard to do in a convincing way. Economists have various models that, while data intensive, are not so awful that they couldn't be done in simplified form on an average home computer.

My guess is that the hard part is giving the player the feedback and control to influence the economy in ways that are also responsive enough to be fun. I think there's a reason that we've never seen SimFederal Reserve, for instance, and that's because raising the interest rate but a point and then waiting several months to see what happens doesn't sound terribly entertaining. That said, I bet Will Wright could find a way.
posted by Schismatic at 6:38 AM on May 31, 2007

Games like Civ probably should have some sort of economic model IF and only IF the designers are looking to make you pay attention to your economics on that sort of level. Most games place attention of building, growth, combat and the like and just don't add economics because it would be one more thing to complicate the game. I've work on some small scale game projects, it would not be difficult to put in a realistic (or at least dynamic) economics model. But really how many games can you name that have free markets?
posted by magikker at 6:48 AM on May 31, 2007

M.U.L.E. is a sophisticated economic sim.
posted by johngoren at 7:08 AM on May 31, 2007

Master of Orion III when out of its way to be realistic simulation of interplanetary government and economics. It was also frighteningly dull.

The goal is to have a fun game, not a realistic one.
posted by JDHarper at 7:31 AM on May 31, 2007

Star Wars Galaxies had an engaging economic model, so far as they even had a main profession of Merchant. Resources had stats that varied randomly, the resources spawned randomly across the galaxy every two weeks. You had to survey for the resources (time consuming) employ the services of the architect profession to manufacture harvesters for you to to gather the resources (money consuming), you had to tend the harvesters and pay maintenance on them, and then when the resource went away after two weeks, you started the cycle over again.

Each of the stats on the resources affected the quality of goods that were produced with those resources, and there were NO npc vendors in the game and drops sucked, so that meant that if you were a weaponsmith, and you devoted a lot of time and energy into acquiring a stockpile of some excellent materials your weapons were the best in the galaxy, and more than likely, there was no one else that used the same resources as you, therefor your weapons were about as unique as they possibly good be.

This applied to all of the crafting professions, so you had people that were expert weaponsmiths, expert armorcrafters, expert medical supply manufacturers, etc. A new subgroup also sprang up. You had speculative merchants who traded their time by going to each of these vendors and buying the best of the best, and setting up department stores and malls where people could come to one location and buy the best things made, without having to spend all their time travelling. There was even a guild on our server named UPS, who wore brown uniforms and specialized in delivering goods from shops to players in game for a small fee.

To the heart of your question about why more interesting / deep economic models are non-existent in modern computer games, I think the answer is pretty simple. Economics isn't normally 'fun'. I'm not a game designer, but I'm betting that it's a very dificult task to take the somewhat frustrating aspects of depressions, resource scarcity, market disparities, and make them understandable or palatable.

You had to survey for the resources (time consuming) employ the services of the architect profession to manufacture harvesters for you to to gather the resources (money consuming), you had to tend the harvesters and pay maintenance on them, and then when the resource went away after two weeks, you started the cycle over again.

Each of the stats on the resources affected the quality of goods that were produced with those resources, and there were NO npc vendors in the game and drops sucked, so that meant that if you were a weaponsmith, and you devoted a lot of time and energy into acquiring a stockpile of some excellent materials your weapons were the best in the galaxy, and more than likely, there was no one else that used the same resources as you, therefor your weapons were about as unique as they possibly good be.

This applied to all of the crafting professions, so you had people that were expert weaponsmiths, expert armorcrafters, expert medical supply manufacturers, etc. A new subgroup also sprang up. You had speculative merchants who traded their time by going to each of these vendors and buying the best of the best, and setting up department stores and malls where people could come to one location and buy the best things made, without having to spend all their time travelling. There was even a guild on our server named UPS, who wore brown uniforms and specialized in delivering goods from shops to players in game for a small fee.

To the heart of your question about why more interesting / deep economic models are non-existent in modern computer games, in my opinion is because there isn't a market for it. Not saying it couldn't be incorporated into a game, but it will probably always be a bolt-on afterthought for most games, especially non-MMORPGs.
posted by voxpop at 7:46 AM on May 31, 2007 [1 favorite]

I suppose this would be possible in a non-MMORPG game like Spore where the actions of individual users playing their individual games feed into a centralized universe-database.
posted by Afroblanco at 8:19 AM on May 31, 2007

As mentioned before, Sim City 4 does have some sort of scary economic underbelly that typically puts your towns in a depression/recession at one or two points in their life cycle, the interesting thing about SC 4 is that you dont really affect it with your buying and trading, like the other games people mention, but with taxes and development choices.

I seem to remember Railroad Tycoon 2 having an interesting but simple supply/demand stock market type economy that actually reminds me of a more complicated version of the classic MULE economic sim.

There are some space-trading games, like the classic starflight all the way to more recent games like Freelancer and Space Rangers 2 that have very simple economic systems, but probably fall squarely into your 'sucky' category of economic models. I know I'd like to hear about one of these games that have good economic models that aren't as dull/overbearing as the recent Master of Orion.
posted by yeahyeahyeahwhoo at 8:27 AM on May 31, 2007

Becaus games are forms of entertainment. You might as well ask "Why do movies show unrealistic scenes?" or "Why do computers in movies do all these cool things I cant do with my PC?" "Or why are there loud explosions in space?"

A good game has gameplay as a priority. Adding in an unpredictable and costly (in terms of development) economic engine simply isnt worth it. It also can ruin a perfectly good game. Why wait many hours for the economy to pick up to buy something you need to get to the next level?

Non-games like Second Life have a hands-off economic engine, but SL is not a game, its a type of simulator.
posted by damn dirty ape at 8:38 AM on May 31, 2007

Its also worth mentioning EVE online, but just like SL, its very much an aquired taste. Those players really dont have a problem spending hours generating credits. Most gamers find this tedious and probably prefer some kind of top-down control to make sure that they can afford to buy the sword of destruction that lets them kill the fudmonster in level 18.

Arguably, when a game is a "profession simulator" it stops being a game and begins being a game-like thing called a simulator.
posted by damn dirty ape at 8:42 AM on May 31, 2007

I'll also defend WoW's economics as a real in-game market. There are no black markets but the auction house and selling individually to players shows real life economic dynamics.

I personally took advantage of the fact that not many players played early in the morning and priced items to sell at a high premium knowing that the lack of supply at 4am worked in my favor. I also targeted a wealthy profession (enchanters) by mining dust, thus knowing that if anyone out there is willing to pay double to get something they need right now at 4am its a wealthy class.

I wouldnt call the WoW system sucky at all. Its a real working economy that even young kids can figure out without dumbing down the system.

Personally, I hated all this selling and buying. Its such a serious waste of time. I'm not alone in feeling this way. IMHO a good game treats economics like it treats going to the bathroom - not at all. For those who like this kind of thing there is afore mentioned EVE and SL.
posted by damn dirty ape at 8:52 AM on May 31, 2007

You might find trading sims like Port Royale 2 or Patrician III to your taste.

You buy and sell goods and move them around from places where they're produced to places where they're consumed. Basic resources are available in different quantities in different places. Supply and demand has strong effects on prices. Towns grow or stagnate based on the materials available to them and how well trade satisfies their needs.

I don't recall any black market aspect, but I think they address the rest of your list. (It's been years since I played either game.)

As to why these games aren't very common or popular, I think it's that they're repetitive and require a lot of careful play. I find it more fun to play something like Sid Meier's Pirates, where the focus is on adventure, rather than economics.
posted by aneel at 9:41 AM on May 31, 2007

There's not any meaningful distinction between a real market and a "simulated" market. Both are subject to the same tools. You could very well see meaningful emergent phenomena in a virtual market but a computer game wouldn't be the right venue for it since markets are unpredictable, uncontrollable, and difficult to understand. Your average gamer who's more interested in seeing what happens next wouldn't appreciate having to stay up all night and acquire an advanced degree just to improve his game state. Simulating markets tends to be a question of variables and probabilities combined with social dynamics but note that such simulations rarely tell you all that much about the real world.
posted by nixerman at 9:50 AM on May 31, 2007

Truthfully, all economic models suck.
posted by fusinski at 1:08 PM on May 31, 2007

I played a stock market simulation game back in the eighties, but there aren't a huge amount of games that focus on simulating these kinds of things when you can abstract an approximation of them. I've been very interested in economic modeling in games, especially MMORPGs for a while, and also in mathematical models for infinite space ever since "Elite Plus" came out.

The main problems with economies in MMORPGs and MMO games in general have been described above, but the main one's I've seen in the ones I've played are a lack of entropy, and the unlikeliness of spending money on anything not related to power leveling. Some games offer items for purchase in game that are purely character developing rather than improving the power of a character. But there is usually no food tax, rent, etc... so all income is disposable. There are also no investment opportunities in most games.

But the biggest issues I see are that self-correcting features of actual market economies such as price inflation or devaluation of over-abundant physical items will have a greater effect on the players who log on the least frequently. It's therefore very difficult to model an economy accurately without having unwanted side effects crop up that make your game unpleasant for the non-hardcore gamer. This, and the fixed monthly subscription cost as opposed to a time-played model are what tend to drive me away from most of the commercial games.

While I agree with Aloysius Bear in general that simulating economies is possible, I wonder how easy it is to throw in bubbles and depressions and also account for fabulously wealthy players deliberate attempts to perturb economies. My ideal game (as a player) would allow a wealthy micro-managing player to attempt things like George Soros' currency speculation in 1992, or allow a player to not get involved in the nitty-gritty of the market.

The guild system in games seems to correct for a lot of the economic deficiencies, but more often with communist or marxist sensibilities than capitalist ones. I guess no one wants to track loans.
posted by BrotherCaine at 7:20 PM on June 25, 2007

Regarding my above comment, maybe automatic cost of living adjustments could be made to a player's gold balance while they are logged off?
posted by BrotherCaine at 7:22 PM on June 25, 2007