AGC: MMO economies
Rough notes, sorry for typos!
Sam Lewis, MMOG economies. Lead designer for Cartoon Network MMO. ex-Kesmai, SOE, MBA and BA in econ as well.
What is economics? Study of forces of supply and demand and how they allocate scarce resources in a society.
MMOs are social games, that’s why you use economics. It helps analyze current phenomena and argue with marketing. And it helps you make design decisions, worldbuilding decisions. it isn’t the only lens to look at things, but it’s a useful one.
[shows a graph of the prices of components to build a spaceship in Eve]. Nice and flat, then suddenly there’s a demand spike and prices go through the roof. What do you think happened? Mor e players got qualified to fly that ship, so the demand curve went thru the roof.
Natural market places: pic of Seal Online. Lots of chat bubbles of people saying “sell sell sell.” How many of you played EQ? You all went to sell things in the tunnel. You think anyone designed it that way? No, it was a natural phenomenon. Econ can explain why. Eve Online, there’s a natural place. (Many games get listed with examples of this). If you know why, you can build to encourage it and allow it to happen.
Final reason to use econ: “You can’t screw it up any worse than it already is.”
Money Supply
The classic problem is that the money supply gets out of hand. Asheron’s Call: dupes and other problems caused people to move to alternate currencies (though it is back to currency-based now). You do need to have sufficient cash for lubrication, but you want to avoid a monetary collapse. You do not want your fiat currency to become worthless.
You also have to avoid the new players being “poor” when they first log in.
Monetary collapse could actually be considered fun. OK, but you should be ready to set up a commodity currency. In AC it was keys and shards. It needs to be
– divisible (something you can make change)
– uniform (no differences)
– storable
– durable
– compact (high value for its weight)
in EQ we started seeing peridots being used. (more examples got called out, dark angel feathers in M59 etc. Even a Kingdom of Loathing example).
Quantity Theory
P=MV/Q the quantity theory of money. When marketing comes in and says “let’s give the payers more money” use this to say why it won’t work.
P is the price level
M is cash in circulation <-- note, circulation, not in game. Thr trillions zapped from UO had no effect because it was not circulated.
Q is the quantity of good.
V is the velocity of money. It is the rate of exchanging things., The propensity to save. You have little control over this. In UO the velocity was small. The minute the currency was zapped, the velocity went up (and prices DROPPED).
Inflation is delta(P) = delta(MV/Q), the rate of change. Inflation itself is not so bad.
Healthy inflation: holding cash maintains your wealth
Hyperinflation: holding cash loses wealth.
Depression: Holding case increases wealth. NO liquidity in system, and the economy crashes.
In MMOs, we control price via the money supply. We solve the "poor player" problem with market prices.
How does money get into an economy? Faucets and drains. In other words, player-to-game exchanges. Player hours help set the size of the faucet, but it's loot, it's newbie gold.
Faucet: player selling to the game
Drain: player buying from the game.
The maximum amount of money per time period is based on max amount of activities in a given timeframe as they play.
The problem is that you do not control the cash creation, players do. If they are short of cash, they mint more by killing things. But then the prices go up, so they kill more things. If they need more cash, they open the faucets.
But the drains are fixed typically. The end result is hyperinflated, it's just a question of when.
Solutions:
- tie the drain to the price level (NPC prices rising and falling based on the cash supply). Exploitable, but Ted Castronova thinks he can do it. You SHOULD be able to do it, but...
- auctions work well, because price levels float.
- transaction tax is starting to be used more often in MMOs; a fee charged for something on the auction house as a fee. Used in EQ2, Cloud...
Two player economy example of a transaction tax example. Adventurer needs bullets to kill mobs for cash.
Crafter makes bullets.
Crafter sets price based on demand
Crafter pays fixed charge and percentage sales on the commission.
When graphed, this results in the currency, price of bullets, and so on stabilizing over time.
"Poor" player issue.
Say the economy has been going for 2-3 years. All your crafters have cool goods they are selling, and then new players show up. In the original design of SWG, the new player needed to get involved into the economy quickly. Marketing, customer service come in and say "new players don't have enough cash." But this is the wrong solution, it's inflationary.
As the value of cash declines, the faucet stream declines
So you can't buy good stuff.
Makes new players dissatisfied
Non-inflationary solutions:
- Have NPCs provide basics. Work s to an extent but is not as engaging and you get resentment too ("I got a rusty sword and a rock??!? it's crap!")
- make elder players dependent on services or starting commodities that newbies provide. In EQ, high end reagents were farmed out of the newbie yards. Commodity is good, service is pretty good but it needs to be an ongoing service.
UO had the scroll that matched players together. But there was the problem of scams. AC2/AC new players were valuable because of allegiance...
"Is gold farming severe enough to disrupt your economy?" We know duping is, If the rate of inflation isn't so high that you get issues, then there's no issue from an econ standpoint. There's questions of access to spawns, etc... "Couldn't gold farming have positive effects or other effects?" Yes, but we have also seen racism, etc,. but now we're out of the scope of the talk.
In M59, there are mechanisms to adjust the gold drops, change the faucet instead.
Market Structures: how people buy and sell
What market structure should be used?
Real world economic objective is "Pareto optimality." Resources optimally allocated. This is the gold standard. This is achieved by structuring a perfect market.
We don't want that. Perfect markets are boring. You need imperfection if you want fun. What you want is something called oligopoly.
Structures approach perfection.
——————————————————-> ^ ^ ^ ^ | | | | barter monopoly oligopoly perfect markets
Perfect market has many buyers and sellers, and homogeneity of product. No time energy or money to buy or sell. Everybody knows the price, there are no opportunities for triangle trade, low or no transaction costs. Secure transactions. Infinitely divisible good (solved by cash).
Oligopoly — almost all market structures int he US. Few sellers, many buyers., Substitute goods — there have variations in quality. There are barriers to enter and exit the market. It costs money to get to one store versus another. Information isn’t perfect. I don’t have the same info about prices that you do. There are some forms of transaction cost. You still want divisible goods and secure transactions.
How to get fewer sellers?
– fragment sellers by location. Having one auction house like EQ, no. EQ2 two. WoW has three. SWG originally one planetary commodity market but many player vendors.
– another solution is to fragment sellers by time. In EQ, you have to be logged in. So you get a second account. In WoW it’s based on limited time — the offer can only stay up for a limited amount of time.
Many Buyers: tougher.
– Having an NPC vendor of last resort increases the money supply, and it provides a ceiling or a floor to the market. I like players to decide how valuable things are.
– consolidate buyers by location. In SWG I could see everything in the galaxy on the market, but I had to go pick it up. iN EQ2, I would purchase at a distance for a premium.
– consolidate buyers by time. Posting offers to buy, as in Eve. This makes buyers be “on all the time”
Substitute goods: differentiate goods based on
– effectiveness tradeoffs.
– Aesthetics/fashion
– scarcity (done something remarkable, etc)
Barriers to entry:
– high fixed costs. In SWG you make an investment in a hop, a house, your tools, a location. IN EQ/EQ2 it’s time investment.
– specialization. In WoW you can get one resource and one manufacturing skill. Need 3 or 4 to make an item. In SWG (old), max 2 crafting profs, but need 3 or 4 for making something.
Imperfect info
In conflict with the few sellers many buyers thing,
– But you want arbitrage opportunities. WoW market tracking software…
– heads up future economic events. In SWG the resource base changed periodically. In There players work the payroll cycle. They know that people do not have cash on Thursday, so they sell low on Thursdays and pump prices on Friday when checks come in
**
Transaction cost is player time.
– make it expensive to change vendors.. SWG, have to pick it up, and search everywhere. WOW, time to get to the location. This makes location based competition a valid choice.
Security and division
– allowing players to cheat on a deal will just cause cs headaches. Admittedly, Eve is doing a great job. Eve is working on a secure contract system though.
– making goods indivisible means you are going to a barter system. But someone will make an alternate currency instead.
If you play with these, you risk destroying the business game.
Industrial organization
This is how relationships between crafters works… it’s a study of the structures, the manufacturing chain, and competitive practices.
Typical MMO:
– adventurers gather raw material
– Others supply sub-components
– crafter creates final product
– sold to adventurer
Design issues:
– manufacturing chain
– overproduction
– elder dominance
Chain solutions:
– no independent crafters who can do it all alone. Allowing a crafter to make some subcomponents and final products is necessary, because else it is too narrow.
– competition o the four P’s (marketing): price, products (substitute goods), place (location of sale), promotion, and MMO unique Predation — allow someone to go in and screw up the other guy. Business games are PvP games after all.
Overproduction. Caused by “oligopsony structure” which is many many sellers and few buyers.
Happens because we tie level advancement of crafters to making goods. No one wants to buy CDEF pistols in SWG because the market was flooded and you needed to make a zillion to advance.
– Find a different levelling mechanic.
– Destroy overproducing factories.
– Finally, make items consumable.
Elder dominance
Caused by a natural monopoly structure. A guy comes in with the investment, and he can always undercut everyone else.
Result of overly high fixed costs to get in.
A tremendous first in advantage.
Constantly declining marginal costs. As I get better it’s cheaper for me to produce. This includes things like success rate in crafting. If masters fail less, they have declining marginal costs.
Solutions:
– cap quality.Elders make the same quality goods as a newbie. –
– Cap cost. Elders cost the same, including wastage.-
– Use comparative advantage: the elders can make more money selling high end stuff to a different market, so they move on naturally,
References:
Check out Terra Nova http://www.terranova.blogs.com
Sam Lewis’ site: http://www.flyingscythemonkey.com — has all the GDC stuff and papers and this material.
Eve developer blog, Dan Speed, https://myeve.eve-online.com, analyzing Eve economy
Synthetic Worlds, book by Edward Castronova.
Personally, I think the whole model of infinite resources is inherently flawed and ultimately is the root cause of most of the economy issues (aside from duping bugs and exploits).
What I see as two key components to avoiding rampant inflation are as follows:
Localized resources with mass transportation difficult enough to keep a those resources from becoming overly globalized. Done in such a way that a local resource isn’t overpowered to the point of being regarded as necessary by the player-base (IE, Oak might be the best thing to use to make a house, but bamboo will serve equally well in forming a basic hut)
Finite resources with the ability to ‘renew’ some of them (plants for example) and to discover new caches of others (digging deeper into a mountain to find more ore). Included in resources would be any currency, whether it be gold coins or colored beads. NPCs shouldnt just be able to pull millions of gold coins out of their seemingly bottomless pockets.
Tholal wrote:
Huh? Economics concerns the allocation of scarce resources to satisfy unlimited wants, not the other way around.
Except that this model, paradoxically perhaps, doesn’t apply to MMO’s. After all, you can create as much resources as you want and allocate them as you will. But as the presenter discusses with “talking to marketing”, creating more resources to hand out isn’t necessarily a good strategy.
StGabe wrote:
I don’t necessarily agree, which isn’t to say I think that’s wrong. The virtual space allows us to experiment with scenarios that normally wouldn’t be encountered in real space, but we can also control that virtual space in a way that models real space. A virtual economy can mimic a real economy. It’s just a matter of design.
Perhaps resources should be distributed on a per-player basis?
Except that no one is going to let you add real death* to the design. Without death we don’t have needs, just wants. This ripples into the supply and demand model because there is no true demand, just some desire. This is why on many MMOs finished goods are valued at lower than materials. Players want a Sword of Orcslaying but the real demand is for crafters who want to make a Sword of Orcslaying. The fun of having a sword they made themselves out ways the extra cost of making it themselves.
*Real death is not being teleported a short distance from where you were fighting. That said, I think WoW proves what I’ve been saying for some time, that the penalty for dieing doesn’t need to be much higher than a “You have failed” message. Players know that their real stake in the game is time, having to do something over again is cost enough. I can see benefits for higher costs, I just think they are outweighed by how a player feels when he dies when it wasn’t his fault, such as a server crash, or the “tank” going link dead, or forgetting to switch to fire armor. (OK, the last one does sound like his fault but it may not feel like an honest death to the player because it’s not a mistake his character was likely to make. )
From the same author, the Hierarchy of Pain:
http://www.flyingscythemonkey.com/Pain.htm
Morning ramble on the subject – I wish I could have seen this presentation:
I don’t know that infinite resources are a bad thing, since ultimitely it is a game and players probably won’t find “gold rushes” very fun unless they happen to secure their stake during one. But at the same time, if you make your resources too easy and painless to get you can end up with problems later on. That is to say, instead of limiting the amount of resources that come into the game, limit the rate of collection of those resources.
EVE does a good job of this where asteroid fields will be mined out and then take days to respawn.
Since resources are consumed by crafters to make goods, which are then purchased by consumers and used up or eventually destroyed, this allows anyone wanting to gather resources and craft to get started, but also insures that oversupply doesn’t become a very bad thing.
On the opposite end you have EQ2’s crafting model, where you can very easily farm tons of resources with little risk involved. Over time this has caused resources to become very cheap commodities, and thus goods made with those resources have fallen in price.
A bigger economic question for any MMORPG is the loot vs. crafting one – how do you want to supply your players with cool items? A loot-based economy is always going to be prone to speculation, farming, and so on. There will be a huge price difference between rare items and more common items, and not much in the middle. In a crafting-based economy prices will tend to scale more evenly with rarity/difficulty/quality.
You can combine the two but you still have to pick one over the other. Earth and Beyond used a mechanism where players could loot an item and then a crafter could take the item and attempt to figure out how to make it. Once the crafter learned how to make the item they were able to make the item with better stats than the looted version. So for most items the crafted version was the one you were striving for.
Specialization is important. Players like to try to do everything for themselves. If you let players master every crafting area then demand drops over time as more do it. If you enforce specialization in some way you keep demand going because players have to rely on each other. Likewise if you make crafting a true profession/class rather than being a side set of skills that people pick up you insure demand.
Significant barriers to entry are a bad thing that will discourage players – it is a game. This depends a lot on how itemization is done in the game, but if players have to go through hell and high water just to be able to get started crafting it will decrease the number of crafters available, which will cause supply to stay low and prices to stay high.
Great read! Like Econ 101 all over again (albeit 20 years later)!
Sorry – I know this doesn’t offer much to the discussion, but I just really enjoyed that write-up.
I’m scared that economists still actually think in terms of perfect markets and “Pareto”. Apparently, being inapplicable to the understanding of real-world events hasn’t diminished its appeal. “Oh, this could be a perfect market except that it involves human beings!”
I always found the resource gathering in WOW a little odd. Other games which have gathering involved (eg. EVE and SWG) make it into a profession; mining and artisan respectively.
In WOW it always felt strange because everyone was a combat player, and the game itself doesn’t really represent a persistant world as the only thing to do is combat. The gathering part of WOW is always a side profession to the main task of whacking kobolds ^^
Having said all that, I can’t work out which kind of economy works better, because both WOW and EVE are gaining subscribers at a steady rate (I discount SWG as they are draining subs due to other factors unrelated to economy).
However, it cannot be overlooked that a game economy gives almost all players a reason to keep playing. Even with a lack of dev created content, ecenomics gives the true player created content that we all strive for. One only has to read of the recent EIB scam in EVE, which resulted in players losing 750,000,000,000 in game credits….and making the news across the net; I can’t begin to imagine how many new players that brought to the game!
I agree with what you’re saying near the end. But you see, the fact that you have to build in constraints shows that you’re not really “maximizing allocation of limited resources” when you create an MMO economy. You want to create a constrained environment with some limits on resources that then allows to people engage in economics. And you want to make sure that the economics is FUN.
It’s a different task entirely than figuring out how to increase the GNP of your country because you could increase the GNP whenever you want just by tweaking some numbers. Designing MMO economies happens at meta-level to economics and has to introduce constraints instead of overcoming constraints as real-life economics does. It’s real-life economics turned upside down.
Because the game is purely artificial. The only reason not to satisfy all needs and wants is that it wouldn’t be fun and not because someone might starve or die of exposure.
Very glad you were able to spell “fiat” currency correctly in the liveblog.
For some reason, Sam, with his “MBA in Economics” insisted on spelling it “fait”.
He also totally misunderstands the “Invisible Hand” of Adam Smith — but then he has this in common with many, many economists, mostly because they rely on bad distillations of Adam Smith rather than reading the original.
Adam Smith’s “Invisible Hand” that supposedly directs the free market, if you bother actually reading Adam Smith, is the hand of God, which Smith felt directed and controlled human affairs so we’d all be ok in the long run. God wouldn’t allow market fluctuations that would let us all starve, for example.
Clearly that won’t work in an MMO, unless your Invisible Hand is the hand of the game designer… just like he then suggests, apparently without irony, with fluctuating NPC prices. I agree that simplistic formulas for doing this would be easily exploitable, however.
But apart from these basic snafus, there are some excellent points. Personally, I especially appreciated how clearly his talk lays bare the old “gold farmers cause inflation” lunacy.
One resource, money, is infinite in the real world because most money is digital. We can print as much money as we want; however, our economic systems impose constraints for economic reasons. We can mimic a real economy in a virtual world because the mechanics are the same. We’re just using a window to another environment populated by real people with the same unlimited wants.