[What makes MMO economies tick? Here's a look at several different games and real-world economic theory to explain the mechanisms that govern value to players, and how to design an economy that best serves your players and you.]
This article basically grew out of my somewhat "rantish" review of Final Fantasy XIV. One of the biggest issues that I -- and many other commentators -- have taken issue with is the way FFXIV handles its in-game economy: the Market Ward system.
My intent with this article is to first address the broader issue of virtual economies in MMORPGs in general, and apply those theories to FFXIV to better explain exactly what is wrong with the Market Ward system.
Wayback Machine: Asheron's Call
I'd like to begin with a journey back to my very first MMORPG: Asheron's Call. AC was one of the three "golden oldies," the MMOs that started it all, as it were. Along with Ultima Online and EverQuest, many of the standards in today's online games found their origins here.
Of particular interest was the development of the virtual economy in Asheron's Call. There were a few major factors that shaped AC's economy: firstly, that money had weight and was, either in coin or writ form, cumbersome and risky to possess; secondly, that the game had no market system whatsoever and didn't even have a secure trade feature (it was added much later); and thirdly, that some of the game's most sought-after weapons and armor were crafted from ubiquitous Pyreal Motes and Crystal Shards.
Putting aside some of the more flavorful quirks of early Asheron's Call -- things that led to the commonplace inscription of item stats and the cataclysmic destruction of Arwic -- I'd like to focus on how players engaged in the most basic economic activity: trading.
Because there was very little economic motivation to amass money in Asheron's Call, economic activity in the game was minimal. Furthermore, the lack of any crafting system and related resource-gathering system meant that players, by and large, didn't acquire much "stuff" to sell to others; most of what was looted was vendor trash. Rarely, a particularly good piece of equipment would drop, and players figured it would be more valuable traded to another than sold to the vendor. But what would they want to trade it for? Not money.
Players instead defaulted to a more limited barter system. The "currency" of the market became the desirable motes and shards, items that everyone wanted. These became so common that players would list the items they wished to "sell" on forums and attached to each a price in motes and shards.
Even more enterprising economists would analyze the market and the frequency of motes and shards and would maintain a "currency exchange rate" between coin, motes, and shards, for those wanting to make sure they were getting a fair deal. In the absence of monetary value, motes and shards became the de-facto currency of Dereth.
Prior to an actual exchange taking place, players need a means of communicating their desire to buy or sell an item to one another. This communication took place in two main formats: through external means (mainly forums where people posted what items they had for sale or wanted to buy) and through internal means (mainly players standing around in Arwic /shouting the items they wanted to sell). Players interested in "shopping" would either visit the forums or Arwic and hope to find something they were interested in buying.
In Asheron's Call, this economic activity was generally viable thanks to a very significant factor: there wasn't very much to trade. The items being exchanged were randomly-generated loot, within which there were a relatively small subset of factors determining the value of those items.
For example, crossbows would have damage modifiers; the most desirable crossbow had the highest damage modifier. All magical effects on items in Asheron's Call were spells that could be cast by players, so magical properties were of lesser consideration than the base numbers. As a result, when players went to market, they basically had one item type they wanted and basically just wanted the best available of that type.
And so, long before the days of auction houses or bazaars, the motes and shards would flow in the rather primitive Derethian economy.
Intro to Economic Theory
Economic theory is based on the notions of Perfect Market Competition, the ideal conditions for economic behaviour that, sadly, elude us in the real world.
However, in virtual worlds, many of the otherwise untenable features manifest themselves: there tend to be no barriers to entry, as every player has the same intrinsic capability to engage in any of the game's internal economic activities (harvesting, crafting, trading, etc.); factors of production tend to be highly mobile in most games (once you can craft something, you can do so anywhere, at any time, without an investment in capital); products are homogeneous (all healing potions are exactly the same; all crafted Scorpion Harnesses are identical); and there are constant returns to scale (no matter how many potions you make, it still costs the same to produce each one).
This means that many of the core theories of economics, those of supply, demand, marginal revenue and the like, can be applied fairly directly to MMORPG virtual economies. The key issue -- and, as we shall see, the dominant one in virtual economic theory -- is how a game designs its market around the concept of perfect information.
Perfect Information means that all producers and consumers know the price and quality of all goods. In the case of quality of goods, most games do present that exact information to players. In World of Warcraft, you can hover your mouse over an item and a pop-up box lists every relevant statistic for that item. Is this sword better than that sword?
The player has all the information necessary to make that judgment. There is no question of material quality, branding, or any other factors other than actual performance of an item when determining its value (aside from aesthetic purchase decisions). It doesn't make any difference whatsoever who makes or who sells you an item: it's the exact same item every time. The issue, then, is whether or not players know the price of an item.
The price of anything is what its purchaser will pay for it, as we all know. However, the value of an item is its opportunity cost: the cost to obtain that item. In MMORPGs, the "price" players inherently "pay" for everything is time: the cost of an herb that takes 10 minutes to harvest is 10 minutes of time.
In order to extrapolate the value of time, players merely need to determine what activities they could be spending that time doing: if, in 10 minutes, you could harvest one herb, five pieces of ore, or get 10 gold, then the value of 10 minutes is one herb, five ore, or 10 gold. Most things in an MMORPG can be derived from their time-based cost, because most things can be acquired merely as a matter of time.
A rare item that only has a 5 percent drop rate from a rare monster that only spawns once an hour has an opportunity cost of about 20 hours. Because chance is involved, the actual cost to any given player will vary: some people will "pay" a lot of time and never get it; others will stumble across the monster and get the item with very little cost.
In any case, if players value things in terms of time, then they can determine what they would be better off doing with their time. Continuing from our earlier example, if five ore sells on the auction house for five gold, then you would be better off getting gold and buying ore than harvesting the ore yourself. This notion of "better off" is another of the tenants of perfect competition: each player should be a profit maximizer.
Players are not necessarily profit maximizers. Ultimately, players are playing a game to have fun, and this "fun factor" cannot be discounted in virtual economics. If it is significantly more fun to mine five ore than to get 10 gold, players might prefer to harvest the ore themselves rather than buy it, simply because they find the activity more pleasurable, even if the economic outcome is unfavourable.
This fun-seeking behavior can cause market irregularities: for example, if the majority of players enjoy mining a lot, the market for ore might become depressed as many players mine it and attempt to sell it -- given there are few players interested in buying it.
This example does raise a crucial point: the deflation in the value of ore does not come from a glut in supply, but rather from a glut in suppliers. The prices are not low simply because there is a lot of ore in existence, but because players are harvesting it and placing it on the market at a higher rate than players are consuming it and buying it from the market.
What's Gold Worth?
The notion of supply rate becomes even more important when one considers the basis for many MMORPG virtual economies: the gold piece. And yet, that I've made it this far without ever mentioning "gold" should indicate an important point: a game's currency is not actually the most important element in a virtual economy. My opening example of Asheron's Call's mote and shard market should paint that picture fairly clearly.
But more importantly, gold in a game isn't actually currency at all; at least, not a fiat currency in the sense of the money we use in real life. A fiat currency is something that has no intrinsic worth; you can only trade it to other people who are willing to accept it.
However, currency in a game does have intrinsic worth: the NPCs that buy and sell items in exchange for the game's currency set the intrinsic worth of it. If a healing potion costs 100g, then 100g is always worth a healing potion. No matter what happens to the game's player economy, 100g will always buy that same healing potion.
The idea that gold isn't a currency should be even more strongly enforced by the methods in which players acquire gold. In my earlier example, the player spends 10 minutes harvesting, and gets either herbs or ore or gold.
Gold is, in this sense, a harvested resource, gathered from the corpses of monsters or by turning in quests, just like a player would gather any other resource in the game. The value of gold is the same as the value of any other item: it's derived from the opportunity cost in time to acquire it.
This applies to all transactions: even if there's a super-rare sword on the auction house for 20,000 gold, you can determine its value by comparing the time it would cost you to accrue 20,000 gold versus the time it would take you to acquire the super-rare sword.
There are, of course, contingent factors. For one thing, acquiring gold is a very low-risk behavior. Players completing daily quests in World of Warcraft can predict with decent accuracy how much time it takes to do them and how much gold they'll get for their effort. I've often heard things like "Oh, it'll take me six days of grinding dailies to buy epic flying for my alt."
On the other hand, acquiring the sword can be a very high-risk behavior. If it has a very low drop rate, players may try and try and try to get it and never actually get it at all. In the time it takes them to gather 20,000 gold, they might never see the sword at all. Thus, a player's risk-aversion could significantly increase the perceived opportunity cost of a risky item.
Even if it statistically takes the same amount of time to gather 20,000 gold as one would acquire the sword, avoiding risk might actually cause the sword to be worth more than its strict opportunity cost to a risk-averse player. Conversely, a risk-seeking player might prefer the challenge of getting the sword for him- or herself, decreasing the value of the sword below its opportunity cost.
Another contingent factor is money in the pocket. Players regularly gather items and money during the course of their adventure at no opportunity cost. A player may want to complete a quest and, in the process, gets attacked by a hostile monster, which they kill and loot a rare item. Lucky! The opportunity cost of this item was zero: the player never actually intended to obtain this item nor spent any time in acquiring it.
Likewise, players regularly get lots of little bits of money from killing monsters and selling "junk" to vendors, slowly increasing their pocket coin without really putting any specific focused effort into getting money. This pocket money can have a strong influence on purchasing items, especially if players have a large amount of money "burning a hole in their pocket" and stumble across an expensive item on the market to buy.
The opportunity cost -- getting the item versus unloading pocket money -- might (depending on the personality of the player) lean very heavily towards giving up some pocket change for a new toy.
One key consideration with money is that, in most MMOs, it doesn't take up inventory space. There is no penalty to carrying it. Therefore, there is no cost in simply holding it, compared to having to decide which of many items to store in one's inventory. My blog on Gamasutra about Value Maximization goes into further detail about inventory management, but because of its special nature, gold is spared the cost in terms of limited carrying capacity. As a result, players can often amass large fortunes in their pocket without paying any storage costs.
Money In, Money Out
The influence of pocket change is the result of the "MIMO" balance in a game's design. MIMO, or "Money in, Money out", refers to game systems that produce or consume the in-game currency. Using World of Warcraft as an example again, a daily quest produces money (money in) whilst buying epic mount training removes money (money out).
One of the inherent challenges to balancing a virtual economy in an MMORPG is balancing the MIMO to prevent the accumulation of excess currency. Many players mistakenly think that the mere presence of money in a market leads to pocket-change-caused inflation. However, it isn't actually the amount of money itself that leads to this kind of inflation, but rather a problem in the MIMO balance.
A core concept in understanding MIMO is to understand the opportunity cost of gold again. Recall that it might take a player 10 minutes to get 10 gold or five ore? What happens if the player also gets gold while harvesting five ore? This gold has no opportunity cost: the player did not seek to acquire it, but merely got it as a side effect of pursuing their ore-harvesting objective. This gold is "money in". This can have huge repercussions when large quantities of money are generated as part of a non-money-seeking activity.
When World of Warcraft introduced Sunwell Island in the Burning Crusade expansion, Blizzard introduced a series of daily quests that provided players with reputation gains needed to unlock access to the Sunwell Plateau raid zone. Players who were working on gaining access (the opportunity cost in time for gaining reputation) also gained a lot of money from completing the daily quests (zero opportunity cost incidental rewards).
Thus the intentional inflationary effect of daily quests (players now had a way to make more money in less time) had an unintentional further inflationary effect (the extra money from players who were doing dailies not for the money but for the reputation). Money in; but no money out.
MMORPGs often use MIMO as part of their content gating system. If players get, on average, 1000 gold while going from level 1 to 40, but players who "rush past" content end up with less gold, then putting a 1000g content gate at level 40 forces "rushing" players to slow down, harvest more money (possibly by doing quests they skipped) and getting "back in line" with what was intended by the developers.
If you'll forgive the incessant WoW references, vanilla WoW had this sort of content gate for its original level 40 mount riding training, though many did not have enough money to meet the content gate at the time they reached level 40. This sort of "gating" design is very risky because different players treat money in extremely different ways: some hoard it, others spend every last coin.
Unlike direct progress gates (e.g.: experience points, level limits, gear requirements, etc.) money gating leads to unpredictable results. It's also of interesting note in World of Warcraft because the money gates are the only ones that can be done socially: other players can give you money to buy your way past the content gate. All other content gates in WoW are generally ones that a player has to do him- or herself: getting attuned to a raid zone, completing a quest, or gathering bind-on-pickup gear.
MIMO also tends to be a "drippy faucet" problem. Players tend to continually gather incidental money while playing an MMORPG. As long as that money keeps dripping into their pocket faster than it leaks out from incidental expenses (e.g. repairs), they'll face lower opportunity costs when making purchasing decisions from other players.
Consider two players with identical play styles and spending personalities; however, one has been playing for a year longer than the other and has accumulated an extra 10,000 gold in MIMO drip. When an item goes on sale on the market, the longer-term player faces a 10,000 opportunity cost discount, merely on account of having been playing longer.
Thankfully, MIMO problems aren't the inflationary beasts that most players make them out to be. For one thing, once that 10,000 gold is spent, the opportunity cost discount is gone. The person who has collected the money does not face a 10,000 gold opportunity cost discount; they paid for that gold in opportunity cost from the item the sold.
Again, consider WoW's daily quests: once players are actually doing them for the gold, rather than for some other reason and getting the gold as an incidental reward, then these are not part of the MIMO equation.
At that point, the gold earned from the daily quest directly represents the opportunity cost of the time it took to do the quest. Even when players are "just doing them because I always do them" they're still paying their opportunity cost specifically to get the gold reward.
One very notable game to be examined for MIMO would be EVE Online, a game without MIMO. EVE Online has a strong, powerful market, and is considered by many to be a shining example of a stable virtual economy. One of the primary reasons for its stability is the lack of MIMO in EVE.
EVE's MIMO Solution
In EVE, players tend not to gain much, if any, money from activities that are not purposefully money-gathering. In fact, most of the time players incur significant monetary losses from non-money-gathering activities.
EVE's activities can basically be split into two broad categories: getting ISK (currency) and PVP. Although there are occasionally "loot bonanzas" when engaging in PVP, these are quite counterbalanced by the common occurrence of losing one's ship and possibly one's life (clone) at the hands of other players. Some players have turned piracy PVP into a profitable activity, but for the majority PVP is operated at a monetary loss in order to facilitate higher profit money-gathering activities. By eliminating the "pocket change" effect, players always calculate the opportunity cost when making purchases.
The examination of EVE also brings the final and major topic to my discussion: the actual virtual economic system designed into a game and the economic ideal of perfect information. Asheron's Call had none; players could only trade and had to use methods beyond the game system to organize those trades.
Most mainstream MMOs have "brokers" or "auction houses" which allow players to list items for sale, and other players can view the list and chose items to buy. EVE Online's market system is the most comprehensive and will be the focus of my examination of an idealized MMO market system.
Recall the notion I presented earlier about perfect information. In an idealized economic setting, a player will know all the prices for all items. If we put aside EVE's physically separated markets (which are meaningful within that game due to the risk of travel, but not meaningful in many other games) then the EVE market provides players with perfect information (at least for the items that can be traded on the market itself).
EVE's market functions as a bi-transactional order market: players can place both buy orders and sell orders for any given item. This is a crucial difference from most Auction House-style markets, where players can only place sell orders for items they have and want to sell. In most other games, players cannot declare their intention to buy an item through the game designed marketplace.
In a market with only sell orders, buyers have no power to influence the price point of items. Their only actions are to buy or not to buy, and not-buying does not indicate a desire to buy at a lower price any more than it indicates a desire not to buy at all. Prices in these markets are purely supply-driven. As a result, if there is insufficient exchange to drive the price of an item down to its opportunity cost, buyers either suffer unfairly high prices or the market stagnates and transactions fail to go through.
By enabling both buy and sell orders, buyers compete to make the lowest highest offer and sellers compete to offer the highest lowest price. Yes, you read that correctly, and no, it's not as confusing as it sounds. Buyers want to pay the lowest price possible, but must offer a higher price than other buyers in order to make a successful transaction (otherwise the transaction will go to a higher bid). Likewise, sellers want to get the highest price possible, but must offer a lower price than other sellers.
These markets are both supply- and demand-driven. When both buyers and sellers compete within the same marketplace, any entrant to the market can see both the upper and lower bounds for transactions within that market. They can (presumably) make as much money as the lowest seller or pay as much money as the highest buyer should they wish to compete in that market. However, most players don't want to compete in the market: they just want to make their fair transaction and get on with the game. This is another point where a dual market really shows its strength.
A Superior Market Strategy
Considering again the sell-only market as implemented in most MMOs, players who want to sell an item must compete in the selling market. They must place their item up for sale and wait for a buyer to come and buy it. However, most buyers will sort by price and simply purchase the cheapest item, even if it is only one unit of currency cheaper than the next lowest item.
Final Fantasy XI's auction house even enforces this behavior: the prices of all items are actually hidden from buyers, who must place "bids" to buy an item. If their bid is higher than the lowest priced item, then they buy the lowest priced item for the amount of money they bid. So, if you try to sell an apple for 1000 gil, and I undercut you by selling an apple for 999 gil, when someone comes to the Auction House and enters a bid of 1000 gil, the system sells them my apple for their bid price.
This system of undercutting, especially if there are multiple "market players" willing to continually undercut for a small price difference, makes it very difficult for a casual seller to simply sell an item and get on with the game, unless they are willing to significantly drop the price of the item so that it is not competitive to be bid against.
What frustrates most players is that, when they drop the price significantly, the "market players" will often buy the item and re-list it and make themselves a tidy profit from the seller's frustration!
There are two ways to counter undercutting, the most obvious being to add fees to the use of the market services. If players must pay a fee to list an item, then they pay a small opportunity cost every time they want to shift the price to undercut. Another method is to impose a time limit preventing rapid undercutting. Many games do both of these.
A far more effective method is to include both buy and sell orders, and, by forcing competition at both ends of the spectrum, drive prices as closely as possible to the opportunity cost of the item (ideally, the price range settles at opportunity cost plus or minus the listing fee).
When there are both buy and sell orders, the casual player who wants to make a quick transaction and get on with the game can sell to the highest buyer rather than compete with the lowest seller. In other words, the player can either pay the opportunity cost in time to compete with the lowest seller in the sell order space, or pay zero opportunity cost to accept the highest buy order. If the two prices are close together, then the cost of taking the highest buy instead of competing for the lowest sell is profitable.
Of even greater interest from a game design perspective is that a buy/sell market actually encourages "market players" to play their game (in my opinion, it's always good to encourage players to have fun playing the game in the way they want to play it, provided that such behavior is not detrimental to the experience of other players, as happy players are subscription-paying players.)
In a bi-transactional market, players can place both buy orders and sell orders for the very same items: buying them and re-listing them for a profit. The behavior that seemed detrimental in a sell-only market now serves a valuable service: these market players act as brokers that provide instant (zero opportunity cost) transactions to players who aren't interested in spending time and effort on the marketplace.
If there is sufficient competition among market players to keep the cost differences between buying and selling low, then players who take their buy orders or sell orders pay a relatively small "fee" in exchange for an instant market transaction.
Furthermore, if fees are only charged for the listing of orders, the "market players" suffer all the fees charged by the market system and order-takers pay or get paid only the price of the item.
The market-takers feel like they're getting a good deal because the market isn't nickel-and-diming them, whereas the market players feel like they're getting free money for doing nothing when, in fact, they're putting their resources forward to ensure rapid market transactions, absorbing excess supply of market goods, and spending their time (opportunity cost!) to make all this happen. It's a win-win.
A Case Study in Doing It Wrong: Final Fantasy XIV
Now that I've said all this about markets, let me make a few broad presuppositions about typical players. For the most part, when players engage in the market they want these things:
- To feel like they're getting a fair price for anything they buy or sell
- To be able to sell things promptly
- To know what is available for sale and, if something's available, buy it promptly
- To be doing all this with the least amount of time, money, and effort as possible
- To not have to resort to using external systems to engage in the game's market is they can avoid it (they will do so if they feel it gives them a competitive advantage in the marketplace)
- Finally, players do not want to be advertised-to or be "spammed" by people shouting about what they want to sell.
Final Fantasy XIV's market system is based on the "bazaar" system seen in many Asian MMORPGs, including its predecessor, Final Fantasy XI. The bazaar allows players to place a number of items (in FFXIV, 10) for sale at a listed price.
Once created, players with a bazaar bear an icon next to their name indicating the presence of a bazaar. Other players can click on that player and chose "browse" from the list of player interactions (along with trade, invite to party, and so forth). This opens a window just like an NPC store with the items for sale and their price.
The purchasing player conducts the transaction just as they would with an NPC merchant: selecting items and quantity and, if they have enough money for the transaction, they get the items and the bazaar-bearing player gets their coin (minus a transaction fee, of course).
Square Enix has declared that it wanted to rely on this system because it encourages "creative selling" and makes players feel personally engaged in market transactions. This is all ridiculous head-in-clouds dreaming, of course.
The actual reality is to find areas crowded with AFK players all sitting around with their bazaar icons up, and other players browsing the crowd in the hopes of finding an interesting item for sale.
This has two negative consequences: the crowding of player areas, which causes lag and increases server overhead for the game operators, and it irks players who would rather be playing the game instead of standing around trying to sell stuff.
In its effort to really drive the bazaar system into peoples' hearts, Square Enix introduced Retainers and the Market Wards. Retainers act as a player's personal storage in FFXIV. Just like the player, the retainer also has a bazaar of 10 items, set up in just the same way.
However, retainers can only be seen by other players (and, thus, sell items) when they are placed within Market Wards, special marketplace-esque instances in each major city. There are a number of wards, each featuring a category of item to be sold, however players need not obey these directions and they do nothing to help players trying to sell multiple categorically different items.
Unfortunately, it works rather poorly, especially since the game can only render a very limited number of NPCs at a time. And it doesn't stop players from doing the AFK bazaar thing too. Why limit yourself to an NPC no one will check, when you can <