This is a short article I wrote for the server forums in 2006 about arbitrage on the inelastic vendor market in World of Warcraft. As mentioned in the article, this kind of trading is hard to do on a large scale without software assistance, and several of the required functions were subsequently removed from the user API space.
I keep this article online as, to my surprise, it got was cited in a handful of academic papers as evidence of behaviors like arbitrage emerging in video games:
Although I would love to take this article down and rewrite it to take advantage of a further 12 years of gaming and personal experience, I know what it feels like to have a cited source vanish, so this page is kept up at a fixed URL: http://progsoc.org/~curious/economics.html.
The Argent Dawn ( World of Warcraft US/RP )
economy is, like many MMORPG economies, a hybrid of an elastic player run
economy where supply and demand vary as in classic capitalist systems and
a computer run "inelastic" vendor market, where supply and demand are
fixed as in idealistic centrally planned economies.
Unlike in real economies, there are no real consequences for players who employ poor economic behaviour. Players will not starve, be forced into bankruptcy or have their possessions repossessed. In addition, the economy is infinitely open ended and inflationary due to the never ending "respawns" of monsters, treasure stockpiles, herb patches and ore outcroppings. In such a responsibility free environment, various forms of pathological trading can take place which can give rise to some unusual effects.
This article looks at some of those effects, how they can be avoided, and how some of them can be leveraged to economic advantage. It is dedicated to Myrdoch <Blood Cabal> (Argent Dawn (RP), Horde), whose writing on the server forum was an inspiration to start thinking critically about the economy of massively multiplayer games.
Economics in World of Warcraft is Serious Business. Poor economic decision making means bad returns on playing time invested, and that in turn means that more playing is required to reach financial and gameplay goals. Some players find the style of play required to make a session profit unenjoyable ( "grinding" / "farming" ) and see it as a chore that must be endured to fund more entertaining parts of the gameplay experience. As some players also buy and sell game currency for "real" money, the economy of these games also touches tangentially upon the economy of the world around us. So, it's fair to say that everyone has an interest in trying to maximise their own personal profitability.
Inspired by reading Myrdoch's posts on making money using the crafting system, I decided to see if I could fund an "epic mount" ( this was at the time the most expensive item sold in the computer controlled economy ) by trading. Using thirty gold seed capital, I was able to collect my funding goal of nine hundred and thirty gold pieces, a large pile of enchanting reagents, equipment and training fees for my stable of other characters, spin off a Horde branch of the same trading business and buy my way into the good graces of one of the computer controlled factions ( Gnomeregan Exiles Exalted ). It would seem it was indeed possible, but perhaps not in the fashion Myrdoch had originally suggested.
At first, I tried to follow his advice directly. One of the things he wrote about which seems obvious in hindsight was that crafting players have to stop thinking about what is useful and think instead about what is profitable. He advised people to look for for crafting procedures which would turn a profit even if all the ingredients were purchased through the player market - freeing the craftsman from having to spend their playing time farming for reagents. A good example of this would be the Onyxia Scale Cloak. These cloaks seem to oscillate around a median retail price of one hundred gold, and can be crafted in under a minute from a thirty-eight gold Scale of Onyxia, and a thirteen gold Cindercloth Cloak. I was sceptical that the game had [m]any items like that, so I wrote an addon named "Profiteer" to calculate margins and profitability ratios and display them in the tradeskill panel alongside the recipe names.
Despite being a little buggy, it showed some results - for example, that Dreamweave Circlets could be sold for something like 150% of their manufacturing costs. I purchased bulk amounts of materials and hit the bricks on the trade channel and Auction House. This worked well the first three or four times, but afterwards, sales slowed to a crawl. I found myself having to buy up cheaper competitors products. Ignoring the fact that the player market is elastic, I had saturated it for my timeslot, and I now had a lot of wildvine and truesilver to dispose of.
The lesson here was that it is hard to do sustainable business in the player economy unless you're selling consumables - durable goods are not good at generating repeat business. Some of you might remember posts by tailors early in the game complaining in a similar vein that the common practice of handing non-soulbound bags down to alts and friends was killing their market, this is a particularly obvious example. Trade in this market also requires a lot of patience as players level into and out of the item band for what you can make at a profit, and if you are trying to sell via the Auction House, those deposits will soon start to add up. The act of production itself also locks up materials, and while a compromise is to stand around in the cities offering to sell what you can make with the mats on hand instead of making first and selling after, it can be very expensive in terms of time.Sadly, the general rule is that trade ingredients in World of Warcraft tend to devalue as they are processed by players [ note - 14 Mar, 2008: I never explained this properly - the goods are devaluing analogously to 'opportunity costs', once a piece of leather is a headband, it cannot become any one of thousands of other and possibly more useful things that you can make from a piece of leather ] - the opposite of the real world, but a consequence of the lack of real skill, time or facilities required to craft, which leads to a glut of tradesmen with largely the same repertoire of products.
There is an obvious question as to whether we can do anything about it. Craftsmen could attempt to unionise, but there is little one player can do to compel another to sacrifice their own profits for the good of the community. Another approach is to try and establish supply agreements with certain game organisations such as raiding guilds. Finally, there is also a limited inelastic, unsaturatable market in the game: the vendors. Unfortunately, I don't know of any trade recipe that can be produced cheaply enough to profit on vendor sales using purchased materials.
The vendors seem to be a pretty unappealing demographic - they won't buy enchanting reagents and they pay out low. However, they have a few points in their favour
The secret to making a profit using the vendors seems not to be selling your own goods there, but selling those created by other people. This is called arbitrage.
A lot of older, traditional thinking about pricing, particularly Marxist thinking, revolves around "the labor theory of value" - that the labour invested by the creator of an object imbues it with worth and that it should set the agenda for pricing, wages, etc. When people argue for a higher sale price based on the amount of time they had to farm materials, or the quests or other activities they had to do to obtain tradeskill recipes, they are making an appeal to the LToV. However, there are other ways to create value which are not as immediately apparent.One is by controlling the flow of availability information. If my friend Alan is starting a canning plant and tells me he will buy all the bauxite I can supply at ten thousand dollars a tonne, and then my friend Chris tells me he is getting out of the bauxite mining business and that he wants to clear his stock at five thousand dollars a tonne, I could notify them of each other and collect a small finders fee, or I can buy Chris' stock and sell it on to Alan, netting a five thousand dollar arbitrage profit per tonne. People who do this kind of trading for a living are called 'arbitrageurs'. The ( fatally? ) confident can even pursue this kind of trade on credit.
In real life, arbitrage is hard. Due to such conveniences as the internet and telephones, it is very easy for buyers and sellers to make contact directly and with market instability and the amount of competition, "perfect arbitrage" ( no chance of negative cashflow, some chance of positive cashflow ) is almost impossible. Worse, failed arbitrage is not uncommon. In 1998, when Russia defaulted on the ruble and domestic dollar debt, Long-Term Capital Management, a major player hedge fund with two Nobel Economics laureates on its board, lost four point six billion U.S. dollars ( including a three billion U.S. dollar bailout organised by the Federal Reserve ) in four months attempting to perform convergence arbitrage of government bonds. Ouch.
Resale of equipment itemised for Shamans or Paladins "across the fence" via the neutral auction house, at least until the arrival of The Burning Crusade, looks like a good opportunity for arbitrage, but it is actually quite risky due to trade barriers erected by Blizzard. Not much of this equipment makes it to an inappropriate Auction House in the first place, with a lot just being sharded, selling directly from the neutral auction house attracts ridiculous costs of business and a very reduced market, and if you want to move items to an other-faction auctioneer for sale on the main house, you need either two accounts or a trusted compatriot - and are at risk of having your one copper auctions intercepted by other traders.
I have seen limited arbitrage of the following variety on Argent Dawn: a player sees a desirable widget on the Auction House for two gold and chats, "WTS - Widget, 4g". If they can find a buyer, they buy the object. I think we can all easily see the risks involved.
Luckily, the presence of the inelastic vendor demand market we discussed earlier opens the door to limited perfect arbitrage. You just need to convince people to sell you things for less than the vendor will pay you for them. This turns out to be easier than might have been thought. You can also practice "statistical arbitrage" ( on average profitable ) on the enchanting goods market if you have access to statistical data about what a given item can disenchant into and the market value of these fragments. This is of course a riskier proposition with correspondingly better rewards, but the fact that enchanting reagents can be listed without a deposit helps to mitigate the risk for patient traders.
So, why would players sell objects on the player market for less than the vendor price? Well, the vendor price of goods is not, by default, information that is available to players in the field, or for that matter, at the Auction House - it usually has to be provided by an external addon ( Wowecon, Lootlink, etc ) or by visiting a vendor and taking notes. If only one party in a transaction has this information, they are in a state known as "information asymmetry", which makes profitable trading much easier. Basically, a lot of people don't know what certain goods are worth, because they've never looked, or there's just too much to remember. They will also chase prices downwards to try and get bidders to choose their otherwise identical goods over those of other providers. The price ends up stabilizing and becoming habitually set at a value below that of the vendor market.If you don't believe me, I suggest looking at the prices of three frequently listed commodities
Fig 1., Frequently traded commodities, prices taken from WowEcon.com for Argent Dawn, Alliance Auction House, 28 October 2006.
Product Median Price # Samples Vendor Price Thick Spider Silk 6s76c 205 7s50c Shadow Silk 9s12c 345 10s Ichor of Undeath 3s80c 88 4s
Observing the prices of Shadow Silk when coming back from the Burning Steppes one night was the second catalyst for me to look further into these ideas. Note that interestingly, despite being white objects, these three are nearly all totally useless outside of a handful of quests and one or two low demand crafting recipes. The lack of demand makes these poor choices for auction sales, and the low median sale prices reflect this. People also occasionally trade cloth at prices which become profitable when the cloth is spun into bandages and then vendored. It's also possible that you may be able to cook raw goods into a more valuable product.
From looking at the prices above, you can see that generally the margins on these trades are small - penny stock stuff, but occasionally you do land a big win, and the sheer quantity of trades performed ( I place about three hundred gold of bids and buyouts per week ) means that the money does add up. You could make more money solo farming Dire Maul, but if you can write or download software to manage the large amount of information required to perform vendor arbitrage, this method only requires you to log in twice a day for fifteen minutes, during which time you can read the paper and have a cup of tea. The rest of your time you can enjoy playing the game, including supplementary farming, if you like. For the purposes of the experiment, I ended up writing two addons to assist, "hyperteer" and "baron".
Unfortunately for the person who sold me the item, they not only received less money than they would have at the inn, but the auction house then goes and removes another 5% of the value of the sale as commission. However, I want to be clear that the effects of traders like myself are not all negative - we do drive close rates and general auction prices up by outbidding people who are genuinely interested in purchasing the item, who will either place competitive bids over ours or buy the items out. Either way the seller still receives more money.So, this is the pathological trading that was discussed in the abstract. People will auction a good for below vendor price, or so close that the 5% commission ( I don't even want to discuss the neutral house commission - it is fiscal suicide ) annihilates their profits, while at the same time suffering the inconveniences, travel time loss of opportunity and delayed payments involved with utilising the Auction House.
People will offer goods which only have one possible disenchanting result for less than the price of resulting material - for example selling Mugs of Hurt for three gold at a time where they could be sharded into seven gold Large Radiant Shards ( this market has flattened in an interesting fashion since ) instead of making the one time thirty minute inconvenience of leveling their bank alt or another alt to level five to learn (Dis)enchanting. This is somewhat akin to selling envelopes with ten dollar bills inside for five dollars.
People will attempt to sell stacks of goods in awkward sizes, e.g. a stack of twenty bronze tubes or three sheets of silk cloth and then wonder why nobody bites. Who needs twenty bronze tubes who can't make them for themselves? Personally, I think reagents should be sold as singles, so the customer is not scared off by buying more than they need, but where you want to list stacks, it would make sense to try and list them in sizes which correspond to how the reagent is normally used. Software exists, or you can write your own, to split stacks and auction individual items for a fixed price.
An aside, but a pertinent one as it seems to be the main association with "WoW economics" is about people trying to establish monopoly control of commodities in the auction house. With substantial capital reserves, and an autobidder/lister, you can attempt to capture entire markets. I have dabbled with this, for example in attempting to bail out the price of Vision Dust, a staple trading item, which was slipping down towards a median sales price of fifteen silver. One of the more obvious problems is, if for example, you wished to reset the price of Vision Dust to thirty five silver, the Auction House cut ensures you would only be receiving thirty three silver twenty five copper for each successful sale at this new price. So, if you want to break even, you can't afford to buy up any stock that is already over this price.
Unfortunately, there will almost certainly be some auctions that fall into a gap between this profitable purchase price and your new selling price. Other sellers whose listings are in this thirty three silver twenty six copper to thirty four silver ninety-nine copper "deadzone" as well as anyone who was already selling at thirty five silver is now your competition in selling an extremely large amount of stock which is about to come into your possession: I ended up with something like fifteen stacks. The deadzone sellers will obviously be more appealing with their lower prices, and they provide a psychological impetus for other people to restart the race downwards. Of course, there will also be a constant inwards flow of new stock coming in from the outside gameworld as well, and buying out the competition will rapidly become prohibitively expensive.
Finally, let's have a look at the differences between the Horde and Alliance auction houses. As you might expect with the Argent Dawn population distribution ( about 2:1 ), the economy on the Horde side is somewhat smaller, with an average of 75-85 pages of goods as opposed to the Alliance tending in the 190-210 range. Unexpectedly however, the yields of below vendor price goods are disproportionately small. It would seem that the players of the Horde are financially better informed than their Alliance brethren. At the same time however, the economy in general and especially for enchanting reagents seems quite depressed. I can only assume that the lesser demand for enchantments is a more powerful factor than the scarcity of reagents caused by fewer magical items being disenchanted. If you want to make money on vendor arbitrage, you need to go where people aren't paying attention, and on Argent Dawn, that is not Orgrimmar.
In summary, an ignorance of the prices offered by vendors, and of the probable outcomes of disenchanting blue items by players using the elastic economy of the Auction House leads to a gap between market values in favour of the inelastic vendor market. This allows other players with knowledge of these values ( asymmetric information ) to perform risk free "perfect arbitrage" and moderate risk "statistical arbitrage" in the market for small, but not insignificant amounts of money. As Myron Scholes described the aforementioned LTCM, vendor arbitrage can be "a giant vacuum cleaner sucking up nickels that everyone else had overlooked."
The profits are real, but arbitrage, like any free lunch, is not without it's problems. While there will not be a market convergence from the vendor side, awareness by the Auction House community that arbitrage is taking place will be disastrous to it's future prospects. Opportunities for profit seem to shrink in direct proportion to the shrinkage of the information gap between arbitrageur and victim. Arbitrage also scales poorly - you'll rarely be able to find more than thirty or forty gold worth of viable bids on the Alliance auction house at any time. Also, arbitrageurs are a solitary breed, and the market cannot really support more than one, particularly in an auction environment where the majority of trades are non-buyout, without them eating each others profit margins to the bone. I know a few other players are wiggling their toes in the water - nobody else could want to bid on that many pairs of Nightscape pants at once unless it was for resale or disenchantment.
In that light, some readers might be wondering why I wrote this article in the first place. Well, I've reached my goal, and I'm ready to get out. Now, as I'll be switching back to a salesman of raw and processed goods, it makes more sense for me to work in a sane market where my competition won't price too low. I also have some more ideas about playing with Economics in WoW, which you may or may not see or feel the effects of in the future. Of course, the usual base motives of wanting to show off how clever I am also play a part, but I don't think what I've outlined here required any special insight.
"Lead Traders", Dancrat ( Alliance ) and Hayek ( Horde ) [these were my trading Alts].
for AedhanCorp [and Aedhan was my main].
( -- Bryn Davies, email@example.com )