Virtual Economies: Is less information better?

In the first of his promised weekly reports, Yanis Varoufakis, Valve’s new economist-in-residence, released his first analysis of a Valve game economy on Friday. In his inaugural post, Mr. Varoufakis analyzes the persistent opportunities for arbitrage in the game Team Fortress 2. I rather enjoyed the in-depth, data-driven discussion, which also raised an important question about the game industry’s exposure of internal economic data - the main focus of this post.

To summarize Varoufakis’ analysis in brief: Valve’s popular free-to-play, tongue-in-cheek first person shooter Team Fortress 2 (TF2) allows players to barter with one another, trading hats and guns for keys, earmuffs, and and sundry other virtual items. Mr. Varoufakis wondered whether there was evidence that this economy was anywhere near an equilibrium. If so, then there is no room for arbitrage, i.e. there’s no way to “sell high” or “buy low”. To put it another way, if the TF2 economy were in equilibrium, then every hat would be worth exactly x number of keys (say), no matter who you traded with. TF2’s economy, it turns out, is nowhere near this happy state.

There are two  related sources of persistent disequilibrium here: one is the barter economy on which trade in TF2 is based, and the other is a lack of reliable information on trades. Mr. Varoufakis’s post includes a detailed discussion of how barter depends on a “double coincidence of wants”: You need to have what I want and I need to have what you want and we both must be willing to give those things up to one another. But when there are 35,000 items to choose from and trade, as in TF2, the composition of bargains can be quite complex. Even if information about every trade were available, it would still be hard for the average player to utilize it. Nevertheless, players are nothing if not innovative, so maybe they could find a way to take that raw information and make it useful. And so Mr. Varoufakis concludes with a question to players and readers: Should Valve tell you what TF2’s arbitrage opportunities are in real time?

Though aimed at players and readers, what Mr. Varoufakis has really posed is a fundamental question about the level of virtual economic disclosure that is most valuable to Valve. It’s a question that other successful game publishers have struggled with, as well. Blizzard-Activision broke ground last year when it started a community API for World of Warcraft that exposed massive amounts of information about the game world and the avatars in it. Yet information about Azeroth’s economy is limited to hourly snapshots of supply schedules -quantities and offer prices of items currently up for sale. With the release of Diablo III’s highly anticipated real money auction house (RMAH), community developers were expecting Blizzard to at least expose the same information they could get about the WoW economy. But they have been sorely disappointed. There is to be zero exposure of data from the RMAH. Meanwhile, CCP has also released historical information for its game, EVE Online, but only average prices per item per day, and it will not expose anything in real time. Sony Online Entertainment has made detailed economic data available only to a few researchers, who are bound by an NDA.

I've been able to find two arguments against this exposure. One comes from the game publishers, and it goes something like “Exposing sale prices would increase the amount of data that we would have to host, increasing storage, management, and bandwidth costs.” Fair enough, yet there’s an absence of compromise here: Could this information not be exposed for 24 hours, and then deleted?

The other group is the small cadre of players whose favorite past-time is playing the game’s economy. Whatever these players’ money-making methods are, economic success ultimately comes back to exploiting opportunities for arbitrage, and the greatest threat to arbitrage is a well-informed market.  Their ability to profitably invest in particular markets would be severely hampered if information about going sale prices were actually available. Since these people are paying players and get their fun from playing the market, their (real) money is on the line for the game publisher.

Methinks, though, that there is some other, more subtle relationship driving this opacity, and I think it ultimately comes down to a not-so-obvious fact: Game publishers rely on enterprising players to mitigate the ups and downs of their virtual economies, and to recognize and provide for other players’ wants.

See, these market players don’t actually go out and farm for trade goods and other virtual items themselves. Instead, they create the incentives for other players to do so, and they expect a return on their investment. For example, by buying cloth when it is priced too low and posting it for higher prices, wealthy players encourage others to go out and farm cloth themselves and increase supply. Wealthy players may also take long-term positions in heavily traded items, so that they can take advantage of any price spikes. Such investment is incredibly important for major game events, like patches, which often significantly change the value and availability of certain items. And then there are the multiplicative effects of the actions of wealthy players: Every gold piece they spend is spent again several times over, raising the fortunes of everyone in the economy. There might even be a "fun multiplier": the market player has fun looking for opportunities, and the players who supply the market go out and quest, hang out with friends, etc., which is incidental to their collection activities.

So wealthy players ensure goods come to market, increase the fortunes of others, and prevent prices from spiking. But it’s not like the game company can’t perform this role instead… or is it? Early versions of Ultima Online’s economy (and Origin's business) suffered mightily at the hands of players whose vision of fun was different from what the developers intended. In fact, this seems to be one of the cardinal sins for game development. Developers are tool-makers and story-writers. They provide means and premises and environments. But there’s no way they can guess the outcome of the behavior of hundreds of thousands of individuals. To think otherwise is have the same pretensions as the architects of the Soviet economy.

Still, to justify an information blackout, wealthy players must be worth more than their subscription prices and item purchases to game publishers. Are they? I'd bet they are. Dmitri Williams is doing some cool stuff with his analytics engine over at Ninja Metrics that works on the same premise. Specifically, he has come up with a way to measure the value of players vis á vis their effect on the experiences of the other players that they come in contact with. For example, griefers reduce revenues by discouraging players from staying, while friendly or popular players increase revenues by encouraging other players to stay.

What I’ve been talking about falls exactly into this sort of groove. Wealthy players increase revenues by making it easier for players to access the things they want that make the game fun for them as individuals. So they increase revenues. But they also keep costs down for the publisher by recognizing impending shortages and getting the game economy to respond appropriately without any direct intervention from the publisher.

And so we return to Yanis Varoufakis's question: Should Valve make information on arbitrage opportunities available? Well, based on the preceding analysis, perhaps not. But there are two vital differences between most game economies and that of TF2: 1) it is a barter economy, and 2) game play is not vitally dependent on that economy. But markets are finicky things, and even changing this one little variable could mean millions of dollars in revenues for Valve. But will those millions in gains or millions in losses? Rest assured that Valve's billionaire-in-residence will have some thoughts on this one.


Virtual Economies: Is less information better? by Isaac Knowles, unless otherwise expressly stated, is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.

5 Responses to “Virtual Economies: Is less information better?”

  1. leTjuhl says:

    There's to ways to approach this:
    1) full disclosure of all trade information
    2) full nondisclosure of all trade information

    And it's either or. Disclosure or nondisclosure of information would include not only pricing, but also trade participants, time of transaction, willingness to pay, anything that can identify buyer and seller.

    Option 1) will make the game less game-y, as players get full economic efficiency feedback on how and what they are doing in-game. That will move playing the game more towards "work" and steer it away from you as a user "playing" it, intruding into the magic circle of play, but on the other hand allowing the user to determine his free-time earnings.

    Option 2) will make the game more game-y, as players would need to learn to play the economic portion of the game, correctly guessing and estimating price points, "playing trade" with uncertain yet also potentially rewarding outcomes.

    More info how option 2) could work technically / procedurally here:

    • I think I understand, but I don't think you're accounting for one important point: Virtual worlds are products that people can choose to stop consuming.

      If you go with option one and require full disclosure, the result will be efficient within the context of the game economy, but it will not be a profit-maximizing choice for the company that creates the game. That's because people who like the economic game will leave (lost revenue source 1), reducing supply of virtual goods and causing others to leave the game (lost revenue source 2), and increasing costs. This was the thrust of my post.

      If you go with option 2, per your link, I think you will just see frustration on the part of consumers because they can not immediately acquire that which they are willing to pay for. More than that, information blackouts can be just as discouraging to people as full disclosure. You're also assuming people want to play the economic trade game. But most people don't. Most of them want to play the raiding game, or the pvp game. The economic game is incidental to them. They will not start enjoying it just because they have to. They'll just quit.

      There are infinite alternative disclosure rules that could maximize a game company's profits, not just two. Which one is best for that game is going to depend on the very subtle relationships between the game economy and the rest of the game systems.

      • leTjuhl says:

        Isaac, thanks for your reply. "I think you will just see frustration on the part of consumers because they can not immediately acquire that which they are willing to pay for." IF users should be able to immediately acquire what they are willing to pay for is a decision to be made by the game developers, as they are providing the base system. There are reasons not to, e.g. the game developers deciding to fully protect (Huizinga's) magic circle of play or developers being interested in the effects of a undisclosed trade system on the virtual world / player society. An undisclosed trade system could be monetized as well, e.g. via conversion rates favorable to virtual currency purchased for real money than virtual currency acquired through gameplay.

        • I agree that the properties of the game are the decision of the game developer. The game developer could just as easily require players to stand in line at the bank to access their inventory. And I'm sure there is some group of potential players just waiting to play this waiting game. But unless there are enough customers who want to do that instead of playing Maple Story or WoW or even Cow Clicker, there is no game developer. And the question is whether or not a mechanic of complete non-disclosure and waiting to trade in the virtual economy is something that will be sustained by the real-life economy. I think not. And if not, then you can not monetize it, either.

          Beyond all that, a virtual economy is not just a virtual economy. It is also a way to deal with systems design problems. It compensates for balance issues in other realms of the game by allowing players to apply and view the values of stats, time, and trade skills. For that to work, players need to know how much things are worth, and they can't do that under non-disclosure.

  2. I was reading this the other day and I think it presents an interesting case for not having perfect information in an economy. Every time I play games that have something like this (i.e. Galaxies, FFXIV) I can't help wanting to murder myself. The problem I always run into is that they make complex economies then leave this huge coordination problem for the people who just want to get on with the game. - I want a sword.... why can't I find a damn sword. I'm sure there has to be an economy that exists between WoW Auctioneer and what Raph is talking about. What would it look like and would it be fun for producers and consumers?

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>