Sunday, October 27, 2013

Spontaneous emergence of barter economies

This may be jumping the gun a little bit, not having fully described the features of virtual economies, but I think a more concrete example may be useful for those who wonder what types of events one could hope to study. In this post I look back to the release of the Trials of Atlantis expansion of Dark Age of Camelot, where for a relatively brief period the top end of a monetized economy devolved into a barter economy.

Prior to the release of ToA expansion, there were relatively few very rare items that were considered very necessary to compete. Those that existed were seldom available for money and the more common needs a player might have amounted to a large one time expenditure after reaching the maximum level to pay for equipment crafted by other players. The most common normal recurring expenses arose from deaths or from expenses related to sieges in Player vs. Player combat (buying siege engine materials and door upgrade/repairs), the latter being significant, but generally borne by a small subset of players. Players wanting to develop their skill in one of the crafting professions also needed to sink a fair bit of money into materials, which came from non-player merchants, but this was also a one-time expense that did not need to be paid as a lump sum. In short, with few exceptions there was little need to have a large amount of cash on hand.

Moreover, in those days Realm vs. Realm combat, the principal activity for many players who had reached the endgame, was not a good source of income. The amounts of money earned were trivial. Later changes addressed this, but for the most part in pre-ToA DAoC acquiring significant quantities of money was a very deliberate act.  

The ToA expansion introduced a powerful new type of item, artifacts, which required among other things that a character turn in a set of three scrolls in order to be able to use. Some of the scrolls were very common and easy to obtain while others, due either to their high desirability, low probability to drop, or obscure and out of the way source were fairly rare. In the initial phase, these were quite valuable. For a few months, it was very difficult to buy these for money. Instead a limited barter economy arose involving just the scrolls considered rare.

The rarest scrolls came to be exchanged 1 for 1 as no character needed multiple copies of the same one, except as a potential trade item. There were few or no other items of similar value and thus it was difficult to arrange any remotely fair trade around them. Thus the rare scrolls never had the opportunity to truly become a second currency.

In the end the system was unstable. Demand for the rare scrolls was not infinite, as each character needed them only once so only growth in demand came from new characters nearing the end in their gear progression. Supply grew much faster. Thus as we expect, the value of each scroll fell and the list of rare scrolls that were difficult to obtain for money or exchangeable 1 for 1 for other rare scrolls gradually shrank as the less rare ones hit the tipping point where enough people could simply offer money for them.

Longer-term changes to the game made RvR a net profitable activity and the average player could be expected to have a lot more cash on hand at any given time. New mechanisms for scroll acquisition were also later added that significantly reduced their rarity, but by the time both of those changes happened the barter system was long dead.

The interesting question is: Why did the system arise in the first place? There are several factors that I believe contributed. The pre-expansion economy did not have any items whose value was comparable to new ones and most characters were well established and had much of what they needed. The need to have money in that economy was limited to characters who had freshly hit the maximum level and incidental expenses for everyone else, which were relatively minor. Together with the unprofitability of the dominant activity for much of the player base that had reached the endgame this lead to an economy where there was not a lot of money lying around waiting to be spent.

With the ToA expansion, there was a variety of new items introduced to an economy where most participants did not have enough cash on hand to buy some of the most expensive of the new things, but they could find them directly. Moreover, the items in question were similar enough that a 1 for 1 exchange between different rare scrolls seemed fair and natural despite some differences in their actual rarity that would have been evident had they been bought and sold for money instead. Had the new must have items been more different, perhaps no barter system would have arisen.

The events in question happened almost ten years ago now, so it would be unlikely that complete logs of trades performed in those days are still around for study, but the nature of the changes that gave rise to this temporary, limited barter economy have doubtlessly been reproduced in other games. With better data some of the anecdotal observations such as the pre- and post-expansion distribution of money among the player base could also be quantified.


Game expansions frequently add highly desirable items and commodities to a game (making owning the new content “necessary” to compete helps sales) and assuming this happened because the game’s pre-expansion money supply inadequate to deal with the new goods available rather than some other idiosyncrasy inherent to DAoC, similar circumstances should be expected produce similar results.

Saturday, October 26, 2013

Currency Basics

One of the most obvious features of virtual economies is the currency or currencies used. In most one currency is the primary currency in which the vast majority of inter-player transactions occur. In a handful, multiple currencies are used concurrently, which can lead to interesting dynamics.

Typically in the most basic currency in the game will be represented by something setting-appropriate like gold coins in a fantasy setting. It will be a strange hybrid between a fiat currency and one backed by commodities, in that it is generated by in game activities like completing quests or killing monsters, but the rate at which it is created is controlled by the game developer. It’s as if there were a country with a gold-backed currency where the government could set the effectiveness of various extraction techniques at will. Technically, it could also seize or issue more of the currency at will too, but the outcry from players this would generate tends to minimize the risk of that.

A second common type of currency is one purchasable with real world money at a fixed rate (With very few exceptions the reverse transaction is not officially allowed). Its most obvious purpose is to effectively facilitate microtransactions by separating the time one has to go through the process of entering payment information and depositing money and the time one buys an item in game. In some games the real-world purchasable currency is tradable among players and/or exchangeable for the more basic currencies. This can allow it to become the de facto primary currency in which transactions are conducted even though in the story of the setting, non-player characters pretend like the more basic currency is important.

A final common type of currency in virtual economies is a useful, valuable item for which there is a steady and widespread enough demand that most players will accept it in exchange for other goods because either they wish to use it personally or trust that it can be traded for things they want later. Typically these are not official currencies, but ones that have emerged for various reasons, which makes for some interesting case studies.


Introduction

Although economists have many theories about macroeconomic phenomena, they do not generally have the means to perform experiments in the real world. The only courses of action available are to construct theories and models or to wait for major events or policy shifts that can serve as a natural experiment. The latter are fairly rare and generally dictate what can be studied, while the former are only as useful as their assumptions.

The virtual economies present in Massively Multiplayer Online Games present a perfect opportunity to study economic interactions on a relatively large scale, which might be difficult to model. Many new ones are created every year, and most feature at least some form of rudimentary economy. Not only does the multitude and variety of game designs and player populations allow one to learn about the functioning of economies, but the detail and completeness of data one could collect in an online game where every exchange and most interactions between players can be recorded is simply unmatched in the real world.

While there also exist a number of single player games that contain simulated economies and multiplayer games which have currencies that can be earned and spent only on non-tradable items, this blog will primarily concern itself with games that feature at least some basic form of an inter-player economy.

There are, of course some caveats. Care will have to be taken in each situation to determine the applicability of the observations. Multiple experiments and complementary real world observations may be needed to truly validate findings.

Some situations have no real-world analogues or are at best similar to very rare occurrences and thus have few practical applications beyond the realm of game design. Some players will almost certainly behave differently in a game than they would in the real world, and sometimes behavior learned in similar previous games or real life may influence interactions in new games. Finally, unintended consequences of changes to the game world might disrupt the economy in ways in which while potentially interesting in and of themselves may complicate the answering the original question being studied.

In the following posts, I will attempt to offer a primer on the similarities  and differences between online and virtual economies, speculate on some past observed events, and offer some theories that could be tested if similar situations arise in the future and data is collected about them.