What Killed the MMOG?

[The actual document will be preceded in italics by an explanation of what it is, and followed by a brief analysis, also in italics]

As a top World of Warcraft player in 2004, I became increasingly concerned about the "gold farmer" problem. I was one of the five most-read independent Western gaming journalists in 2004 and had access to many top game producers and designers world-wide. They all told me the problem had no solution. In 2005 I withdrew from the public and set about to solve this problem double full-time, hoping to solve it within a year. It actually took me 4.5 years so after nine man-years of research the result was the 34 page "Sustainable Virtual Economies and Business Models". It was written in six days at the end of August, 2009. Prior to now it has only been read by two USC professors, Mike Zyda and Henry Jenkins, who took the time to read the entire document at the end of 2009.

Dr. Zyda "valuated" the systems and methods contained in the document at a number that has never been matched so I have kept the contents secret. As the solutions described in the paper still have yet to be reproduced anywhere else, the technology is still "bleeding edge" eight years later. As my primary goal is to improve the gaming experience for consumers, not personal enrichment, not publishing the paper has put me in an uncomfortable position. The size is too large to be reasonably posted here on Gamasutra. I also have concerns that the methods I've developed could be used to harm the interests of consumers if deployed ruthlessly.

I have decided to publish one or more chapters here on Gamasutra, one at a time.  Two of the three "soft" chapters on Immersion and Cheese were already published on LinkedIn in 2011. What follows is the first of the remaining eight chapters which are much more technical. This chapter explains the problem in some detail, in preparation for the solution which is presented in the other seven chapters. 

The document is unedited for historical purposes.

Real Money Transfer (RMT)

For the sake of full disclosure I must state that I was a very successful RMT professional gamer back in 1999 in the game Everquest. The money was just so easy that I became concerned that this would lead to RMT on an industrial scale that would cripple game play. By 2000 I had contributed to a major mainstream article on the subject and had become possibly the world’s first anti-RMT activist. Given that I will try to objectively explain how RMT affects game economies and in later sections will discuss some effective countermeasures.

RMT became industrialized in the year 2003 by my estimation with the formation of large factories using thousands of computers aimed at the most popular online games (there weren’t that many back then). While these RMT clusters were scattered all across the globe, the biggest and most organized ones were located in China, as was the primary RMT hub company, IGE. Today the RMT industry has expanded to cover hundreds of products from several countries and has achieved remarkable levels of efficiency and profitability.

Most of the process is now automated with bots running on thousands of computers with minimal human oversight. A human operator will walk past the rows of computers and if someone or something is interfering with the bot, he or she will attempt to get it going again. Some operations still fully employ humans (some fairly young) working in shifts but that really is not all that efficient by today’s standards. In all cases coin is sold at set exchange rates. In many cases a “leveling service” is also available. This is almost always bot-driven. Some products also have account exchanges available where you can purchase a used account with all the trimmings.

RMT is not all bad so first I will discuss its merits. RMT exists because it meets a demand. Given how much money is being made in RMT, that demand is substantial both in the number of customers and the size of each transaction. By some estimates the money being made in RMT is greater than the amount of money being made by the companies that create these worlds in the first place. Exact figures are difficult to come by as this process occurs in a virtual shadow economy where money can be laundered quickly and effortlessly for any purpose.

Some numbers would help explain why this business is so lucrative. The cost to run a bot account in one of these worlds is by some estimates between $2 and $2.50 per hour. Since the accounts are optimized for profitability, they tend to bring in perhaps ten times as much coin per hour as a maximum level account played for entertainment purposes, and hundreds of times as much as an account at half the level cap or less.

So let’s set up a hypothetical monetary unit called the Casual Online Income Unit, or COIN. One COIN = One hour’s work for the average casual gamer in terms of income in the virtual world. A hardcore player might make 5 or even 10 COIN per hour. Let’s assume for the purposes of this example that the well optimized bot makes 50 COIN per hour. If the RMT company charges $5 per 50 COIN that is at least a 100% profit for the RMT company. For the casual consumer, 50 COIN represents 50 hours of work. If this person feels that gaining wealth in the virtual world is tedious or too time consuming, they can just buy COIN instead from an RMT source. If the consumer makes $10 per hour at their job this means with one hour of real work they can purchase 100 COIN or 100 hours of game work.

Clearly any time a consumer can trade one hour of labor for 100 hours of equivalent labor, it is perfectly logical for them to engage in this type of transaction. Even if the transaction was illegal, profits on this level will tempt many if not most people to engage in the activity. Common crimes in society like prostitution or petty drug dealing typically pay only 10 times as much as “legitimate” work or less and have much higher risks than engaging in transactions with RMT companies.

There is of course a downside, but this downside is a bit more subtle. The easiest part to explain is that when an RMT farmer goes to the places where the best loot is, whether that is a mithril node or a rare dragon spawn, they remove a good from the economy that most surely would have gone to a legitimate player otherwise. Real players will run around and wonder “where did that mithril/dragon go?” Or they may be standing right there when it happens and not have to wonder at all. This is the virtual world equivalent of a denial of service attack. They can log in but they are prohibited from participating in their entertainment of choice that they paid for. This won’t be true for newer players as only max level players will generally be in competition with max-level farmers for world goods. The RMT group can then turn around and sell you the goods they took from you or others for coin, and then sell that coin to you or others. This is the equivalent of a hacker group putting a virus on your computer that advertises a “better virus protection program” that will remove the virus they just infected you with if you give them your credit card number.

Another problem with farmers is that they maximize their production while minimizing their consumption. Whereas a player will attempt to maximize the abilities of their avatar by regularly upgrading their gear, a farmer will only utilize gear upgrades that cannot be sold. Farmers are typically the richest players on a server but conversely the worst-equipped. If you inspect them they look totally impoverished for their level. The reason this matters is because it causes hyperdeflation, a condition so economically radical that I believe it has never occurred in real space macroeconomics.

As an example let us assume you are the only person on a server with a horse or a battleship, and you worked very hard for this prestige, then you can feel a tremendous amount of pride and self-esteem from such an achievement. People with talk about you with respect and even gather around to watch you if you enter a public space. Having the only such item in a virtual world makes it almost priceless. Once a few more players have the same achievement, the value of the item drops a bit but generally that loss is filled by some camaraderie, and the new guys still know you were first. Once 100 players have this item it becomes common and now of relatively little value. Once everyone has the item it becomes almost worthless, especially if there is a better item that players are upgrading to.

When a farmer floods the world with an item you have, they strip your item of its value. This causes you direct economic, and probably psychological harm. If you find a ruby, at first you might feel a rush as you realize you are rich! If you go to your friend to brag and he pulls out a bucket filled with rubies and says “Oh yea, I got a bunch of those, do you want some more?” you are likely to feel instant ego deflation as you realize that a ruby in this world is a common item and that you are not rich. In the real world as I write this the economy is suffering greatly from deflation of housing prices which is quite painful to those already with houses.

Farmers also deflate the value of coin as they flood the world with it. In the real world you can do the same thing by counterfeiting. Germany tried to take out England’s economy in WW2 by counterfeiting their currency. If it had occurred long enough it would have devalued the pound to such an extent that its value as a means of exchange would have been lost and no one would have used it anymore. Nations will go to war to prevent this if they are not already at war, but this goes on routinely in current virtual worlds with developers claiming “they have no defense.” What a dire situation to be in to surrender without firing a shot. Defenses will be discussed later in this work.

The clever reader is probably thinking “But doesn’t this deflation make things easier for newer players?” If you need a horse purely for transportation, and you can buy that horse easier, then yes it will help you get that horse. Please remember however that that horse initially also had prestige value. Economic advancement is one of the two primary avenues of advancement in virtual space (the other being leveling). When you strip the prestige from ownership of items in a virtual world you are removing the thrill of economic advancement. Players quickly get bored with this path and will be left with leveling. As leveling to the cap generally does not take very long, this reduces the longevity of interest that players have in your game and your player base will dwindle. An abundance of cheap gear also makes leveling easier, reducing the time/thrill/prestige from leveling. This weakens the appeal of your remaining achievement path. Unchecked RMT attack on your world can turn it into an elaborate (and hopefully pretty) chat room.

A final point is that when money that a consumer has budgeted for gaming goes to an RMT site instead of the company that produced their game of choice, that money is removed from the development cycle. Some estimates have put this amount as high as 50% of the total spending by online gamers. This money does not go to paying developers, producers, stock-holders, and other investors and thus does not promote future product development or advances in interactive media technology. Some producers are aware of this problem but since they have already surrendered to the RMT industry, their countermeasures have mostly taken an “if you can’t beat them, join them” approach which has to date yielded limited if any success.

"Joining them" meant developers would not wait for third parties to kill their economies. They would do that themselves through the use of microtransactions and other methods. Thus what I would later describe as "First Generation Free to Play" would be adopted as the gold standard business model for online games. This class of business models, while having some commercial advantages over previous models, is not enjoyed by consumers. Product quality is reduced by their use as key challenge elements are undermined by pay to win elements. The original 2009 paper predicts this situation and explains the problems with 1st Gen F2P models. An example of a 2nd Generation F2P business model is contained at the end of the paper as it was determined that the choice of business model was just as important for MMO design as the economic systems employed. 

In 2012 I published my Supremacy Goods microeconomic model here on Gamasutra, which explained in detail some of the key characteristics of 2nd Gen F2P business models. I pointed out that two products, League of Legends (RIOT) and World of Tanks (Wargaming), had incomplete 2nd Gen F2P business models and I predicted that these products would easily outperform their competitors. Which they did. Even an incomplete 2nd Gen F2P business model is going to easily outperform a complete 1st Gen F2P based game. 

By 2014 I had expanded my stable to include 17 unique 2nd Generation F2P models. One of these was deployed to "World of Tanks: Blitz" (Wargaming.net) but was not 100% pure as I was not allowed to remove premium ammo from the design. Still, this "almost clean" design has performed very well as the top mobile F2P esports game for the last three years. A recent patch (I was not involved, I left Wargaming in 2015) altered the business model to make it more like a 1st Gen F2P product, and was almost universally rejected by consumers. I anticipate some sort of rollback will be attempted. 

Following my work on Blitz, I spent four months in Saint Petersburg working to deploy the prototype economic and monetization models for World of Warships (Wargaming). I was able to remove the last elements of 1st Gen business models from this design, making it the first "clean" 2nd Gen F2P model to reach market. If the assertions in my 2009 and 2012 papers were correct, it should outperform all equivalent 1st Gen and incomplete 2nd Gen F2P models. This is exactly what we observe with World of Warships having the world's highest global ARPPU of any F2P game. The product is slated to be expanded to mobile devices, where I anticipate it will outperform the previous leader in mobile F2P esports, WoT Blitz.  

None of these products are "Massively Multiplayer". As it increasingly becomes clear that 1st Gen F2P has run its course, developers and investors are looking to see what alternatives exist. Alternatives do exist. Consumer demand for MMOG's is still intense, but if powered by 1st Gen F2P or subscription models, they are almost certain to be commercial failures. Even the aforementioned RIOT and Wargaming products using incomplete 2nd Gen F2P models have been very difficult to improve because "what's missing" is not well understood internally. Using data science to run millions of A/B tests (the current industry approach) could still take tens or even hundreds of years to arrive at a solution without an understanding of the fundamental problems involved. 

Solving those problems was the goal of the 2009 Sustainable Virtual Economies and Business Models paper. MMOG's have not been failing commercially from lack of consumer demand. They fail because they are thrown together almost randomly (what I call "Frankenstein Style") without an understanding of the requisite systems for success.

The Wargaming production teams in both Minsk and Saint Petersburg that I worked with are very talented, cohesive, and hard working. In citing the success of the products that came out of these studios using my models, I am in no way attempting to diminish or take credit for their hard work.