Census Addon Complete…

July 2, 2010

… more or less. It’s actually not that hard, and it served as a great introduction to the more complicated addons I’ll have to put together. The next on my list is the auction house scanner. This will take some more exploring.

In the meantime, I’m taking a completely random sample of the game world by zone. This is a ‘first step’, and it includes unlikely places such as old raid zones, arenas, and five-man dungeons – such as Scholomance and Botanica – that are never used anymore because they were designed for the endgame of the previous expansion (or the original game).

From this, I’ll be putting together a list of zones to sample that will vary with server time (probably by the hour, but perhaps only by every 4 or 6 hours). In particular, to identify players it is only useful if an average of 2 characters is observed in a zone at any one time. So i’ll probably exclude zones from future samples if they average less than two characters, and especially if I never see more than 2 characters there. I’ll also have to design a search-text parsing algorithm in order to break out the population in cities, where the number of characters in-zone is greater than the social interface maximum of 49.

Search terms: rearing their ugly heads again

June 17, 2010

In my last post I outlined the three addons that I will be designing. Recently I’ve focused on designing the census addon, because before I can write it I need to be able to determine the frequency and breadth of samples as well as the total time to run my experiment. This is a vexing problem. I commented a long time ago on the importance of finding appropriate search terms before you can get your project off the ground. Today I had a minor breakthrough after throwing up my hands and consulting a former colleague at the FTC. I had wondered whether there weren’t situations analogous to mine, where one can view a characteristic of interest directly, but not its underlying source. He gave me a vital moment of inspiration when he pointed out that web analytics faces a similar problem when attempting to identify unique visitors to a webpage. After 45 minutes of scratching in the googly dirt, I hit upon a search phrase that produced myriad valuable results. The phrase was: “Unique visitors” probabilistic.

WHAM! That did it. I uncovered a slew of papers on a matter that is mostly of interest in database design: probabilistic counting.

This is some intense stuff. It will take a while to comprehend. But the main theme of this post is not the discovery, but the search! Added to my previous strategies for uncovering search terms, one must include “Thinking of analogous situations”. If you’re stumped on a problem, chances are high that someone else has been in a similar spot and come up with a cool new way of dealing with it. Analogous situations in fields that are far-flung from one another show up all the time, but their solutions are grounded in the same mathematics. By searching for papers on those analogous problems, you may find some very helpful ideas in designing a methodology to deal with your own.

Scraping Warcraft Data: First steps

June 9, 2010

It’s time for me to start writing my addons. I need at least two, possibly three. I have downloaded the API, but I am a novice programmer, so I’ve also purchased a book that I hope will get me to the skill level I need.

So, what addons will I be creating? The first one is an auction house tool. The second one is a census-taker. The third one, if/when I get around to building it, is also a census taker, but a very specialized one.

Auction house tool: This would be a totally stripped down version of Norganna’s Auctioneer. What we are interested in is just scraping everything – including character name – and storing it to a file. We can take this data and save it for later analysis. It needs to do this as frequently as possible. Because we want character name, the addon will be more complicated (probably the most complex I design). Otherwise, we could makes use of the much simpler Blizzard-defined function QueryAuctionItems().

Census-taker addon 1: This addon writes /who output to a SavedVariables file. It will scan the major cities and other known high population areas for characters and append the results to a file along with a time stamp. It will scan as often as the game will let it. That’s it. It can also be used for initial surveys to determine what constitutes ‘high population areas’.

Census-taker addon 2: This addon is a little trickier, and may not ever be implemented. What we’re looking for here is an addon that allows a correctly placed character to “observe” individuals who are most likely utilizing the auction house. The addon scrolls through friendly targets in the character’s field of vision, and for each target it records the name of its target, the target’s level, the locale in which the target is observed, and a time-stamp for the observation. Designed properly, this addon affords us two augmentations to our study. First, it can be used to increase the total number of characters seen at any one time in populous cities like Stormwind and Orgrimmar. Second, it can be used to capture auction-cancellation effects.

This latter point is a major obstacle. While my WoW player’s instinct is that few auctions are cancelled, the truth is that a character’s likelihood of cancelling an auction is probably positively correlated with their participation in that market. I haven’t yet worked through the preliminary math on this, but the basic strategy rests on the realization that if a bunch of auctions disappear between scans, and the player whose auctions they were is at the auctions house, and [insert some sort of decision-making calculus here], then there is a high probability that the missing auctions weren’t purchased, but rather cancelled. In the absolute worst case scenario, we can just ignore any sales that take place while the seller is online, but this is overkill. The point of this addon is to try to finesse things a little more.

Player as firm: The unit of analysis

June 7, 2010

An important part of the research process is identifying the smallest feasible unit of analysis. So what is the unit of analysis in WoW? Well, it’s the player! This may seem obvious, but to my knowledge it has never been explicitly recognized in any of the economics work I’ve seen. Let’s take a moment to understand why.

WoW players often have multiple characters on a single server. The reasons for this are many, but there are three that are most important to understanding the WoW economy. First, players like to create bank and auction house characters, and often these roles are mixed. The bank character becomes a store of cash and valuable items that it can quickly mobilize to participate in the economy. For some players who particularly enjoy dealing with the economy, the bank alt is more like a main. Second, players like to have access to multiple professions. Not having to tip other players for their, say, enchanting or blacksmithing skill, can save a little cash. (This is notwithstanding that the cost of actually learning and levelling the tertiary professions could be quite high for the player). Lastly, players who want to experience a different perspective on the game will most likely create alts on the same server as their more powerful high-level characters. That way, they can participate in a sort of nepotism with their alts, moving them through the game more quickly.

What should we make of this? A player’s characters on a single server are jointly engaged in production, and each has access to the same information about the market. They constitute the productive units of a ‘firm’. (One could also argue that they are individual firms in perfect collusion with one another, but I think this overly complicates the picture). The CEO, board, and shareholders are all embodied by the player. Players, not characters, are the smallest productive units in the game economy. On this view, competition within the economy should be viewed as taking place between players (“firms”) who have interests in multiple sectors of the economy. The organization of firms is heterogeneous and occurs on a spectrum: at one end are players who do not engage in a lot of inter-character interaction, as if they were wholly-owned subsidiaries of a conglomerate. On the other end are players who actively engage their characters with one another to meet a single economic end, and this would include both gold/item farmers and players who get the most fun out of participating in the economy.

We are left with several practical and theoretical questions. On the practical side is the very real problem of identifying players when you can only directly observe the actions of single characters. This is not easy, but there are ways to do it. On the theoretical side, a whole host of interesting research topics presents itself: How are player-firms organized? What is the nature of competition in the economy? How economically organized are guilds? What role does arbitrage play in the economy? How influenced are players by day-to-day fluctuations in the market? I hope to take up these and other questions in future posts.

WoW account hacked: Resolution

June 7, 2010

I sought and executed many strategies to find and remove the keylogger I suspect found its way onto my PC, but ultimately I had no choice but to nuke the drives and start from scratch.

I ran scans using Windows Defender, Microsoft’s Malicious Software Removal Tool, Malwarebytes, and two types of AVG scans – one after booting windows, and another using an AVG Rescue CD to do a pre-boot scan. I found nothing.

In the meantime, Blizzard got back in touch with me and restored my Battle.Net account to my email address. Using a clean (I hope!) PC, I immediately created a new email address with a completely new password, and then switched all my account information to that email address. I will probably break down and get an authenticator as well.

An altogether interesting, frustrating, and time-consuming interlude in the broader mission to conduct economics research in WoW. Serves to remind a person that although many companies are working to incorporate RMT into their business model, there’s still plenty of demand for the third party kind – and plenty of people using unsavory methods to supply it.

WoW Account Hacked, Redux

June 3, 2010

…. and as if that weren’t enough, my attempt to log on to my WoW account was again thwarted last night, for someone had just changed the email address associated with my Battle.Net account.

According to CNET, there is evidence that thousands of players have had a keylogger installed on their computers. The keylogger is apparently selectively targetting WoW accounts only.

This is the only plausible explanation. The password I created yesterday for my account was very strong, and there were no unauthorized users logged into my email account this time. I scanned my password with no less than four anti-virus/anti-malware scanners last night, but found nothing. As I understand it, this means the thing is probably working at the kernel level, which means that my best bet is to completely wipe my hard drive(s) and start fresh. Oh joy.

Well, I suppose I’ll just continue to fight the good fight. Updates forthcoming when I have more to go on.

Update: I should point out that the hack of my account coincided almost perfectly with my download of the WoW Addon Studio. I’m not sure if there’s any actual chain of causation, but the coincidence is suspicious.

Update: I am now getting ready to run a LiveCD antivirus and antimalware scan of my PC. Results to be posted when I have them in hand.

WoW Account Hacked!

June 3, 2010

My WoW account was hacked early this morning.

The story:
After waking up this morning, I got on the gmail account associated with my WoW account and saw two very unsettling things:
1) A big red warning at the top of my inbox stating that my gmail account had been accessed from China the night previous, and
2) An email from Blizzard notifying me of a password change on my account.

I immediately attempted to log into my WoW account to no avail (wrong password). Whoever hacked the thing didn’t change the email address associated with the account, so I immediately changed my Battle.Net password and tried to log in again. This time, a box popped up requesting a six-digit number from the new authenticator keys Blizzard has released. Except I don’t own one of those; the perps associated one with my account to keep me from logging in while they were doing their thing. Very clever.

Apart from the 45 minute wait to speak with a Blizzard Rep, the resolution was more or less satisfactory: they’ll be restoring my account as best they can (they claimed 95-99%) as soon as they investigate the case, which takes 1 to 2 weeks.

Let me just say that I never participated in RMT, have never shared my account info with anyone, and have never succumbed to any phishing scheme. I did, however, briefly join a vanilla WoW private server using the same email address and password. I can only assume that this is where they got my account info… but who knows how they got my gmail password. Admittedly, it’s a personal email account – nothing important besides my WoW account info was transmitted to it – so the password wasn’t very secure.

Anyway, its a minor interlude that can only serve to remind researchers of a very important step when collecting information: BACK UP YOUR DATA!

Why WoW? A note on world choice.

June 1, 2010

At State of Play VI Dimitri Williams received applause when he stated emphatically that “You can not study virtual worlds if you do not use them,” – a point reiterated in his paper on mapping and which I discussed with him at Terra Nova.

As noted in that last link, I am skeptical of this point. However, I think I understand why it received such praise at SoPVI: serious gaming has long been the object of perceived attacks in papers on media effects. Nonetheless, if you’re getting ready to be a serious scholar with a focus on MMOs or VWs, you might get discouraged that your work will not be taken seriously if (again like me) your experience in them is limited.

So: Why WoW? My experience is limited solely to playing World of Warcaft. Its where I intend to perform my studies. I have tried other games (EQ II, Final Fantasy), but none has ever drawn me in like WoW, and this is mainly because I play WoW to socialize with my brother, who is also an exclusive WoW player. Because of this, I know its history, I’m intimately familiar with its trends and mechanics, and – most importantly – I understand its economy extremely well. I have thought of creative ways to get at the data I need, and given careful consideration to the many and sundry pitfalls in my way. Your world of expertise may be different than mine.

Why not WoW? From an economics perspective, there’s no good reason. There’s nothing in my toolbox to make me think that gamers, once in their particular world, will react to incentives in functionally different ways between worlds. Some might argue that WoW is not a rich enough economy to attract serious study, but I disagree that richness of environment qualifies a world for study. Rather, its all about the richness of data and the ease of acquiring it. And it is data, ultimately, that an economist needs in order to study something.

Finally, a brief note on applicability beyond the game. One might question the inevitable heavy focus on WoW in games research because the results will have some sort of unobservable WoW-bias. Maybe this is true (though I am skeptical). Do not be paralyzed by this concern, however. Let others repeat your experiments in other games and try to reconcile their results with yours. In the mean time, priority one is producing research with robust results that merit publication in your field of study.

Virutal economics revisited

May 31, 2010

I started this blog a long time ago, in part to serve as a forum to discuss my research in World of Warcraft. In my first serious undergrad research paper, I explored the effect of the mass account purges that constituted Blizzard’s first major foray into defending its economy against RMT. I found that price levels on the auction house dropped precipitously in response to these bans; that is, the economies experienced mass deflation.

With notable exceptions, research into virtual economies has similarly focused on macro-phenomena, and this seems to have followed from the belief that such research serves to position VWs as media for policy research. The macro focus seemed to be a consequence of the fact that it was difficult to get microlevel data, but this problem has diminished: Dimitri Williams, who was given amazing access to data from Everquest 2 data by Sony Online Entertainment, worked with Edward Castronova to produce a paper on the stylized facts of the EQ2 economy. Yet while this has served as yet another interesting think-piece in the limited bibliography of virtual economics studies, it has gone basically unnoticed in the field of economics. In fact, the best papers I have read so far on the economics of virtual worlds have been produced by undergrads (here and here). An honorable mention must go to a high school student who is now an undergrad at MIT, and whose ambitious paper, though flawed, evidences a remarkable amount of careful thought for a high school senior.

What makes these papers valuable is that their authors are not trying to shift paradigms, but rather to explain in precise detail how virtual economies actually function, and the ways in which player participate in the markets they create.

In fact, the most important paper on economics produced by the Gods of Game Research is not an economics paper at all, but rather an exposition of Dimitri Williams’s mapping principal, in which the author lays out the pitfalls that one faces when entering virtual worlds to perform research. Williams thesis is that researchers must carefully establish the way in which activities in virtual worlds map to those in real life before drawing inferences. From an economics perspective, this mapping has yet to be carried out in any systematic way. One of the issues an economists face, for instance, is establishing the way in which a player’s participation in markets for virtual goods maps figures into their production of ‘fun’ – that is, their utility function. Absent such information, any study that seeks to study macro-phenomena can not establish a map to the real-world economy, where participation in markets is fundamental for existence.

What appears to be missing in the game studies arena is a set of individuals with sufficient amounts of micro-level market data and the will to analyze it. With little to no help forthcoming to the average researcher from the companies themselves, the researcher is forced to produce the data on his own – a very costly venture in time and money.

As I once again seek to jumpstart this blog and define its direction, I will post further on interesting research questions and on methodologies that will allow research economists to enter the field at a relatively low cost. I look forward to producing a blog that is useful as a research diary and forum of ideas for other interested individuals.

A brief note on experimental design

February 26, 2009

On the Metro ride home I was just starting into an article (on matching estimators, if you must know) when I had a moment of inspiration that got me around a major blockade standing in the way of my hopes for a research project.

I won’t get into specifics on the off-chance that the solution I “discovered” gives me a competitive edge in the short run, however:

Experiments in social sciences are difficult because you can rarely control all of your variables – i.e. the humans whose behavior you are trying to understand. Nonetheless, we often run across situations that are very nearly controllable experiments except for some minor detail which, lacking the means to control for it in some way, renders our scientific foundations too shaky to merit the pursuit (or, at least, the avenue of that pursuit). The way around this problem (if such a way exists) often eludes us because we focus on what we see in our frame. Stuck in that state of mind, one is prone to forget that a subject can not affect an experiment if it is demonstrably absent from the experimental frame while measurements are being recorded.

It’s ceteris paribus in action. Except sometimes you realize that you don’t need to hold everything else constant by sheer statistical will and force of data. If you’re paying attention, you will see that sometimes things will hold themselves constant for you, and you just have to show how they did that.