« South Korea | Main | World of Madden-Craft »

Jun 20, 2006



I haven't played much as alliance in the last month but on my server (Hyjal), though there was some activity in Stormwind, it seemed to still be Ironforge.

I've played more as horde, including just hanging out trying to get people to buy enchants, and I can say the main horde city is still Orgrimmar. I tried setting my hearthstone to Undercity and after the third time I ran to Orgrimmar to sell an enchant or buy an item, I set it back.

Note that the horde has much more reason for Undercity to become "the" city, since the most popular endgame instances are on its continent. But it's not.

This is a classic case of network lockin. Since some interactions require personal presence (selling items quickly without an auction-house fee, crafting, enchants, unlocking), the place to be is where everyone else is. And the place where everyone else is today, is the same place as yesterday. The horde hasn't switched to Undercity for the same reason I'm using Ethernet instead of Token Ring.


While the expansion of auction houses might not have change which city is used for trading, I believe it still may have had some effect. Instead of depending on what other people may have noticed try running a census program in the main cities at peak time. Too bad you don't have data from before the multiple auction house patch because then you could really prove your point.

Even if your theory isn't correct WoW still is a good tool for social experiments.


Something similar might have happened in pre-NGE Star Wars Galaxies.

SWG was a particularly interesting place for economic study, given its relatively rich crafting game and fully player-run economy. In the first months of SWG, traders could be found trying to do business in all the NPC cities.

Within a year, virtually all trading had coalesced to the open areas just outside the starports of the major cities on the major planets: Theed on Naboo, Coronet on Corellia, and (to a lesser degree) Bestine on Tatooine. For a trade-oriented player, you simply had to go to one of these locations -- that's where the buyers had learned to go, and the buyers went there because that's where the traders were.

By the end of SWG's second year, trading had collapsed down to one location: Coronet. The common explanation (among players, anyway) for why this happened boils down to two beliefs:

* As player populations declined, there weren't enough players left to fill up three trading locations.
* Nearly all direct travel routes between planets went through Corellia.

It's not clear that these two reasons really explain why trading narrowed to Coronet. For example, although most travel lanes did run through Coronet, the release of the Jump to Lightspeed expansion meant that players could fly their own ships wherever they wanted instead of having to rely on the public transportation network.

And yet, while I don't know what happened on every server, it seems that this pattern was repeated on most of SWG's servers -- all trading went to Coronet. That indicates that some force was operating, or some information was available, to cause rational actors to reach similar conclusions about the likely rewards for specific economic behaviors.

That SWG no longer exists, but I hope there will someday be another game world offering the features that reward economic study: deep crafting and sales subgames, a fully player-run economy, and enough players to guarantee opportunities for economic interaction.



Jamie McCarthy wrote:
Note that the horde has much more reason for
Undercity to become "the" city, since the most
popular endgame instances are on its continent. But
it's not.

Actually Orgrimmar is currently a better choice for a "home base" than Undercity for the Horde side. Endgame raiders typically have their Hearthstones set to Kargath to ensure quick travel to Molten Core and Blackwing Lair. Zul'Gurub is equidistant from either OG or UC (in terms of travel time) due to the presence of zeppelins outside of both cities which can take players to Grom'gol. Also, Onyxia's Lair and the Ahn'Qiraj instances are much more accessable from Orgirmmar.

It will be interesting to see if the addition of the Naxxramus instance in Eastern Plaguelands will cause an increase in the Undercity "population".


Building on Bart's comments, SWG had some notable changes in its support for player transactions, and the fact that Coronet remained the main focus throughout is interesting.

First, trade routes alone can't explain the appeal of Coronet. Corellia had four starports (IIRC) that all shared that "nexus of trade routes" status. Kor Vella had a strong Imperial theme and was closest to the rebel theme park, but that didn't attract tradesfolk.

On every server I visited, Coronet was surrounded by a semicircle of poorly-organized, poorly-stocked, and poorly-managed "malls." The merchants seemed to understand the adage of 'location, location, location" but absolutely ignored such things as customer service and maintaining their wares. At any one time, a random selection of vendors would find over half were empty- just a waste of space from dead accounts.

You'd think that the annoyance would have been sufficient for the player base to start "migrating" to a cleaner site, but I never noticed it.

Player cities had the potential to make more "localized" pockets, where community members traded among one another. They seemed to, for the most part, but cities often operated like a second skin to guilds, and it was often assumed that crafters offered "special rates" to city members (much as people discount guild members), and that the items in the city-based vendors were priced for outsiders.

During one of my many returns to SWG, I'd heard that they'd added galaxy-wide search to the bazzar terminals. Now you could search for "fizzlepop" and find the cheapest fizzlepop for sale at any location at any time. Finally, I thought, the great Coronet malls had to be feeling the pinch.

Not a dent. Don't know if that's because of some emergent behavior, pure stubbornness, or the fact that the darn search seemed so buggy as to be nearly useless.


Trading collapses to one location because there's almost no cost to move to that location, and goods are accessible from any location equally. I don't have to carry 10 suits of armor to Orgrimmar; I can just retrieve them from my magical universal vault. I therefore go to the highest population density, because the odds of finding a customer there are highest.

This is amplified as the cost of travel is reduced. In Guild Wars, which has instantaneous teleportation built into the game's interface, while buying and selling sometimes happens in large cities around the world, it mostly focuses on one place -- Lion's Arch -- and one district (instance), America 1.

The only way to create any sort of trade distribution that reflects real-world markets is to create a significant cost for travel. In FFXI, where travel to the 'real' city, Jeuno, is not possible for new players, there are thriving local markets in each of the three starting cities. While this distinction fades as the playerbase matures, it's a good example of travel 'friction' increasing the value of a local marketplace.

Market functionality aside, though, the real reason to add an auction house to all the major cities in WoW is that the Undercity and Stormwind were basically useless without them. The predominance of loot-drop based gear in WoW means that the auction house is not just a city service; it's the city service. The number of Night Elves making the cross-country trek to Ironforge is evidence of that. Playing a Night Elf before the AH expansion was... grim.


Getting back to the topic of using synthetic worlds for experimentation, there's no doubt that developers have inadvertantly conducted experiments in the past. (God knows I felt like a tortured lab animal in SWG.)

There's likely a wealth of information that can be mined from previous changes in the game cycle, if we had access to that kind of data. (Something I know the PlayOn team has been hoping for...)

Changing game elements specifically AS an experiment would be only a little different that what developers do now- observe a behavior, make a change, observe the result... (granted, if SWG is any example, they don't have the same ethical constraints or effective oversight)

One shouldn't be deluded into thinking that these virtual worlds significantly reduce the effects of unknown variables. These people still live offline, beyond the reach of the experiment.

Plenty of offline fan forums, guild sites, and other services will still color their play behavior outside the game. Churn will change your population, particularly as the early-adopter is distracted by the next "shiny object" that comes along and the casual player picks up the game from the discount bins.

If you're running different code on different servers concurrently, you still must account for the self-selection of the server populations. At his GLS fireside chat, Rich Vogel remarked on how the east and west coast servers developed differently and had different demands.

I also wonder how the "griefer" play style would take to an announced experiment. If I were a griefer, I'd have great fun in making anyone's life hell- but there's something special in the idea of messing with the powers-that-be.

None of this is to say it shouldn't be done. There are likely fewer uncontrollable variables in there than in real life.


This reminds me a bit of "Candid Camera," and various other "Psych 101" experiments, in that it is interesting to see how large systems are modeled "in a teacup," or how kids act like grown-ups when given the tools, or how sports can seem very business-like, etc. etc. The point isn't usually to see how funny it is when people do weird stuff, or do the same stuff in weird environments... but to see if regularly observed beahviors are the result of a more "natural inclination," or are environmental. We've seen some good comments, above, that suggest that some behaviors observed in RL relative to travel are, obviously, not going to translate 100% to VWs; i.e., "going somewhere" is always going to be an issue of time, not distance. And while that is, frankly, more the case in RL than we often realize, it's a great observation. The shortest distance from "Point A" to "Point B" in a VW is "click." If the rules/code allow for "click." If they don't, the shortest distance is...

In medical research, these types of experiments are often grafted into and tracked along side each other as parts of longer, longitudinal studies. The point being to give health-care professionals (and us... what-do-you-call-us... er... patients... yeah) good, solid, "Do this, not this," recommendations. Otherwise it's just a bunch of poking at rats. Which is fun, sure... but a waste of good rat meat.

So. What's the point of these studies? What measurements are we going after here? There about 120 years' worth of "We like this" vs. "This sucks" data in the medical community on which journals like NEJ and The Lancet and lots of others then pile up new stuff. And then Newsweek and Time report on 'em when it gets into something common-sense enough like, "You should eat less red meat!" or "Even 15 minutes of exersize a day can lessen your chance of heart attack."

If we're gonna use MMO's as petri dishes, either for RL experiments or for case studies of what to do or not do in future VWs, I'd rather talk less (at this stage) about specifics of one study, and more about;

1. The different goals that different audiences might have for such studies

2. Where we might collect and peer-review such work

3. How the work might then get "crop rotated" back into the gaming industry, so we don't have just us "beardy intellectual types" drinking our own bathwater. But rather provide a service that makes games more [See goals from #1]er for all the [See audiences in #1].

Or does that exist, but I've just missed the link somewhere... If so, please point.


This is off-topic, but I'm not sure of a better forum to ask:

Are there any muds/mmorpgs which encourage botting? and if so, would you point me towards them?


If you want to study really complex virtual world economics, go to EvE Online. WoW and the rest are sandboxes compared to what you get in EvE, and I'm not speaking of PvP play, even though the developers like to say that EvE is about PvP on all levels, not just combat, but industry, science and trade as well :)


Edward, I've long wanted to ask: why don't you ever study/write about Second Life? Is it because it is not synthetic *enough*?


*clicks stopwatch* 11 comments before someone mentioned SL, a new TN record! ;)

Regarding SWG and how even after Jump to Lightspeed, everyone still went to Coronet. Players went to Coronet because they had always gone to Coronet, and they didn't like change very much. In SWG, change is almost always bad. Change in SWG means someone gets whacked with the nerf bat.


*clicks stopwatch* 11 comments before someone mentioned SL, a new TN record! ;)

Add the random mention of "Go play EVE, because it's been done there." every few threads. Not to say it's untrue, but I love how it just pops up nearly as regularly.

But just to make sure this isn't a wasted post...

I want to point out that it's foolish to just sit around and talk about how MMOs would be great for this or that. Instead, someone (like, say, Andy Tepper) ought to just do it and see how it goes. We can debate the ethics and applications of this all day and get nowhere.


Fascinating topic. Since we have some PlayOn data that's relevant, I crunched out some of the city migration figures yesterday.

The city splits are stable September through December, and we see a noticeable change on January 3, when the other two cities open up auction houses. I haven't taken the time to see if there is any significance, but to the eyeball it appears that at least on this one server, the changes are ongoing even to the present.



I am glad someone had some census data stored up for the main starting cities. If you do find some significance you should post a seperate topic here on TN/Playon. I am especially interested to see if the population on each server will eventually migrate back to their race's starting city for their trading.

I am not a statistician but I think this would be proven if the racial demographics match the same trends seen in the number of level 60 characters in a city.



Migrating back to one's race's starting city seems very unlikely, imho. As a player, I don't see a strong correlation between a character's race and their actions in-game, except for the relatively rare RPers.

In the PlayOn post, I include both a graph and the data underneath, in case anyone else wants to play with some of the numbers.


As someone mentioned Andy Tepper...that's a good point. Why not go ask him? As a former player of A Tale in the Desert, I can confirm he's done plenty of things that at least half qualify as experiments. He's done plenty of things to the game "just to see what happens".

So...go ask him.


Hmm. Judging by how difficult it is to get into the industry to begin with, maybe a different method would be better.

Maybe the researchers in this field could get together and take a "Space Shuttle" approach -- identify the most promising, interesting data you would like to collect and throw it out there for the developers to collect and send back?


I also find it somewhat odd that all the emphasis is on WOW for these types of "studies". Yes, I'm an EVE player. -getting that out of the way. :-D

Although WOW is huge, numbers-wise, the actual number of participants is not only small per shard, it is segrated by nationality to an extent.

We EVE players are not segregated (other than the ongoing china server which is another ball of wax), and our markets span the world...literally. There are IPO's, scams, theft, skulduggery, wars, alliances, and just recently a VERY big IPO fraud very reminicent of the ghosts of Enron. The talk of implementing a stock market is back on the front burner over where we interact.

If you want a study of "people interaction" 15 to 25 thousand online in the same "shard" seems a much better barometer and an interesting petri dish than the sharded 'guided' mmo's such as WOW or its ilk. This seems even more relevant when considering the 'hands off' approach is used by the developers of the game to allow the players to more or less dictate the course of events.

(I refer in this case to a recent ipo scam that was 'reversed' by the GM's only to raise a massive stink on the forums, which then was RE-reversed in FAVOR of the scammer.)

As far as the 'centralization' of markets, it is natural that markets will localize. For example, in EVE changes to the jump system by CCP attempted to spread out the markets, and while this was accomplished to an extent, it only created 3-4 central hub points that the players (traders/missioners) gravitated to, basically negating the attempt at diffusing the markets through regions to a greater extent.
--The law of conservation of energy" applies even to these worlds. My way of saying: People are naturally lazy and shorter, faster, quicker will win the day every time.


Heh heh. Sevarus said, "Ilk."


Interesting things on WOW or SWG, but I wonder if it's a consistant phenominon. Was it Coronet on all servers?


It was the case on all the servers I had characters on, and was reported anecdotally (on the official SWG forums) as occurring on other servers as well.

Basically too many times to be coincidence.



But couldn't the 'Coronet' case be just a jungle drum type of thing? A trend that is spread in forums, in-game etc. As soon as this has been established in the community, why do it differently on separate servers? I have never played SWG but at first glance I find it hard to believe that all players migrated to Coronet at the same point in time.



espie, I'm no game theorist, but from what I've read it sounds like this "jungle drums" effect is another way to say "n-player asymmetric coordination." Buyers and sellers individually used market information to make rational decisions about their economic behavior.

Players didn't all suddenly migrate to Coronet overnight -- it happened over several months. You could probably call that "at the same point in time" as long as your definition of "point" is sufficiently broad. ;-)

What's interesting is why multiple trading locations weren't maintained. My guess is that there was some critical mass of players required to support multiple loci. Once the population of traders itself perceived that the critical mass no longer existed, Coronet somehow became the obvious choice.

That might have been due to structural rules (travel lanes), or from a widespread perception among players that Coronet was somehow "central," or from some other reason. But traders on most (if not all) SWG servers definitely migrated to Coronet (and over roughly the same period of time) more often than can be adequately explained by random chance.

I agree that there might be some value in studying such behaviors over in EVE Online, as its economy is mostly player-run. SWG's, being entirely player-run, was still probably more interesting from the economic study viewpoint, though. (No NPCs making stuff out of thin air!)



In reference to Andy Havens post (Jun 20, 2006 5:08:28 PM):
I mentioned something similar to this a few weeks back (Guilds and Government June 6, 2006). I had in mind, as Mike Sellers (Guilds and Government Jun 6, 2006 5:47:18 PM) noted, is a Nomic of sorts but designed from the ground up for research purposes as much as for entertainment purposes. As they stand now research within current games is limited due to the design of the games and limited access to the data reports.
A few games, such as Eutropia, Eve and Second Life have demonstrated the possibilities of VR economies and persistent universes, as a researcher I don’t have access to the data that I need to properly evaluate the behavior. This God’s eye view perspective would be enormously useful (possibly mandatory) in determining larger more subtle behavior patterns that would otherwise be lost. Personally I feel that the ethics and privacy concerns it would be necessary and probably must less of a hassle in the long run, to build a game from the ground up. If you set up your economy similar to Eutropia’s it is quite possible that this research project could be self sustaining. However, I suggest that RW academic (and probably government with a real cash economy) oversight would still be required.


Before the Jump to Lightspeed expansion, all interplanetary shuttle routes converged on Corellia (well, I guess they still do). To get to Dathomir from Tatooine, or Endor from Talus, you had to stop over on Corellia. Hence it was a popular planet to rebuff and heal, being only one hop from anywhere.

Of the starports on Corellia, I believe it's only Coronet that has a large open square in front of the starport. It was in this square that players would line up to buy buffs from Master Doctors setup there with their medical droids. Additionally, Coronet's layout is such that the cantina (Entertainer buffs) and the bazaar (AH for WoW players) are both in close proximity to the starport. The choice was clear.

Between Jump to Lightspeed, which removed the need for most players to take interplanetary shuttles, and April 2005's Combat Upgrade, which nerfed Master Doc buffs to the point where nobody bothered getting them anymore, Coronet became more deserted. The focus of player activity on most servers became the Mining Outpost on Dantooine, which was the most popular group grind spot for a multitude of reasons I won't get into here.

After the NGE of November 2005, the only NPC city which could boast any sort of regular player population was Mos Eisley on Tatooine. The reason? That's where all the new players landed, in a game that had alienated many of it's veteran players into cancellation.

So over time, the player "hotspot" in SWG has migrated around in reaction to each of SOE's radical changes to the game. The hub of player activity changes to meet the patterns of the majority of players' needs.


To Todd, per his comment above:

Todd said: "This God’s eye view perspective would be enormously useful (possibly mandatory) in determining larger more subtle behavior patterns that would otherwise be lost."

Although that would certainly be enormously useful/helpful (as you say) for studying certain VW/MMO behaviors, trends, data, etc., we should keep in mind that that is rarely the kind of position we have when studying RL societies, economies, trends, business fashions, goals, user needs, etc. Much research done in RL requires (often with a heavy sigh from the researchers I've known) various layers of glass, fog, mirrors, dirt, language, red tape, crud, money, time, paper, email, etc. between the "pure" experiential or market data, and "what comes out the other end."

And then, years later, you find out that the 100% fantastic, iron-clad conclusions that were made (calcium is good for making strong bones and teeth!) aren't always true, because you didn't know everything there was to know about how calcium binds in the bones of people at different stages of development and aging. Whoops! It ain't never perfect... Layers of crud.

My main suggestion wasn't for one game or one world where various "petri dishes" could be set up, but for one (or several) central sources of peer review and/or understanding about what it is we are trying to accomplish with our studying. Besides having fun with the yakkity-yak, blowing of steam and chatting about something we love.

Which is what "scientists" were doing back when they were called "alchemists." And then the Royal Society got together and said (and I'll paraphrase 50 years of scholaraly inquiry and debate. Ahem...):

"If we want to stop being burned as witches every couple years and be taken seriously, we have to get our s**t together, use the same units of measurement, agree on what counts as 'methods,' agree on terminologies, not kill each other in duels over results, pool some resources, publish, agree on what counts as 'published,' and generally behave. Or we can keep getting burned as witches. I call the vote..."

Thus, modern science was born. Or at least one branch.

We've got lots of great people in this field doing lots of great stuff. But much of it (that I've read) is somewhat repetitive (because it's the fun stuff). Or it's of the "freezing stuff in liquid nitrogen high-school chem expriment" variety, which is the "this is interesting... but so what?" kind of study. It gets some attention because it's got a, "Gee whiz" tacked on there, but... so what? And, in the end... what is it we're trying to accomplish?

I'm asking again, because I'm just not aware of it. Is there a place (more august than TN, he said, eyebrows-akimbo) where such learning is regularly collected and reviewed?

Has anyone thought of "purposes of inquiry" for *why* (I'm big on "whys") we should do game research? In the natural sciences, there's an assumption that all knowledge will fit into the greater scheme of things and thus inform the whole. Which is never bad. You find out weird crap about the fruit fly, and years later, it turns out you needed to know that for 728 other reasons. Cool. I don't think that's the case for gaming research, because games aren't ever "natural." Game study is a bit like sociology, a bit like psychology, and marketing and biz studies and comp sci, with some law, communications, art, and other weirdness thrown in.

Why should we study it? I'm not asking rhetorically; I think we definitely should? But there are clearly a bunch of reasons, and if they were categorized as such as part of a "VW/MMO games study" collective, it might be an easier/better way to evaluate individual studies, whether they were made in "petri dishes," or from simply interviews with gamers after-the-fact.

What are the "root whys" of studying VW/MMOs? Here are some initial thoughts.

1. Build funnerer VWs/games. By "funnerer" I mean quantitatively and qualitatively better from a pure play perspective. As measured by user satisfaction surveys, level of use, term of use, degree of heterogeneity of player base, etc. There can be a variety of factors to score games on, but if anyone wants to argue the value of the statement, "better games are better," well... that's fine.

2. Build more efficient VWs/games. By efficient, I mean games that are less expensive to publish, yet return more profit per unit to the owner. Why would I word this goal as more efficient as opposed to simply more profitable? Because profit can be a factor of size/scale and marketing and advertising. You can have really crappy profitable games. I'm not sure we're interested in that. Maybe. That might be a sub-genre of this study about game marketing, but I'm more interested in how to make the same "better, funnerer" games mentioned above, but for less money/resources.

3. Understand the relationships betwen VWs/gaming and other social/psychological issues. This is the antropology area. All the "sex in games" and "gener bending" and neat "is my avatar evil" stuff.

4. Understand the relationships between VWs/gaming and other entertainment venues and media.

5. Study the use of VWs/games as educational tools. Compare them to RL models in terms of effectiveness, efficiency, etc.

6. Study the use of VWs/games as training tools.

7. Study the economics of VWs/games in terms of RL economics. This is not, "I make money as an employee of a game company," but "I make money as a player of a game; doing RMT; buying/selling virtual land in SL."

8. Study the legal ramifications of VWs/games as they apply to each other and to RL law.

9. Study the pure "art" (visual, interactive, video, musical, story, etc.) created by and within VWs/games.

10. Record the history of VWs/games.

Ten's a good place to stop for now. If there was a place to collect, review, rate, store, comment on and generally collate the various VW/MMO studies in such a way, and in categories something like this (just a starter set of ideas, folks), it might be helpful.

It might help grad students evaluate where their proposal "fits" on the totem pole (or if it's been done before... alot), it might help the community not beat the same dead horses again and again, it might keep us on track for some particular goals...

And I might get to play better, funnerer games soonerer! : )


Andy, those seem interesting, but they also seem pretty utilitarian. What about "the proper study of Mankind is Man"?

"[G]ames aren't ever 'natural'?" I would argue the exact opposite; they're entirely natural in that they're played by human beings. Games -- including these newfangled online, computer-mediated varieties -- are one window among many for attempting to understand the human condition, which is the greatest adventure of them all. (To a human, anyway.)

Would you be willing to add an eleventh reason (which is distinct from your #3 reason): "Study the behavior of people in game worlds (where limits to action are removed) to understand what being human is really all about"?



This may be going back a ways in the discussion, but earlier we saw a quote that

"Note that the horde has much more reason for Undercity to become "the" city, since the most popular endgame instances are on its continent. But it's not."

This is not necessarily true. Blackrock Mountain (home to many high level instances, including: Blackrock Depths, Lower Blackrock Spire, Upper Blackrock Spire, and the ultra high end raid zones of Molten Core, and the Blackwing Lair) is located closer to Orgrimmar. In addition to Blackrock Mountain, the continent also has Dire Maul (a 3 part high end instance), and Ahn Qiraj, itself home to two high end raid instances. (though AQ is at the far end of a painful mass-transit ride.)

Compare that to The Undercity Continent, which has only Zul Gurub, Scholomance, and Stratholm (arguably 2 instances) for high end content.

In addition, it would seem that Raiding guilds (those with enough members and commitment to tackle the highest-end content) may in fact help drive the markets, as they have the largest buying / producing power. For example, many of the most coveted items and enchantments can only be obtained by those who venture deep into the game's raiding content. (Molten Core and Blackwing Lair being the two most well conquered)

Lastly, back to Edward's point, 3/4 of the Horde races all have a "home" on the same continent as Orgrimmar. SO his original point seems to hold, and in face Orgimmar remains the likely trace hub over Undercity. Wether this is due to player's personal attachment to "home" or to the high end content concentration is still up for examination.

While this will not prove the point as well as a move to Stormwind, it does not contratdict it either.



I've since run numbers on which cities are being used by players on the 5.5 servers that we closely track. I there is enough interest, I'll post the graphs or even the raw data somewhere.

But the upshot is that since 1.9 came out 1/3/06, there has been a slight shift away from the previously-sole auction house city, but that city still remains dominant today on all servers viewed. Moreover, the shift was gradual, and appear to be ongoing on some servers, while other servers have reached a steady state. Stormwind made the most gains in the shift, and Undercity is second. Darnassus remains desolate, and Thunder Bluff nearly so.

I think all this confirms the OP: these worlds make interesting petri dishes.


Bart: I think that's a great 11th thing to study. Probably could be a "meta study" that goes in as "Number Zero." And I only stopped at 10 'cause... well. 10's a good number for a list. I'd also argue that while my list begins with some utilitarian items, the study of art within games, and the nature of "fun" itself are, in fact, part and parcel of what makes us human.

And by "natural," I meant simply that there's not such thing as naturally ocurring "laws of games" or "rules," where there are natural laws that science can track down and say, "Here there be quarks."

Not that people don't behave "naturally" in games. Of course they do. To quote my favorite GMism from my dear friend Ed King, "The animal is never wrong." And studying behavior, be it singular or group behavior, in games is clearly going to yield some really interesting results as to what we're "all about."

I do, however, think it's important to start with some basics and work outward. If you want to understand neurochemical psychology, you first have to have studies of biology, chemistry, neurology, anatomy, pharmacology and psychology. You can't just jump in and go, "Well! I wonder what the effect of free endorphins in the brain would be if stimulated by mild jolts of electricity, in the absence of familial and/or pair bonding production of such from natural sources within the brain?" You don't have the language for it.

Similarly, if we jump in and talk alot about what games "mean" from psychological, sociological and anthropological points of view... when the manufacturers of the games/spaces we're basing those studies on are 99.2% interested in making a buck, and the players are 88.9% interested in killing 45 minutes (on average) per week as a way of passing some time... we're spinning our wheels. You look at which gasses burn brightest and hottest, and then you work to refine them. You don't study the "everything" properties of "everything." I'm sure the social psychology aspects of "Puzzle Pirates" is fascinating. But unless we have some basic terms and structure around which to "study our studies," we are (to a degree), being less than efficient.

It's fun to talk about stuff. Movie critics get paid to do it about films. I, for one, though, think that VWs/MMOs are something more important, and that we're getting into some stuff that could be very, very different than what went before. The sooner that all the various Our Gangs stop putting on various shows in back lots on wooden stages and has some kind of agreement as to what and how to talk about this stuff, the more likely it is that we'll have better learning from our discussions, and that mistakes will lead not to more mistakes, but to fewer.


In reponse to Andy Havens list of 10 possiblw research goals:

When you say "we should keep in mind that that is rarely the kind of position we have when studying RL societies, economies, trends, business fashions, goals, user needs, etc." this is percisely the reason why a game/research program such as the one I described should be done. With RW social science research methodologies, it is often impossible to control for the variables. however a research project such as the one I described controling for important variables might very well be possible. For this reason alone an attempt should be made. This is after all why qualitative research is often criticized as being un-scientific. And without giving credence to that point of view there is something to be said about controling variables in research.
While discussing the various merits of Eve's economy vs Etropias economy with my brother I asked the question, "is it possible to design a game in which the players could be thrown into a situation in which they could completely control the social environment as opposed to the gamming companies?" what if for instance on the day the game offically started there was only untamed wilderness in the VW and random groups of players would be collocated in random geographical areas. The players would then have to work collectively together to build and maintain a city of there own design. they would be responsible for everything from architecture to social customs to local laws, taxes, and governemnt. then once the city has reached a predetermined level of development they could then transfer there economy to a real cash economy similar to Eutropia's. Imagine this you start the game with the above preconditions and are colocated with say 200 other players will all of you try to build a city together or with the original 200 break up into smaller groups. say it breaks of into 2 or 3 groups will these groups war if they do what are the patterns of conflict how is the conflict resolved? (remember this game has perma death) What about legal studies what laws do the players set up for there cities and how do they enforce them? will similar patterns develop across many servers or will there be no decernable patterns?



I think that would be a fantastic game and study, and I'd not only read the study, but would volunteer to play the game. But I would challenge you, before starting, to define very clearly what it is you hope to learn. Otherwise, you will come away with a set of observations that are purely *yours.* Without setting up various controls, all kinds of motives and reasoning could be ascribed to user behavior and end-states.

Watching people monkey around in a sandbox is fun... But not really research. You could use some of the 10 "measurement" ideas I suggest (or others), to start coming up with ways to differentiate a number of separate shards of the type of world/game you suggest.

For example, if you wanted to test the hypothesis that "Teamwork is more fun than solo play," you could write in rules for one shard that say that resources are gathered more quickly when more than one player is in certain proximity to others. Or you could have skills that develop as "team skills" such that if four folks get up to Level 5 in "Lumberjack" (felling timber), one of them is promoted randomly to "tree spotter," and his/her presence in the group later will cause lumber to be felled more quickly. In another shard, you could simply let people develop resources all at the same rate, giving essentially equal bonuses for individuals that would add up the same, but not based on real "group play." So 10 players working together on Shard 1 at felling trees for 1 hour would still get the same number of units of wood as 10 players working independently on Shard 2... but would the Shard 1 people report as much "fun" being had? And would the "shared skill" elements, long term, contribute to more loose kinds of group behavior?

You could have one shard where all avatars are generic, and all names are generated randomly, and any use of gender pronounds is automatically rendered neuter by the game engine. I type in, "What does she think?" But when I hit enter, what comes onto the screen is, "What does it think?" How would the removal of any gender cues affect social issues in the game?

You could start one shard with a promise of a prize for the group of X that makes the most in-game money in a certain amount of time, where X can be as small as 2 players, or as large as 100 (in a shard of 200). Does the promise of a real-world prize stimulate economic competition, or warfare? What if there was also a larger prize for the team that made the most in-game dough between several shards doing the same thing? And a 2nd prize that was also sweet? Would one shard try to come up with two 100 player teams and try to totally work together? Or would everyone go all out for themselves? And what would happen if you changed the rules mid-way to say that the score would be based not on total group cash... but cash-per-player in the group? Time for some culling? Gulp.

The petri dish is a great idea. But it is in even more need of definition of goals, because it is such, I think, or you'll end up going on (like I just did) for a long, long time... studying... whatever.

The main question is, I think: "What do we want to know?"


Response to Andy's "What do we want to know?"

That of course would depend upon the researcher’s particular interest. My personal interest would lead me to look at conflict and conflict resolution. What are general incentive and disincentives of both but this is besides the point what I had in mind was a more general idea for a VW. (Though I must admit it is rather vague in my own mind) I see a VW that is like a huge research laboratory in which many different researchers are working on many different problems but whose work could be interlinked to form a larger macro-research project. Being an anthropologist I am most interested in culture creation and development therefore the variables that I would attempt to control would be environmental in nature. For instance there are many theories out there that attempt to explain why one culture group is pastoral and nomadic while another is sedentary and whose subsistence consist of widen agriculture. This all fine and good but these theories were developed after the fact. By this I mean these theories are often developed with an incomplete knowledge of the decision processes that individuals faced as they developed into one cultural form or another. With the VW I am envisioning it would be possible to gain a fuller understanding of these decision processes because we (as researchers) would have the ability to look at the past from the perspective of "God" if you will. We will have a nearly complete history of what went on during a given period of time.
As for a specific research question that would depend upon the specific research topic of the individual researcher. For instance would it not be possible to use one group of individual players to answer several questions at once? (I know this sound counter intuitive as it relate to controlling the variables but give me a chance here) Why is it not possible for one group of players to act as test subjects for an economist to say study economic value while at the same time an anthropologist is looking at cultural exchange? Both of these issues are closely related in fact, they may be so closely related that to artificially separate they would be to skew the results of your experiment as it relates to real world experience. More generally the game universe as a whole could serve as a clearing house of sorts that matches up researchers (with similar research goals) with test subjects (players). Think of it as a laboratory for rent in which as a scientist I can rent out the whole laboratory or just say a small corner of the laboratory. Because the game universe would be operating under some sort of monetary scheme (either subscription based or under a RCE scheme similar to Eutropia) it could do this at minimal cost to the researcher. Not only that but the researcher could be located anywhere there is a high-speed hookup. Like I said what I am seeing is still in its infancy (I have only been working on this idea for a couple of months) and is rather vague (even to me) but hey that’s what TN is for right?


horde ftw.



I don't see anything wrong with what you're suggesting per se. It sounds interesting and fun. But it does boing back to a couple points I made earlier.

1) You won't really get sociological data bout the "real world" from game play, except as the one intersects the other. To try to make "stiff" correlations between how people might farm, hunt, care for young, etc. in a game vs. how it was done in a natural setting is... well... going to get you in a world of academic hurt. That's what I meant by my statement that no game world is "natural." The game "gods" are literally that. And until there is a game that is capable of putting people into a sim that is an exact duplicate of a situation you are trying to study... what you are studying is a game. Which is good mostly for making observations, deductions, predictions and statements about games. Which is what we're talking about mostly here anyways, I think.

2) Unless you get very specific, very long-term commitment from some hard-core players up front, you won't ever get something even approaching "culture creation" in a game setting. You may get sub-culture, at best. You may get something that some people might call (on a charitable day) "art." Certainly some "content" or "media" or "interesting, entertaining stuff going on." And I don't mean to belittle it, because I have lots of fun playing games. But, thusfar, our games/VWs exist *within* other cultures. They aren't cultures per-se. We talk about the "gaming culture" and "sub-culture" a lot, and there are geek cultures and various sub-cultures that thrive on all kinds of game and tecnho related media... but "cultures" as a whole won't rise up out of lab experiments. They can't. They take too dang long. They require buy-in at a macro level and aren't coaxable. You can pay somebody to take an experimental drug, wear a weird beanie or eat monkey chow for a week... but you can't pay people to change their geshtalt.

3) On the other hand, the idea of having a university or grant-based, "free to players" set of MMOs with a set of variable starting conditions, with the caveat that "the machine is listening," recording all the data (not personal data about the player, but game data about what goes on in the "world") as the price of entry is intriguing. Like they say, "Money talks, BS walks." If you set up 10, 20...100 slightly different VWs with subtly different rule-sets, with the express purpose of studying both the play styles in each, AND which one got the most "votes" in terms of who said, "This shard ROCKS! Come play here!"

That would be cool as hell.


Andy Havens> If you set up 10, 20...100 slightly different VWs with subtly different rule-sets, with the express purpose of studying both the play styles in each, AND which one got the most "votes" in terms of who said, "This shard ROCKS! Come play here!"

What if this game were designed to constantly mutate its rules?

Create a population of 100 servers, all with slightly different rules. Let each one run for two months -- at the end of that time, the 20 servers with the lowest populations are culled, and 20 new servers are created with minor modification of the rulesets from the top 20 servers, and everybody runs for another two months.

Question: How long would it be before the Perfect Game evolved?




Bart: Niiiice. It's "The Darwin Game Game Generator Engine Engine." DG3E2. Or, for marketing purposes, "The Meta Sandbox."

I think you'd need more than 2 months. And maybe a slightly different equation for "voted off the island" than just pop -- some combo of population, playing time and a satisfaction rating. But with pop weighing highest, perhaps. That's just niggling though...

As Artie Johnson said on Laugh In.... "Verrrrry Interesting."


My experience is that genetic algorithms can take hundreds or thousands of iterations to develop any accuracy on even basic problems. To begin with, we're already making a very big assumption that a perfect game exists within the domain of the genetic game design algorithm to find, which is an iffy proposition, and there's something of an assumption that the permutations allow for radically different game design solutions to be put together in forumulaic ways to get anything you want.

In the end, I imagine you'd frustrate most of your playerbase with constant tweaks and silly game design changes, and the resets would remove their reasons to carry on and put up with it, at which point they'd quit, and your investment wouldn't have tought you as much about the perfect game as how you can piss off your playerbase ;)

I'm just poking fun pragmatically, since there's sometimes a tendency to forget about how things might or might not be practical around here.


Of course there'd be practical challenges to making such a design work. As the ":-)" implied, the notion wasn't meant in all seriousness.

But is it really so Out There?

You'd need educated players. They'd have to understand that "the game" wasn't just the specific ruleset of a particular server, but was the larger meta-game of finding a/the ruleset that satisfied the most players. A Perfect Game for those players might not be perfect for anyone else... or maybe it would be. Wouldn't it be interesting to find out?

And to bring this conversation back to where it began, isn't that the whole point behind Petri dishes? You take a control organism and pop it into a large number of different environments and conditions to see which ones are the most and least encouraging to growth. And that provides useful information about your organism.

Absolutely there'd be practical issues to resolve in actually implementing this concept as a playable game. That's called "an interesting challenge." ;-)



There was a great little Jr-High level sci-fi book by Ben Bova called "The Dueling Machine." Micro version of plot: it involved a device that allowed two opponents to strap-on a shared, virtual environment and "go at it" in order to settle disagreements "as gentlemen." The environment could be... anything... and was generally designed by the "challenged." So you could have "Gunfight at OK Corral," or "Space War," or, in one hillarious case, a duel in a giant physics lab where expriments were running... and deadly.

Maybe you don't need whole "shards" for an experiment like this to work; not whole "petri dishes," but "petri shot glasses." Single or short-term (daily? weekly?) shards where some number of players could choose what "rules" were in place... hmmmm...

Or what about a massively concurrent world with "rule islands?" Different rule-sets for each island. And where a record of how much time your character played in each setting was maintained? 10 minutes on "goofy, upside-down, money-works-backwards, funky-70's-shoes" island... and I'm outta there.


Instead of being only a negative behavior to be avoided, "voting with your feet" would suddenly become a valuable effect in a gameworld like that, wouldn't it?



Andy I know what you mean about your first point.

1) You won't really get sociological data bout the "real world" from game play, except as the one intersects the other. To try to make "stiff" correlations between how people might farm, hunt, care for young, etc. in a game vs. how it was done in a natural setting is... well... going to get you in a world of academic hurt. That's what I meant by my statement that no game world is "natural." The game "gods" are literally that. And until there is a game that is capable of putting people into a sim that is an exact duplicate of a situation you are trying to study... what you are studying is a game. Which is good mostly for making observations, deductions, predictions and statements about games. Which is what we're talking about mostly here anyways, I think.

It would be extremely difficult to correlate VW phenomena with the RW. But what about this, for get about the RW lets focus on this as a tool to study VW phenomena solely while keeping an eye out for nay RW correlations.
As to Bart's comments: "What if this game were designed to constantly mutate its rules?"
this is kind of what I have in mind but instead of the rules being changed and set by those running the servers what I have in mind is a player controlled rule set. One in which the players change the rules as they see fit. Let them argue over what is fair and proper. the object is to see how they resolve these conflicts and how those resolutions are affected by the game environment.
As for the perfect game, “How long would it be before the Perfect Game evolved?”
If you mean perfect by a perfect set of static rules I am not so sure that there is such a thing. If however you mean perfect in the sense that the game evolves to meet the demands and expectations of the players more effectively then you might have something here. I would suggest however, that the rules be defined not by how many players are on a particular server but by the players directly defining those rules through there collective actions.
The question that interest me the most is this, how will the players define the rules and what will the processes be that they use to define those rules. Additionally this will allow the players to collectively define the history of the VW. Some of this has already started to happen with Eve online. (player defined histories). The difference of course is that the starting point of the storyline is artificial where as in this game the storyline would be player determined through there collective actions.


Before the Jump to Lightspeed expansion, all interplanetary shuttle routes converged on Corellia (well, I guess they still do). To get to Dathomir from Tatooine, or Endor from Talus, you had to stop over on Corellia. Hence it was a popular planet to rebuff and heal, being only one hop from anywhere.
Love Valentines Day, Gift Ideas

The comments to this entry are closed.