« Independence Air Files for Bankruptcy | Main | Sex on the Margin »
Synthetic worlds
Question: You’ve proposed the creation of an academic center for the study of synthetic worlds. What would such a center try to accomplish?
Castronova: It would try to develop some sense of the policy options in this space, try to parse out the opportunities and the risks. It would develop educational applications. It would develop the following research idea: If we built a series of comparable synthetic worlds, couldn’t we unleash slightly different policies in each one? It seems to me that this would be an unprecedented opportunity to study human behavior at the social level. No one has ever been able to study such things before, not in this way, not using the tried-and-true experimental methods of the natural sciences. Imagine how a few well-designed experiments about socialism in 1870 might have affected world history. We have a lot of pressing questions about societal behavior right now—human population response to disease, for one example; community response to natural disaster, for another. These things could be directly studied in synthetic worlds; without them, all we have is pure theory, and historical data (in which policy causality is basically impossible to untangle). The idea would cost $20 million to $50 million, but it would also dramatically improve business as usual in a large chunk of the university.
Here is the interview. Here is the web page for the book. Here is Castronova's blog. Here is the Amazon link to the book.
I file this one under: Matters I am Probably Too Old to Understand. Isn't the ever-versatile Jane Galt addicted to Civilization? I once tried the game and was too addled to play it, or even to figure out what playing it would mean, much less to use it for research. I brought it on my laptop to my Mexican village, where they shrieked each time the cow went "moo."
Is this experimental economics but with "synthetic," programmed people? Is this experimental economics with real people playing a computer game but in a richer synthetic environment? Is this experimental economics but cheaper because you use the "status money" of a game instead of real money? Some mix of the above? Comments are open...
Thanks to www.politicaltheory.info for the pointer. And speaking of synthetic worlds, John Fowles has just passed away.
Posted by Tyler Cowen on November 8, 2005 at 05:48 AM in Economics | Permalink
TrackBack
TrackBack URL for this entry:
http://www.typepad.com/t/trackback/3576/3528496
Listed below are links to weblogs that reference Synthetic worlds:
Comments
(1) the human species is very likely to go extinct before reaching a “posthuman” stage;
(2) any posthuman civilization is extremely unlikely to run a significant number of simulations of their evolutionary history (or variations thereof);
(3) we are almost certainly living in a computer simulation
Well there goes #2.
Posted by: joshg at Nov 8, 2005 8:37:15 AM
Regarding your last paragraph: That is one reason why I dislike the term "synthetic world," it's misleading. The more popular and term (I believe better) is "virtual world” because there is nothing synthetic or fake about them. ‘Virtual’ simply distinguishes between that which exists in 3D space, the real, and that which exists inside computers. Civilization is not a virtual world in the sense we discuss. Only one or a few people play it. Virtual worlds are games where thousands of people play together, like the popular World of Warcraft.
These are worlds where thousands of real people interact in a macroeconomy much more simplistic than in the real world. They use real money, specific in game currency (except for some worlds, where hyperinflation has made the official currency worthless and the people began using rare crystals as a new form of currency!)
As far as experimental economics is concerned: Many worlds use 'shards' or different servers. That is, they have ~200 copies of the world, and their subscribers are split up among these (This is usually because of technological limits). So what you end up with are a ton of possible control and experimental groups.
The only thing stopping us from performing macroexperiments on these worlds, tweaking the economy on one server, but not the other, is that we don't own the games! If some researcher teamed up with one of these production houses, that would likely be fruitful. And as Castronova has noted, the cost of building a good world is extremely high, so building a world from scratch for research purposes is probably not very likely.
There are however, still research projects we can perform now. Perhaps one server has a vastly different demographic makeup than another: we could look for any possible economic differences this might cause.
Another benefit to studying these worlds is that, simply because they are in the computer, precise data is available, unlike the real world. We can actually sit down and know all the prices for every good in the world at any given point. Of course we run into the barrier that usually only the company who owns the game gets this kind of data...but still--it's there.
Posted by: Matt Masten at Nov 8, 2005 9:19:33 AM
Another problem. I doubt a researcher could develope a game anyone would want to play.
The people who play the games like World of Warcraft, normally play a lot. That means the games had better be fun.
Posted by: cube at Nov 8, 2005 10:08:23 AM
"Civilization" has little to do with economoics. The simulation operates at a higher level of description where phenomena such as riots and inefficiency do not emerge from agents but are simply mandated as assumptions. Judged as a theory, it is utterly non-parsimonious. Judged as instruction in history, it is seriously (even critically) misleading. But judged as a game it's like crack.
The synthetic worlds Castronova is discussing are a part of the real economy, just a very small part.
Posted by: michael vassar at Nov 8, 2005 10:35:36 AM
Michael Vassar, I'm curious what you think is "seriously (even critically) misleading" about Civilization's take on history. Not because I necessarily disagree with you, but because I grew up playing (one of my fonder childhood memories is the morning I woke up to discover that the graphics in the original DOS-based Civilization had changed markedly when we installed DOS 6.0) and it probably affects my interpretations of events in ways I don't necessarily recognize.
Posted by: jadagul at Nov 8, 2005 10:56:53 AM
To echo Mr. Masten, not all computer games are suitable for this type of research. Civilization allows for multiple players, but only up to 8 or so I think, and for finite. It's a lot like the board-game Risk, in that sense. The Massively Multiplayer Online Roleplaying Games (MMORPG's) are a very different creature because they have no beginning or end. They're an infinite series.
The MMORPG's are actually a great way to study these issues because of multiple instances of very similar worlds. For instance, the Lineage game has tens of thousand of players on each of their 8 (I think) mostly identical servers. It would be so easy to enter in small changes in some of the servers but not others and observe the result.
The biggest hurdles I see here are commercial ones. The MMORPG's are businesses whose primary business is providing a fun couple hours to their paying customers. They wouldn't let you do stuff which upset their customers/ made people feel like guinea pigs.
Posted by: Brock at Nov 8, 2005 11:27:28 AM
jadaqul,
I would second Mr. Vassar's assertion that Civilization is misleading as to history. As just onc example, check out TCS's review of Civilization IV, which was just released.
http://www.techcentralstation.com/102805B.html
Of course, the Civilization titles have their differences as between each other, so they would have their own inaccuracies as well. I never played CivI, so I can't say how it might have effected you.
Posted by: Brock at Nov 8, 2005 11:33:11 AM
Castranova has written quite a bit in the past about how virtual worlds tend to follow precisely the type of economic changes as we would expect in a similar real world. For instance, there are a dozen or so major MMORPG's. They each have "increased the money supply" (by making certain goods more common, for instance) at different rates, and the cost of those items in the real world (when sold on ebay or other forums) tends to increase at the same proportional rate as real world goods in an inflationary economy. There are many interesting hypotheses that can be tested simply by examining virtual world data, even without control on the game server.
Posted by: cure at Nov 8, 2005 12:14:11 PM
Castranova has written quite a bit in the past about how virtual worlds tend to follow precisely the type of economic changes as we would expect in a similar real world. For instance, there are a dozen or so major MMORPG's. They each have "increased the money supply" (by making certain goods more common, for instance) at different rates, and the cost of those items in the real world (when sold on ebay or other forums) tends to increase at the same proportional rate as real world goods in an inflationary economy. There are many interesting hypotheses that can be tested simply by examining virtual world data, even without control on the game server.
Posted by: cure at Nov 8, 2005 12:17:08 PM
Virtual they may be, but running controlled experiments in these worlds would be as unethical as in the real world. From what I have read, players place great value on their virtual lives. I remember China Daily reporting that one Chinese player killed another, after the second guy borrowed a sword, online, and sold it to another player (for real money).
Still, it offers a much larger set of environments from which to study natural experiments.
Posted by: Pedro Bento at Nov 8, 2005 12:52:05 PM
Virtual they may be, but running controlled experiments in these worlds would be as unethical as in the real world. From what I have read, players place great value on their virtual lives. I remember China Daily reporting that one Chinese player killed another, after the second guy borrowed a sword, online, and sold it to another player (for real money).
Still, it offers a much larger set of environments from which to study natural experiments.
Posted by: Pedro Bento at Nov 8, 2005 1:02:28 PM
Virtual they may be, but running controlled experiments in these worlds would be as unethical as in the real world. From what I have read, players place great value on their virtual lives. I remember China Daily reporting that one Chinese player killed another, after the second guy borrowed a sword, online, and sold it to another player (for real money).
Still, it offers a much larger set of environments from which to study natural experiments.
Posted by: Pedro Bento at Nov 8, 2005 1:04:04 PM
jadagul:
The first lesson of Civ is completely true: Nations have no friends, only temporary allies. After that, it all falls apart.
1) National leaders have far, FAR less control over thier economies than the games would suggest, and are NEVER immune to replacement.
2) Trade benefits both parties, not just the one initiating the trade. (They finally got that one right in SMAC, Civ II was particularly obnoxious on that one.)
3) Societal response to changes in governmental policy is non-linear to the point of being chaotic.
4) Redirection of inputs or outputs requires substantial infrastructure change. (Imagine having to pay credits every time you made changes on the city map or adjusted the $/!/:) bars.)
Posted by: Nathan Zook at Nov 8, 2005 1:15:18 PM
As a long time player of MMORPGs with an interest in economics, I have always found Castranova's writing interesting. But a word of caution, much of what he writes is flat out wrong. He has a tendency to draw sweeping conclusions from very sparse amounts of data and his own personal biases.
Are all academics the same?
Posted by: monkyboy at Nov 9, 2005 5:10:47 AM
It's calibrtaed simulation - which is theory, not empirical work - but with cooler graphics than usual.
Posted by: Tom at Nov 9, 2005 3:05:05 PM
Followup - That is, it's calibrated simulation if you have artificial agents acting in the artificial environment (A-Life, etc.). If real people are playing the games, then it's...hmmm....
Posted by: Tom at Nov 9, 2005 3:10:59 PM
You'd only want to consider the worlds with real agents, since no program is going to adequately mimic the range of human creativity and learning ability. (This criticism applies to the "agent-based" literature as well; I am uncertain as to how much we'll really be able to learn from those studies.)
A question I have is whether the agents populating these worlds are "typical" enough. Evolutionary pressures (the less profitable die off) will still function just fine, but will the behaviors we see tell us enough about the real world? (Perhaps the "entrepreneurs" are too busy to play these games.)
Experiments being held within the game might attract certain kinds of players, but perhaps that's what you want to avoid.
The ethical issues are not, I think, unsurmountable. Environmental changes (like increasing the money supply) can actually keep things interesting for the game. The counterargument: my strategy is ideal for this game and I become king; rules change and I become a pauper; now I am pissed off because I had no idea the rules might change, or because I have no idea how to play under the new rules. Creating rules for how the rules might change in interesting ways may well be entirely too challenging.
I am exceedingly confident that Alex and Tyler could come up with an exceedingly interesting game to play. But the opportunity cost would be too high.
Posted by: randy at Nov 9, 2005 3:18:03 PM
I might be inclined to direct those interested in "virtual" realities (I would not agree with Castronova about the synthetic/virtual but perhaps he does have a point.) to these items:
To Second Life, a non-game environment. http://secondlife.com/ Although not a "member" myself, I find it interesting from a market pov.
From the Commerce page
http://secondlife.com/commerce/
"• An Internal Currency with Real Market Value
Millions of Linden Dollars change hands every month for the goods and services Residents create and provide. This currency may then be bought and sold on third party sites for real currency. Many of these sites even offer in-world "ATM" machines to facilitate transactions. Some of the sites offering currency exchange services may be found here."
From the Business page
http://secondlife.com/commerce/business.php
"Taking Your Business With You
The business opportunities do not stop at the virtual world. In Second Life, all Residents retain Intellectual Property rights over everything they create in-world—in Second Life, and offline. Turn a series of screenshots into a graphic novel, and sell the rights to a real life comic publisher. Prototype a fashion line, create real world versions of it, and sell them in a local clothing boutique. Shoot a machinima movie short in-world, and sell it as a pilot for a cable network. Not only is all this permitted—it’s encouraged."
Cory Doctrow also wrote an interesting treatment of these markets and those who inhabit them in, "Anda's Game" It was posted at Salon under "Technology and Business" just about a year ago.
The manipulation that Castronova would introduce would probably not work. However, that does not preclude serious study. And soon, those who study markets, will begin to study the implications of these markets. Castronova's focus on gaming versus other simulated environments leaves me thinking he has some catching up to do.
Posted by: Johnny at Nov 9, 2005 3:19:27 PM
Regarding Ethics: The design team for all of these games already tweaks and toys with their worlds to a huge degree. This is no different than the tweaks that would be made for an experiment. There are certainly some problems, which a few people brought up. First, is the sample of virtual world players representative of the real world enough that we can make inferences about it? Consider one possibility: Poor people are perhaps unlikely to pay $15 a month to play some of these games, let alone own computers , etc. Thus our results might not be relevant.
On the other hand, this might not matter. We might not need a representative sample to make useful inferences. No research has been done on this at all, so we just don't know. As far as Randy's comment on entrepreneuers: There are entrepreneurs left & right in these worlds. Everyone wants to make money, just like the real world. Everyone is searching for the quickest, cheapest, and best way to do it. It's only what we expect from the human spirit.
Another comment regarding ethics: Some people might be upset if they feel they are being "experimented with." But this is really just an issue of perception. Like I said before, they already are being experimented with, only not with academic research in mind.
I believe this speaks to tom & randy's posts. Again, these worlds are not agent-based simulations, which are nowhere near complex enough to useful.
No offense to Tyler & Alex, but it is difficult for anyone to make a game people want to play. This includes people with years of experience in the video game industry. My opinion: Academics should NOT try and make their own worlds. Instead, academia needs to team up with the commercial producers for research. This is my best case scenario.
Randy also brings up a good point, which is highly relevant to developers: the game must be fun, the economy must be fun, or else people will leave. This will definitely be the major constraint when performing experiments. But also, consider the implications of what it means to have a "fun" economy. Think about it.
And finally, in response to Johnny: Second Life is a frequent topic at Castronova's blog, Terra Nova, it has not gone unnoticed. This is because of the points your brought up of course, it is an extremely flexible and free virtual economy. The study of virtual worlds is not restricted merely to games, but it includes worlds like Second Life, and a few other social worlds.
Posted by: Matt Masten at Nov 9, 2005 9:11:24 PM
Don't let your desire for a particular outcome affect your reasoning.
In an MMPORG, rule changes being made by the developers to increase the number of player-hours devoted to their game are being made _for the benefit of the players_. Whether the effect is good or not, the goal is to improve the gaming experience.
Suppose I was paying to play an MMPORG. The developers are now working for me, and part of the value proposition is explicitly based upon that fact. The proposal to subvert this exchange without compensating or informing the players is alienation, pure and simple. This is product placement in a grocery store or radio station payola.
I does not matter if the changes proposed are well within the bounds of modifications which normally occur. The fact that these changes are being made at the behest of someone other than the one paying for the service makes it theft, bait-and-switch, or whatever.
I quit a free MMPORG because the developers were more interested in their vision of the game than the player's vision. The devs were treating the game (that is, the thousands of hours being spent by their players on the game almost daily) as their personal plaything. No thanks.
I think that a huge amount of useful data could be gathered simply by observation. I think that a statement up front that a game might have different rules on different servers as part of an academic experiment would solve the ethical problem, and present a low economic hurdle. If the nature of the possible scope of these changes were specified, that would be a very low hurdle. (Hey, if you tell me that a genuine economist is influencing rules changes & not just for kicks, I would call that a GOOD thing!) If players were allowed to change servers (with limited loss of stored value), you might actually have a significant class of people assign positive value to the proposition. (Including this data point.)
Posted by: Nathan Zook at Nov 10, 2005 8:54:18 AM
tekenfilms orgy ^^^ vreemd schoolmeisje ^^^ desiderio giovane ^^^ allievo urinate in pace ^^^ timoree jeune ^^^ mysterieux asiatique pipes ^^^ quarante et un jeune ^^^ maman dans le vent ^^^ sconfinato cameriera gruppo ^^^ selvaggio amatoriali dildo ^^^ sinnesrorelse ung tillverkningen alska ^^^ tyst som en mus tonaring ejakulation ^^^ shy amatoriali schizzate di figa ^^^ poor spirited cameriera anale fotti ^^^ raflott trent ^^^ anselig gallerier ^^^ het seng ^^^ uvanlig paps ^^^ tharaleos magoula katourima ^^^ nosokoma omadiko prothalamos ^^^ mman bas ^^^ plus chaude fils galerie ^^^
Posted by: levan at Sep 5, 2006 6:57:07 AM
bad credit mortgages bad credit mortgages
bad credit loans bad credit loans
bad credit auto loans bad credit auto loans
bad credit home loans bad credit home loans
bad credit personal loans bad credit personal loans
Posted by: bad credit repair at Feb 12, 2007 5:28:24 AM
apply for a credit card - apply for a credit card
apply for master card - apply for master card
credit card offers - credit card offers
bad credit report repair - repair bad credit
apply for credit - apply for credit
Posted by: apply for master card at Feb 19, 2007 5:30:52 AM
google排名
wow power leveling
wow power leveling
wow powerleveling
wow powerleveling
wow power level
wow power level
hongkong hotel
hongkong hotel
beijing tour
beijing tour
翻译公司
翻译公司
上海翻译公司
上海翻译公司
北京翻译公司
北京翻译公司
门禁
wow gold
wow gold
powerleveling
powerleveling
wow powerleveling
wow powerleveling
wow power leveling
wow power leveling
power leveling
power leveling
wow power level
wow power level
窃听器
福彩3D
搬家公司
香港六合彩
Ddo plat
Ddo plat
FFXI Gil
FFXI Gil
GW Gold
GW Gold
GW Power Leveling
GW Power Leveling
RS Gold
RS Gold
RS Power Leveling
RS Power Leveling
Lotro Gold
Lotro Gold
Lotro Power leveling
Lotro Power leveling
china tour
china tour
猎头
猎头
货架
货架
货架
google排名
google排名
翻译公司
翻译公司
上海翻译公司
上海翻译公司
北京翻译公司
北京翻译公司
仓库
仓库
仓库出租
仓库出租
物流园区
物流园区
监控
监控
激光打标机
软件工程硕士
工程地毯
工程地毯
集团电话
集团电话
激光打标机
激光打标机
打包机
打包机
拓展训练
拓展
拓展培训
香港六合彩
搬家公司
搬家公司
窃听器
Posted by: 水无痕 at Jul 7, 2007 2:27:49 AM
wulvyou 07年7月11日
google排名
google排名
wow gold
wow gold
powerleveling
powerleveling
wow powerleveling
wow powerleveling
wow power leveling
wow power leveling
world of warcraft powerleveling
world of warcraft powerleveling
wow power leveling
wow power leveling
cheap wow powerleveling
cheap wow powerleveling
powerleveling wow
hongkong hotel
hongkong hotel
beijing tour
beijing tour
翻译公司
翻译公司
上海翻译公司
上海翻译公司
北京翻译公司
北京翻译公司
门禁
集团电话
集团电话
货架
货架
货架
拓展训练
拓展
拓展培训
google排名
条码打印机
机票
搬家公司
北京搬家公司
仓库
仓库
仓库出租
仓库出租
物流园区
物流园区
条码打印机
猎头
猎头
虚拟主机
红外测温仪
红外测温仪
超声波测厚仪
超声波测厚仪
超声波探伤仪
超声波探伤仪
频闪仪
频闪仪
涂层测厚仪
涂层测厚仪
电火花检测仪
电火花检测仪
beijing
beijing
china tour
china tour
great wall
beijing travel
beijing travel
搬家
搬家公司
搬家公司
北京搬家公司
北京搬家公司
摆闸
痤疮
痤疮
色斑
色斑
黄褐斑
黄褐斑
青春痘
青春痘
口腔医生
口腔医生
口腔
口腔
口腔医院
口腔医院
口腔科
口腔科
三星集团电话机
呼叫中心
电话交换机
交换机
呼叫中心
测厚仪
测温仪
停车场
道闸
Posted by: asdd at Jul 10, 2007 11:20:20 PM
