Raph's Website » Treating players like numbers

Treating players like numbers

January 4th, 2006 (Visited 11025 times) Tags:

There’s been a lot of heavy, arguably negative stuff here lately. So I thought it might be interesting to talk a little bit about possibilities instead.

In general, I’m a big proponent of treating players like people. The reverse is what game companies are often accused of, of course: treating players like numbers. Dehumanizing them. Viewing them through a purely statistical lens. All very mechanical.

But wait… what can be done if we really treat players like cogs in a machine?

Way back when, at the Austin Game Conference not last year, but the year prior, I suggested that we ought to be looking at “games that cure cancer,” to somewhat mixed reactions — for two different takes on that panel, see here and here. A year goes by, and I am sitting at a different conference and someone (Byron Reeves) presents exactly that.

What he showed was a mockup of a Star Wars Galaxies medical screen, displaying real medical imagery. Players were challenged to advance as doctors by diagnosing the cancers displayed, in an effort to capture the wisdom of crowds. The result? A typical gamer was found to be able to diagnose accurately at 60% of the rate of a trained pathologist. Pile 30 gamers on top of one another, and the averaged result is equivalent to that of a pathologist — with a total investment of around 60-100 hours per player.

I am reminded of the powerful ways in which collective intelligence is harnessed by games such as Alternate Reality Games. Seemingly impenetrable puzzles can be solved fairly quickly. In SWG, the first event we did involved finding objects on spawns that had messages that looked like this:

NNnrrNpbqgNNa,NlNxnbgpteuuunNqNNsNfqNvrvNNvNNnffqNNsNogu)NNN:e:N’nbayaNnNfrNcatbrNv’NrsNNNNe.NnbreNN
PCvEErxvzkxviqissi,elsEEhEesXeFEEesEvvrEgErsieelerriie(EEijEXrrvxeismovwEaExkEEeeeglErEaEieEEl:arctan(1/2):eEe
SwszegsSmFjdsxogSSszjqs,sswSSlSSswGwwsSqSjSSzsasfgksSfSSOwSsfSvSSSzqSfsswksSlSsSeSzSyS:3!:vzfqsSsDSJL
-PygCvcrgkcjCkveCccg-CCgcCgCWrcgqvvChCt’cpkVppechvCCgqcgoCkcPCmCcugccwgcqrk-ce:0:Cntucc’Cgp’CuxCCuw,

There were 40 of them, representing two completely different messages. Players collected them all and cracked the cipher in less than 48 hours. (I stuck one sample puzzle at the bottom, if you want to take a crack at it). This one is pretty trivial compared to some of the stuff that the ARGs have thrown out there.

The approach here is to treat players as a distributed parallel processing machine, filtering information and each node contributing its specialized knowledge: cogs in a machine.

Speculatively, there’s many approaches to be had here. I’m not speaking about doing datamining on our game populations — there’s highly fruitful stuff to be had there, but a lot of it is about the specific ways in which people behave in the game spaces. Rather, I am speaking about ways in which we can externalize the computing power of large populations using the incentive structures of games.

Consider:

We use tools like SETI@Home to distribute highly technical tasks like signal processing. However, there are cognitive tasks we could outsource in this fashion, as in the cancer example. Could, for example, jury trials be improved upon if assessing the truth of evidence were a distributed game? Could airport security be improved by creating scenarios about infiltration in our MMOs?
Idea markets are already in use in a wide array of fields, generally constructed as games. Arguably, Second Life is being used as an idea market for the design process for furniture and clothing. While the Pentagon’s attempt to use a futures market for forecasting terror attacks floundered due to politics (”betting on terrorism??!!?!”) there’s clearly possible approaches that can be taken within a game.
Players are the best alife agent we’re managed to find. Why rely on Axtell and Axelrod-style simulations when we can build environments and loose players in them? The trick would be to get the environmental stimuli to match the ones provided in the real world. (I’ve argued before that this isn’t really possible, but hey, it’s worth a try). The old landscaping trick when deciding to put down paths was to wait a while, see where people trod the grass, and then pave where they had walked them ost; that sort of thing would be trivially easy (and cheaper) to do in our games, given a decent topographical map of the landscape.
There’s a university right now using a major MMO to do total-immersion teaching of a foreign language; by putting the students in constant contact with the foreign language and forcing them to learn it in order to succeed at advancing in the game, they hope to see as hig has a 30% greater fluency on the part of the students.

I’m not really making a pitch for “serious games” here — rather, I am suggesting that our non-serious games can be put to serious uses on the side, thanks to their scale. Being a cog in the machine can serve higher purposes even while each individual cog goes on its merry way enjoying its gameplay.

There’s a lot of possibilities. Got some more?

The SWG Puzzle

The below is just one of the two messages, and I have spared you the effort of gathering the pieces individually and not knowing how many pieces there are. If you despair, you can see how the community solved it here.

NPpPpPttpPuPbpPptPpPpPcjkginxchacwPctPsPPc.tPtptgPddpPtiPPPjctwstnpPvPapPxti:pi:psd’PuP.cpPphppwpPuPP
VIiwiIIIII’IIIIipIsbIk’vei.IVzspiwbIb:1:kImiwwIIkIIIgIliuaniiimtiI.izlIIiIgziwioniwIibIwbmpnIciIiIpO
mepsyaqddgmkqmiqmmMrtmMs:sqrt(5):MiIMmuMmMmsamMPMsmMmmzwuxemMmMMfzxzmmgMIumuaMgdyqgMpfzmMmmsmmm,mMwmMaym.S
DOrgddDh’qDhkpkowpbDvwdsqdDbVDDhxDz:sin(1):drsh;DbrzwDfDDhduuDDvhqqq’nlEDDuDDfdD.ri-DkdDhDDDvlzrvkDhdwDwD
JydRzsRkiRvrRzfxcRRRkRRuRRmlvz,vRjRdiRrRgfruRRcRrxiRuRrrRyvRfrirjrk:5:rkrRRyxRrjrRcRrRrlkexzCRiwRcVL
OG.GlGggGGkGGxGkgGguGgnGgjoGgnkGGggngprGt:gamma:GgGGjGGugGugiaGGjxGgcGgkvzkG,GoGxGkoGxqGntsGgGgGGyGoyhGY
Q.iftymnjxm.wjdxF!FFjFXfFnFzFFwffFFyFtyfffF,qF.jFFihFFhffjt:cos(1):fFtljmFxskdsmfjfmtfGFpFftfFfFfwFfgnFfF
-PBhesBnxmBBblBBboxz-b.fBpibNvBtzvB,:i:BbbfBbuBjbbBgBbbiuBijbBdPbpbBjuBBmBBpBBpBu.mc’BBfbgBsb’bbb’pO
LYJJJrjjqjaCLjtJnKnJvxbmnJJmXJ.jwdqjqcjnJawfxJUejnxbannJaJccnJcJJbJjhJhJnJd:zeta(3):JJJtfUaJjbqjnowwtjJjnW
UuufinCUvquUMuuUUsylUoxuhmUUU.UUiuyuUjsoUuuUgPyiuUUnUqmfUuUXUiUmUh:00001111:myUuunuunmyyuUgylbUUuyUubUUnUCu
SoOhOofxO.oooOoOf.OfOhccosoOuszswh.sggotowhboaOv:3:soqfvmvo.oopofOfqzOoOosrozssOsMhgwhsszorwOaOO.soo
HhHa:phi:Hhashhfhh,hHHhHbhHlbHlzhHlhdjiHlHlfzHtHhhvhhmHhyHvyvAHtnHHhrHll,hHhuNHfHupluhl’lsHfHvhwaaHlaP
-PygCvcrgkcjCkveCccg-CCgcCgCWrcgqvvChCt’cpkVppechvCCgqcgoCkcPCmCcugccwgcqrk-ce:0:Cntucc’Cgp’CuxCCuw,
NNnrrNpbqgNNa,NlNxnbgpteuuunNqNNsNfqNvrvNNvNNnffqNNsNogu)NNN:e:N’nbayaNnNfrNcatbrNv’NrsNNNNe.NnbreNN
tTghinTtfhTeTT:#00000A:TmtgTxrhTTTTkttrttTTgTmiXhtfTtnTdTT,xTthtlxTtTxe.btbbutTTTTTTbkuTthpxLatTmzwTtkxtET
OkVkcxKKoloryKxdbboksxKoKkkyDkKdKdsKoKKKnoKoxk:sqrt(2):ykKkpKoKKkuks,diykZykKyiykkmKlkk.KidkKdobbKouoKdKbS
UjQqQqjsQqjsuQxQqqdqQquduuue:2^2:ljeituqqsqQqkvQuqtQtqqkchqQQqjmjoQqQiQQcusvbjdtQQQkbqQqQeiqkaQouQqkTE
SwszegsSmFjdsxogSSszjqs,sswSSlSSswGwwsSqSjSSzsasfgksSfSSOwSsfSvSSSzqSfsswksSlSsSeSzSyS:3!:vzfqsSsDSJL
PCvEErxvzkxviqissi,elsEEhEesXeFEEesEvvrEgErsieelerriie(EEijEXrrvxeismovwEaExkEEeeeglErEaEieEEl:arctan(1/2):eEe
LdjLzqdalzLpospsLpaLrelclLL,LLlLlLynLl’ltLlLLLzyLcLlLllLtwzLLzzLcyLdfzf:2:LdzlfLLLloLLdLLdLlnLyetnpY

[?]
Type in a relevant tag, and click the button, and help organize this blog's information.
[More Help]

You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

mmplay - Massively Multiplayer Online Games wrote on January 4th, 2006 at 11:57 am:

[...] Ok, nicht wirklich. Aber Raph Koster macht sich in einem Blog Eintrag auf seiner Webseite Gedanken darÃ¼ber, was mit der gigantischen Ressource an kreativen Spielern in MMO’s alles mÃ¶glich und auch unmÃ¶glich wÃ¤re. Er vergleicht das Potenzial mit Projekten wie SETI@Home, nur das es hier keine Computer sind, die parallel an einem Problem arbeiten, sondern Menschen.

Er fÃ¼hrt das Beispiel von Byron Reeves an, der ein System Ã¤hnlich zum Medical Screen in Star Wars Galaxies entwickelte und mit realen Daten zur Krebserkennung fÃ¼tterte. Im Ergebnis war ein typischer Spieler fÃ¤hig die Krebsart in 60% der FÃ¤lle korrekt zu diagnostizieren. Nimmt man 30 Spieler zusammen, sind die Ergebnisse gleichwertig mit denen eines ausgebildeten Pathologen.

Raph Ã¼berlegt auf rein theoretischer Basis in seinem Blog “Treating players like numbers” wie sich dieses Potenzial, dass den Spielern zudem SpaÃŸ macht, weiter genutzt werden kÃ¶nnte. [...]

Ripples in the Sea of People wrote on January 4th, 2006 at 8:28 pm:

[...] http://www.raphkoster.com/?p=242There’s been a lot of heavy, arguably negative stuff here lately. So I thought it might be interesting to talk a little bit about possibilities instead. [...]

MMORPG - MMORPG Spieler heilen Krebs wrote on January 5th, 2006 at 7:00 am:

Eclectic Mayhem: Ben Darnell's blog wrote on April 17th, 2006 at 8:10 pm:

make this an enjoyable part of play, without privacy risks. Harnessing the wisdom of a MMOG crowd might prove useful–while one player might deliberately sabotage the screening process, 100 unrelated players are not likely to do the same. Raph Kosterintroduced this idea on his blog in January, 2006. Airport Screening Is A Badly-Designed Game written by Tony Walsh for Clickable Culture copyright (c) in whole or in part 2006, Tony Walsh Tagged: Design Gaming Life Mixed Reality Technology Watchdog

Clickable Culture wrote on April 17th, 2006 at 8:10 pm:

Game Tycoon wrote on November 12th, 2006 at 6:47 pm:

[IMG] Itâ€™s taken a while, but I want to return to my article on using games to tap the wisdom of crowds. First and foremost, Iâ€™d like to bring attention to the writings of Raph Koster. Raph informed me that heâ€™s been thinking about this idea as well for quite some time now. However, Raph one-upped me: he found someone who actually tested the theory! From Raphâ€™s blog: What [Byron Reeves] showed was a mockup of a Star Wars

Confessions of an Aca/Fan: The Official Weblog of Henry Jenkins wrote on November 26th, 2006 at 11:08 pm:

can develop, as Edery proposes, mechanisms for linking game play mechanics with real world data sets. Indeed, Raph Koster — another games blogger who has been exploring these ideas — does Edery one better, pointing to a project which actually tested this concept: What [Byron Reeves] showed was a mockup of a Star Wars Galaxies medical screen, displaying real medical imagery. Players were challenged to advance as doctors by diagnosing the cancers displayed, in an effort to capture the wisdom of crowds. The

Smart Mobs wrote on December 1st, 2006 at 12:18 am:

The Cooperation Blog â€” CooperationCommons wrote on December 9th, 2006 at 1:39 am:

Timothy Burke said on January 4th, 2006 at 11:00 am:

This is a bit of what I was thinking about in my paper for DiGRA–the potential fertile meeting ground between Axelrod-style work on artificial societies and virtual-world research.

One of the practical necessities for making headway, though, would either be successful research-oriented virtual worlds that were sufficiently appealing to players that they would play even though they were also being research subjects or the willingness of big commercial developers to allow open datamining of their worlds (or to share their own in-house datamining). Do you think the second prospect is likely? Because I see continuing barriers of time and expense to the first.

Mandrel said on January 4th, 2006 at 12:13 pm:

Looks like your link to the solution was swept away when the SWG boards were recently archived. I bet this was a realy fun activity for the community to participate in. If more challenges and puzzles were developed, mayhaps the game wouldn’t be where it is today?

Wilfried said on January 4th, 2006 at 12:19 pm:

What you describe is basicaly the concept of a Think Tank, so I guess you can find lots of ideas with looking at what Think Tanks are working at

Back when I was a teenager in school my English improved fairly much by playing Infocom games like “Wishbringer”, “Trinity” and “A mind forever voyaging”.

However, it makes me fear a little to think of what can be done with the wrong intentions in mind. Like manipulating political or historical knowledge or even spreading ideas of terrorism. Luckily by know this is mostly science fiction. But looking at the China Communist Youth League funding “Anti-Japan War Online” MMO, it seems pretty clear to me, what the intention already is.

Raph said on January 4th, 2006 at 12:40 pm:

The link worked this morning (the thread is still in the archives), but you have to be logged into the forums. Once you log in, the URL is http://forums.station.sony.com/swg/board/message?board.id=Development&message.id=323382.

The response was mixed, in large part because it was a collaborative puzzle-solving exercise, but the reward was given to the first person to send in a solution. (everyone who participated in the event got a separate reward, which probably would have sufficed). Several folks, including some on the team, disliked having an out-of-game event, which is something that I personally have no hang-ups about. Lastly, the more story-oriented folks felt a bit left out.

The people who did get into it seemed to love it. I think that ARGs prove that married with ongoing interesting fiction, both story fans and the puzzle-solvers can greatly enjoy this sort of thing.

Mandrel said on January 4th, 2006 at 1:34 pm:

Looks like I’m getting an error, maybe need Dev level permissions to see it. Anyways, now that I really think back, I seem to remember this, though I didn’t take part. I was too busy crafting, working on Weaponsmith at the time.

Why can’t something like this be made persistant? Include the clues into loot tables, and to solve it a player would speak to an NPC, input the text and surrender the clues for a reward. Maybe the first person on the server to complete it gets the “special” reward for finishing first, but the content is there for the life of the game. Dev hours aren’t seen as wasted on a “one- off” event, and the world’s history and lore is enriched.

MikeRozak said on January 4th, 2006 at 1:38 pm:

We use tools like SETI@Home to distribute highly technical tasks like signal processing. However, there are cognitive tasks we could outsource in this fashion, as in the cancer example. Could, for example, jury trials be improved upon if assessing the truth of evidence were a distributed game? Could airport security be improved by creating scenarios about infiltration in our MMOs?

This reminds me of my internal “rewrite” of the Matrix…

In the Matrix, the machine kept the humans around as an energy source, which is a silly premise. However, if the machines kept the humans around because humans were better at solving some problems than the machines, then the human subjugation in the Matrix made sense. All the pointless stuff you do in life is really part of solving a larger problem.

The Hitchhiker’s Guide also used this idea.

User-created content includes this idea as well.

Raph said on January 4th, 2006 at 2:32 pm:

Mandrel, the in-game quest part is indeed still there, and I believe people can still take a whack at solving the puzzle.

xaldin said on January 4th, 2006 at 3:45 pm:

The total immersion to learn a language would have been really helpful for me. I’d like to see more stuff like that honestly. I do think that online worlds/enviroments (which is what we’re talking about really not games by the current sense) have a lot of potential the issue becomes trust in the results. I’d be very nervous about letting say the population of SWG diagnose me for possible cancer. Your results may average out to even but I typically don’t go to an average level doctor either for those same reasons.

Mandrel said on January 4th, 2006 at 7:00 pm:

I know you like discussing generalities and theories here Raph, but this was bugging me, so I went back into the SWG forum archives for some info. I found a readable post about the storyline here. In fact, the clues for this stopped dropping about 2 years ago.
As far as your main idea, I think it’s an interesting way to approach design, but the implementation may be where the difficulty lies. How does one parse the “good” or “successful” data from the bad?
In your example, a typical player (what is that anyways?) was able to diagnose cancer correctly at a 60% rate of a pathologist. Was the sample size large enough to make those results meaningful? Was it a simple test of “is this cancer a. or cancer b.?”
I would go so far as to say that there are large groups working on problems, albeit not IN the games, but in the community on the forums. Their “incentive” is a better, more enjoyable game. Many games have passionate, well informed community members that spend hours tracking game mechanics, and writing proposals for improvements.
Here we run into that implementation problem, of parsing the good from the bad. For every one constructive, well- written proposal there may be hundreds of unrelated, off-topic posts (or data points) that drown out the good.

Raph said on January 4th, 2006 at 7:09 pm:

The Wisdom of Crowds theory says that you’d want to tackle specific problems that are easily quantifiable, not ones that are vague like “improve this game system.” The trick then becomes asking the right question.

“What should the damage be on this sword?” might a better example. Note, you’d take the AVERAGE of all the answers given, even the stupid ones.

Darniaq said on January 5th, 2006 at 5:58 am:

While the SWG puzzle was far beyond my cognitive abilities, the debate about its merit reminded me of the older UO events from, err, February 2001 I think it was. Every two weeks there was a new event, and one I remember featured a server riddle/puzzle and then a mass mob assault. Very good time that everyone could partake in.

The only issue I personally had with the SWG puzzle was that the reward dispensation was counterintuitive. Using people as parallel processors would seem to tie to giving a reward to all participants in some form. What I seemed to get from the SWG puzzle event was that it wasn’t so much the first person to figure out (anathema to Wisdom of the Crowds anyway, I feel), it was the first person to report it. Everyone deserved a reward for being part of the machinery. You don’t just grease one cog in a clock

And as to those who didn’t like it because it was an out-of-game event, I can understand that, but I personally don’t have a problem with it either. Heck, for all the reading and posting that goes on around these parts, we’re already taking the games with us out of the game worlds

Otherwise, parallel processing of any form has always fascinated me. It seems to counter the assertions of insular American culture (everyone’s a hero), but I think that just means people need to recognize where they are an individual and where they can be part of a larger whole on occasion. Maybe it’s just in the packaging of it. Few seem to have a problem doing the wave at a stadium.

Darniaq said on January 5th, 2006 at 6:00 am:

Oh, and for some reason, in these sorts of discussions I’m always reminded of Hari Seldon’s Psychohistory from the Asimove series (and which was picked up in a Star Trek book series, and I’m sure other places).

Raph said on January 5th, 2006 at 8:32 am:

Actually, Darniaq, the precise process was to get five fragments, combine them, and turn them in at an Imp or Reb recruiter for 500 faction points and a badge. The puzzle was above and beyond that. So everyone who participated at all in gathering clues (whether they shared them or not) did in fact get a reward.

There was also a large-scale “the faction that turns in the most gets a boost next month” thing.

We probably should not have had a reward for the puzzle solving itself — all it did was deliver some fiction.

Pyr0 said on January 6th, 2006 at 2:24 am:

This reminds me of Amazon Mechanical Turk, which is almost fun.

The users complete simple tasks that cannot (yet) be solved by computers, they call it artificial artificial intelligence.

Tasks include selecting the picture which best represents a business storefront or rank your Top 3 establishments in a given city.