For the last two years I've been designing an experiment in play: PMOG, the Passively Multiplayer Online Game (said P-Mog). Justin Hall, Duncan Gough, and I founded a company called GameLayers with investment from O’Reilly Alpha Tech Ventures, Joi Ito, and Richard Wolpert. We quit our day jobs last summer and got serious about bringing play to the world wide web.
The impetus of my design is my lifelong desire to play with the layers of information that, unseen but forceful, impact our real and online lives. I want to see the invisible world, or at least become more aware of it. Perhaps ironically, PMOG is not a visually intense game. I don’t know that we rank 2D, much less 3D. This is the game HUD that persists from page to page in the Firefox 2 and 3 browsers:
We're now in the beta of our second public version. Both versions were implemented as Firefox extensions that follow players as they surf the web. The players provide the game with access to their browsers; the game provides the players with weapons, writing instruments, a gifting system, and a self-generating RPG character.
We started out to make a casual, massively multiplayer online game that took place alongside the rest of a player's online life. To do that, we had to answer two questions. One: what kind of interaction that occurs alongside the Everyday can we provide to players that they'll accept? And two: how can the game provide players with a set of behavioral summations that they could reasonably attribute to their decision-making process?
In the first public version of PMOG we ran the game through a sidebar on Firefox. People forgot (or never knew! ha) that the game was happening. And while the sidebar had a lot of space, it was problematic precisely because the game could be so easily ignored once the sidebar was closed.
The current version of PMOG injects HTML overlays onto any html document with a http://. These events, as we call them, stack up in the bottom right corner of the Firefox 2 and Firefox 3 browsers and are easily dismissed. A lot of game data is delivered to the player via image and text, in those HTML overlays.
Game-to-player messages (You leveled up! You unlocked the Torch badge!) and player-to-player messages (hey dahling, will you be a dear and trade me 10 crates for a mine? kthxbai!) are delivered as HMTL overlays as well.
Fitting Into the Browser
Currently, there are three stages
of intensity to the appearance and persistence of the HUD. I'll begin
at the least involved level and move on to the most involved.
At its most invisible, players can completely pause the game. The
player ceases to earn datapoints from each new top level domain they
visit and they do not receive any in-game events. They can also hide
the heads up display, leaving a small "P" favicon in the bottom right
edge of their browser as the only reminder.
The second level of integration is that the player can choose to hide the heads-up display but continue playing, in which case events still occur and the player still earns currency but their tool inventory, level progress bar, and other game statistics are not present in their browser. In-game events occasionally appear in the bottom-right corner of their browser.
At the most involved stage of PMOG, the player has a styled tool bar at the bottom of their Firefox 2 or Firefox 3 browser that displays all their game statistics. Dismiss-able game events (tripped traps, invitations to mission, player-to-player instant messaging) stack up in the lower-right corner of the browser. Additionally our extension tracks their movements over URLs that begin with http://, but it only attach that location to the player if they engage with game content. This is the default setting.
Passive-ists aren't asked to roll for initiative and then take part in a full-on turn-based combat. Rather, moments of combat and gifting invite the player briefly into the gameworld. This is not the same as the strategic blow-by-blow that makes Dungeons and Dragons style combat so engrossing. PMOG's fun is often the fun of discovery and misdirection.
Making the RPG Casual
Role-playing game metrics are
already inherent in a lot of social networks, as Amy Jo Kim has pointed
out. In both role-playing games and social networks, players posture
within the space of the avatar and explore their identities online.
Players balance social relationships with their own goals and values,
and use these relationships to promote the exchange of virtual goods.
However,
distinctly RPG metrics of dexterity, strength, charisma, et cetera,
don't serve the browser-based internet in the same way they've served
the kitchen table. Attributes such as these exist in RPGs largely to
determine if an event has occurred and how the player can or does
respond to it: Have you been spotted by the lacedon hiding in the reef?
Do you attack first or are you attacked? Are you prone? In a gameworld
largely driven by the collection and reinterpretation of the players'
data streams, these metrics are unnecessary.
The landscape of
the internet is no place to stall a player's momentum. Rather than
spending the first hour of game time rolling a character, the play
simply begins. Over the course of hours the player will accumulate
characteristics. Though we don't want players to belabor the choice of
race and class, I do rather like "types" in games and wanted to use
them.
Associations
There are currently six associations in PMOG, and at any time each player is described by three of them.
Pmogeon associations reflect what game tools the players use and
determine what array of game tools can be purchased by the player.
Because the flexibility of the player within the game is determined by
their associations, I wanted player identity to be flexible as well.
Now associations are determined by the tools that the players have used
in the game, and by the associations of the players whose missions they
take. In this way players can both determine their designation and have
it be determined by the content they choose.
This may be a case
of adding additional elements when I should have subtracted (why not
remove a designation like class?), but I wanted to both have
associations and allow them to be flexible.
Initially, character associations were determined by attributing
characteristics to websites themselves. The player’s type was generated
from the number of times they’d visited those hundred URLs. Players
accumulated the sites’ characteristics. Reading the Wall Street Journal
made you more of a Destroyer; visiting Flickr made you more a Hoarder.
But
the players themselves resisted the approach. The landscape of the
internet has already been established by destination logic: websites
are places. Players expect to affect the environment rather than have
it affect them. The game is watching, sure, but the player decides when
and if to engage.
This is the point in the design process at
which the internet really became physical for me, and the aesthetic
decisions stem from that. Mines, the first tool I'd designed, were
initially meant to be crafted by players from flotsam they'd collected
on websites like so much primordial goo. Now they're prefabricated
tools you can buy, trade, set, or detonate. The imaginary world of the
internet in my mind became a city that had been built over and over
again, a sprawling maze of secrets.
Becoming Massive
Now
that people are playing the game again I have another set of
fascinating problems and questions. Scaling PMOG successfully is
paramount among them. PMOG would be less interesting if it lived on
shards; much of the gameplay relies on displaying the internet as an
uninterrupted field of play. We've also been warned against asking our
servers to calculate long strings of relationships like "friends of
friends"; this is supposedly one of the reasons that MySpace ate
Friendster's lunch. But if we become as massive as we aim to be, it
would be foolhardy to attempt to show players everything that happens
in the game. So we can limit visibility to approved relationships, in
the model of Facebook and MySpace.
Rather than using "friends"
to describe online relationships we use one meta-designation
(acquaintances) and two sub-designations (rivals or allies). Below is a
screenshot of one of my allies, Pixielo. The top of each player's
profile looks like this, with buttons for you to attack, gift, or
message them, and three buttons that display the tier of relationship.
These relationships will be a lot of fun to add features to. Maybe you can hunt your rivals in real-time games of tag; maybe your weapons deal less damage when your allies trigger them. There are tons of ways to continue to layer play within the existing structure.
Tomorrow afternoon at 2 during Emerging Technology, Justin Hall and I will be presenting some existing and imaginary surveillance-based game designs. If you're around, please stop by to argue with us.
Having decided to have associations, how did you decide how many you'd have and what they'd be?
If you're stressing the role-playing aspects of the game, why are you asking people for real-life data such as gender and age?
What's the aim of the game? I don't mean why is it fun, I mean where's the perceived finishing line?
Can I take back my last few visits, for when I forget to switch the game off when I visit the porn site? Or when I accidentally click the link on the poker ad?
How do you make money from this, or don't you?
Richard
Posted by: Richard Bartle | Mar 04, 2008 at 03:12
The last two paragraphs, of what Richard said, is precisely what I was pondering about.
I'm no privacy fanatic, but I think that the guys who oppose Facebook, for this very reason, would start falling unconscious when they see this page. Storing usernames, encrypted passwords, all browsing data, etc. Effectively combined with Facebook you'd have more than enough material to blackmail pretty much everybody. How are you planning to address the issue?
Posted by: Nicholas Chambers | Mar 04, 2008 at 05:36
The associations evolved from noticing the behaviors that people participate in online, which are very much like the player types that you've noticed in MMOs. I applied one "type" to each behavior, at that point there were four. That is also when I was first told about your work, actually. Now there are 8 associations and each has one tool. As the game progresses the associations will be used in combination with each other to unlock other tools in the game.
I think I misused the phrase "RPG" in the article, when I should have said "Making the MMO Casual" - they've become rather inseparable when they should not have. DnD is an RPG; World of Warcraft seems not to be. So I don't think we're so much stressing the role-playing aspect of the game as we are playing with the role-playing aspect of playing online together. It seems like the avatarization of identity online has, at least for the Facebook generation (of which I am a member), made each of us characters, removed a bit from our other selves that we display in school or with our families. The associations are very changeable, and I think that cuts down on the expectation for role play from the player. In PMOG most players play under pseudonyms, so far, and age seems to be neither here nor there.
I'm continuing to develop the leveling system (we have 20 as of now) and increasing players' agency as they level up. I don't know... I don't perceive a finishing line yet. That is partly the freedom of not running a 3D client, being a lightweight Firefox extension makes the game malleable. It can grow with its players and in response to them in very core ways. Ways as core as the point of playing, I guess! :)
You can always erase your history without losing the resources that you've gained from surfing, you can have a completely private profile, or hide the tools that you've used. Here's a screenshot of the information you can choose to reveal about yourself.
The game doesn't care if the site that you mined is a porn site or an Evangelical charity, it only cares if you've decided to participate in the game in that place. Otherwise the site itself is just a way to accrue resources. We used to attribute characteristics to the sites themselves, and at that time I considered porn and other sex sites to be a little too general as a type to apply characteristics to it. Perhaps if we'd continued down that path there would be a difference between the players that like furry porn and the ones that like BDSM, because that create a meaningful or interesting difference. But before we abandoned that system we'd instituted "genres" of sites, and sex in general seemed to only say that the player was, in fact, a human being.
Right now we don't make any money on this. I can foresee a time when players may choose to earn badges or temporary powers by opting to participate in sponsored content from media companies. Customizable characters that a lot of the free-to-play casual games employ wouldn't work for PMOG, since we're more textual than graphical. I suppose we could create a "space" online a bit like Animal Crossing, but I'm relatively unsupportive of taking the game too far outside of its natural habitat, the web. We've been advised that business models are emergent, and that the best one may not be the most obvious. That sounded true at the time. :)
At some point we'll need to make money to support and grow the game. I think that what we're doing is new enough as a medium that it's okay not to know exactly how to make money; overall we're committed to the player experience and maintaining the enjoyable lightheartedness of the world however we make money. So if anyone has ideas, I am all ears!
Posted by: Merci | Mar 04, 2008 at 05:37
"Can I take back my last few visits, for when I forget to switch the game off when I visit the porn site? Or when I accidentally click the link on the poker ad?"
The short answer is yes. The longer answer is that once we started playing PMOG for ourselves, we quickly realised that there were a number of privacy controls we needed, such as those Merci posted and also these ones that allow users to delete their entire browsing history, or their entire account.
With regards to the concern that we're "Storing usernames, encrypted passwords, all browsing data, etc.", it's a valid concern. However, PMOG only stores urls that you visit and even then only http urls, no https or IP addresses. You can delete this information from our servers at any time (see the privacy screenshots in my previous paragraph).
I think that these are the basic issues we should be expected to address. Even if it did get access to them, PMOG doesn't infer any gameplay mechanics from encrypted passwords, usernames, and so on, so the most sensitive data (from a privacy perspective) is actually the least interesting to us, in terms of making a game out of that human data.
Posted by: Duncan Gough | Mar 04, 2008 at 06:21
I think, I'll have to reword the statement:
My point is not to say that you can easily abuse the data entrusted to you, but rather that you may not be able to provide the security needed for this sort of thing. Here are just some things that you might want to consider:
- Are your IT security systems efficient enough to protect the data from hacking attempts?
- Was the system tested by a reliable party?
- Was the system written by a party with previous experience and expertise in the field?
- Can you provide basic phishing protection?
- Which data can I re-acquire from your servers, should I gain a foreign ID?
- Is the data sent to and from your server encrypted?
- Which employees can access the data? Can the access be backtracked?
- etc.
As I've said, I do consider the idea behind your project very interesting, however with the sensitivity of data you are acquiring, you are bound to higher security standards than an average MMO (exception: banking details). This applies just as well for your IT security, lines of authority within the group and the legal declarations you make on regards of selling or giving away the information to third parties. Currently, your privacy page is still rather slim for a program that sends my passwords back and forth across the net.
Posted by: Nicholas Chambers | Mar 04, 2008 at 07:15
This seems like one of the coolest ideas I've heard in a long time. I hope to get a chance to play with it and find out how it works, in the mean time:
How many people are playing? How casual is it actually? Are people having fun? What kind of innovation (if any) have the players exhibited in manipulating the game resources that really got you jazzed?
Posted by: Christopher Weeks | Mar 04, 2008 at 13:39
Have any of you "played" google image labeler?
Its a two minute timed round where you try to match tags of pictures with an anonymous partner and you get more points for more detailed key words and more images tagged in the period.
Is it fun? It provides a dopamine rush in terms of association and score motivation and the meeting of the minds with hidden person part of the fun even though there is no direct contact before, during, or after.
Perhaps there are other moderately useful and slightly interesting tasks that could become quests. Typing a few lines of text from images of old books like recaptcha ? Vetting and scoring websites for original content, topic, versus advertising spam. Participating in a survey?
Posted by: shander | Mar 04, 2008 at 17:02
How many people are playing?
As of today we have more than a few thousand people playing in our closed beta and we have over 15,000 people queued up to play.
How casual is it actually?
The casual-ness of the game depends on how the player decides to play it. Pixielo, the player whose profile I used in the article, has taken all but three of the missions in the system. She is obviously, a less casual player than most. I actually play very casually, since I'm pretty busy and don't have a lot of time to take the game up on the offered invitations.
Are people having fun?
Here's an example of the fun players are having. From what I've heard, people are having a great time. Here's another example from Techcrunch of Michael Arrington having a good time playing.
What kind of innovation (if any) have the players exhibited in manipulating the game resources that really got you jazzed?
So one of my favorite examples so far comes from a player called the1joebob, who wrote a script that automated crates filled with datapoints to re-up on a site he owned. He wanted to be radically generous and it was really inspiring to me, that he was customizing the game in order to gift other players without cheating even, just within the context of the game currency. Another player, Zous, helped to fund the1joebob's generosity by giving him datapoints when he (inevitably) ran out. That was pretty awesome.
Posted by: Merci | Mar 04, 2008 at 20:49
I don't get the "passive" part of the whole thing. Could you elaborate?
Posted by: BCKing | Mar 05, 2008 at 03:41
The resources (datapoints) are accrued passively by hitting top level domains. The associations (they determine which tools a player can freely buy as they level up) are generated by the game for the player, rather than the player choosing to be a certain race or class. Additionally, you can hide the toolbar and forget you're playing the game until you accept an invitation to take a mission or otherwise participate. You'll still be gathering resources; this removes the grind from direct player action/attention.
Posted by: Merci | Mar 05, 2008 at 04:04
very exciting Merci!
"The game doesn't care if the site that you mined is a porn site or an Evangelical charity, it only cares if you've decided to participate in the game in that place."
are there other ways to classify not only sites but online activity, so that PMOG could more meaningfully understand/account for what players are doing? e.g. it might not have to know which sites users are visiting in order to know that they are uploading content, downloading content, writing, scrolling, streaming media, etc. I'm fascinated by this, but would Love to elaborate ways to more meaningfully correlate what players are doing online with how they fare in the game.
as a side note: this is Stephanie, a friend of Howard's - we still have a date in late March, right? just checking ;)
Posted by: Stephanie Gerson | Mar 07, 2008 at 01:22
If only you could see our feature list... One of the most interesting problems we face in the future is how to expand the game to become deeper and richer without alienating our players. Chances are, nobody likes a page-scraper. And not everyone knows what good people we are. ;)
Posted by: Merci | Mar 07, 2008 at 01:28
This is really fascinating. Browsing = grinding. Well, I was browsing anyway, eh? Turns StumbleUpon into QuestUpon. Some thoughts...
If you are visualizing the Internet as place, and sites as sub-geographies, one way you could identify origin for players is through association with a particular site, either automatically or by number of visits (which wouldn't, of course, work if you're not storing that data...). Let's say, for example, I want to make Terranova not my "home page," but my "home." OK... everyone else who "homes" TN is, by definition, my ally. Not that we can't argue... but we aren't enemies, eh?
The trick is in calculating distance and (eventually) url-o-political (e-opolitical?) relationships based on this "location." How far away is LifeHacker from TN? How about Kotaku? Do sites with much in common aggregate or compete for resources? There might be some way to mine inbound and outbound linking information to determine distance and relationship.
I'm thinking that some way to self-identify with other players as "us" and "them" that doesn't involve "me knowing you," would help the play. It would also trim down the numbers when you get lots more players.
As far as making money... yikes. You've got a game that rewards going to a web site. Allow site owners to sponsor unique quests and items or to become specific "places" on the map, ideally related to what they do in RL. Have pre-made packs of items/quests/activities for players related to classic game locations like inns, training areas, forests, fountains, mountains, sports venues, etc. and sell them to interested sponsors.
So, for example, Killian's (and Heineken, Bud Lite, Absolut, etc.) could pay to sponsor in game pubs/inns where health can be regained, messages for friends can be left, people can play mini-games (darts), etc. A company that provides GPS tracking systems might want to sponsor items/quests for hunters. A security company might want to sponsor thief stuff.
I can also see licensing a mod-able version of this for academic use where a university department puts items/quests on pages with information about class and major-specific resources. I'd love to see the guild wars between the psych department and the architects...
Posted by: Andy Havens | Mar 07, 2008 at 10:37
We're also presenting a big breakdown of PMOG at SXSW on Monday at 2pm. Here is more information on that. :)
Posted by: Merci | Mar 08, 2008 at 23:45
Very intriguing; I'm waiting to see what turns up!
Posted by: Thuranor | Mar 09, 2008 at 07:31
I'm not a developer but I've been a player on PMOG beta. I love how this article addresses more of the direction the game aims to take and some of the methods used for gathering the data which makes it possible.
The information presented here does raise some security questions but I have faith that this band of developers have the combined knowledge to make PMOG happen or they would not have taken the challenge in the first place.
Changes will be made and security measures will be reinforced before PMOG opens to the masses without invitation. As a beta tester I know there is a chance something may not work properly and it is my responsibility to report it. I accept this for the sake of taking part in something which harnesses information never before used for gaming purposes. It's clever and fresh and unlike anything else out there right now.
Andy Havens- I love your spin on this. I would love to have the acquaintance system based on nesting points that we tend to visit most often. As these sites do not earn additional datapoints, they should be used for something. What better way to determine our allies than through the data we feed into the game. I also like the idea of sponsors setting up "shops" for us to stop at along the way on our missions. The game will have make money eventually, and this seems a good way as long as it doesn't get ridiculous.
Richard Bartle- What's the perceived finishing point of any MMOG? I'm not being facetious. I play and have played a number of "active" MMOGs and MMORPGs and I thought the goal was to meet people and track achievements through an entertaining outlet. With PMOG we have the acquaintance system, we have badges to strive toward, and we have an appealing steampunk style backdrop for our adventure.
BCKing- I haven't visited the PMOG site for a few days or taken any missions. I've also had my toolbar minimized. Therefore, I've been as passive as possible while still participating in the game because I was signed in and not paused. I've still earned datapoints and added to my surfing history. Other times I'll visit the site daily, visit new websites strictly for the purpose of gaining datapoints, create missions, lay portals for the purpose of guiding other users, and take user created missions. This would be an example of the more active Passivist.
In-game I'm Ellavemia, I've got nothing to hide. Look me up, I'm pretty chaotic neutral most of the time. Cheers!
Posted by: Aimee Valle | Mar 09, 2008 at 14:20