Rod Humble"Just as a poem doesn't need pictures and a painting doesn't need music, a game needs nothing else apart from its rules to succeed as a work of art." Rob Humble looks at the artistic aspects of game rule design.
Issue 41 - Game Rules as Art
- Thread starter The Escapist Staff
- Start date
Original Comment by: Patrick
http://www.kingludic.blogspot.com
Great article, a fairly clean breakdown of game design and its possibilities. I'm excited to see and be involved in the creation of the first games that classify as type 3a.
"Easier first steps can be made by creating an artificial umpire who can weigh competing emergent outcomes and make a rule out of the one which would best suit the game."
This is a rough estimate of how Facade's drama manager works, where the emergent outcomes have to do with how the language processor breaks down the input (and occasionally just breaks down).
"One can imagine a Type 3a game that inherits from pen and paper roleplaying games or umpired war games. The players would be motivated to do things "beyond the rules" - and an artificial umpire would generate new rules in response to this desire, in real-time. We can call this a "judgment system."
I think here you're spilling over into type 4 territory. For a 3a game, you can get around the need for this sort of new verb generation by simply offering lots of potential verbs in the background, and only allow specific sets to be decidable for a given context. Storytron operates in this manner, providing a hefty sea of verbs to flesh out the storyworld (hundreds up to thousands, breaking The Sims' record of about 50 verbs.) In this manner the user gets a sense that there is adaptivity going on, because the designer should make provisions for options like stealing cars when appropriate, and then the next choice will have different verbs as a result, but this is more in the designer's efforts than in an automated umpire.
Rules in a storyworld tend to be varied and numerous, and in the Storytron enginet his involves crafting a linear equation that weighs different reaction options an NPC might have. I think in designing and tuning these formulas, there is plenty of room for esthetics and ideological implication.
Storytron and Facade both employ a "drama manger" which acts as umpire. In every specific storyworld using the Storytron engine, designers can come up with their own metric framework that provides feedback to this umpire and allows it to adapt the storyworld accordingly. So many of the potentials you describe in this article are already underway.
Great article, keep it real over at EA.
http://www.kingludic.blogspot.com
Great article, a fairly clean breakdown of game design and its possibilities. I'm excited to see and be involved in the creation of the first games that classify as type 3a.
"Easier first steps can be made by creating an artificial umpire who can weigh competing emergent outcomes and make a rule out of the one which would best suit the game."
This is a rough estimate of how Facade's drama manager works, where the emergent outcomes have to do with how the language processor breaks down the input (and occasionally just breaks down).
"One can imagine a Type 3a game that inherits from pen and paper roleplaying games or umpired war games. The players would be motivated to do things "beyond the rules" - and an artificial umpire would generate new rules in response to this desire, in real-time. We can call this a "judgment system."
I think here you're spilling over into type 4 territory. For a 3a game, you can get around the need for this sort of new verb generation by simply offering lots of potential verbs in the background, and only allow specific sets to be decidable for a given context. Storytron operates in this manner, providing a hefty sea of verbs to flesh out the storyworld (hundreds up to thousands, breaking The Sims' record of about 50 verbs.) In this manner the user gets a sense that there is adaptivity going on, because the designer should make provisions for options like stealing cars when appropriate, and then the next choice will have different verbs as a result, but this is more in the designer's efforts than in an automated umpire.
Rules in a storyworld tend to be varied and numerous, and in the Storytron enginet his involves crafting a linear equation that weighs different reaction options an NPC might have. I think in designing and tuning these formulas, there is plenty of room for esthetics and ideological implication.
Storytron and Facade both employ a "drama manger" which acts as umpire. In every specific storyworld using the Storytron engine, designers can come up with their own metric framework that provides feedback to this umpire and allows it to adapt the storyworld accordingly. So many of the potentials you describe in this article are already underway.
Great article, keep it real over at EA.
Original Comment by: Rod Humble
Thank you Patrick,
I think the work you are doing on Facade is great! That?s insightful on how the Storytron and Façade is being crafted. Both works sound somewhat different (although related) and more ambitious than the kind of systems I am talking about because they involve storytelling. My head spins when I think of the challenges interactive storytelling brings in to play, you guys are climbing a mountain, the pay off would be wonderful.
For me, the goal of game rules being crated and modified on the fly is hard enough, but the rules are overtly just that, game rules rather than dramatic or narrative simulations.
It?s an exciting time!
Sincerely,
Rod
Thank you Patrick,
I think the work you are doing on Facade is great! That?s insightful on how the Storytron and Façade is being crafted. Both works sound somewhat different (although related) and more ambitious than the kind of systems I am talking about because they involve storytelling. My head spins when I think of the challenges interactive storytelling brings in to play, you guys are climbing a mountain, the pay off would be wonderful.
For me, the goal of game rules being crated and modified on the fly is hard enough, but the rules are overtly just that, game rules rather than dramatic or narrative simulations.
It?s an exciting time!
Sincerely,
Rod
Original Comment by: Duncan
http://ghostsinthegame.blogspot.com/
There are games that are stepping in to fill the roles of 3a and 4a type games already. The computer-driven models don't exist yet because AI is still hard to synthesize. But player-run systems yield a first look at how computers can act as frameworks for dynamic play.
For a 3a example, look at Neverwinter Nights. Players are able to create content, ranging from levels to characters, to art, models and even rules. The whole system in mutable. Beyond what can be created, the game can also be played in a dynamic mode. A player still has to take on the role of the DM, as in a PnP version, but has nearly as much control over the scenario. Rules can be created, enforced, bent, and discarded at whim by the DM.
4a represents more of an open system. Multiple players within a world, creating rules and content. This is then expanded upon, or redesigned by the next people to come along and play. Virtual worlds are starting to show us this kind of play. Second Life, comes to mind. So do Eve Online and A Tale in the Desert. Players create and enforce rules, the computer doesn't have to. Spore may also show us some of this sandbox style play in a single-player context.
Games have always been about interaction with other people. Until the computer can fully represent a person, the nxt best place to look is actual people.
http://ghostsinthegame.blogspot.com/
There are games that are stepping in to fill the roles of 3a and 4a type games already. The computer-driven models don't exist yet because AI is still hard to synthesize. But player-run systems yield a first look at how computers can act as frameworks for dynamic play.
For a 3a example, look at Neverwinter Nights. Players are able to create content, ranging from levels to characters, to art, models and even rules. The whole system in mutable. Beyond what can be created, the game can also be played in a dynamic mode. A player still has to take on the role of the DM, as in a PnP version, but has nearly as much control over the scenario. Rules can be created, enforced, bent, and discarded at whim by the DM.
4a represents more of an open system. Multiple players within a world, creating rules and content. This is then expanded upon, or redesigned by the next people to come along and play. Virtual worlds are starting to show us this kind of play. Second Life, comes to mind. So do Eve Online and A Tale in the Desert. Players create and enforce rules, the computer doesn't have to. Spore may also show us some of this sandbox style play in a single-player context.
Games have always been about interaction with other people. Until the computer can fully represent a person, the nxt best place to look is actual people.
Original Comment by: Rod Humble
Thanks Duncan that is insightful.
I agree human player based systems are tremendously useful and your examples are great ones. I do think that we can make progress towards artificial umpires without a full blown AI however. A ?dumb? artificial system which is modifying itself based on player feedbacks and requests doesn?t seem impossible. As you note several current games can provide learning on patterns. The informal and temporary social rules that players create in MMO?s may be a starting point for perhaps a more formal and permanent system. When discussing the article internally here at EA time and time again comparisons were made to how societies create and discard rules. Artificial societies are a separate topic, but the notion of agent based rules systems was interesting and possibly applicable back to plain games and their role as an art form which is my interest.
Sincerely,
Rod
Thanks Duncan that is insightful.
I agree human player based systems are tremendously useful and your examples are great ones. I do think that we can make progress towards artificial umpires without a full blown AI however. A ?dumb? artificial system which is modifying itself based on player feedbacks and requests doesn?t seem impossible. As you note several current games can provide learning on patterns. The informal and temporary social rules that players create in MMO?s may be a starting point for perhaps a more formal and permanent system. When discussing the article internally here at EA time and time again comparisons were made to how societies create and discard rules. Artificial societies are a separate topic, but the notion of agent based rules systems was interesting and possibly applicable back to plain games and their role as an art form which is my interest.
Sincerely,
Rod
Original Comment by: Mark
3a and 4a, I thought, are both being well-expressed in MMOGs these days. Patches and balances form a weak, vestigial sort of umpire-based balancing, but you can't deny that the social culture emerging from the likes of WoW is anything but a pure expression of 4a. In fact, I'd go so far as to say that the only way to get more "4a" than Second Life is to play Calvinball.
3a and 4a, I thought, are both being well-expressed in MMOGs these days. Patches and balances form a weak, vestigial sort of umpire-based balancing, but you can't deny that the social culture emerging from the likes of WoW is anything but a pure expression of 4a. In fact, I'd go so far as to say that the only way to get more "4a" than Second Life is to play Calvinball.
Original Comment by: Rod Humble
Mark,
Thanks! I agree MMO?s are instructive in how a future artificial system could work. I should have noted more stridently in the definitions (although hopefully its clear in the notes afterwards) that I am referring to artificial umpires or players.
Computer games that make themselves as you play them or change themselves as you play them.
I just checked out Calvinball, that?s awesome
Mark,
Thanks! I agree MMO?s are instructive in how a future artificial system could work. I should have noted more stridently in the definitions (although hopefully its clear in the notes afterwards) that I am referring to artificial umpires or players.
Computer games that make themselves as you play them or change themselves as you play them.
I just checked out Calvinball, that?s awesome
Original Comment by: Patrick
http://www.kingludic.blogspot.com
I should note that I had no role in the creation of Facade or Storytron, I've only been studying the two for a while. I think everything you said about the artistic integrity of rule-sets and their ideological implications applies to interactive drama.
Storytelling doesn't have to be a black box, its basically about peoplen Technically stories are sequential events in a narrative, but stories have always been produced by human cultures. The idea is that stories demonstrate rules, so for centuries human storytellers have buried rule systems in their works, rules that have profound implications on society and what-it-means-to-be-human. Those sort of systems usually didn't have much to do with how rockets fly through space, or how many hit points the protagonist has, or the exact dimensions of the places it puts its characters. They had to do with people and their inter-relations. Those are the raw elements of a new form of challenge; still the same medium as the games that have come before, but engaging forms of cognition, social and verbal, that have largely gone untapped.
It still offers lots of hard design problems, but a few are being solved already, so theres a definite possiblity we'll see a refined second generation of products in three years. By this time it will explode, and I don't mean "hey, casual games are making 800 million, thats an interesting trend," I mean everyone is going to flip because games will start selling thirty, fourty, even fifty million copies. Interactive drama has the potential to grab more and more market prescence as the product gets refined. Stories and human drama appeal to almost everyone with money to spend on entertianment.
In the meantime, I'd be happy if Storytron makes a 10% profit in 2007.
I think by the third generation of drama engines, we'll see research done in pure ludology and clever agent-based AI methods converge with the design foundations of interactive drama and bring us platforms for highly adaptive, improvisational stages and characters.
You should note that Calvinball was really about Calvin's relationship with his imaginary friend, the meta-play of rule changing anarchy is really a vehicle for a social dynamic going on in the kid's head.
http://www.kingludic.blogspot.com
I should note that I had no role in the creation of Facade or Storytron, I've only been studying the two for a while. I think everything you said about the artistic integrity of rule-sets and their ideological implications applies to interactive drama.
Storytelling doesn't have to be a black box, its basically about peoplen Technically stories are sequential events in a narrative, but stories have always been produced by human cultures. The idea is that stories demonstrate rules, so for centuries human storytellers have buried rule systems in their works, rules that have profound implications on society and what-it-means-to-be-human. Those sort of systems usually didn't have much to do with how rockets fly through space, or how many hit points the protagonist has, or the exact dimensions of the places it puts its characters. They had to do with people and their inter-relations. Those are the raw elements of a new form of challenge; still the same medium as the games that have come before, but engaging forms of cognition, social and verbal, that have largely gone untapped.
It still offers lots of hard design problems, but a few are being solved already, so theres a definite possiblity we'll see a refined second generation of products in three years. By this time it will explode, and I don't mean "hey, casual games are making 800 million, thats an interesting trend," I mean everyone is going to flip because games will start selling thirty, fourty, even fifty million copies. Interactive drama has the potential to grab more and more market prescence as the product gets refined. Stories and human drama appeal to almost everyone with money to spend on entertianment.
In the meantime, I'd be happy if Storytron makes a 10% profit in 2007.
I think by the third generation of drama engines, we'll see research done in pure ludology and clever agent-based AI methods converge with the design foundations of interactive drama and bring us platforms for highly adaptive, improvisational stages and characters.
You should note that Calvinball was really about Calvin's relationship with his imaginary friend, the meta-play of rule changing anarchy is really a vehicle for a social dynamic going on in the kid's head.
Original Comment by: Rod Humble
Thanks again Patrick, Your post has given me a lot of food for thought. I had previously put interactive storytelling as a different area completely but your description really helps me imagine it as a set of rules systems. Hmmmm.
I would add just one thing. I wouldn?t want to make the benchmark of success for interactive drama or a game as art, the amount of money it makes. They are much more important than that.
Thanks again Patrick, Your post has given me a lot of food for thought. I had previously put interactive storytelling as a different area completely but your description really helps me imagine it as a set of rules systems. Hmmmm.
I would add just one thing. I wouldn?t want to make the benchmark of success for interactive drama or a game as art, the amount of money it makes. They are much more important than that.
Original Comment by: rt
I think it should be pointed out that the character-simulation approach of Facade and Storytron can't be looked at narrowly as a solution for storytelling, because stories are shaped, really, by their own medium. A poem, a song, a movie, and a game may all contain stories; but they are told in different ways and emphasize different elements. ("The book is better than the movie." vs. "I don't have time for the book, I'll watch the movie.") At present the only devices we have to tell the story in a game are to attach, with great effort, a movie to the game(or all too often a game to a movie), or to embed it into the rules and let the story be expressed through the events of play. The former is easier to visualize at a high-level that to actually implement, and vice-versa for the latter(the genius of games like Civ, Tetris, etc. certainly doesn't rest within the art assets or physics engine). But it should be noted that in all cases these stories have been based on some kind of authorial control, some statement or commentary.
To introduce character simulation means that random and non-meaningful events are likely to occur and dilute the power of the story. This already happens with our simplistic mainstream rulesets. It happens in Facade, and it will happen in Storytron. That, coupled with the fact that players know in advance that they are dealing with automatons and not real people, make me suspicious of the claims of the interactive dramaticians. What is the motive to follow the story? And how would it be different from what the pioneering of "cinematic games" did in the 1980s and early 90s? The ID technology makes it more *detailed*, but to avoid degeneration, it still has to follow some predictably chunky pathways. So all we seem to have done is replace the cutscene-actionscene pattern with a cutscene-talkscene pattern where you manipulate the other characters with your smooth emoting and you manage the resources of lies and betrayal. That's great, but it's window-dressing in the same way that car-chases and gunfights are window-dressing, wheras the good stuff, the real story stuff, is in the decision-making of each character - what they do and why they do it. That's the part that has to have some meaning; if it's too predictable and has too much closure, we lose interest. If it's too vague, we lose interest for different reasons. That's the part that seperates the great/successful artists from the imitators. It's great tech, but it's not a silver bullet and it's not a new market: this isn't a competition between "social" and "asocial" entertainment. It'll be absorbed by the game industry in some fashion. You'll get a new genre and the tech will be integrated in bits and pieces into the old genres, just as how, say, 3d and CD-ROM led to some new genres that got integrated incrementally into the old ones. People will profit, but it won't be blowing off doors. And if what I've seen and heard is any indication, it's a pretty labor-intensive thing to create an ID story even after the tech is there. Sounds like something developers should be avoiding where they can.
I sense far more power lying in the advancement of online, multiplayer social gaming. There is a genuinely new frontier lying in that, and most of it is still touched upon only by text-based interfaces. WoW is very successful, and Second Life is *interesting*, but the really good stuff is going to lie somewhere between "you're in OUR world now, follow the rules and have fun grinding" and "build whatever crazy shizzle you like. No there is no particular game here"
I think it should be pointed out that the character-simulation approach of Facade and Storytron can't be looked at narrowly as a solution for storytelling, because stories are shaped, really, by their own medium. A poem, a song, a movie, and a game may all contain stories; but they are told in different ways and emphasize different elements. ("The book is better than the movie." vs. "I don't have time for the book, I'll watch the movie.") At present the only devices we have to tell the story in a game are to attach, with great effort, a movie to the game(or all too often a game to a movie), or to embed it into the rules and let the story be expressed through the events of play. The former is easier to visualize at a high-level that to actually implement, and vice-versa for the latter(the genius of games like Civ, Tetris, etc. certainly doesn't rest within the art assets or physics engine). But it should be noted that in all cases these stories have been based on some kind of authorial control, some statement or commentary.
To introduce character simulation means that random and non-meaningful events are likely to occur and dilute the power of the story. This already happens with our simplistic mainstream rulesets. It happens in Facade, and it will happen in Storytron. That, coupled with the fact that players know in advance that they are dealing with automatons and not real people, make me suspicious of the claims of the interactive dramaticians. What is the motive to follow the story? And how would it be different from what the pioneering of "cinematic games" did in the 1980s and early 90s? The ID technology makes it more *detailed*, but to avoid degeneration, it still has to follow some predictably chunky pathways. So all we seem to have done is replace the cutscene-actionscene pattern with a cutscene-talkscene pattern where you manipulate the other characters with your smooth emoting and you manage the resources of lies and betrayal. That's great, but it's window-dressing in the same way that car-chases and gunfights are window-dressing, wheras the good stuff, the real story stuff, is in the decision-making of each character - what they do and why they do it. That's the part that has to have some meaning; if it's too predictable and has too much closure, we lose interest. If it's too vague, we lose interest for different reasons. That's the part that seperates the great/successful artists from the imitators. It's great tech, but it's not a silver bullet and it's not a new market: this isn't a competition between "social" and "asocial" entertainment. It'll be absorbed by the game industry in some fashion. You'll get a new genre and the tech will be integrated in bits and pieces into the old genres, just as how, say, 3d and CD-ROM led to some new genres that got integrated incrementally into the old ones. People will profit, but it won't be blowing off doors. And if what I've seen and heard is any indication, it's a pretty labor-intensive thing to create an ID story even after the tech is there. Sounds like something developers should be avoiding where they can.
I sense far more power lying in the advancement of online, multiplayer social gaming. There is a genuinely new frontier lying in that, and most of it is still touched upon only by text-based interfaces. WoW is very successful, and Second Life is *interesting*, but the really good stuff is going to lie somewhere between "you're in OUR world now, follow the rules and have fun grinding" and "build whatever crazy shizzle you like. No there is no particular game here"
Original Comment by: Mark
http://www.frontal-lobe.net
I'd guess that Spore is a type 2a, though depending on where you draw the line between "game" and "metagame" it might be a 3a.
http://www.frontal-lobe.net
I'd guess that Spore is a type 2a, though depending on where you draw the line between "game" and "metagame" it might be a 3a.