next »

[Alex Vostrov]

Yesterday I posted a comment on Twitter directed at Paul Eres. “You know Paul, it just kills me to see you pouring all the work into levels. Why do you focus on linear content production?” I think that I may have struck a nerve, because his response was, “well, to me, the heart of game design is content, not mechanics. most of what you believe about game design is a lie.” I can see how my original message could be taken the wrong way, but it wasn’t meant to offend. I've written this post here and on my blog to explore the issue further.

If you follow Paul’s blog, you know that he periodically posts messages about having finished some area of Saturated Dreamers. The game must be huge, because he’s been at it for a while now. One time he put up a video of the level creation and it looked like a painstaking process. I think that he places all of the game objects manually, essentially hand-crafting the levels.

This really bothers me at some level. Perhaps it’s because I really liked Paul’s last game – Immortal Defense. The mechanics were well made, but it was fundamentally another tower defence game. The compelling thing about it was the story – a tale of an immortal protector being trapped in an alien void. I suppose that it’s not surprising that Paul is a staunch defender of the “content” approach.

What’s my problem with this approach, anyway? If it works, it works, right?

Linear Output

What bothers me is how much work goes into making a game. Think about it. A smart, talented and passionate designer like Paul goes underground and cranks out a game in two years. That’s TWO YEARS of someone’s life. How long does it take to play; 8-20 hours? That an incredibly worrying ratio. Here’s the root of the problem:

Content = Time * K

The content that one designer can produce is essentially proportional to their time and the K is skewed against you really badly in this case. The good part, I suppose, is that the case above is at least somewhat feasible.

Diminishing Returns

Other approaches aren’t so lucky.  For example, this is what it takes to create a branching storyline.

Content = Log(Time) * K

Because of the branches, creating content actually gets harder as you keep creating. This is a really bad situation to get into – a combinatorial explosion. It’s the primary reason for us not having true interactive storytelling yet. Remember how in Mass Effect 2 the characters more or less ignored your actions in the first game? The above formula is the reason.

This is why seeing any designer chip away at the content bit by bit is disturbing to me. I want to see more games like Immortal Defense and that’s hard if designers have to toil at every pixel. So, what’s the solution? One is leverage – get better tools so that your K goes up. That’s another topic. I propose to turn the formula around.

Harness the Explosion

Content = K^Time

In other words, the more effort you put in, the more content you get – the effect is exponential. How is this possible?

Instead of thinking about the game as a series of events or a narrative, what if you shifted your frame of reference? What if you thought of it as a set of interacting mechanics? The advantage of this stance is that each element that you add to the game has potentially interesting relationships to all others. The designer harnesses the power of combinations instead of being a slave to them.

I’ve briefly glanced this mode of design in my work. Attack of the Paper Zombies, for example, has a random zombie mutation mechanism. I tried to make it that different combinations of upgrades would feel unique to the player in significant ways. A better example of this is Tarn Adams’ Dwarf Fortress. The stories in that game are entirely an outcome of the interacting game elements.

Problem Solved?

I think that I’ve made a good case for why working primarily on content is not the most efficient approach. Now let me discuss some reservations that I have for this method. Firstly, just throwing in more game elements is not the solution. They must interlock with other parts of the game in an interesting and comprehensible manner. Fortunately, this is just a mater of craft. A more serious problem is that the designer still has to keep the entire interaction set in mind. In other words, while you have this amazing creative power, your judgement ability doesn’t necessary scale to compensate.

The above concerns may or may not be significant. Roguelikes have been using the power of combinations for decades and they don’t show any signs of hitting theoretical barriers.  The best course of action, I think, is to try this approach on a lot of games to see what comes out.

[Aquin]

You know, the topic title is a good one for a game.

Speaking of, I can see what you mean.  I'm a bit surprised Paul kinda jumped down your throat with-- no wait, hang on.  Never mind.

But what you suggest...isn't this what metroidvanias and roguelikes (when they do it well) are already for?  I don't see how a story-driven game with heavy narrative could possibly benefit from this structure. 

[Alex Vostrov]

You know, the topic title is a good one for a game.

Speaking of, I can see what you mean.  I'm a bit surprised Paul kinda jumped down your throat with-- no wait, hang on.  Never mind.

But what you suggest...isn't this what metroidvanias and roguelikes (when they do it well) are already for?  I don't see how a story-driven game with heavy narrative could possibly benefit from this structure.  

I admit that doing something as meaningful as Immortal Defence would be very hard right now.  Is this because it's impossible or because we haven't tried hard enough yet?

I see glimmers of hope in DF - you periodically see players telling stories of dwarven tragedy in their games.  Perhaps that element could be isolated and refined.  DF spends, what 90% of it's time simulating spleen physics and stuff like that?  What if we spent the time simulating the emotional responses of characters?

By the way, I don't mind Paul's response.  We need more people who put a lot of thought into their design.  If they happen to be opinionated, then it's that much more fun. 

[Seth]

What bothers me is how much work goes into making a game. Think about it. A smart, talented and passionate designer like Paul goes underground and cranks out a game in two years. That’s TWO YEARS of someone’s life. How long does it take to play; 8-20 hours? That an incredibly worrying ratio.

Is that a worrying ratio?  Seems reasonable to me.  Take a look at how much time and energy is often put into novels (can be read in 8 hours) or movies (watched in 2).  Or, hell, what about a painting?  Seen in one second.

Not that I don't like games like Dwarf Fortress--I really do love them.  But it's true, I think, that procedurally generated games can often lack a certain consideration that painstaking and extremely well designed games can have.  And generally I don't think it's good to be thinking in terms of how many gameplay hours you can squeeze out of something.  Honestly I would prefer it if most games kept it around a really really solid ten or something.

[Alex Vostrov]

What bothers me is how much work goes into making a game. Think about it. A smart, talented and passionate designer like Paul goes underground and cranks out a game in two years. That’s TWO YEARS of someone’s life. How long does it take to play; 8-20 hours? That an incredibly worrying ratio.

Is that a worrying ratio?  Seems reasonable to me.  Take a look at how much time and energy is often put into novels (can be read in 8 hours) or movies (watched in 2).  Or hell, what about a painting?  Seen in one second.

Not that I don't like games like Dwarf Fortress--I really do love them.  But it's true, I think, that procedurally generated games can often lack a certain consideration that painstaking and extremely well designed games can have.  And generally I don't think it's good to be thinking in terms of how many gameplay hours you can squeeze out of something.  Honestly I would prefer it if most games kept it around a really really solid ten or something.

You're right about the books and movies.  They are in the same boat.  The difference is that they can't escape.  Games can.

With regard to length, think of it in terms of cost.  Right now, games are more expensive to make than books.  Because of this, books have a much greater reserve to draw upon.  This leads to there being more great books.

If game designers get more bang per design hour, the average quality of games will rise.

[Melly]

It's an interesting way to think, certainly. And I always enjoy thinking about things in a different way.

I think what you're getting at using a system like this is emergence. Something like this is probably not tailored for very linear, directed games, but more for experiences where gameplay and story can simply emerge from a system of interconnected mechanics. Like Dwarf Fortress. The more complex such a system becomes (hopefully without sacrificing too much usability) the more varied the emerging possibilities are.

Another thing I like about your suggested approach is that it forces the developer to think about his mechanics in a more meaningful way. He can't just throw them there as gimmicks, they have to interact with all previously implemented mechanics. This raises the complexity of implementing each new mechanic, higher the more you already have, but in turn it means less menial labor.

[Seth]

If game designers get more bang per design hour, the average quality of games will rise.

I'm sorry, I don't follow this?  I thought you were just talking about using stuff like procedural generation to make more hours of gameplay.  To me, more hours are not always better than fewer and that randomized content is hardly ever better than well designed content.

(also, I'm actually not sure Dwarf Fortress is a good example of making a better ratio.  Sure, most people can play it for a long time, but do you have any idea how long he's been working on that?)

[ஒழுக்கின்மை (Paul Eres)]

the 'lie' reference was a reference to the saying 'everything you believe is a lie' and not to be taken literally -- i forgot that english wasn't your first language, should have been more careful, i didn't mean to make it sound as if i was offended; you didn't strike a nerve per se, it's more that i was dismissive because pretty much 99% of game devs nowadays (through the theoretical influence of crawford) feel that gameplay/interaction is the most important thing about a game, and that anyone who thinks otherwise is a heretic.

so i've come to be somewhat brusque about the subject, since there's absolutely no way to convince someone that they're wrong on this topic, and arguing over it is pointless. your OP is an example of this; you prove theoretically that you're right, at the cost of ignoring case studies and empirical evidence to the contrary (such as many of the games many people, including probably you, most like are content-heavy games).

i've presented my thoughts on this topic elsewhere (most recently tigirc and in my early questions on formspring, and in these forums in a thread i don't remember the name of right now, but i remember alec had a big post in it where he presented an opinion similar to mine) so won't go into it much, but basically i believe separating gameplay from graphics, sound, story, etc., is a bad idea and that believing in that separation and favoring gameplay over the other inseparable parts of a game is the primary reason for how bad most games are (both aaa games and indie games).

so yes, i really do believe that most game design theory is false, from its basic axiom, from first principles. a realistic theory of game design would have to start from scratch, and treat games as a whole, as organisms where every part is crucial and changing one graphic or one sound effect can cause the entire game to fail or to excel, rather than as machines with interchangeable parts and "gameplay" as the core system and graphics, music, and story as the chrome that surrounds the core system and is just there to be a pretty interface to the world.

note that this is opposite to the way it was when crawford started out, so we've come full circle: he used to have to struggle terribly to get anyone to care about gameplay or interaction, whereas now we have to struggle terribly to get anyone to think there's more to a game than its interaction and core mechanics.

anyway, because i see games as organisms where any little bit can ruin or compete the rest, i feel that the most important part of game creation is content creation, or more precisely, that everything in a game is content, and that game design is game content creation and arrangement. the core game design rules are like the outline to the plot of a novel, whereas content creation is the actual writing of sentences. the outline is important in its own way, but it really isn't a part of the novel. likewise, the rules of a game are important in their own way, but they aren't the game, they're just the rules which the entities and content within the game follow. so the "game" is really the content, not the rules the content follow, just as we are our individual cells, not the laws of biology or our DNA code.

[Alex Vostrov]

I'm sorry, I don't follow this?  I thought you were just talking about using stuff like procedural generation to make more hours of gameplay.  To me, more hours are not always better than fewer and that randomized content is hardly ever better than well designed content.

(also, I'm actually not sure Dwarf Fortress is a good example of making a better ratio.  Sure, most people can play it for a long time, but do you have any idea how long he's been working on that?)

Now that I think about it some more, you're right.  If all that you're looking for is a fun 10 hours, the fixed content creation is probably superior.

My original thought was that you could scale it down to still be 10 hours with less work.

[Alex Vostrov]

the 'lie' reference was a reference to the saying 'everything you believe is a lie' and not to be taken literally -- i forgot that english wasn't your first language, should have been more careful, i didn't mean to make it sound as if i was offended; you didn't strike a nerve per se, it's more that i was dismissive because pretty much 99% of game devs nowadays (through the theoretical influence of crawford) feel that gameplay/interaction is the most important thing about a game, and that anyone who thinks otherwise is a heretic.

so i've come to be somewhat brusque about the subject, since there's absolutely no way to convince someone that they're wrong on this topic, and arguing over it is pointless. your OP is an example of this; you prove theoretically that you're right, at the cost of ignoring case studies and empirical evidence to the contrary (such as many of the games many people, including probably you, most like are content-heavy games).

i've presented my thoughts on this topic elsewhere (most recently tigirc and in my early questions on formspring, and in these forums in a thread i don't remember the name of right now, but i remember alec had a big post in it where he presented an opinion similar to mine) so won't go into it much, but basically i believe separating gameplay from graphics, sound, story, etc., is a bad idea and that believing in that separation and favoring gameplay over the other inseparable parts of a game is the primary reason for how bad most games are (both aaa games and indie games).

so yes, i really do believe that most game design theory is false, from its basic axiom, from first principles. a realistic theory of game design would have to start from scratch, and treat games as a whole, as organisms where every part is crucial and changing one graphic or one sound effect can cause the entire game to fail or to excel, rather than as machines with interchangeable parts and "gameplay" as the core system and graphics, music, and story as the chrome that surrounds the core system and is just there to be a pretty interface to the world.

note that this is opposite to the way it was when crawford started out, so we've come full circle: he used to have to struggle terribly to get anyone to care about gameplay or interaction, whereas now we have to struggle terribly to get anyone to think there's more to a game than its interaction and core mechanics.

I wasn't aware that idiom was in wide circulation. *shrug*  I think that it has less to do with my language and more to do with Twitter's brevity restrictions.

As far as your stance goes, I thought that Jesse Schell put the case forth pretty well in The Art of Game Design.  He's of the holistic experience school of thought as well.  Besides, I think that you're overstating Crawford's influence.  If it was as great as you state, he wouldn't go into exile to work on Storytron alone.

If anything, the mainstream industry is on the other side, Paul.  They've been tripping over their feet to cut out the gameplay from games since the early 90's.  Witness the love affair with Hollywood.  We have designers like David Cage who would switch to director at the drop of a hat if they let him.  What about the cutscene-fest that's the latest MSG?

To be honest, I wasn't really looking to wade into this discussion again with my initial comment.  My point was that there's a way to get non-linear returns on your time as a designer.  Admittedly, I have to concede that this would be very hard for games that you create.

[J. R. Hill]

It's possible to create something emotionally and intellectually stimulating (and fun) that way, but if you want everyone to have the same experience it obviously wouldn't work.  And let's face it, a good scripted game that you only play through once can have a far greater impact on a person than a procedurally generated game you play three hours a day.

They each serve a purpose, but to compare the two is like trying to compare reading a book and playing basketball.  In the end it's going to come down to personal preference more than which is most useful.

[george]

I think the OP is very interesting and worthwhile, but what always strikes me about this idea (time = content * k) is that it seems to forget this:



I think the multiplier is not just in the creation of the material, it's in the experience of it as well. Does 'linear content' have more weight in that experience than mechanics?

[ஒழுக்கின்மை (Paul Eres)]

@alex - i feel that it's the publishers in the mainstream game industry that favor content, not the programmers or game designers; i feel the 'interaction is king' theory dominates game programmers / game designers, as well as games journalists / game reviewers, and most intellectual gamers. besides, while most games in the mainstream industry favor heavy content creation (especially in japan) that doesn't mean that they theoretically understand why it's good to favor that, or are favoring it in the right way.

and yeah, the issue is narrower than what i mentioned, but i feel that the reason you believe what you wrote in the OP is because of the theoretical underpinning that content doesn't really matter, that once you create the game design rules most of the job is done and the rest is just tedious work. to me, creating good content is almost everything. creating the content for games is the most fun part of game design for me.

i see a lot of developers create 'tech engines' with an example level, and shy away from content creation (more levels, more enemy types, more spells, etc.) -- i think that shying away is a consequence of the gameplay is king theory. because if core game mechanics are the things that really matter, then creation of interesting content (even interesting level designs) is necessarily unimportant, and thus boring tedious work. but that isn't really the case, creation of content is the heart and soul of game design. you're just as much creating a game when you work on a particle effect as when you work on the scoring system. if someone believes that particle effects are not every bit as important as the scoring system, likely all they'll create will be tech demos, and never finish games, because they'll see finishing games as working on unimportant parts.

(i think i repeated myself just there, but i guess i could say it was for emphasis, since i think this is important; it's important that people realize that designing a game is more than designing the game rules, that creating level 3 or level 30 is even more important than creating level 1 was, not less important.)

[Alex Vostrov]

@alex - i feel that it's the publishers in the mainstream game industry that favor content, not the programmers or game designers; i feel the 'interaction is king' theory dominates game programmers / game designers, as well as games journalists / game reviewers, and most intellectual gamers. besides, while most games in the mainstream industry favor heavy content creation (especially in japan) that doesn't mean that they theoretically understand why it's good to favor that, or are favoring it in the right way.

This is a good point.  Programmers tend to be a lot more system-minded, it's true.

and yeah, the issue is narrower than what i mentioned, but i feel that the reason you believe what you wrote in the OP is because of the theoretical underpinning that content doesn't really matter, that once you create the game design rules most of the job is done and the rest is just tedious work. to me, creating good content is almost everything. creating the content for games is the most fun part of game design for me.

I've created both content-heavy games (Pandora's Gearbox) and mechanic heavy (Swarm).  To me, when I'm making a mechanic-heavy game, the system IS the content.  That's the value that I'm delivering to the player.  I don't shirk the responsibility of creating the game, I just do it at a different level.  It's a noun vs. verb thing.

Curious that you consider spells and such to be "content".  I'd say that they're a bit of both.

You're correct that my perspective of the interactivity debate informs my views.  I suppose that for counter-balance I should create a story-heavy game.  I have come to respect the power of storytelling recently.

Also, I'm curious about something.  I saw the story of ID almost as a comment on the meaninglessness of tower defense in the larger scope of things.  In other words, it was an anti-immersion game.  Most games pretend that their world is the real world and the player is meant to ignore the jagged edges.  ID used the very jagged edges of the TD genre to make a point.

Was this intentional or was it more of a coincidence?  The reason I ask is that I thought that it was a very effective example of using the gameplay to support the story.

[ஒழுக்கின்மை (Paul Eres)]

i kinda hate when people ask me questions about ID because i made it like 3 years ago and hardly remember anything about it (i haven't even replayed it since the last big update to it, in oct 2008). so i'm just going to have to answer with 'i don't remember whether it was intentional or not' -- wynand (who wrote most of the story) could probably answer better than i could. however, you could read the interview he had with patrick dugan, here (it's down now, so here's the archive.org link):

http://web.archive.org/web/20070610175242/http://kingludic.blogspot.com/2007/06/q-with-immortal-defense-writer.html

he goes into the issue a little there.

and i do agree that "gameplay" should support the "story", but it's more that i see the two as inseparable: there is no such thing as gameplay or story in isolation. every part of a game should interconnect with every other part (in ID, this was called indra's net). another simple way the two support each other is that whenever a new enemy type appears, it usually has a role in the story. or how each of the tower types is also sort of a character, etc. or the whole love/hate final secret hell level.

[shig]

Team Fortress 2 is doing this thing OP is talking about very well with their updates. I think TF2 is probably a better example of this strategy than Dwarf Fortress, since the game didn't initially have the updates's content when we bought it, and we can see the increase in players and gameplay variety within every particular update being released sporadically.

[gimymblert]

Procedural is generally bad, not in idea, but because they are design with generic content creation in mind, not specific. On contrary Tiling is design to cover generic as well as specific arrangement.

Even with content heavy, you can still remain on the right side of the explosion. Modularity need to be design with specifics in mind.

[SirNiko]

Sometimes the nonlinearity IS the game, though.

Megaman does this by making the acquisition of the weapons a key game mechanic. "Who should I fight first?" is an important, meaningful decision, as is "Who is weak against this weapon?".

Metroid Zero Mission did the same by making the Low% game have multiple avenues to attempt, so the player was really open to find what works best (like if you could bomb jump well enough to skip the speed boots).

But yeah, sometimes you need to look at your non-linearity and question what it gets you. Final Fantasy 13 was totally linear (Except some parts near the end), and it worked out well. Compare that to, say, FF7 where the non-linearity mostly meant you could do optional sidequests and totally overpower the final boss, or Oblivion where the nonlinearity just meant wandering around disconnected from any sort of plot while you did the root picking sidequest.

I don't think I've ever played any of Paul's games, or if I have I didn't realize he made them, so I can't comment on whether his titles would be improved by more linearity.

-SirNiko

[Alec S.]

I believe that both handmade content creation and procedural emergence have an important role in games and should be used in conjunction.

I would tend to agree with Paul on the importance of content.  A game should not be considered to be mechanics first and everything else far last.  All the elements need to go together, and handmade content goes a long way to making a game feel interesting and well-made.  More love goes into a hand made level, as it were.

Procedural elements and emergence are important for making the gameplay experience personalized for the player and interactive.  There are a couple of things that I would say about procedural elements, though.  The same care that would go into traditional content should go into the procedures.  Also, I think that procedural elements of the game should try to interact with the player in some way, ie. the player has input into the procedure.

I guess a hypothetical example that would bring this all together would be a game with a pre-made world which can change and develop through player interactions with procedural systems.

[BlueSweatshirt]

I agree with Paul, in that games are like organisms. But you have to remember that organisms have key underlying functions, just like games. They also have key overlying functions, just like games. The underlying functions, are say, the engine and the guts. Not so important for a player-- Very important for the one constructing the DNA of their game. The overlying functions, though, are the core mechanics, and what the player interfaces with every minute.

You can also see it as games are kind of like buildings. You need a solid foundation before you can build up and create your structure, and the finally decorate your new 45-or-so-story-building. The foundation is crucial to keep your building sturdy, but if there's nothing inside, there's really no point to the building, is there? Ultimately, you need a good foundation and a well-placed interior to have a successful game. Everything that goes inside contributes to the interior, from the reception desk to the elevator music, to that plant in the back corner of the Boss's questionably over-sized office at the top floor.