Notgames Forum

A goal for our (frictional game's) upcoming game is to remove any sort of puzzle like obstacles, and instead just base the entire game on interesting interaction. The idea is that the player simply progresses by engaging interaction and does not encounter any game stopping puzzles.

Now the problem is that we need some kind of choke-points, ie milstones, in order to craft a narrative and to make the production feasible. What I want to have is some kind of clear goal for the player, that they can move forwards to, a goal that is pretty obvious yet provides exploration along the way. In mario-like platformers the goal is simply to go to the end of the level, many times simply by moving in a certain direction. This "move to the right"-approach is pretty nice for simple 2D games, and games like Everyday the Same Dream use this nicely.

However, we are going to have complex 3d environments in our upcoming game and a simple move to the right approach is ruled out. One could use something like Half-Life 2, where there is only a single way to go forward. However, this makes the journey very linear and forced, something I want to avoid. So what we have been discussing is having something like a gps-system that points the player in the right way. The problem I have this is that it does not allow players to think for themselves. It is very easy to fall into a play style of simply going directly to the assigned goal location. I had this problem in games like Dead Space and Bioshock, where I just followed the directions, unaware of what I was supposed to be doing.

Another idea is to have some kind of bread-crumb trail that leads to the milestone, for example blood trail or a pipe. This forces the player to investigate the environment in order to find the milestone and thus creates greater sense of immersion and encourages exploration. However, this can be quite hard to use for all milestones in a game, and will only fit for a few.

Finally, one can use a todo-list and simply list one or more tasks the player must do in order to reach the milestone. The problem is that you normally need a sort of map to go along with this, and thus you get the same problems as with the gps-like approach. Also, spelling out to the player exactly what to do can be really immersion breaking and end-up making the player not think about their actions. If vague hints are used the player becomes stuck anyway and it is not helpful.

The question is now what other ways there are of leading the player? I am curious to hear what other examples from games you liked or how you approach this problem your self.

Sounds great!
I loved when this sort of thing happened in Amnesia, when I was doing things because I could. The only problem there was that I would often do things that were absurd in and of themselves (picking up a piece of pipe just because you can) because they only make sense in the context of a puzzle. I imagine for the new game you will make interactions rewarding on their own?


My first question would be "Are you sure?"
If you haven't done so already I'd suggest you make a little list of the reasons why you need these milestones and then check if there's other solutions. Maybe there's a better solution.


I had the same experience in Bioshock. But I was kind of thankful for it, because I didn't really care about the gameplay or the story. At least it prevented getting stuck in a place where you don't want to be stuck.

But: what if it's fun to get stuck? Maybe the problem could be turned into its own solution.


I find this problematic also because it requires you to explain this trail in the narrative. And the explanation might be hard to believe, which would break the immersion.


An alternative to this would be to simply require that the player does 5 things (out of a given 10 e.g.). It doesn't matter which 5 things. So you don't have to give anything away. This way players can also time their progress.


If you can't find a solution that can be part of the fiction or the narrative, maybe you could consider a completely artificial, symbolic solution.

When the game is satisfied with what the player has done, it could simply say "End of chapter 3", no matter where the player is. And then the player gets an option continue exploring chapter 3 or to move on to chapter 4. If the latter, you fade out and zap to the new location.
Additionally, you could have a symbolic representation of the number of "meaningful interactions" they can do in a given chapter and tell them how many they have done.

Is it required that the player walk all the way to the end of the level?
Are the chapters of the story separated by a door in the game world?
Is it necessary to map the linearity of the story to the space?

(I added a whole chunk of ideas to that and my browser ate them all  :( )

The best one I can remember is to put entry and exit of a level right next to each other. So if the player remembers where they came from, they know where to go. If it's necessary, you can simply keep the exit door locked until the game is satisfied with the performance of the player. But personally I would leave the exit open at all times. And if the player wants to run through all chapters without doing anything, let him. It's his loss.

Other ideas were to add an element to the narrative that justifies leading the player. I guess these are alternative versions of your breadcrumbs idea.
- the player is following someone who left marks
- when the time has come, a strange creature appears that taunts the player so the player runs after it; even if he couldn't catch the creature, he'll still be closer to the exit
- when the game wants to move on, the rooms furthest away from the exit start collapsing, or filling with gas, or flooding, forcing the player to move towards the exit

Another, rather silly, idea I can remember: all levels are connected to each other by different means of transportation. A bus that only arrives when the game wants you to leave (the sound of the engine leads the player in the right direction). An elevator but the bell boy is asleep, dead or drunk. An airplane but you always arrive at the very last minute and when the game doesn't want you to leave yet, customs officers harass you so you miss your plane. A conveyor belt that only starts running when the game is ready to move on.

I'm a bit uncomfortable with the idea of requiring the player to have done certain things before they are allowed to progress. Is this what you are thinking? Is it really necessary? Can't you allow them to progress at their own risk of missing things?

Maybe future levels change when you have done a certain thing in former levels. And this change gives access to other parts of the story.

Would it be possible to craft a bare-bones version of the linear story that players can play through in five minutes but that isn't very satisfying (or scary in the case of a horror game -is that what you're making?)? And everything they do -voluntarily!- simply makes the experience richer (and more scary ;) ).

Maybe different endings become available after doing certain interactions.
That way, you don't need to give away any surprising plot twists in the bare-bones version of the story.

I agree with Michaël on the choke-points. The question is what the aim of the game is in the sense of what it tries to achieve with the player. But within the present cultural context that is difficult; just like smart AI can be there without being noticed (Dan always quotes Halo on this), so can rewarding detail. I know people who ran through the Graveyard because they could not register the woman was limping for not expecting it. Placing the exit next to the entrance and giving the player an unrewarding experience is from my perspective completely fair... but it will not register with some players if they expect the arbitrary puzzles.

In my oncoming game I have a fixed length and two 'choke points' - which I make possible by not giving the player the freedom to miss them; there is no freedom of movement. If the player causes the character to linger the character will chose faster actions to arrive at the same fixed point at the same fixed time. In a way I believe this eliminates both the 'have I spent enough time here' sentiment as well as the 'quickly as possible' sentiment. The player 'pushes' the boundary of the story in a way and it elastically snaps back.

Not knowing much about the content of the game it is hard to say, but something like Michaël's creature could have such an elastic function; if the player goes about too slowly it will pull him further.

Although aware of the danger I like the-entrance-is-near-the-exit idea. I remember the book The House of Leaves had a house in which the interior dimensions did not match the dimensions on the outside - and there appeared a door on fixed times which lead to an impossible labyrinth of corridors, all pitch-black. The main characters are never coerced to go in there but slowly got an obsession to find out what was there to the point where they took increasing risks to get a little further before returning to recharge their batteries and film. That is in some forms an ideal scenario for the player: because if he can exit at any time all his attention is voluntary. If not, you are coercing him in some manner from the start and that can lead to taught helplessness.

Another point is that if you go for actions that are rewarding in themselves a lot of things become less useless. The first concern is to communicate this in such a way to the player he will understand it. That is like watching the likes of Koyaanisqatsi - to really appreciate it you should have seen it before and there is always a first time. The second concern is that the interactions must be very interesting. For me that involves a considerable richness of culture and detail.

Good suggestions, gives me something to think about!

My main reason for having these  milestones is that they basically serve as events that unlock certain part of the game + make up the most crucial parts of the story (this might be meeting certain character, finding certain item, investigating a certain location, etc). The minimum any game can have in terms of this are spacial based milestones where simply getting a certain distance from the start (or similar), triggers the milestone. Of course, milestones could be removed altogether, but that would mean there could not really be an end state (that being a milestone). And I am honestly not confident enough to try and design that :)

So I still want to rely on some sort of milestones, and I agree when you say:
This is exactly what I am trying to avoid, by making the milestones in such a way that it is always up to the player if they want to progress. One way to visualize it is that I want big glowing "NEXT CHAPTER" button, that the player can press anytime they please and then surround that with interesting points of interaction (something you covered). To make this engaging, the collection of interactions would change when the player chooses to press "next chapter" and thus needs to consider what to do with more care. Also future states can depend on past interactions.

Now I am willing to do it this obviously, but want to have it more woven into the story. So i need to have some sort of device / design thinking that provides this structure within the game context. That is really the core of what I am getting at.

We are considering having a fairly short "base game", where most of the engaging and interesting things happen when the player strays. But there still needs to be milestones placed. And these milestones should:
1) Feel seamless to the player and be part of the game world. In a perfect situation the player does not even notice that there are milestones, but simply progress in a way they like.
2) The way of a achieving the milestones should have a very low probability of halting the player. Meaning that the player should feel that they are not sure what they should do next.

The thing that are closest to what I want is to have something like main road the player can always travel forward on. This main road is then filled with side roads that the player can choose to detour too. I am actually not sure that this might actually be the way to do it :) I just feel that it would take away the sense of exploration, by always forcing the player through a pre-designed route. Instead I want the player to feel as much as possible that they personally discover the things.

So just to state it outright: The biggest challenge is coming up with the condition for reaching a milestone, and not really the design of the milestone itself. Eg, the problem is to figure out what makes the conveyor belt to start and to figure out conveyor belt in the first place.

Perhaps I should not give up the thoughts of leaving everything totally open just yet. I just have a hard to time figuring out how to plan the basic plot points or having the player accidentally stumbling across the ending.

This is interesting, because a way to bypass is to make the stuck state more interesting, perhaps even rewarding. In most games getting stuck comes with large negative feedback (as all fun stuff suddenly stuck). But what if getting stuck would provide extra experience in some way? Of course, one can not provide endless extra content for stuck players, but it is a interesting way to think about it.

This sort of design thinking can probably eliminate a lot of the chance to get stuck. On does not have to say in advance that only 5 was required either and make the player believe that they did as much as they could.

Jeroen D. Stout:

I like this rubber band kind of approach. The lego games actually use something similar that they call "springy path" (I believe), where the player always have the simple route available but if deciding to go out and explore they game gets harder, and harder pushing them back. Applying something similar to overall narrative structure seem interesting, and having the game draw the player back in if they are not being "productive" (what ever that might be at a certain point in the game).

The hard part is only setting up these elastic functions for various parts of the game.

I just recalled that we did something like this In Penumbra black plague. At a certain puzzle the voice in your head (which was an actual separate character) gave you simpler and simpler clues the more time passed, at the end simply shouting out the solution. This was a very simply example though, and choosing the right amount of time proved quite hard. Still there where not many complaints to this method and I cannot recall any players complaining.

In The Book of 8 -one of the prototypes we're currently working on- we unlock the play space chunk by chunk as well. This is a leftover from the original version of 8 -our first, unfinished project. Neither is a linear story. But we decided to use an unlocking structure because we were afraid that players wouldn't be able to handle a wide open space from the beginning of the game. We were afraid they would get lost and not know what to do, and thus not be motivated to do anything. So we decided to guide their experience by only making the space available bit by bit. At the end of the game the entire play space is unlocked and available for exploration.

Anyway, how we unlock chunks of the space it is through (implicit) secret corridors. When you find a special object, you return it to a certain place and that "opens a door" in that place to a secret passage through which you can find another part of the game world space that you couldn't get to because it was locked, blocked or otherwise separate from the part you did have access too. Players do not need to use that passage right away. There's opportunity to open more than one before entering. In that sense the game remains non-linear and only gets bigger.

In The Book of 8 we have stylized this a lot, though. Partially because of budget considerations, it is no longer a complete open world. Instead we're using the book metaphor where every page represents a room. Rooms that are not accessible yet are represented as pages that stick together (it's an old book and the entire game consists of cleaning it up). So we get the added advantage of the linearity of the book to aid navigation.

This is quite literally what happens in The Path. Hence its name!  :D

There's a straight path from the start of the main part of the game to the end. It takes five minutes to run it. Alongside of the path there's a spooky forest with here and there a light in the distance.


Don't worry. Very few players could resist the temptation to enter the forest in The Path. There was nothing to do on the path itself. If you wanted to have any fun, you had to take the risk and enter the spooky forest. It takes less than a second for a gamer to take this decision.

The only way we caused confusion (quite on purpose), was by telling players that they should go straight and not go off the path. That was the only rule of the game. And some players actually followed it. :D But if you don't play with them like that, you should be fine.

Just for comparison. While our game The Book of 8 is completely non-linear in principle, it does have a specific ending. The condition to achieve this ending is that you have have cleaned up the game world sufficiently. And I guess the different choke points simply prevent players from cleaning it up right away (because at every choke point they get a new tool to do other types of cleaning -a tool they can use on rooms they have been in before).

I never really thought of The Book of 8 in this sense. Because as a designer I don't really think in terms of progression and I'm far more concerned with the design of the moment than with the flow of the experience. But I guess we ended up with a similar structure anyway.

And you can cheat!
If the player has spent enough time in the chapter you can just open up the door without any reason. And if there is one narrative point that they really need to get before continuing, just make it happen automatically, if it seems like the player can't find it.

So I have been thinking about this so more and changed my approach to this problem.

I started out with the hope of coming up with some kind of magic bullet. I wanted an idea that I could apply throughout the game that always lead the player on the correct path. But as I have been reading your suggestions I have been coming to the conclusions that this is probably a bad idea. When we started designing Amnesia we (at least I) were obsessed with the iea of finding ONE game mechanic that could drive the entire game. After several months we gave up on this idea because it constantly clashed with all the other things we wanted to do (atmosphere, building an immersive world, etc). I think I am barking up the wrong tree again, with having a ONE mechanic that solves the player guidance.

Instead I will embrace the fact that creating this kind of experiences are hard, and try and shape the guidance mechanics depending on the situation. This way I can instead focus on what it is I want to convey and then design the virtual in a way to achieve this, by using a mix of the techniques discussed above and more.

I do think that Jeroen's idea of having a stringy narrative can act as a good guideline, but that I should not try and find a specific mechanic to accomplish for the entire game. So my goal will be that the game always steers the plaeyr in the right direction, using different methods each time, and perhaps different methods in the same part, so it caters for different play-styles. This is actually a bit what we did to the better designed puzzles, we simply implemented whatever the solutions the testers tried out, not worrying about having out own "ideal" solution intact. I think this should work with guidance towards a milestone to, adapting the progress to how the player plays. Of course, one cannot cater to all play-styles, but just adding a few options could make the feel a lot more intuitive and fluid.

With this in mind, do you recall any moments in games (Amnesia would be nice, but suggestions from any games are welcome), where you where stuck in progress and suggestion on what would have helped you move on? Would be of great help!

So I finally bought The Path this week and intend to play it next week during my "vacation"! :) I know I am bit slow with this, but better late then never eh? ;)

Also hope to buy and play Dinner Date too! But time tends to run faster than you expect :)

Interesting - I have been asked in an interview recently what the difference between a literary character and a game character was. I still have not sent the interview back because I cannot, for the life of me, think of one attribute which is not based solemnly around 'games are interactive'. It lead me to think about characters differently - in unemotional terms, I see them as a 'cloud' of attributes through which you street the reader/viewer/player. A good character forms a sort of harmony and that can be done in any medium in which the medium itself has the power of steering.

Now you mention this, I realize my own question I use to force people to think of their game concept quicker was atomic - your one mechanic. You are right, it is multiple things, changing, but elegantly. I have no idea how valid this is, but perhaps start thinking of it as a concert, with an anthem, different leitmotifs, a crescendo. Not just in terms of content but of actions itself. You could have quiet interludes without control (spend 2 minutes in this room before it unlocks), free-roaming led-by-hand experiences, tight scenes. I imagine the challenge (one I also now am struggling with) is that you have to communicate the gameplay language to the player before he can use it, so the teaching of new mode is a difficult element if you start building symphonies.

I was spurred to use this metaphore because you used hard as-if a per-case thing... and while symphonies are hard they are an intellectual puzzle and perhaps an approach not entirely dissimilar to games in abstract.


I'll have to dig for this - but I will report to you when I have played Amnesia early Januari when my new PC arrives.

The last thing I remember was the ladder in the cistern in Amnesia. It didn't occur to me that I could push the lever in two directions. Instead I was throwing some more rocks at the damn thing.

The solution, however, for my taste, is not to have such obstacles at all. In a game, they quickly feel very artificial. Even if it would be likely to experience obstacles like this in real life, when you experience them in a game, the feel artificial. Just because you know you're paying a game. And also because there usually is a solution (unlike in real life).

[rant mode]
Perhaps a bit off-topic, but for the last couple of days I have been thinking more and more of this hardness property and feeling that it might work as a good guideline for making better games. Basically, the harder the game is to make, the better it is. What I mean by hard here, is basically the same thing as computer science use it when saying that certain groups of problems are hard. Proving a mathematical theorem, composing a symphony and writing a poem are all hard problems - there is no simple short cut. What this means for game design, is that if you can easily construct extra content for your game or even even "easily" procedurally generate it, then your design is simple and of lesser "quality". There is bound to be lots of special cases here and so on, so it is nothing I take that literally.

For both Amnesia and the Penumbra games, I have always been a bit annoyed at how hard it was to make extra content. I have seen other devs being able to create add-on packages,etc and been jealous of this. However, now I think I should actually embrace this and instead be worried if any part of the game is too easy to design and replicate. It is when game contains something that I know will be impossible to reuse or extend that I should feel most proud of it and this should be something to strive for.

I find that the same is true for great music, literature, movies, etc. Sequels and follow-ups are very rare.

Now this does not mean that games with simple mechanics that allow easy creation of content are bad by any means. I just think that it might be a  good thing to have in mind when designing. This could be especially true for "not-games", where the whole point is to go beyond the normal games that now flood the market. In these games, it is also very easy to see how the content is simple (as apposed to hard) design, with repetitive gameplay and a bunch of DLC and the like coming out close after release. It is games that do not get sequels or where game-play on repetition where we see the some interesting things happening.
[/rant mode]

I highly recommend building your game world first and then seeing what opportunities arise. I find it very difficult myself to imagine such things on paper. When you build a world, the opportunities for interaction quickly present themselves. And if you still lack inspiration, just observe a test player for an hour. :)

I'm kind of obsessing lately over the concept of simulation as a further extension of figurative painting and cinema. I'm working on a blog post about the two different kinds of rules in games: the rules that serve the simulation and the rules that are part of the game. And I want to make a case in favor of the former and against the latter. Because the former serve the simulation and the make-belief experience, while the latter destroy it.

This is something I am drawn to too. My own is goal is to try and approach a virtual world that is as coherent as possible, making the game more about "living" than "playing". Will be very interesting to hear details on your thoughts and approach to this!

I think there could be a solution here for your current problem as well. Maybe the gameworld simulation can change, guiding the behaviour of the player. Instead of adding a puzzle or evaluating performance.

Maybe the player can have control over parts of the simulation. Perhaps the player can make the avatar say certain things or stand in certain poses that the game can interpret as "he thinks he's done here, he wants to move on", upon which the game can decide to either move on or lead the player to something they have missed.

I don't think I've ever seen that in a game: the ability of the avatar to express the feelings of the player about the game. Maybe it's a bit like asking for a clue in casual puzzle games. But then embedded in the simulation: the things the character does when they get bored or confused could be different depending on the personality of character in the fiction, and even of the situation in the story.

I strongly believe in giving the player this kind of agency in a virtual world. I think one of the best features of Grand Theft Auto is honking! :)

I think that is a very good viewpoint. But furthermore, if you choose to view it as a hard problem rather than just hard work it may become easier when you find ways of doing it. An ideal game is like a the ideal symphony - not a single not can be removed or added. But I think we are far from this stage of development. We are still making folk music, so to speak.

This is a good rant, though. It very much confirms my feelings and inspires me to see the next game for what it is, a big challenge.


Very interesting, I am very much looking forward to that. It does sound like there is a mild danger you may put everything which 'feels natural' in 'simulation', however... the rules of football are evidently part of the game but also are 'non-negotiable' and become transparent to the point where one just behaves like that. The mind optimizes the question of realness away, or in social thought, rules becoming transparent is the essence of a game's half-realness.

But in Callois' terms, if you are speaking about agôn games, transparent rules that govern your success in achieving goals mean a very 'thin' interactive space (to speak with Harpold), which is benificiary. If your goal is memesis then that rules are consequent in simulative terms rather than agôn-favouring terms is more important. But that induces a thicker space for goal-achievement.

Perhaps that is part of the problem as well (reflecting back to positive and negative freedom); the moment a person wishes to achieve a goal in a negative freedom space he starts playing the rules rather than the game. It is the problem in law of drivers trying not to get caught rather than obey the law because to them the former is more beneficial (all the advantages of breaking the law).

It does mesh quite well with one of the points of notgames, then, of course, that of having no implicit goals. But perhaps it is important to note whether you as a designer are responsible for making me not play the rules or whether I as a player should, like when reading a novel, know how to behave.


Interesting. You could structure that by making a game-in-a-game. You and an AI (with whom you have a tie so strong that you do not perceive him as a game) 'play' a game together which you do allow the player to think of as a game.

Thank you for the warning. I'll certainly keep this in mind.

My concern, however, is art, not reality. And in art, playing with the rules, is one of my greatest pleasures. When we see the humanity of a beggar, the eroticism of the Madonna, the evil of Aphrodite. When we brush against the limitations imposed by rules and habit and perhaps every so slightly suggest a gentle transgression.

I do think an author should be prepared to take responsibility for anything a player might do with their work.

Maybe it's a question of implicit rules versus explicit rules. In a way, the former cannot be broken, or even doubted, by the player. Because they enable the simulation: gravity, three-dimensional sound, navigation, etc. The explicit rules are the ones that require a sort of negotiation. The player needs to accept that it is required to kill all monsters before the door of the dungeon opens.

You're right that once you accept these explicit rules, they may become like second nature. But I think this is a dangerous terrain for the author. And we should really ask ourselves if we want such behaviour to feel natural. Is that the message we want to send?

Oo, I'm excited to see you write about it this way.  A while back I wrote a blog post about the same thing (http://www.godatplay.com/2010/05/on-what-makes-videogames-distinct/).  I make a distinction between videogames as an art form - your "simulation" - and what I call "abstract structures for meaning" such as story and game rules.

Some snippets:

---

Thomas, I think a possible way to think about milestones would be to frame them situationally instead of spatially.  Your milestone then would be to cause the world to reach a certain overall state that allows further progress, instead of finding a key to unlock an area.  Your "key" then would be a situation.

One example of a situational milestone would be a character getting angry and throwing a tantrum.  That could be caused by any number of things, or even triggered automatically (if it was within character to spontaneously throw a tantrum).  If you had spatial limitations, the character would storm off and unlock whatever areas were needed or take some sort of route throughout a space.

I suppose this almost is a question of medium transparency and whether you want the reader to read a book or read a story.

For instance, I am learning to make a point in discussions to emphasise I do not make 'protest' games and that Dinner Date does not 'make fun of' powerful protagonists, because I want to prevent people start playing it as an artefact and rather play it as a game. I want the rule that you do not change Julian to be natural because the rules are what creates the experience. I suppose the same goes in Amnesia - that the screen blurs should be 'part of it', not something which feels unnatural. Rather Daniel's freight should feel unnatural and something to be avoided.

So perhaps some explicit rules (and non-standard implicit rules) should be pre-negotiated to make sure the player can engage the system in a meaningful way? That someone knocks me down in Zeno Clash feels like something undesirable because I do not 'accept' that characters can knock me down (or rather, they aught not to). But that experience is only possible because I force myself to live 'in' the rules which make the game possible.

This always leads to some meta-data which 'grows' on players. So perhaps, yes, I want part of the experience to be natural (and thereby transparent) because it is the gateway to better experiences.

I am beginning to believe that these rules becoming transparent is a growing process which force of mass has more influence on than cleverness.

@Thomas and God at play

In light of this, perhaps keys and valves are just rules we have accepted (even though they often make no real sense) and most situations are not?

I agree that part of the issue is to convince the player that certain things are just required. When playing mario, it is never questioned that you need to reach a flagpole or collect stars, the player just accept this as a condition for progress. I think these sort of things only work to a certain extent though, and especially when you slow down the pace these become harder to maintain. In mario you are so focused on the action, that any thoughts on the actual world are no of interest, so player never questions why there is a flagpole / star in the first place.

However, as you slow down the pacing of the game, and encourage the player to think about the game's world, all of this things become annoyances. Why do I have to collect stars? who put them there? and so on... This is a problem we faced instantly when in Penumbra: Requiem, we tried to loosen on the world/story coherency and just add fitting stuff. People started wonder why things were in that sort of state, and then in the were just pulled out of the experience.

I find this a huge difference between gamey games and more notgamey games. You must be so much more careful when world building. This is a very very good thing though. The will to encourage your audience to really think about and carefully experience your work, is the same that differs good and bad food (any art really!).

There is of course stuff that you do get away with, for example nobody complained about loading screens in Amnesia and similar, even though they are hardly supported by the fiction. The room for these kind of things is way smaller though.

The danger of conventions is that they are only useful for people who are familiar with them. Only to them can they become transparent enough to experience the rest of the fiction,

Gamers might think of Ico as an artistic masterpiece. But I once witnessed an non-gamer play it and she couldn't get into the atmosphere and the story at all. She was laughing all the time because the game conventions simply made the world seem absurd. It was embarrassing.

I want to take the term "design" in "game design" seriously in the way that furniture designers do. A chair is for sitting. Your reference as a designer is the body of the user. Not other chairs.