Notgames Forum
November 10, 2024, 07:06:09 PM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Search Calendar Login Register  
Pages: [1] 2
  Print  
Author Topic: The self-playing game  (Read 34883 times)
Jeroen D. Stout

Posts: 245



View Profile WWW
« on: April 04, 2011, 11:51:37 AM »

Since this topic popped up in a lot of conversations I thought it might be worth a topic of its own.

I think in 2009 I made a prototype game (Arrival by Train) with two AI's having a (mimed) conversation. They were rigged to have a dreadful time because their interests clashed. As a player you could overrule the choices of one and ensure that they had a better conversation. I remember back then that a game that plays itself was a bit frowned upon by fellow game students.

Now with the new game (alias Berlin) coming on, I have been thinking on doing a self-playing AI again. I suppose it is quite a wide topic but as relevant and 'nouveau' as having a game without challenges.

One of my ideas for Berlin primarily was to have two AI's, like in Arrival by Train, and let the player influence the inner workings of one of the AI's (through Dinner Date-like bubbles) and so be influencing the game and be a part of it. I had the catch that if the player would do nothing the game would play itself and for quite some time I was thinking of ways to punish the player; have it be lacklustre if he did nothing, or offer him very similar content each time. But the game is all about the inevitability of good romance. In essence it is about the opposite of what the play in Heavy Rain suggests: in Berlin there are situations which are going so well there is nothing you could accidentally do wrong, you can only accidentally do things right.

So punishing the player for letting to of the controls through making the situation less energetic is like me saying: "you have to focus, player, don't you dare defocus and relax. This is a game! I will let you fail, you hear me! (etc)!" If anything it is me as a designer arrogantly saying you 'have' to play my game or I will, like some despotic teacher, 'fail' you.

Another problem I realized is that if you start offering the player 'special content' in an artificial manner just for bothering to do more than just watching events unfold it is perhaps self-evident that there is no actual reason for playing the game. And ignoring the player doing nothing turns it into the unrealistic idle-animation excuse. And some players quite simply do not feel like doing something all the time - I have heard from playings playing Dinner Date without doing much so they could really listen to the story. And I want to have that far more as a viable option.

So my new thought is to see interaction/non-interaction in Berlin not as a requirement, but rather as a continuum the player can voluntarily place himself on. The game always lasts about three minutes, but events play out differently based on what the player focusses on: and, now I realize, also based on whether he chooses not to focus at all. But the choice not to focus is not value-judged, it is rather a different way of engaging the game.

I will see whether I can reward the player both for doing something as well as doing nothing. When the player acts, his character becomes more dominant in the situation, when he does nothing, the other character takes that role and many of the same things will still happen; but from a different point of initiative. Either way will be pleasant and rather differently flavoured (so to speak). In other words, a player who plays the game four times acting a lot will be rewarded for doing nothing as the game will offer him things he could never experience if he acted a fifth time. And a player who is just genuinely too shy to act on his first go will be rewarded for doing small actions and seeing things 'go his way' in later play-troughs. And after finishing the game 'months ago' you will be able to start the game and just leisurely stare at the screen and do small things as you please.

One part in which this choice is most material is a situation which will always occur in the game because it progresses the story as a whole; and here if the player is dominant his character will just suggest and do it, if he is staying away from control the other character will propose the same thing and they do it likewise.

I thought this is interesting, ever since Arrival by Train I had a little forgotten about voluntary player involvement and now it popped up again. It also solves, in my view, many problems inherent to a game that fails to admit the player might not do anything.

From what I know some of us are all taking somewhat different angles on this matter; I know Thomas more from puzzle solving and Auriea and Michaël already have done a lot in this direction with Drama Princess - so I hope these observations and thoughts may be helpful.
Logged
Michaël Samyn

Posts: 2042



View Profile WWW
« Reply #1 on: April 04, 2011, 12:27:00 PM »

in Berlin there are situations which are going so well there is nothing you could accidentally do wrong, you can only accidentally do things right.

That's a great concept!

The exact opposite of the requirement of failure in most games: the inevitability of success! Smiley

More importantly, I think: a chance for being charmed by the responses of an AI.
Logged
Michaël Samyn

Posts: 2042



View Profile WWW
« Reply #2 on: April 04, 2011, 12:30:50 PM »

But the choice not to focus is not value-judged, it is rather a different way of engaging the game.

Excellent!
A perfect attitude towards a design that favours being over seeing! Smiley
Logged
Michaël Samyn

Posts: 2042



View Profile WWW
« Reply #3 on: April 04, 2011, 12:35:18 PM »

Considering the AI controlled by an active player simply as more dominant, also makes the game more interactive in the Crawford-sense of the word: it emancipates the AI that is fully controlled by the computer by allowing it to become dominant when the player is less active. Brilliant! Smiley
« Last Edit: April 04, 2011, 12:37:08 PM by Michaël Samyn » Logged
Michaël Samyn

Posts: 2042



View Profile WWW
« Reply #4 on: April 04, 2011, 12:39:00 PM »

By the way, I really admire your putting so much effort and thought in a game of only three minutes long. It's very courageous.
Logged
David

Posts: 52



View Profile WWW
« Reply #5 on: April 04, 2011, 03:38:54 PM »

Very interesting, thank you.

I was just wondering if the player's acts in my current project would make the end of the story change, and i didn't like the idea to fall again in "this end is bad, you failed" or "this end is better" or "this end is the most beautiful end, here is a piece of sugar little doggy"... Now i will think about your idea: always the same overall end, whatever who are the characters who act.

I like the comparison that you make between most of games and normal school : in both we have to do what a "teacher" wants us to do, and as we don't do it then nothing happens, we just wait until we decide to follow the instruction. Your idea to both 1/ allow the player to not do anything and 2/ not to pause the game nor make the player lose, is very interesting. I will think about it for my current project.
« Last Edit: April 04, 2011, 05:58:42 PM by David » Logged
Jeroen D. Stout

Posts: 245



View Profile WWW
« Reply #6 on: April 04, 2011, 05:36:15 PM »

Thank you both kindly!

Because of the way it may seem (and the tweet) I probably ought to mention that these ideas came from a few directions and conversations and I presented it in the 'I' form to have it reflect on Berlin, rather than as a product only of my own thinking. It rather is a path I am interested in.

(This said for fairness' sake.)

Good to hear the thought goes down well - I am looking forward to implementing it and seeing how it goes!
Logged
Thomas

Posts: 384



View Profile WWW
« Reply #7 on: April 04, 2011, 06:04:46 PM »

Have previously discussed this with Jeroen too and also been thinking about it. I have been thinking about it from a different angle though.

One of the problems with designing environments is that the more open and free you make them, the harder it is for the player to progress. But if you make a small easily navigational, you loose out on a lot of the details you can add. I think this is a major problem for enjoying adventure games, as you can get so easily stuck that you loose immersions and the will to role-play.

So the solution (that was a bit inspired by The Path), is simply to have the protagonist continue to where he/she wants to go. If implemented in a similar system to the one The Path Uses (where you give up your control for the avatar to interact) this could come very fluently into the gameplay. It could almost be an action meaning "go where I am supposed to". It would be interesting to add this to a normal adventure game, or just to The Path Smiley

This is not something I intend to add in our next project, but I would very much to see it in action. Very interested how it works out for you Jeroen!

And oh, regarding rewards depending on what you do, the thinking I am leaning towards is to not see it as rewards but "confirmation of play style". What this means is that the consequences of actions / choices try to make what the player ought to like as much as possible. We tried this in Amnesia's multiple endings a bit. For example, players that just rushed through the ending and did not listen to the dialog carefully would get an ending where they simply come out as the hero. More careful players would find themselves in a more thought-requiring ending. Hopefully people should not see endings as good or bad, but "just right". It was far from good really, and response was not like, but at least it serves as an example to work from and illustrates my idea Smiley
« Last Edit: April 04, 2011, 06:06:27 PM by Thomas » Logged
Chris Platzer

Posts: 10


View Profile
« Reply #8 on: April 05, 2011, 12:41:21 AM »

So my new thought is to see interaction/non-interaction in Berlin not as a requirement, but rather as a continuum the player can voluntarily place himself on. The game always lasts about three minutes, but events play out differently based on what the player focusses on: and, now I realize, also based on whether he chooses not to focus at all. But the choice not to focus is not value-judged, it is rather a different way of engaging the game.
I love the idea that the game goes on, with or without the player.

I imagine this could provide a much deeper feeling of involvement and interaction. Because the game doesn't wait for you to trigger the next scripted event. You have to pay attention to things happening around you and to timing.

I also like that you don't want to impose on the player whats right and wrong / how to play the game.

And oh, regarding rewards depending on what you do, the thinking I am leaning towards is to not see it as rewards but "confirmation of play style". What this means is that the consequences of actions / choices try to make what the player ought to like as much as possible. We tried this in Amnesia's multiple endings a bit. For example, players that just rushed through the ending and did not listen to the dialog carefully would get an ending where they simply come out as the hero. More careful players would find themselves in a more thought-requiring ending. Hopefully people should not see endings as good or bad, but "just right". It was far from good really, and response was not like, but at least it serves as an example to work from and illustrates my idea Smiley
That sounds really great.

Out of curiosity (since I haven't played Amnesia yet): do you just measure how long people stick around when somebody talks (and then just present a different final level based on that measurement)? or do you hide breadcrumbs in the dialog that allows people to find the hidden path to an alternative, more intellectual, ending?
Logged
Michaël Samyn

Posts: 2042



View Profile WWW
« Reply #9 on: April 05, 2011, 08:54:42 AM »

I think the difference between Jeroen's desire to allow the game to continue playing without input and Thomas's desire to confirm a player's style, has to do with a different approach to dealing with content. Thomas seems to be working very much from the idea of telling a story while Jeroen seems to be more interested in description of a situation.

I'm very interested in the latter, in the idea of simply designing worlds for the player to "be" in. But the tricky part of such an approach is how to keep the player interested.

Even I, with such a strong desire to lose myself in virtual worlds, wasn't motivated to play Grand Theft Auto 3 anymore after finishing the final mission, or playing Ico again after solving all the puzzles. And even now, while thoroughly enjoying Assassin's Creed: Brotherhood's immersive qualities, I am grateful for the treasures I can find and and areas I can purchase.

Jeroen's solution to this problem is clever and daring: just make the game 3 minutes long.
Thomas solves the problem by adding narrative evolution/discovery. Which is probably smarter vis-à-vis gamers' expectations.
But I can't help but feel there must be another way. Another way to just give a simulated world to the player. A world they want to be in for its own sake and a world that they can stay in forever, if they choose to.

As you can see, through video games I want to solve my big frustration with novels that I like: the fact that they end. Maybe that's just a phantom grail, though. Maybe I should accept the transience of art, as I accept the transience of life.
Logged
Thomas

Posts: 384



View Profile WWW
« Reply #10 on: April 05, 2011, 09:53:30 AM »

Chris:
Amnesia has extremely simplistic design in the ending. It is actually just a matter if the player solves a puzzle or not in a certain time-frame, while a character is talking. I am not very pleased with it, but at least it shows how not to do and gives an idea of the basic idea Smiley

Michaël :
Plans of creating an experience machine, eh? Wink

On a lil more serious note, I think it is a very interesting path to pursue. Just trying to figure out the how we can make engaging videogames without any superfluous awards (counting treasures, story, etc), can lead to very interesting insights. I just think you need to add some kind of extra reward structure on top to make it work though, some kind of designer implemented discovery to be made, and these will eventually run out.

It is interesting to juxtapose this with part of the experimental film movements (at least how I understand it), where the goal is to NOT entertain, and instead constantly remind the viewer that they are watching a movie and to make them understand that process.
Logged
Jeroen D. Stout

Posts: 245



View Profile WWW
« Reply #11 on: April 05, 2011, 12:58:15 PM »

Re: Hinting in adventure games

With all the countless old adventures there must be some (now open-source) adventure into which you could put this simply by giving the AI a walkthrough and a few timers. It is an interesting thought and I wish this had come up when I still made adventure games. It would be a great way to circumvent the 'look up a hint' addiction I always got with difficult adventure games.

I am actually quite intrigued whether any adventure game has done anything like this at all. I know some put 'super knowledge' in their dialogue with 'no need to go back there', but it would be interesting to see this in body language.

(Though I remember a puzzle in one of the Monty Python games which took place in a zen garden and you could only solve it by doing nothing for three minutes.)

I think the difference between Jeroen's desire to allow the game to continue playing without input and Thomas's desire to confirm a player's style, has to do with a different approach to dealing with content. Thomas seems to be working very much from the idea of telling a story while Jeroen seems to be more interested in description of a situation.

I'm very interested in the latter, in the idea of simply designing worlds for the player to "be" in. But the tricky part of such an approach is how to keep the player interested.

Even I, with such a strong desire to lose myself in virtual worlds, wasn't motivated to play Grand Theft Auto 3 anymore after finishing the final mission, or playing Ico again after solving all the puzzles. And even now, while thoroughly enjoying Assassin's Creed: Brotherhood's immersive qualities, I am grateful for the treasures I can find and and areas I can purchase.

Jeroen's solution to this problem is clever and daring: just make the game 3 minutes long.

I think the original idea of the game only became doable as soon as I chose the short form - I had some contending ideas of longer lengths but they would have been too much like evolving Dinner Date as they were more linear concepts. The short length makes it possible to suggest their meeting, conversation and evolution of feelings started recently, but well before the start of the game, and means I do not have to return to a worldly context; it can stay their thing.

A requirement is that you can play it a significant number of times, though.

Speaking about 'carrier waves', I suppose that the sensations of a lot of places you want to be in is carried by a narrative, belief or activity. But at the same time I have always felt that if there was a 'sit down' button Beyond Good & Evil would have been a lot better; if there was some way to do nothing yet still gain impressions of Jade... right now if you stop doing anything nothing happens. Perhaps if 'something' happened beyond an idle cycle the world have been more accessible.

That is another 3-minute argument; I hope with Berlin I have figured out 'what ought to happen' for 3 minutes, beyond that mark I have difficultly tying it together. Perhaps this is like starting with compositions with new instruments - once you get to grips with it longer stretches become possible; or I hope so, because the 3-minute format makes narrative hard to do. But right now all I could imagine for a 3-hour piece is dividing by chapters, which just means I am taking the same short-length risk but somehow try to make up for it by doing it a lot and throwing away player involvement every scene switch...

I think part of the challenge for Berlin is selling the length as a concept that is novel rather than strange (as you hint by your juxtaposition), which I figure is more presenting style and PR than actual play when it comes to it. I keep thinking of it as 'variations upon a theme', in the sense that the theme and content is relatively similar, but that does not make 'variation XI' inherently less interesting.

I love the idea that the game goes on, with or without the player.

I imagine this could provide a much deeper feeling of involvement and interaction. Because the game doesn't wait for you to trigger the next scripted event. You have to pay attention to things happening around you and to timing.

I also like that you don't want to impose on the player whats right and wrong / how to play the game.

I think I still tell the player to 'act romantically' (or rather, say that this is the way I believe you ought to interpret the game), but it is rather that I do not relate his actions to romantic success... so you can sit there and watch the events and then get involved if and when you feel like doing so. But doing so should have no risks. Of course, if you want specific things to happen you can play and try to make them happen; and it may be that I never anticipated them and they do not work. But that should not cause failure, as such.

Considering the AI controlled by an active player simply as more dominant, also makes the game more interactive in the Crawford-sense of the word: it emancipates the AI that is fully controlled by the computer by allowing it to become dominant when the player is less active. Brilliant! Smiley

Yes - the symbiotic-AI is more tied to the player than 'filling in' the space between player and computer-AI. I have been thinking on how to do this without computer-AI (or with many computer-AI's, like in a group conversation) but it is quite difficult. I suppose that filling the power balance with a computer-AI who also wants positive things to happen is the easiest; without this AI the symbiotic-AI would communicate with the player too much (in a sense of the word); with multiple computer-AI's they could actually exclude the player if they all became more dominant. (Though the later could also be pleasant as in Hugo's 'drunken talkers' scenes, I suppose.)

This is an interesting route, though. It is (if I may venture to be so bold) an interesting idea to give a group of students the task of making a game that can finish itself yet feels engaging to play yourself. Getting some short rough experiments would be interesting in a lot of different game types.
Logged
Michaël Samyn

Posts: 2042



View Profile WWW
« Reply #12 on: April 05, 2011, 11:21:22 PM »

(Though I remember a puzzle in one of the Monty Python games which took place in a zen garden and you could only solve it by doing nothing for three minutes.)

Smiley
That's how you acquire the shapehift spell in The Endless Forest: allow your deer avatar to fall asleep in a mushroom circle (takes five minutes of not touching the controls).
Logged
Michaël Samyn

Posts: 2042



View Profile WWW
« Reply #13 on: April 05, 2011, 11:30:03 PM »

The most inspiring game AI that I have interacted with was the "creature" in the first Black & White. If you haven't played the game, I highly recommend it! And while I'm recommending, in terms of "ambient behaviour", Yorda from Ico is superb. Both greatly influenced our ideas during the Drama Princess project. The creature because he is so lovable and Yorda because her randomness adds to her personality.
Logged
Jeroen D. Stout

Posts: 245



View Profile WWW
« Reply #14 on: April 07, 2011, 10:41:20 PM »

The most inspiring game AI that I have interacted with was the "creature" in the first Black & White. If you haven't played the game, I highly recommend it! And while I'm recommending, in terms of "ambient behaviour", Yorda from Ico is superb. Both greatly influenced our ideas during the Drama Princess project. The creature because he is so lovable and Yorda because her randomness adds to her personality.
Black & White is a game I did love dearly - the animal was a singular good part of it. I must have watched him just do things for such a long time and thought about how I could teach him. It was a very interesting relationship. It would be interesting to see a Black & White where it is more the implied goal to raise an animal than to battle other gods. B&W certainly is a game which did not always know where his strengths were and it is a shame that Milo never went through, it had the same interesting AI going on.

That the creature was a big silly animal probably saved him from a lot of accusations of unnatural stupidity which AI has to survive Smiley
Logged
Pages: [1] 2
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.20 | SMF © 2006-2008, Simple Machines Valid XHTML 1.0! Valid CSS!