If gameplay were plot, would there BE story?

Started by
78 comments, last by GameDev.net 18 years, 9 months ago
Quote:Original post by Wavinator
What makes you a hero, and what makes the approach hero-centric versus character simulation?


I think I know what you're getting at, so I'll see if I can explain what I was talking about. Unfortunately, I'm not thinking very clearly right now. [sick] (gee, that smiley is a bit too graphic).

If I remember my beginners guide to writing, there are at least two different approaches to writing stories. One is the plot-based approach, where the story is structured into a series of acts hinging around major plot landmarks. A lot of films seem to be structured this way. The actions of the characters are there mainly to facilitate the set-up and execution of these plot landmarks, although their personalities will shape exactly what those landmarks are and how they are executed.

The other approach is the character-based approach, where the author creates a series of interesting characters, puts them in a setting, and then writes about how they react to each other. Another analogy would be the reality show, where the people chosen for 'Survivor' and 'Big Brother' are carefully chosen, and the resulting footage edited, to get an interesting clash of personalities.

RPGs can be designed in either way, although the plot-based approach is the dominant one today, as that seems to be the structure of the ancient saga style that is the distant root of the story type.

Your approach seems to be the character-based approach, which I also agree will work if done well. With the right bunch of character types thrown in the mix a series of interesting situations will develop. The downside (which can be seen on Australian version of 'Big Brother' [grin]), is that if you don't have the right character types to generate the friction required for the interesting situations, you might end up with a flat non-story. Have you got any good ideas for how to avoid this (as I think this would be fantastic if you could implement it)?

By the way, my theoretical approach to this, which I plan to implement eventually, is more along the plot-based approach. In this case, the NPCs aren't really autonomous, but are more like improvisational actors. The NPCs goals are thus to facilitate the generation and execution of the plot landmarks that are used to structure the overall story arc.
This probably doesn't make that much sense, does it? Sorry!

By the way, why were people freaking out by the Sims?
Advertisement
Quote:__C) An intelligent natural language parsing self-aware AI

I am very sorry [embarrass]

Jokes aside, I honestly like icon-based approach.

About useful NPCs
-----------------
I've said it in other threads. I believe the only way to develop (short term!) attachment to an NPC is to make them useful. By extension, you will only care about relationships if you really *need* someone. Wavinator's example is right on the mark, but I feel that incapacitation is the right way.
Yes, it is very specific, and that's what one wants when designing a game. What would be bad if it *has* to be always the same limitation.
I see this working if you define classes (hey, archer / mage!) and give them strengths and weaknesses. So, let's say you got a thief in the group, and he's constantly bugging everyone... but you know there's gonna be a lock at the end, and he's the only one who can pick it. If you were a thief yourself, then you'd just get rid of him somehow.

About proportional force
------------------------
I don't think it's as important as the rest discussed here, but it's still interesting. I guess your NPCs should be judging you (and everyone else) every time you do something, meaning if you use excessive force they should go weird on you, be it dropping from the group or plotting aganist you or asking to step down leadership.
Excessive force... now what is that? it means something different to all of us, and surely should for the NPCs, so you could have each one measure it as a comparison to what they would have done. This isn't as horribly complex to calculate as it sounds, you could have an AI running on your player that doesn't act, just thinks... and upon each action have it suggest actions using the thresholds/weights of each other player, then measure the distance from this to the path you took. Or something like that.
Again, seems like a lot of work while other basic things remain undiscussed.

About representation
--------------------
I really like Ketchaval's approach, establilishing a conceptual link to combat, which we are accustomed to in games. Although I didn't quite grasp the meaning to those labels, I imagine one could obtain a good hierarchy from it, sort of the good ol' "attack/spell/summon/item/run" menu from RPGs (which expands with details in each action). It also allows for an interesting (and maybe interchangeable with character status!) finite set of previously designed interactions. Also, short scripts with defined effects sounds good. Proper design of the effects would allow the desired mechanics to emerge, and the short scripts would be the icing on top of that.
Working on a fully self-funded project
Quote:Original post by Jotaf
Quote:Original post by ferr
There would be one player, that'd be you as the detective. Everything else of course is an NPC all of which have unbelievable AI.


You're right, UNBELIEVABLE AI :P What you're suggesting is really easy for anyone to throw out, it's similar to the "MMORPG where you can do anything" idea, but implementing it is plain impossible given the current state of technology. It would be a huge amount of work just to point out the flaws in that system :P Maybe you should ditch that and follow Wav's path :)


Yeah,
1) The only reason I posted the idea was because I ditched it already, realizing it was just a pipedream.
2) Are we here to only discuss concepts that can be currently implemented? I would like to think that one day a game like this could be (relatively!) easily developed.
3) What flaws? The only flaws I can think of that you could come up with are technology limitations, and I've already covered why that is irrelevant.
4) I just played a PC game the other day that was recently released (for free) called Facade, which includes chatterbox AI exactly as I described, so my gameplay isn't entirely out of reach.
5) It's not impossible, even given the current state of technology, just a bit out of reach. I'm sorry that you are unable to visualize how it could be done.

edit: coincidentally there's a thread concerning Facade right next to this one. nice how that works out.

[Edited by - ferr on July 10, 2005 9:28:49 PM]
Quote:Original post by ferr
Yeah,
1) The only reason I posted the idea was because I ditched it already, realizing it was just a pipedream.
2) Are we here to only discuss concepts that can be currently implemented? I would like to think that one day a game like this could be (relatively!) easily developed.
3) What flaws? The only flaws I can think of that you could come up with are technology limitations, and I've already covered why that is irrelevant.
4) I just played a PC game the other day that was recently released (for free) called Facade, which includes chatterbox AI exactly as I described, so my gameplay isn't entirely out of reach.
5) It's not impossible, even given the current state of technology, just a bit out of reach. I'm sorry that you are unable to visualize how it could be done.

edit: coincidentally there's a thread concerning Facade right next to this one. nice how that works out.


Cool, I'll check out Facade when I get the time :)

I didn't mean to put your idea down, you're right that some day it will be possible. But "possible" doesn't necessarily mean "easy". Chatterbot AI in a game is hardly proof that this kind of AI is near. There have already been many discussions on this, and I'm getting tired of it. The conclusion is invariably the same: it's just too much for today's technology, and in the world of tomorrow it would still require a life's worth of work.

The first issue that comes to mind is that, even if you develop AI good enough to mimic a sentient being (which is basically what you're asking for, just so you get an idea of how big it is), the AI alone would do nothing. Before anything interesting could arise from the simulation it would have to chew a huge amount of information on how the world and the human psyche works. What's obvious to you is not obvious to a blank, logical entity.

I could talk about other flaws, but as I said, it's been discussed. If you're interested search the forum for some posts on the subject.
I agree that a sentient AI to control NPCs is way over the top, and will not be achievable for games for decades if not centuries. There's no way you can emulate the massive knowledge base that such an AI would require without giving it the same senses as a human, then training it for twenty years. And that's assuming it has access to the same computing power and structure as a human brain, which we still don't understand yet.

However, do we need to have very smart NPCs in order to make compelling intelligent characters? From my understanding of basic human psychology, people are ready to enfuse things with human-like characteristics (what's the term, anthromorphising?). If the AI was to mimic human characteristics to enough of a degree, then our basic human instincts would flesh out the rest of the character to make it compelling. So the question is more: is it possible that we can fake an intelligent AI enough for people to believe that is human, or at least human enough to treat it as a compelling fictional character?
Hehe, what if the game is set up in chapters and in each chapter you play as one of several characters so that it isn't just a case of "I'm all-right Jack" but that how you act in one chapter affects the other character in the next. Ie. If you force someone to leave the group you end up playing them struggling on their own in the wilderness in the next chapter.

I find it hard to imagine this type of game though because I keep trying to relate it back to (combat oriented game like) tactical RPGs. Or ones where it is always a matter of success / fail.

Ie. If you don't have direct control over the way that the party goes, and fail to influence them to go to Luxor, should this be a failure? Or should each different path open up different gameplay / story-play events? Ie. Both directions trigger interesting events.


What sort of feelings (and relationships) would we be looking to produce between the characters?


Some ideas, suspense / unease / paranoia thinking that the other characters are plotting against you secretly.

Mutual? bonding, getting to trust and like an NPC.

Hatred.
Animosity.
Empathy
Liking
Disliking
Being wary of.
Wanting good things to happen to them, ie. achieving their goals


[Edited by - Ketchaval on July 18, 2005 8:18:39 AM]
Quote:Original post by Trapper Zoid
I agree that a sentient AI to control NPCs is way over the top, and will not be achievable for games for decades if not centuries. There's no way you can emulate the massive knowledge base that such an AI would require without giving it the same senses as a human, then training it for twenty years. And that's assuming it has access to the same computing power and structure as a human brain, which we still don't understand yet.

However, do we need to have very smart NPCs in order to make compelling intelligent characters? From my understanding of basic human psychology, people are ready to enfuse things with human-like characteristics (what's the term, anthromorphising?). If the AI was to mimic human characteristics to enough of a degree, then our basic human instincts would flesh out the rest of the character to make it compelling. So the question is more: is it possible that we can fake an intelligent AI enough for people to believe that is human, or at least human enough to treat it as a compelling fictional character?


I think that the answer is no. One of the projects that I now spend some time on is trying to implement a simple knowledge network in a fake RPG setting that translates to natural conversations with an NPC, in the form of asking questions and sharing information. And even in it's early stages, I can see that sometimes it does indeed feel like I'm talking to an individual who lives in that world and knows what it's talking about (instead of scripted text).

I imagine that getting the NPCs to talk to each other and share information (street gossip -- great for an RPG :) ), allowing logical conclusions to arise from this knowledge net ("oh my, the nearby town has been attacked by the orc, we might be next!"), and finally basing some of the NPCs' actions on this knowledge, even if in a very limited way, should be enough to lead the player to believe that every NPC actually has a mind of its own!

Of course, this is just one example. I remember reading Richard Bartle's articles on skotos.net and how they described a simple planning system driving the NPC's behaviour; and I was very impressed, knowing that if I saw an NPC in a game solving a problem in a creative way it would feel pretty much like there was a human in there.
"Consider the case of progressive murder: One story out come should be that everybody dies but the murderer, who gets to claim the treasure at the end of the quest."

Presented correctly, this is no worse than "dying" in current games. This reminds me of quest-driven board games such as The Lord of the Rings and Shadows Over Camelot. SoC, especially, since it has a traitor role that tries to make the heroes lose, and throughout the game mistrust spreads throughout the team, trying to figure out who the traitor is.

Check out my new game Smash and Dash at:

http://www.smashanddashgame.com/

Quote:Original post by Wavinator
Hmmmm... players dependent on NPCs. There's something REALLY valuable with this.

Part of the problem is that the player is so independent that he can normally override every concern except those related to combat (or stealth, or puzzles when allowed).

Now, if you have a traditional RPG that's tactically varied (i.e., teammates have distinct roles), you now make interpersonal conflict, an essential pillar of storytelling, a fundamentally integrated part of the game.


1st up I agree that this kind of storytelling game would be best prototyped as a multiplayer experience. A.I coding would be too difficult at the moment, so why not hand it over to real humans.

Secondly, the type of *gameplay scenario* that this relationships-focus feeds into is important, and I think one of the biggest problems is that we ARE generally thinking of combat/puzzles/stealth/exploration and how the character dynamic affects how this process works.

Well, what if instead we limited and / or removed these things and did make it into a sim type experience. Ie. We limit the location to a small confined area, say a recuperation hospital or a large house with a garden. We removed any 'combat' beyond a bit of slapping. We simplify sneaking around or ignore it. And we put the focus of the game *squarely* on characters trying to achieve the GOALS they are given at the start of the scenario.

Ie. A sims-type family game with conflict built into the goals.

The brother doesn't want the sister dating her no-good boyfriend, so he will answer the phone and only pretend to take a message.

The younger brother wants to get his bicycle repaired but doesn't have enough money, how will he get it? -Dad has left his wallet lying around, will he steal the notes / credit-card? Or do chores?

It would probably be better if they were all after a similar goal, but had to compete, make temporary alliances and backstab to get it.

So the game isn't about combat, but about conflict to get to a specific goal. The characters can't escape each other, and have to cooperate with one or other of the characters to do certain tasks. etc.

It would be kind of like a murder-mystery type "roleplaying" experience, but where each character has their own goals to achieve in the play session.
Quote:Original post by Ketchaval

So the game isn't about combat, but about conflict to get to a specific goal. The characters can't escape each other, and have to cooperate with one or other of the characters to do certain tasks. etc.


sounds difficult to do

This topic is closed to new replies.

Advertisement