Quote:Original post by kseh
First thing that came to my mind was a brain scan of some sort. Then I thought, maybe if you had a 'map' of the character's brain and different areas represented the diferent personality traits you want and actions a character could take. When it comes time for the character to make a decision, maybe do a pathfinding thing from trait to action and the top 5 present themselves as choices. When events occur, alter areas of the brain map. For example, the more violent actions the character does, make the area near the violence attribute have a lower movement cost. If a traumatic event occurs, cut a swath of high movement cost through some area of the map making it tougher to make the same choices the character used to.

This is a very intriguing approach, although pathfinding might be a little like using a drag racer to go to the corner store for milk. Weighting for actions based on decisions might work just as well.

Strangely enough, though, I get the strong feeling that I need to handle this whole idea with a lot of care. If I make it a really stark calculation of numbers, what I'm in a sense saying to the player is that they are nothing but a machine. That would be bad because it would be exactly the kind of mechanistic portrayal of things I'm trying to avoid (heck if I didn't mind that I could just make a combat game and be done with it).

The brain map idea is still very cool, though. But even though it flies in the face of player control and immediate, clear feedback I don't think you should just be able to tweak some region of the brain and get a certain effect. Maybe a CHANCE of an effect, but just like magic in games I think there should be some mystery.

Quote:Original post by Tim Ingham-Dempster
As to using it for gameplay or character being mutually exclusive, if one option is good and one is bad from a gameplay point of view then it will always be a gameplay issue, if all options are balanced it isn't a gameplay issue. You could have some decisions gameplay affecting and some character affecting, but you would have to be very careful to make sure the player knows which is which. It would be very easy to have some players think that all decisions were tactical and miss out on the character side of the game and other players think that all decisions were character based and make poor game choices without understanding why.

Wow. Okay, thanks for explaining this even further. I should have seen this but wasn't really connecting the dots.

This is like the choice to be evil in a game like Fable, where you get some great material reward (a cool sword), or to be good, which changes story (your sister's life is spared). It's clear to me now that the choice can't be a mix of one or the other, as it's awfully tempting to act out of character just to make gameplay easier/more fun.

I think I'm biased toward gameplay changes with what probably amounts to flavor text that describe how your character is evolving.

Quote:
I'm not convinced that a player doing things that go against their alignment on the internal state axes should make them less happy and vice versa. If a pacifist gets caught in a war then it would reduce their happiness, but if they chose to go to war then isnt it more likely that they changed how they feel about the issue? An alternative way to make the player get more set in their ways as they get older could be to assign values for the internal axes to every action a player can take (you would have to do this anyway to determin how happy or sad to make them based on that action). When presenting the player with options you would only show ones which are within a certain spread of the player's position on these axes. As the player ages the spread would get narrower. I don't know how well that fits with what you have planned for the rest of the game, but its an idea.

This might work for what I'm trying to get across. Certainly if the pacifist chooses to go to war they're choosing to change their philosophy, at least on a certain issue. That amounts to a personality shift of sorts.

So it may be best to ditch happiness as you're suggesting because the best place to locate the "soul" of the character would naturally be with the player. I'd thought that happiness would be a positive incentive, but the negatives are starting to worry me.

Thinking about this further if you chose a starting role and then decide to abandon this because of gameplay reasons the game shouldn't punish you, and your idea starts sounding better and better when I factor in the possibility of making the play environment fluid. If, for instance, the player has a home that's threatened with extermination, being a pacifist may no longer be a viable gameplay path.

Quote:Original post by RedFawks
Another thought though, what if during the character creation process, it were to ask you a series of questions about their personality.

Actually this would be a cool way of setting up your character. One of the old Ultima games (Ultima 4?) did this, and I think Morrowind offered this sort of thing as a character creation option. One thing I'd be worried about is in matching the answers to the type of character the player wanted to play and getting in the way of them being able to tailor their character to the ideal they hold coming into the game.

Warning: personal philosophy regarding how games should be designed follow. You may have a different approach, which is cool. (And, based on the thread content so far, I believe it highly likely that you do.)

See, I think you're approaching this in a fundamentally wrong way. In a game, the player and their agent are one and the same entity. The agent is really just that-- a virtual hand with which the governing intelligence carries out their will. In practice, there is only one entity in the expression, not two.

And there's where it gets interesting.

Since there is only the player, you can't really adjust the player's (physical, as it were) "evil bar" and be done with it. For starters, there's no such thing. (though if you do become aware of one please let me/the behavioral psychology community know ;) ) I think it's also very telling that all those sliding scale morality systems you see in many RPGs invariably fall on their ass halfway through. They feel tacked on because, no matter how many branching dialog trees with alignment checks or whatnot that you throw in, that's all they are. Some arbitrary game mechanic/stat to be minmaxed in order for the player to make the plot play out the way they want it to/the PC to act consistently with the desires of their controller. There is no meaning, and never will be, no matter how many levels of Good/Evil Bar(tm) gradation or branching whatever that occur as a result.

That doesn't mean that the game is somehow incapable of affecting the player in any way or indeed that the game cannot react to said affectations. It's also very telling that the one morality system I'm aware of that eschews a lot of the minmaxing bullshit (the Paragon/Renegade system from Mass Effect) has been met with universal acclaim and works exceedingly well in practice. On the surface, the difference seems extremely superficial, (even contradictory, considering that you can max both attributes out at the same time) but if you really look at what happens as a result the whole thing has been turned on its head. The game no longer tells the player what their stance is on some issue, *the player tells the game what's going on.*

So, to bring this full circle, the question really becomes one of interpreting the player's choices by way of examining what he/she does with their agent. Could the trait things work? Certainly. In fact, I think you could really make something interesting by constructing a comprehensive set of personality trait continuums and tossing them into a L4D-esque AI director. I don't think that simple underlying mechanics are the problem so much as the blatantly obvious presentation (i.e. YOU DO NOT HAVE A HIGH ENOUGH GOOD SCORE(tm) TO REMOVE THIS PUPPY FROM THE PATH OF THE TRUCK AT NO PERSONAL RISK) you tend to see in a lot of games, *especially* those that stress 'choices and consequences.' Really, the tricky part would be coming up with a good set of indicators/scales and a 'director' capable of producing scenarios that aren't half-assed mission Mad Libs. (save X from Y with plot twist Z) Certainly no mean feat, but they are fundamentally approachable problems.

EDIT: And looking back it appears I tl;dr'd a bit somehow (though I could swear I read everything :X) and some similar suggestions were made. I like to think I have a more concrete idea of how such a system would work, though, so I'll keep the post content up. Trait discussion would certainly be welcome, though, as this is something I'm wrestling with on a personal project.

Quote:Original post by InvalidPointer
Warning: personal philosophy regarding how games should be designed follow. You may have a different approach, which is cool. (And, based on the thread content so far, I believe it highly likely that you do.)

No worries or need for a warning, I really appreciate this. It's often better to get criticisms than "hey, that might work, go for it and see what happens." At least with a critical reply I've got the chance to test whether I have conviction in the idea or whether I need to evaluate it further.

Quote:
In a game, the player and their agent are one and the same entity. The agent is really just that-- a virtual hand with which the governing intelligence carries out their will. In practice, there is only one entity in the expression, not two.

In most cases I'd agree with you, especially where you're asked to identify strongly with a single avatar (FPS or RPG, for example). But there are quite a number of exceptions, and sim-games, especially the sim-life genre, are probably among the biggest (grand 4x strategy or RTS might be another).

A sim game breaks the direct control expectation by casting the player as a director / manager. They know they don't control the avatar directly, only by proxy, and they do this through the the internal states they must manage. The fact that the internal states are visible and even exist, and often model esoteric or undesirable states (like unhappiness or loneliness), are the biggest clue in terms of convention that the player is NOT directly the will of the avatar. ("After all," the thought goes, "I'm not sad or lonely." The Sims further breaks this convention by giving the player avatars they can potentially lose control over, as with Sims that become too depressed).

But your point is valid in the sense that I'm looking to fuse RPG and sim-life game. How critical the convention is will depend on how far implementation falls on the RPG side of things.

Quote:
Since there is only the player, you can't really adjust the player's (physical, as it were) "evil bar" and be done with it. For starters, there's no such thing. (though if you do become aware of one please let me/the behavioral psychology community know ;) ) I think it's also very telling that all those sliding scale morality systems you see in many RPGs invariably fall on their ass halfway through. They feel tacked on because, no matter how many branching dialog trees with alignment checks or whatnot that you throw in, that's all they are. Some arbitrary game mechanic/stat to be minmaxed in order for the player to make the plot play out the way they want it to/the PC to act consistently with the desires of their controller. There is no meaning, and never will be, no matter how many levels of Good/Evil Bar(tm) gradation or branching whatever that occur as a result.

Isn't this partly due to the heavy investment these sorts of games make in trying to immerse you in the role, story and environment? In a game that pins user experience on immersion the more you see arbitrary, quantitative values tracking nebulous, qualitative traits the more discontinuity I think you experience.

But I've abandoned immersion because it's just simply unfeasible. So by convention when you play what I have in mind there's no question about you seeing internal states simply modeled. Stat bars are simply part of how you would play the game.

Quote:
So, to bring this full circle, the question really becomes one of interpreting the player's choices by way of examining what he/she does with their agent. Could the trait things work? Certainly. In fact, I think you could really make something interesting by constructing a comprehensive set of personality trait continuums and tossing them into a L4D-esque AI director. I don't think that simple underlying mechanics are the problem so much as the blatantly obvious presentation (i.e. YOU DO NOT HAVE A HIGH ENOUGH GOOD SCORE(tm) TO REMOVE THIS PUPPY FROM THE PATH OF THE TRUCK AT NO PERSONAL RISK) you tend to see in a lot of games, *especially* those that stress 'choices and consequences.'

Regardless of how we might look at games differently this I think is the cool part.

You want to tell the player that previous choices mattered, and since the subject is morality / being I think you want to make the previous choices constrain or impact future choices. So how to do this?

"Not enough points to rescue the puppy" (funny) is hamfisted, I agree. But what the game is trying to convey is that "you're a cold blooded SOB and you've made choices that make you unfeeling."

There's no way to capture a lifetime of choices in a few hours of play and moreover I think this concept brushes up against some philosophical constraints, chief among them whether or not your past experiences really do, in a sense, fossilize your perspectives/beliefs and subsequent freedom to make WHATEVER choice you please. I think the issue of agency gets really strained at this point-- a player may find themselves thinking, "I would/would not" do this or that, but what if their capacity is rightly diminished (as with old age, impairment or even addiction)?

Quote:
Trait discussion would certainly be welcome, though, as this is something I'm wrestling with on a personal project.

I'd offer some possibilities but I think they'd probably suck due to how we may be viewing agency. An immersive game which does not model the self (at least not the psychological self--hit points always seem okay) is probably going to have to offload indication of traits into the environment. In a surrealistic sense you'd have Flower, changing the perception / color / gloom of the environment, but in a more concrete world it would probably have to come from the characters or really graphically intensive stuff like being able to see your own haggard face in the mirror.

This actually flows into a broader idea I toyed with. Think Tarantino or good cinematic noir.

One of the passages in The Red Badge of Courage describes how this man plagued by his inner cowardice is enveloped in this crisp clarity in the midst of a desperate fire-fight.

The grass is distinct and vivid, the bloom of a bullet wound on a soldiers chest reminds him of a scarlet flower...

If you ever written a screen play you know the challenge is to show the story not to tell it. Consider this approach with your idea.

So in the midst of battle the gunfire becomes rhythmic, the flames of a burning corpse appear warm and comforting, a pile of corpses doubles as a short flight of stairs. This defines the scene for a battle-hardened veteran whose comfortable in the midst of horror.

The same scene for someone that leans towards the Beauty side will appear(and thus render) differently. The gunfire shakes the ground and the screen and cracks like thunder, making it hard to keep your bearings. The flames of a burning corpse are like a firey devil, it lashes out when you draw near and almost swells your eyes shut with thick smoke, the piles of corpses are a slick mass of meat and a testament to your impending doom.

In a pacific setting the one that leans towards beauty will interpret actions more clearly. A handshake is a gesture of friendship, where for the one steeped in horror it may appear to be a menacing fist. A friendly offer of drink is one experience for the beauty and a possible assassination attempt for the horror character.

If you've ever seen the movie Soldier it was a somewhat corny story of a group of super soldiers; orphans trained from birth to be blood-thirsty killing machines.
There are several scenes where the protagonist (Kurt Russell) misinterprets everyday physical exchange with potential life threatening actions.

They achieve this by subtly altering the facial expressions of the actors as they make the move and altering the camera angle to make the move appear menancing. When they combine this with a low siren sound it becomes a very effective technique to let the audience see a simple offer of a handshake or a hug through the eyes of a professional paranoid that doesn't understand common physical exchanges.

So the larger question I'd pose before you attempt to imagine the perfect abstraction for the game mechanic is to ask yourself how much you can encapsulate through clever rendering quirks and sound effects.

A horror player may get an upbeat tempo "slaying soundtrack" in the middle of battle while a beauty player may get a horror track that has them spinning in every direction for the next abomination to befall them.

A horror player may be deaf to the sounds of the dying but a beauty player will hear every screech in the night.

See how far you can take the mechanic by presenting it in a rich format, and then abstract whatevers left over, if anything.