Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Sowelu]

I hate you, Milkman Dan.

[Nivim]

 Morality...
 You can't divide actions into "good" and "evil" because they're too subjective. Try simpler things like "charity" and "greed". Wikipedia has a long list for both Virtues and Vices, most of which can be defined in the rules of a game, with a great deal of time anyway.
 It's true the game can't know what you plan, or why, but it doesn't need to. You can always choose to do something different when the time comes, completely regardless of previous plans or actions, so such knowledge is useless. Instead, it needs to record your past, as something more than the current state of an integer or five; memory. Since each act overshadows, and is overshadowed by this history, some kind of pattern recognition needs to be created. Something that sees "oh, your 'charity-greed' number makes a ~sine curve; your trying to balance it out, or you're conflicted" or "you've taken everything and given nothing so far, but hey, you just dumped all those gains into this (non-investment, non-refundable) cause; you've either had a change of heart or you were saving up."

 I wonder if DF might do something like this, something that creates a graph for your character's morality. It wont make your character do anything, but NPC will use it to figure you out.

[nenjin]

I sort of described this to a friend in a pipe-dream game talk. It's kind of modeled on what Toady does.

You give the character many morality attributes. Every action taken adjusts one (but probably two) attributes. Every NPC has these attribute ratings, and their AI hinges on how your scores compare to their's. You give them favorable, unfavorable, neutral and violent responses in the dialog. You tinge and color all the voice overs along these lines, so they can deliver the same lines in different tones to show how they feel about the player. 

When you apply that to the AI dealing with itself, and all the NPCs interacting with each other, and you've basically procedurally generated a real community. When the NPC AI's actions can adjust their own morality as well....you get dynamic change. As a PLAYER, experiencing that and being able to manipulate it would be a really gratifying.

It would make for an awesome introduction story to the character, playing little vignettes of their life growing up, making choices, that affects their starting morality values when the main story beings.

[Nivim]

 I think for such a good morality system, the responses would need to be altered by more than colors, and altered from more than those four possibilities. I can't think of something that doesn't lead to a procedural conversation generator though.

 It would be much harder to figure out just what attributes (virtues and vices) are modified based on choices in DF, since there are just so many choices. I guess charity and greed wouldn't be that hard, but it would be dependent on just how much the NPC cared about whatever it is. Giving all your gold away to desert tribes who want your water, probably wouldn't be very charitable, even though a foolish character might think it so.
 I'm not sure if dynamic change would work very well for the NPCs, since their choices are based on the morality, but the choices are meant to change the morality. So it would either mean things would stay just the same or would create loops. There's change by experience, but that isn't dynamic the same way as the player's would be.

 (An iron mace named "Apathy" strikes at your torso, but it slides off your chiton armour!)

[subject name here]

op is way tl;dr but can someone tell me if it's another one of those "things are different from the way they were when i was a kid and change is scary" things?

ps: sid meier rules

[Jack A T]

The best morality system I've ever seen in a game, in my opinion, is Ultima 4's system.  Yeah, only way you could win was by being good, but actions that were good in one way were evil in another way, and that ended up mattering.  It was more than just "good or evil".

[nenjin]

I think for such a good morality system, the responses would need to be altered by more than colors, and altered from more than those four possibilities. I can't think of something that doesn't lead to a procedural conversation generator though.

Well, you have to start somewhere and, from a game play perspective, thinking about your average schmoe NPC you deal with, you don't need more than that, at least for non-major characters. If you want something from any NPC you deal with (even if it's just a reaction), all you really need to know is how that NPC is going to react to you. You COULD decide that accepting but disdainful, snide but accepting, and all those variants are important, but at the end of the day, they're really not for game play. That's all flavor, and could be expressed through a few different tone variants and facial expressions. In a game sense, they still either have to approve of you, disapprove of you, don't care, or want to kill you. A real person is usually limited to those choices when dealing with others. Sort of like if you asked someone to state their simplified, ultimate opinion of another person.

You also don't need a procedural conversation generator for this, because that's still way too complex and micro I think for what most people expect out of an RPG. You just need to know that your choices are having an impact in the way people react to you, and that it can be subtle as well as overt. Right now, we basically have "like, don't care, and kill on sight" and all game play is driven by those three beliefs. So there's really no middle ground. Adding two more states, that THEN play off a lot of individual actions by the player, would add a lot of depth and increase the range of responses, reward or punishments that can be dealt out.

It would be much harder to figure out just what attributes (virtues and vices) are modified based on choices in DF, since there are just so many choices. I guess charity and greed wouldn't be that hard, but it would be dependent on just how much the NPC cared about whatever it is. Giving all your gold away to desert tribes who want your water, probably wouldn't be very charitable, even though a foolish character might think it so.

Again, think macro vs. micro. Important people, yeah, you'd want their perception to matter a lot. But all the people that fill out the world? Not so much. You can abstract their opinions away into much smaller and more manageable chunks. You mix and match changes in morality as well, so you're not just throwing money at poor people to get your "charity" score up.

So something like, there's a criminal in town who stole to feed his kids. You track him down and he says you'll have to kill him to stop him. So you do.

Peace value goes down, violence value goes up.
Compassion value goes down, stern value goes up. (Can't think of a good word that is the opposite of compassion atm.)
Law abiding goes up, and criminal goes down.

So your actions have a large impact on your own morality, and by extension, to the rest of the game world and not in a simplistic one or two combination way. And some morality would be weighed more heavily than others, or the relative difference between them needs to be higher for one to cancel out the other. So you could get to something like Robin Hood for example, where you can be a likable celebrated criminal, even though you're still a criminal. 

I'm not sure if dynamic change would work very well for the NPCs, since their choices are based on the morality, but the choices are meant to change the morality. So it would either mean things would stay just the same or would create loops. There's change by experience, but that isn't dynamic the same way as the player's would be.

Again I think this would be a macro vs. micro kind of thing. I'd want major, important character's morality to be able to shift based on what they do or you cause them to do. The filler? Maybe not, although it would allow for things like a nice peaceful town turning violent because, say, an NPC group of bandits moves in and forces them to start fighting to defend themselves. And this in turn makes a violent player character more acceptable to them. Maybe the player has started killing townsfolk, and that has the natural result of making everyone more violent (not necessarily accepting of murder, or lawlessness.)

That it could create cyclical feedback loops is actually realistic. Assuming the change took long enough to be noticeable and feel meaningful, I don't know if the loop is really a problem. It'd be worse to have moralities that would change with NO hope of changing them back.

[Soulwynd]

Instead of some absolute good/evil system, I like seeing something more geared liked/disliked by a certain nation/city/individual/group/etc.

Makes way more sense than the old almost religious good/evil concept.

[Nivim]

Soulwynd, that sounds like a vote, which doesn't help in this type of conversation, or that you didn't read what was written, which is very annoying.

 The part I'm rather confused about is the "micro vs. macro". What makes an important person different from the average person? Why would there be any "filler" if you meant the morality system, the individual version we were speaking of, to be used to create a procedural community?
 The problem with the loops is that if they included any changes, it would repeat indefinitely, maxing and minimizing the numbers for every NPC involved. Some other element needs to be added other than the morality and experiences.

 I apologizing for not countering and reworking everything point-for-point, like is normal in these conversations, but I'm a bit too tired to go through all the inevitable logical traps, and I probably made them worse by trying some.

[nenjin]

S'ok.

Macro: View all the characters that aren't vital to the story as a large population, rather than trying to focus on the individual. Create a system that addresses everyone, instead of trying to write one that accounts for every specific person as an individual. Think of it sort of like Entity populations in DF, versus the individual dwarf psychology that is in Dwarf Mode.

Micro and important characters: Quest NPCs, for lack of a better term. Party members, ala Dragon Age 2. For them, you'd want a system that's more detailed so it can account for the fact you'll be paying much more attention to these characters than....

Filler: Not talking about the AI or morality here. "Filler" are all the characters that populate the world so it's full and interesting and stimulating. The people on the streets, in Inns. The bums. The guys who say stuff when you walk by. All the filler can be addressed in macro-style. You want to see people reacting to you different, but you aren't striking up conversations with these guys. You need their moods and attitudes to shift, but it doesn't need the finesse you'd have when dealing with an major quest NPC like the King or his Councilor.

[Soulwynd]

Soulwynd, that sounds like a vote, which doesn't help in this type of conversation, or that you didn't read what was written, which is very annoying.

I'm annoying like that. Nope, but I've read it. Any sort of morality can be resumed into being liked or disliked by a certain peer, be it macro or micro. Including oneself. Last time I programmed a mud, about 10 years ago, each unique NPC, faction, guild, side, had an array that contained player id, like/dislike var, and the time when he was last seen for flushing purposes, since it's a multiplayer game. Certain actions done against certain npcs changed how liked/disliked he was and so on. As for the unique npc interactions, the npcs had certain priorities. Would he get past disliking someone if the said someone was liked by a faction? If so, by how much? And so on.

From a programmer point of view, even if you use good/evil or whatever system you want to use, you're basically just playing with a number. As for it being all procedural, it can be done, but it takes a bit more of work to make it seem a bit less computer-generated.

[nenjin]

It's all just numbers. Trying to find a way to connect those numbers to something interesting and dynamic and fun is the challenge.
The more numbers you use and the more inputs you have, the more interesting it is for the game. Many simplistic good/evil/morality models could be more interesting just by changing how many axises you decide to use, or how many actions you could take (that you'd actually WANT to take) that affected morality.

King Arthur the Roleplaying Wargame had a neat system that got me thinking along those lines. You had your "Righteous vs. Tyrant (Good Vs. Evil) on one axis and Sidhe vs. Christian (fairies vs. the Church) on the other. Your morality was a single score that moved across a sphere, describing the synthesis of all your morality actions. At it's core it was still really simplistic (all the sphere described was what abilities/magics you have access to) but it was a move in the right direction. More inputs, more directions.

[Muz]

A friend of mine roleplayed an evil ghoul in a D&D game.  He WAS evil, in his heart.  But he followed the party around and never did anything evil at all.  He said "Hey, I need to eat corpses to live, you guys make lots of corpses, I don't have to eat people, I don't starve in agonizing pain, and if I ever turn bad, you guys can destroy me."

But, he always had betrayal in his heart.  And the paladin ended up flipping out and going his own way.  Because paladins can't associate with evil characters.  Of course, the ghoul knew this...

Games can never know your true intentions when you never express them to anybody in any way!  Hrm.  But then I guess that's okay.  But...I'm thinking that NPCs should never be expected to have ANY knowledge of your long-term plans.  That way you can take them by surprise, which gives you coolness cred.

D&D morality actually seemed to make sense to me, because you had a DM who could play the lawyer over what they meant. A paladin who burns a town down to kill the supervillain who wanted to destroy the world? Rather neutral, because the good and evil crosses each other out. A paladin who burns a town down to kill the supervillain so that he gets the bounty on the villain's head... definite evil. It's why you were allowed to punish 'inconsistent morality', because you could clearly see when someone is simply playing the good guy to abuse game mechanics and use that +5 holy avenger.

But what about when people do logical things on both sides of the moral scale? Take for example, your typical inquisitor types who kill good people because he genuinely believes that they're bad, and that evil is a cancer that must be stopped. He donates his treasure to orphans, he saves villages from dragons, frees people from slavery. But he also kills innocents. And brags about his conquests.

Or what about the orc who follows his leader's orders to wipe out all of humanity? He believes that humans are pests and intend to kill all the orcs anyway. He takes part in pre-emptive strikes and raids, chops off human heads and piles them into a big bone wall intended to scare away humans. But deep down inside, he does it because he doesn't want his wife and kids getting raped and killed by human adventurers.

Computers just aren't very good at tell what is good and what is evil. And I suppose game developers too, because most of them are evil anyway.

It's all just numbers. Trying to find a way to connect those numbers to something interesting and dynamic and fun is the challenge.
The more numbers you use and the more inputs you have, the more interesting it is for the game. Many simplistic good/evil/morality models could be more interesting just by changing how many axises you decide to use, or how many actions you could take (that you'd actually WANT to take) that affected morality.

King Arthur the Roleplaying Wargame had a neat system that got me thinking along those lines. You had your "Righteous vs. Tyrant (Good Vs. Evil) on one axis and Sidhe vs. Christian (fairies vs. the Church) on the other. Your morality was a single score that moved across a sphere, describing the synthesis of all your morality actions. At it's core it was still really simplistic (all the sphere described was what abilities/magics you have access to) but it was a move in the right direction. More inputs, more directions.

Something like this would make for a fun model. Good vs Evil is just way too simplistic for most things. You can have different measurements for intentions, actions, or even more specifically.. actions and intentions against certain people. A person could be fanatically loyal to his (good/evil) country, but that might not mean he's a bad person. Then the game would judge morality based on a weighting of these different things.

[Virtz]

The best morality system I've ever seen in a game, in my opinion, is Ultima 4's system.  Yeah, only way you could win was by being good, but actions that were good in one way were evil in another way, and that ended up mattering.  It was more than just "good or evil".

Pretty much this. Even though you only really dealt with aspects of good (the 8 virtues), they were done in an interesting way.

[Sowelu]

But what about when people do logical things on both sides of the moral scale? Take for example, your typical inquisitor types who kill good people because he genuinely believes that they're bad, and that evil is a cancer that must be stopped. He donates his treasure to orphans, he saves villages from dragons, frees people from slavery. But he also kills innocents. And brags about his conquests.

Oh man, that reminds me of the webcomic Goblins and the paladin Kore.  I mean, that comic focuses on morality a lot anyway (since the goblins are the main characters), but there's a dwarven paladin who is SO far off the deep end...  You don't see him much, but you just know he's going to show up more later.

Spoiler (click to show/hide)

He slaughters a young human child because a group of orcs and trolls had adopted the boy, and he figures they had already planted the seed of evil in his heart.  Sorry, kid.  Evil is a cancer, and it can't be allowed to spread.

I do see this stuff as a great argument for a "psychotic" scale in morality.  Isn't that a Mass Effect thing?  I've never played it, but don't they have a morality bar with "Renegade" on one end, since you're supposed to be the good guy no matter what?  Good and evil is the interests of self versus society.  Law and chaos isn't all that interesting to me, that's just personal playstyle.  But if you turn law/chaos into upstanding/psychotic, then it makes a lot more sense.  That upstanding-selfish merchant who tries to drive all other merchants out of business and then raise his prices, even if it means that poor people go hungry?  Sure, he's a bad guy.  But is he really worse than the selfless-psychotic masked avenger who sneaks in and shanks him in the middle of the night, then doesn't steal anything?  GOOD QUESTION.

And that scale is a lot easier for a computer to measure, I think.  Even looting the bodies makes you more selfish--it might be a "good" act, but it's still a little sketchy; public opinion of that would be questionable (sure, take a dead guy's sword from a battlefield, but don't go collecting twenty swords and selling them).  Society can be totally accepting of people who are even very selfish or moderately psycho, but if you start heading into the extremes, then it will radically change their perceptions.

EDIT:  Okay, question for y'all.  Should your morality be noticable when there's nobody around to see it?  If you shank some guy in an alley (god, that's like my favorite word now), and never get caught...then what?  If you commit heroic deeds and never tell anyone, what's the game effect?  What about lying about doing great deeds and never being found out?