Another thought I just had... what if you convey your motives to the game via an inner dialogue (angel/devil on shoulders style).

Click fireball spell on orphanage...*popup of angel/devel profile*

Angel: Are you sure? Those kids are innocent!

Devil: Those kids must die!

*clicks devil*

Angel: It's for the greater good if I can infiltrate the bandit lair

Devil: Those souls will make me super powerful!

*clicks angel*

game result -> neutral

I'm going to offer some criticism, but just to be clear, this doesn't mean don't pursue your idea, it just means think about what benefits you actually gain, and whether the detriments are worth it.

Constructive criticism is always welcome, you don't improve your ideas if everyone agrees with you :D

It seems the main argument here is that the computer will do a terrible job of determining your true goal, which is very true. Ferrous mentioned a witness system, which I very much like the idea of. Going back to the Order of Paladins example, they aren't going to let you in if they just saw you burn down an orphanage, but if you were well disguised at the time (and nobody tracked you to the point you took your disguise off) then they would still accept you. This could still limit your choices in the game, but in ways that are easier to avoid (rather than stopping you becoming a Paladin because you killed a pet chicken at the start of the game).

Conversely, some NPCs may require witnessing you murdering the children to prove to them that you really are evil.

I'm wondering how far you could take this with "Higher Powers". For example (Paladins again, because they're easy to pick on as a class) if you become a Paladin, but then murder some innocents, do you lose the powers the angels have bestowed upon you? Maybe you fool your way into the order and when you undergo the rights of passage the deity shows up calling you out and you get to kill him, weaking the Paladins' magic (taking them down from the inside). Can you take over as the leader and spread rumours about the (actually good) enemy, so that your follwers believe they are doing good and therefore create a loophole?

There is a lot more complexity in this topic than I first thought about, which was just to have a set of scenarios that could be completed in good/neutral/evil ways that may influence the kind of NPCs that offer you help/try to kill you.

Mmmmh.... I know it doesn't fit the classical good/evil split, but personally, I prefer games which are offering a more "realistic" characterization of people and organizations. Which ends up in "shades of grey" instead of the "black and white" of the classical "good/evil" split, but that can be a very good thing as it gives people a better chance to actually relate to ingame characters.

In addition, just as with player characters, enemies, organization you can join, hell even the NPC selling you items should have a motivation to do the things they do, else the players immersion might suffer.

So, for an NPC selling you stuff that might be easy. He needs money to feed his children, you have money and do not seem to be a thief, hence you are in business. Just make sure the store NPC comes across as running a business, not giving out freebies. Giving the player items for free as a promotion -> yeah, why not. "Here, have an item, its a gift" -> this store would most probably already be out of business.

For me, one of the worst things in RPGs are loot chests in stores. COME ON! I walk into a store, walk to a chest behind the counter in plain sight of the store clerk, take the stuff out without asking or paying. Then I walk to the counter, and the clerk asks friendly if I would like to buy some items. Either he is blind and deaf, my player character is so terrifying that the clerk doesn't want to mention the theft he just saw (though why the city guard is not after my sorry arse the next time I set foot into this town is beyond me), or the store owner is gullibly stupid!

Now, as to the paladin order. I know, the paladins are usually the goodie two shoes of fantasy games and stories. I can live with that. But. Paladins are warriors, not saints. They serve a church, not god directly. And if a church is running a warrior order like knights or paladins, that church is after power before everything else.

So why, the hell, would the paladin order care if your character burned down an orphanage? If anything, that deed has proven you are worthy of being a paladin, as a warrior of the church probably will burn down a hell lot more buildings - and humans. Its unfortunate that you have proven your worth by burning down that innocent children, but then, that orphanage did not pay their tenth to the paladin order to get their divine protection, so a friendly paladin mob would have paid the orphanage a visit soon anyway. And these suckers were pretty stubborn, so the paladins would have burned down the orphanage anyway at some point.

Yes, its not the nice paladin order from the clean and stock fantasy game. But that paladin order never really made sense, or had any kind of depth. You want depth? Might need to add some grey to the white of the paladin order, just as your satanist cultists need some shades of grey to make them interesting.

I really would go in a different direction with your good/evil split.

Instead of concentrating on white ("your character shall not do evil things... oh, you killed that goblin that attacked you, that is actually evil, you know"), and black ("You just saved a ton of human beings by wiping out this village... there was a mass murderer amongst them. -10 evilness points"), which are concepts that will lead to pretty bland and boring characters and stories, as well as a ton of incosistencies (like when is it good to kill someone? So why is your paladin wearing a hammer?). Instead of that, I would concentrate on a heaven/hell split, or order/chaos, making sure neither of both is good or evil in a moral sense. They are polar opposites struggling with each other, and your character can follow them while doing good deeds, evil deeds, or anything in between.

I know there already is an order/chaos axis in the classical alignment system, but people tend to concentrate on the good vs evil axis even though this often leads to extremly flat characters.

If I would have to keep the good/evil axis, I would make sure that you cannot, no matter what you do, max out your good or evil score. Its realistic, its more fun, and it will drive some min-maxers mad ;)

For example there have been some very good games where you had to decide between dialog choices, with clear good / evil choices. Only that some good choices actually made things worse for people and ended up giving you evil points, and the other way around.

Done well, this is a very powerful thing. The player no longer can play "by the book"... "oh, I can rescue those children and get my cookie points"... well duh, those kids are possessed and will now wipe out the whole town, so forget your cookie points. And look more into it next time before carelessly clicking on a dialog option. The player will have to think, and really get into the story to make the right decision, instead of just skimming answers.

Also, it can enhance the story. You want people feel the emotional weight of a story section, while leaving them with the impression that it wasn't a cruel story writer, but their own poor decisions that lead to the sad outturn of the story? Give them options, make the options influence stuff big time, and make sure the influence will turn into a bad thing no matter what the player does. Used carefully, this can really enhance the story.

Another idea is to have a more complex system in place. Want to get into that paladin order? You will need to offend those monk order living next door, because they hate each other. Just because both serve god and are pretty good on the alignment score doesn't mean they agree with or like each other.

Want to max your score? Fine. You still need to decide who should be your friend or enemy, because you cannot be "good" in everyones opinion. Result basically is a seperate good/evil axis depending on faction or even NPC.

Some NPC might have seen you do evil things and is certain you are the devil incarnate, while those monks still consider you good enough to join their order.

Maybe with slow propagation of such values to other NPCs? Maybe you killed kittens today, and the burglar that saw you do that now thinks you are evil. You apply for your paladin order membership tomorrow, and get in as the paladins think you are looking like a fairly decent guy. Next week the burglar gets caught, and during interrogation tells the city guards what he saw you doing. Two days later the city guards find out you are now working for the paladins, and the day after a city guard member tells one of the paladins. It takes another week to reach the head of the paladin order, but a day later you will finally have to answer some questions because of the rumour that you killed some kittens two weeks ago.

Sounds complicated, could propably be simplified into a fairly simple system, where NPCs have certain relationship networks such eyewitnessed deeds could propagate through with some delay.

If you also give the player the option to kill the eyewitness (in order to not get caught, thus stay "on the good side" by "doing more evil deeds"), you could then also implement a background process that would simulate how the network of other NPCs connected to the killed NPC is getting aware of the NPC going missing, and propagating this to law enforcement NPCs, which would then start a simulated search (basically just a delay) which might turn out to find the victim, and if they do, with a slim chance you still get caught (which then is just a random chance).

Thinking about it, that would be a TERRIFIC system... maybe a little bit complicated, still.

It would give players wanting to play good characters the constant lure of "the easy way out", while if they choose this path, also constantly threaten to blow their cover. You can just shortcut this section if you just kill all those guys over there... but that wouldn't be a good deed, would it? Then again, nobody will see it, so nobody will ever find out? Ow sh*t, didn't see that hooded girl over there trembling in fear after seeing me kill all those guys in cold blood... well, no point in my shortcut if I let her escape now. Just need to make sure she will never talk about it, so nobody will find out. What can go wrong?

You see where this might lead, given there are enough potential eyewitnesses around, while the "dark side" way of trying to get good points actually IS almost always easier. You are seducing the player to take the shortcut, while making it clear from the beginning that once he takes the first shortcut, he might need to do it again and again, each time with a bigger chance that he will be found out.

There would be players ignoring this system of course by never taking the shortcut. But the mere existence of the shortcut might actually even enhance their expierience... "damn is that crowd annoying... maybe I should have just dumped them into that abyss over there when I had the chance...."

Of course, I work here from the asumption that this is a story driven RPG where the story trumps everything else. I know there are people who just want to roleplay without much of a story, and not everyone is offended by flat characters as I am. So take my opinion with a grain of salt as usual :)

Sounds complicated, could propably be simplified into a fairly simple system, where NPCs have certain relationship networks such eyewitnessed deeds could propagate through with some delay. If you also give the player the option to kill the eyewitness (in order to not get caught, thus stay "on the good side" by "doing more evil deeds"), you could then also implement a background process that would simulate how the network of other NPCs connected to the killed NPC is getting aware of the NPC going missing, and propagating this to law enforcement NPCs, which would then start a simulated search (basically just a delay) which might turn out to find the victim, and if they do, with a slim chance you still get caught (which then is just a random chance).

Thinking about it, that would be a TERRIFIC system... maybe a little bit complicated, still.

(This should probably be a different post, but since it's on topic here, I'll go with it)

This is actually an idea that I've been developing in my head for a few years now, but as with most of my ideas it is spiralling out of control. It has always annoyed me in games like GTA how the civilians just pop up and fade out as you move around the city, having no impact on the game. If you kill a bunch of them they were just fleeting personas, you get a slap on the wrist (even if that means getting killed and waking up in the hospital or a bit of a fine for mass murder if you surrender to the cops) and their lives meant nothing.

I would like to develop a world where everything is procedurally generated starting from the shape of the landscape to small villages (e.g. on rivers), growing into cities, kingdoms, empires, etc. All of the characters would be born with characteristics (intelligence, charisma, drive, morality, etc.) and they would interact with each other. Instead of fading in and out as you walk around, they would have schedules for going to work, shopping, etc. Relationships between characters would be built up based on their likelyhood of contact with each other and based on their characteristics they would follow different paths and be guided by each other. This could be like a student likes a teacher who is inspiring, so their career is affected in the future.

This would provide an engine for generating realistic worlds. Quests could be given out based on realistic parameters from "real" people in the world. This could generate a range of games:

• GTA - gain reputation starting from simple muggings, getting hired for a bank heist, gaining popularity with a mob boss. Mob boss has rivals with targetted quests like intercepting a rival's shipping container, jail breaks, etc.
• LA Noire - investigate crimes in the city (done by career criminals, random psycho shootings, talking down suicide attempts with real back-story) by questioning witnesses who would have seen things based on their actual movement patterns and knowledge of what looks suspicious in their neighbourhood.
• War games - an evolving battlefield (instead of static maps) with espionage missions, battlefields that change location over time as fronts move, real reasons for countries to form alliances.
• RPGs with proper quests based on what's happening in the world, with a dynamic storyline - if you take longer to get to a location it will be different. Maybe you spent too long leveling up in a forest and you miss the attack on the village.

I'm sure there are many more, but as you can see it would be a very complex system, requiring a lot of up-front processing and maybe a strong server farm to continue to evolve the world. To kick this idea up a notch, the final evolution of this engine would be like a time-cop simulation. You go back in time to prevent crime-travellers from changing the past, but there are consequences when you come back to the present. You happen to have let a person die who did not originally; this was the grandfather of a cop who stopped a suicide attempt. The suicidee originally went on to inspire an architect with the courage to pursue his goals and become famous for his buildings that feature heavily in columns. When you get back to the future, these buildings now have arches instead of columns. Or a sliders style game where each parallel world has a major historic outcome flipped and regenerates from that point (e.g. Hitler won the war (sorry for the godwin)).

Did I mention the idea was spiralling out of control?

So because this idea is super complex and not likely to ever be completed, I thought I'd start it more simply with a procedurally generated RPG world with quests that have multiple choice alignment outcomes and an evolving storyline based on those outcomes. Like I create a bunch of scenario "cards" and a few of them are chosen to generate the world. Based on your actions, different NPCs will offer you opportunities (like the original paladin vs necromancer scenario). I was trying to get some ideas for "evil" outcomes of these cards. An example might be a starting card - your farm is attacked by bandits and burnt to the ground: you could hunt down the bandits and turn them in to the authorities, join them (didn't really care that much about your family, now you want adventure), or slay them, hunt down their families and raize their villages in return.