Quote:

Original post by Kest I'm not sure how RPG-like your game really is, but you should try to avoid making your player talk to worthless characters to determine if they are indeed worthless. There's nothing more annoying to me than having to talk to 30 people in a town to find the two that actually talk back. Considering your game is an RPG / simulation, I'm not sure you'll have this situation. Fallout handled this by making clone characters look the same. In different areas, clones may change. And in one town, an important NPC may look like the clones in another town. But most usually, the clones in one area are easy to avoid after you figure out what's going on. In Morrowind, all NPCs were pretty much clones. The exceptions might be guild leaders and shop keepers. You could do pretty much anything to set them apart. - Make important characters stand still, while clones wander (difficult to talk to someone who's moving around). - Make clone characters look slightly less detailed. - Give all important characters unique mesh/texture/sprite resources. - Have name pop-ups on mouseovers or player distance, and give the clones generic names (lady, gentleman, kid). It could even be something discreet, like clones never turning their neck + head to look around. Just anything that gives the player a clue. Again, sorry if irrelevant. I can't really help much with the question. Never counted.

That's good advice. I remember pplaying Star Wars: KOTOR. The unimportant characters (the clones) would just say something if you tried to talk to them. They would not initiate dialogue, or let you choose an answer back, etc. It made the world still seem life-like, but I didn't have to talk to every single person to find the one I was looking for.

Thanks for the info. It seems my estimate was reasonable. I'll still probably start with just a small village, as the game concept is already too close to impossible to implement as it is, but I'll write some appendicies into the design document for how to expand it to an entire world too.

Quote:

Original post by Kest I'm not sure how RPG-like your game really is, but you should try to avoid making your player talk to worthless characters to determine if they are indeed worthless. There's nothing more annoying to me than having to talk to 30 people in a town to find the two that actually talk back. Considering your game is an RPG / simulation, I'm not sure you'll have this situation.

While I agree with this in a standard RPG setting, this won't be a problem in my game. The gameplay mechanics are more strategy simulation than RPG; I'm thinking it will play quite a lot like Tropico or RollerCoaster Tycoon, except you are building an RPG village (or world). So the unimportant characters will be easily spotted by their stats when you highlight them with the mouse. I'll probably make them have a slightly different appearance from the important characters as well.

Quote:

Original post by Trapper Zoid This is a question that's I've been pondering in the back of my mind for a while, since aspects of my game design depend on it. What would you say is the typical population of an RPG world? I'm only considering the NPCs and PCs here, What do you think? Is my estimate of a population of 500 people accurate?

Few millions. Games however abstracts majority of these to alow easier interaction.

Quote:

I think I could design the AI to model up to a few hundred characters given from what I've seen in other games, possibly more if I use a few CPU saving tricks, but I was wondering if a few hundred were "enough" to correctly model a world. I'm thinking yes, but I was wondering what everyone else was thinking.

You could build up to one million, provided you'd have enough memory, you'd need however something I once invented, and then forgot...
10 GB could be enough to model all this.

Quote:

Original post by Raghar Few millions. Games however abstracts majority of these to alow easier interaction.

I've always thought of it as a couple of hundred "normal" NPCs, and a few million brigands, guards and various ne'er-do-wells who you end up slaughtering on your way to saving the world.[smile]

I've thought about this a bit more, and there's some good design reasons to keep it small, even if the technical challenges weren't present. I'll probably stick to a few dozen as the maximum population limit, and abstract the rest of the world away. Thanks for the help!

One simple you can do to help with the abstraction is to have it so the NPC's alternate locations. Or have some appear only at certain times, perhaps a few that travel from town to town.

I'm working on a similar thing, but rather than it being agent based and having each agent having needs/etc, it's RTS based AI, then another layer on top of that which manages the world ala Civ (There's a limited number of build points per world-square, so it's easier to manage than if you could build wherever you like). Then another layer on top of that which is the story generation (based on Joseph Campbell's 'hero with a thousand faces' monomyth) and hero units will propel the storyline (and may step in if the simulation gets in the way of what it wants to tell).

And each culture will evolve according to a tech tree, where if you decide to go down the secular-tech route, magic things will be cut off, and if you decide to go down the shamanist-tribal route, tech things will be cut off.. and everywhere in between.

The series of books by Joseph Campbell (Again), The Masks of God (the first book is about 500 pages, the last about 700, it's about 2500 in total.. not a brief read), is what I'm basing the myth/culture system on. The books cover about 10,000 years of myth and culture and it's pretty comprehensive about the evolution of myth and society around myth.

And I'm even going as far as, if you go down the myth/religion/magic route, the gods that grant you your power will even change. So you may start off with a shamanist culture where you have a trickster god, like Loki or the Titans, then if your culture becomes a 'planter' culture, the gods of order will appear, like Zeus, and there will be strife between the two with the eventual banishment of the trickster god to a prison (Hel/Hell/etc)... This could also spark off new story generation, with you as a hero making a difference to which side wins.

And I've completely ignored the original question with how big a world should be. In my case, you might have a max of a few hundred units per square, not being able to send any more squads/legions/etc into a certain square until there's less units there. (I realise this kinda kicks the realism, but it adds another dimension to the strategy, on the world level.)

I apologise if this post has been completely unhelpful, but I do recommend taking a look at the Masks of God series of books, by Joseph Campbell, they're very good.

Quote:

Original post by FroZtByte I'm working on a similar thing, but rather than it being agent based and having each agent having needs/etc, it's RTS based AI, then another layer on top of that which manages the world ala Civ (There's a limited number of build points per world-square, so it's easier to manage than if you could build wherever you like). Then another layer on top of that which is the story generation (based on Joseph Campbell's 'hero with a thousand faces' monomyth) and hero units will propel the storyline (and may step in if the simulation gets in the way of what it wants to tell).

You're right; I'm thinking of having a simple agent based model, sort of like a simplified and adapted version of the Sims, to run the characters. Then I'll be adding a drama manager layer on the top of that. I haven't yet figured out how many layers I need (the drama manager might be multilayered), since I'm going to figure out the details of the domain first. The drama will be based on my own symbol system, however, loosely based on Propp's Morphology of the Folktale but adapted to a different domain.

Quote:

And each culture will evolve according to a tech tree, where if you decide to go down the secular-tech route, magic things will be cut off, and if you decide to go down the shamanist-tribal route, tech things will be cut off.. and everywhere in between.

Whoa! That's a little bit much, isn't it? I'd have thought getting a Campbellian based story system working would be trouble enough! [grin]

Quote:

The series of books by Joseph Campbell (Again), The Masks of God (the first book is about 500 pages, the last about 700, it's about 2500 in total.. not a brief read), is what I'm basing the myth/culture system on. The books cover about 10,000 years of myth and culture and it's pretty comprehensive about the evolution of myth and society around myth.

I have read a couple of Campbell's books, although they are pretty heavy going. Instead I've studied Vogler's adaptation of Campbell to film a lot closer, since that's easier to understand and the film adaptation is, in my opinion, a lot easier to adapt to games.

Quote:

And I'm even going as far as, if you go down the myth/religion/magic route, the gods that grant you your power will even change. So you may start off with a shamanist culture where you have a trickster god, like Loki or the Titans, then if your culture becomes a 'planter' culture, the gods of order will appear, like Zeus, and there will be strife between the two with the eventual banishment of the trickster god to a prison (Hel/Hell/etc)... This could also spark off new story generation, with you as a hero making a difference to which side wins.

This is the sort of thing that's the whole reason for me trying this sim, although my stories will be more RPG based and less mythology. But it's hard going to get the whole thing to work, so I'm not expecting to get it finished anytime soon.

Quote:

And I've completely ignored the original question with how big a world should be. In my case, you might have a max of a few hundred units per square, not being able to send any more squads/legions/etc into a certain square until there's less units there. (I realise this kinda kicks the realism, but it adds another dimension to the strategy, on the world level.)

I suspect that the individual characters would have to be very abstract if you have a few hundred units per square. I am wanting a more Sims approach, with each character having a little (if simple) personality of their own, so I'll be sticking to a couple of dozen for now.

If you are interested, I had a thread a where I listed a number of interactive storytelling resources, which I like to link to whenever someone else seems to be working in the same area. Since you're already deep into Campbell you've probably already looked at most of these, but it could be some use to you.

Best of luck! It's always good to know there are other people as mad as me attempting something on this scale.

Quote:

Original post by Trapper Zoid This is a question that's I've been pondering in the back of my mind for a while, since aspects of my game design depend on it. What would you say is the typical population of an RPG world?

I think this is a function of the kind of game world you design, and is not a top down answered question, but a bottom up evaluation. Though there are many kinds of stock characters that traditionally populate a given genre, most of the time, when writing the game, whom your character interacts with either as a primary or stock character can be functions of expositional choices dramatically.

The more clearly your game world is defined, and the more structured your player information necessity exposition line is detailed (what they need to know, where they find it and when, essentially) will be the answer to these questions.

Adventuredesign

Don't forget, you don't really need to do the full processing for NPCs that no players can actually see, as long as you make sure they're all updated with the correct information, are in the correct location, are doing what they should be doing based on global events, etc. when a player does move into the area, at which time you can start fully simulating thier activities.

So, I think if you set it up well it could be technically feasible to have quite a large number of NPCs in the entire world population, which then leaves it more up to world design.

I'd say your ~500 estimate is probably an excellent goal to work with for experimentation, you can always adjust it later if it doesn't work out well for some reason.

Quote:

Original post by Kazgoroth Don't forget, you don't really need to do the full processing for NPCs that no players can actually see, as long as you make sure they're all updated with the correct information, are in the correct location, are doing what they should be doing based on global events, etc. when a player does move into the area, at which time you can start fully simulating thier activities.

That's true, and it's something I've considered (albeit not deeply at this stage). Theoretically there could be a massive number of NPCs, if you can chunk them up into small enough groups (such as a village) and then update each of them when the player is in the vicinity with a log of the world events.

But I've been thinking a bit about the nature of the game, and I'm not sure whether more is better. If the player is expected to care about nearly every NPC, then hundreds will be too many to cope with. I think I'll just mark this down as "theoretically possible" and wait until I've finalised exactly what the player will be doing in the world to see which is better. This design is still in the early stages, I'm afraid...