Exploration of Procedural RPG worlds [was Re: [pygame] Buried In Game Ideas]

["Dave LeCompte (really)" <dave@xxxxxxxxxxxx>]

"Kris Schnee" <kschnee@xxxxxxxxxx> wrote:

> The appeal of the
> simple RPG would be that it's procedural rather than truly random, so
> that you can say to other players, "Check out the island at (42,0); it's
> got neat stuff," and it'll be the same for them as for you. People would
> end up mapping the game world together.

Well, "procedural" and "random" aren't necessarily opposites. A randomly
generated dungeon IS one kind of procedurally generated dungeon. And, if
you use a random generator that produces reproducible results (using a
known seed, and knowing that the underlying implementation is the same),
you get results that can be shared between players.


There was a recent interview with Ubisoft's Clint Hocking at gamasutra.com
where he discusses exploration in RPGs (think "Oblivion"):
http://www.gamasutra.com/features/20070514/ruberg_01.shtml

In particular, he says:

"I can only assume if your open-world game doesn't do well, you didn't
support exploration as well as you should of. If you don't design systems
that support the exploration, you have a giant empty game. Then it?s you
walking around in space for a little while until you get bored and then
you give up."

The discussion of Daggerfall's dungeons being boring and repetitive sounds
like they failed to properly provide incentives for reaching the deeper
parts of the world, and they didn't provide enough variation in their
dungeon generation procedure.

Thinking about Laura's suggestion about commerce - that's one system that
supports the exploration. It's easy to think of a simple economic sim that
implements supply and demand, and the more exotic areas of the world
generating scarce (and therefore expensive) goods.

I recall one of the early massively multiplayer games trying a similar
supply/demand model for their magic systems - if you were using a rare
spell, you would get a greater effect from it. As people started using the
spell more frequently, it would decrease in power. That's something you
can do if you have a shared server to track spell use, but even with a
disconnected game, you could use the magic system as an incentive for
exploration - the most powerful spells require the most exotic reagents,
and the most exotic reagents are, of course, the most difficult to get to.

If you wanted to further borrow from online RPGs and implement a crafting
system, you could have more incentives to explore, with powerful items
requiring scarce resources.


I'm currently in the design stage of my own RPG with large world
exploration, and one of the carrots that I'm planning to use to encourage
the players to explore further is to have certain skills be taught in
certain places - if you want to learn to be an assassin, you have to find
someone to teach you those skills.


One thing to think about when considering procedural generation is that
"Infinite Game Worlds" are not infinitely satisfying. If you haven't
explored your own world pretty comprehensively, you may be surprised to
find the players exploring the island at (42, 0) that has neat stuff, and
then going off to the island at (44, 1), which has a large supply of
pointed sticks, which they use to attack the natives at (47, 4), which are
meant to be really tough, but have an unfortunate weakness against pointed
sticks.

The point of that example is that a simple procedural system used on a
large area will almost definitely have unexpected combinations of
features, which may be exploitable. And, given the existence of the
internet, your players will communicate with eachother and publish
exploitable trade routes or geographically convenient combinations of
features. Which means a certain amount of planning and a certain amount of
time exploring the world after generation time, which means a finite scope
to your game.



Just some thoughts - I'm certainly not discouraging procedurally generated
content; I think it's a crucial tool to deal with the trends in content
size we've been seeing over the past several years.


-Dave LeCompte