How much should I prepare a campaign setting beforehand?

[DiscipleofBob]

For 4E, I plan on writing a fresh homebrewed setting. I don't want to plan so much ahead that I start railroading every bit of plot, I'd rather set the plot points around the PC's to get them more involved in the world, and I want to generally let the players make most any concept they'd want (within reason.)

I've thought of a pantheon (very much different from the official pantheon), a general arrangement of countries for the mainland and some basic concepts for the countries and other continents.

Exactly how far is a good amount to plan ahead without venturing into railroading or creating a bunch of NPC's the players won't care about?

[Gralamin]

It all depends on how your PCs play. If you have a player thats likely to put massive ranks into all knowledges, you probably want a good idea of your setting. If your players rarely even remember which town their in, you probably don't need as much.
You also want to have enough done to have a reserve, just in case the players are quicker then expected.

[Icewalker]

Well, one note about railroading, is that the way to go isn't to force them to do something or to give them totally free rein (that'll lead to them doing the unexpected, and you won't have any interesting responses prepared). Instead you have to guide their actions from above, making it so they are making the choices, but they do exactly what you want them to. Very challenging, I'm not great at it, but a DM of mine actually managed to run a flashback (which everybody else had already seen through other characters) with a person playing a character in it, who he got to do exactly everything the same.

Back on topic:
You can never really overprepare, if you have time. Because, even if you set up things and they never come into play, you can continue adventures in the same environment and they still will. The idea of creating a whole system, then standing back and having the NPCs perform due to their natures works better than making a path for the PCs to follow, and adding in the edges when they drift to the sides.

[Tempest Fennac]

I'd say that sorting out individual cultures and economies for each of the countries; as well as the sorts of creatures who live there) would be a good idea (different economies would lead to certain items being more common or practically impossible to get hold of, and the sorts of creatures who live there could have an impact as well. For instance, in a (largely) CN country where Gnolls and Drows are the dominant specises, males would be largely seen as inferiour to females, which could cause problems in an all male party, which would lead to the players needing to work around the problem in social situations (having countries typically being lawful, neutral or chaotic in general would probably be a better idea then having them as good, neutral or evil for the most part due to how it would probably be more realistic when you factor in what most of the creatures which inhabit the countries would be like, except for some really rare cases).

[Behold_the_Void]

I'd say that sorting out individual cultures and economies for each of the countries; as well as the sorts of creatures who live there) would be a good idea (different economies would lead to certain items being more common or practically impossible to get hold of, and the sorts of creatures who live there could have an impact as well. For instance, in a (largely) CN country where Gnolls and Drows are the dominant specises, males would be largely seen as inferiour to females, which could cause problems in an all male party, which would lead to the players needing to work around the problem in social situations (having countries typically being lawful, neutral or chaotic in general would probably be a better idea then having them as good, neutral or evil for the most part due to how it would probably be more realistic when you factor in what most of the creatures which inhabit the countries would be like, except for some really rare cases).

Honestly, I would argue you don't have to go quite that far (although if you have the time to do it, so much the better!)

When I set up a campaign setting, I tend to focus on one major country, or even a major city in the country. You can build out as needed, so long as you have the area immediately accessible by your PCs prepared. I tend to plan specifically adventure-by-adventure, and even then I tend to just get some stats on paper for things for them to kill. Of course, I'm good at improv, so your mileage may vary. I find my players are generally fairly compliant to stay on the rails, they know they can talk to me out-of-game about what it is they want to do, and in-game they know if they go with what I've got planned, it usually turns out better.

Of course, one group I play with tends to not want me to prepare as much, because when I have time to set down and plot out stats, I tend to overestimate their combat proficiency and beat them down. They aren't as proficient with the game system though, so I should learn to tone down my encounters a bit.

[black dragoon]

I tend to start out at a very vague starting point(look at my thread for an example) and steadily fill in general details and narrow it down till it's just fleshing out the entire thing by adding in locations,NPCs and oddities for the PCs to encounter.

[CTG]

It actually depends more on you than the players. If you can improvise quickly enough, you don't need more than the major parts created when you start, you can quickly mke the rest up as you go along, just be sure to keep notes so when you say one thing, that is the way it is.

This can go either really good or really poorly.

However if you wish to give the players a very realistic view of the world, then it will take time. I too am currently re designing my world (albeit 3.5, not 4), and I am to redesign design maps, kingdons, a lot of NPCs, political landscapes, economies, climates, towns, ect.

I did it before when 3.0 first came out, and I put over 100 hours into it, and I knew almost everything about the world that could ever be desired.

However this is for a serious game, if it is Sunday afternoon drinking (root) beer and eating pizza, then no need to take it so seriously unless you want to.

[bosssmiley]

Good general rule I've used for a while: only ever sketch things in until the PCs get there.

Be prepared to erase and re-draw unexplored parts of the gameworld as the players express an interest in things. Throw lots of background fluff and contradictory rumours about and see which ones pique your players' interest, then flesh those out. Sure, you might get odd, nonsensical and downright kludky elements that come with organic campaign growth (I'm looking at you OD&D Mystara), but that's all part of the fun.

Remember to steal from the best, shamelessly ripping off source material wherever you find it. No-one cares if it's copypasta'ed if they're having a whale of a time.

Maximum fun, minimum effort.

[Pronounceable]

Depends on what exactly you are doing. Are you making a world for the next campaign, or will you be running many campaigns there for the years to come?

If former, not much initial work is needed, beyond the overarching plot. If latter, you can't really prepare enough. There's good advice above for either of those.