I like to run more sandboxy games. I like to make the world interesting and distinctive, with secrets and lore to reward players for learning about the world. The world should be consistent, and the unique elements should fit in logically and inform the players about the world.

For instance, my favorite homebrew setting is a theme park planet orbiting a spacetime anomaly. Learning about the anomaly gives hints about various major characters and the technology of the world. There is, for example, a “mineral” that is ejected from the anomaly and occasionally lands on the planet. That “mineral” is basically raw interdimensional matter, which is used to make “magical” technology. This is the standard unobtanium trope, which makes it easier for players to ask the right questions to learn about it.

And I think that’s the balance: unique enough to be interesting to learn about, familiar enough for players to get a good footing and ask the right questions, and consistent enough that you can intuitively improvise when players are writing their story together.

There is only a problem when your world is not consistent within itself. That means, if the players learn a fact about your world, they must be able to use this fact to predict outcomes of their actions. Otherwise, they might as well be monkeys with a typewriter, i.e. their actions could just as well be random.

But other than that, you are also completely correct that if you have too many things that are different from any established experiences the players could draw from, play also grinds to a halt. Like, let’s say you build a setting that doesn’t run on any physics we know of, there are no 3 directional dimensions, there is no time, etc etc - I don’t know how that would look, but you would basically be asking your players to learn a completely new way to live and make decisions in this space, which is possible, but the simplest thing might require lots of thinking and problem solving.

With that in mind, it’s important to keep most of the inner workings the same as either our current established reality, or some well-known fantasy one.

So long story short, I agree with you, it is important just to have a few things that work differently than “expected”.

However, “expected” can be very different depending on which players you find. So, as always, I think the most important thing is just to have a proper “session zero” or player selection process, where it is discussed how the setting looks and how far away from established norms it is. If everyone agrees to take it very far away from those norms, then you can be more “original”, if not, then you either can’t or need to find different players (and if you can’t find any players, then you probably went too far :p).

I tend to value mechanical fidelity more than originality. I like when my settings have causal relations and explanations, so that I can quickly cascade out the consequences of my players’ actions, making the world reactive to the game.

In service of this I will use any and all tropes necessary!

I play PbtA games, and those specifically tell you not to do too much work beforehand and “play to find out”. As you say, I start with a few core ideas and touchstones. Then, when the players ask a question about the world, I often turn it around on them and say “you tell me. What do you think?” Sometimes I’ll make them roll, sometimes not, but I’ll always incorporate at least some and often all of what they give me. This gives the players more buy-in, I’m not just reading my bullshit to them; and also gives me a lot more jumping off points for later stuff. Way less work for me, players feel a sense of authorship, win-win imo.

Uh well where do I fall on the scale when I have my own setting wiki? >_>

Worldbuilding is actually one of the elements I don’t particularly enjoy, I’m here to hang out with my friends and tell cool stories. The less effort I have to put in the better from my standpoint. My preference tends towards games with their own settings or settings easily explained with a few touchstones.

My favourite game so far has been in the Dresden files RPG, which is very easy to grasp the setting of even for someone who has never read the books. I set the game in our hometown and scattered a few options for problems around for the players to encounter and just built a story from there. My worldbuilding setup was collaborative presenting aspects of the Dresden universe which I was interested in running and asking what my players were interested in.

I have also run a few Mothership games with great success and very little worldbuilding - you’re basically in the Alien universe where “The Company” owns everything. If you have a common touchstone like the Alien films to give everyone a shared idea of the world it makes it so much easier to set the tone and mood of a game with very little effort on my or the players parts.

@juergen_hubert I think the real balance is between how complex the world is, and how much is exposed to the players at any single moment. Then the players can decide what matters to them. The downside is that most of the work of the DM might be lost if the players don’t explicitly go looking for it.