How can I prepare unique locations for my sandbox, when I don't know which unique locations my group will travel to?

Question

I'm making a sandbox campaign using 5e to run a post apocalyptic fantasy Tokyo adventure, so each main location is known to the players and I can't just put in random encounters. Those locations need to have a strong connection to what is there today. The players will explore the world to pursue their goals while a mysterious villian observes from the darkness.

I want to use player background as a motivation and incentive for them to explore, and sometimes run into places they should have left alone. How do I use their backgrounds to help them decide which locations to explore so that I know which locations to prepare?

The sandbox is supposed to take them from lower heroic tier to lower epic, so I want to give plenty of space and time for them to develop and find clues, so I can't put all their motivation into one single location, and I don't want to transform every landmark into a generic encounter.

Here's an example of my map http://community.wizards.com/sites/mtgcommunity/files/styles/large/public/mymap_1.png?itok=wvI9mEOz

I have all the marked locations named, but I don't know if I should map the dungeons yet because I don't know how often should I include something to aid the players find their goal and how often I should make them a deathly nest of aberrations and treasure.

You can see I have many locations, and it's overwhelming to prepare them all. Where should I start?

Answer 1

I'm currently running a mostly-sandbox game in the post-Apocalyptic midwestern United States, although I don't intend to restrict the players should they decide to up and drive to Los Angeles or something. So, I've had to address this issue very recently. Here's how I did it.

Plan for where they are

The players were new to the world, which is a bit of a pastiche of elements from Fallout and (eventually) other settings. They grew up in a Vault underground, and were suddenly thrust out into the world when the Vault opened for the first time. This gave them purpose (the Overseer sent them on a quest), as well as a convenient excuse to not know anything about the world. Said quest was open-ended enough that they could make exploration a priority.

In doing this, since they're very low level and have limited means of transportation/food supplies, they can't just run off into the Wyoming desert if they hope to have a life expectancy beyond ten minutes. Thus, you know where they're going to be for the immediate future and can offer some guided adventure hooks in that area. When they've handled those, you hopefully will have seeded them with additional ideas of where to go, through learning about local geography (an old gas station map told them about the major cities to the north, south, east, and west), and talking to NPCs (a local native told them that her tribe planned to head east to the settlement near Cheyenne).

Plan for where they're going

The trick here is to do this at the end of the session, not the beginning. IE, you ask them what they want to do next session and then you have all week (or however long between games) to prepare. In the beginning, since they don't know much about the opportunities in the region, they'll mostly be looking to find out more and also find their options restricted. This will become more important as they start to network more and have more places to go.

Branch out

Are the players hanging out in a particular region and show no signs of leaving? Flesh out that region. Are they planning to go somewhere else? Ask them at the end of the session and then flesh out that area and the areas surrounding it such that if they decide where they've gone is boring, they probably won't have to go far to find something interesting to do.

The most important part is to be one step ahead of your players, and the best way to do that is to plan things directly in response to what they want to do.

Answer 2

There is no way to prepare for only and exactly what the players will explore in a sandbox. Two major ways of solving this have emerged over the decades:

Prepare Everything

The first and most obvious method is to solve the problem with brute force: prepare everything in advance. This is extremely time-consuming, and is a project that few GMs ever actually complete despite their good intentions. Those that do manage it, do so because they focus on a single world for many real-world years and consider world building to be a worthwhile hobby by itself.

Build the world Just In Time (JIT)

JIT is a relatively recent computer-science term, but it describes the most common successful sandbox-building strategy: build only what you need when you need it.

In practice, this means you start with a small slice if the world and write only very brief notes (as short as one sentence or even one word) for each feature, and then expand your notes only when needed. This expansion is either done between sessions just like in the "prepare everything" method (except limited to just this one element), or is done using improvisation and (semi-)random generation tools and techniques; or more often, a mix of both.

For example, a map might have a mountain on it and the single word "Dragons" beside it. You don't need to know more until your players get within sight, rumour, or near travel distance of the mountain. You certainly don't need to know the name of each dragon and 1000 years of the mountain's history when the players are still apprentice adventurers 100 miles away! You may not even need to know what that mountain and word mean when they're 20th level if they never went in that direction and now they're exploring the political situation on the Elemental Plane of Earth.

When the players start looking like they're going to be nearby that mountain though, that's when you prepare it and flesh it out. "We want to explore Dragon's Mountain next session!" means you know what you should be preparing this week! Even that much warning isn't needed though, with tools and improv. Any idea you could come up with during between-session prep is an idea you could come up with during the game. If you have stat blocks and creativity, you can make up an encounter on the fly. If you have random tables, stat blocks, and creativity, you can improvise an entire complex of mountain valleys, factions, leaders, monsters, names, history, and treasure during the game.

This method of Just In Time game preparation is very efficient, and leaves you a lot of time to search out tables and tools to help you flesh out your notes during play instead of spending time detailing every candle maker and blacksmith in a town they might never visit.

Of note is that the most significant preparation is always before the first game of the campaign, because you need to detail the starting area (often a town) more than with just a single word on the map. But with people-generation tools, a town can be improvised to life from very thin notes just like the larger world can be. Prepare the start location and a handful (two or three) locations of exploration/adventure, and let your players show you where to add detail to the world.