How to Make a Game More Sanboxy?

[Reaver25]

This sounds like a weird question.

I've got an idea as to how I'd like my campaign to go as far as story arcs and campaign stuff. What I haven't figured out is how to make my game more of a sandbox open style of play. I don't like the idea of railroading the campaign. Any ideas?

[Maxilian]

Look at the map of the campaign, look at the different places and ask yourself "What is there?" and add there whatever you imagine, also make the backstory of the PC relevant somehow, maybe the Farmer boy run into a Merchant that their parents use to sell the different products that they produce in the farm, etc...

[JellyPooga]

Set it in the desert? :smalltongue:

[Sigreid]

Throw out lots of plot hooks that you believe will appeal to different characters and describe events that don't directly involve them in ways that might tempt them to want to figure out how to get involved.

[Specter]

Create a detailed backstory for places. A mine stops being a random place and becomes an interesting place on its own when PCs figure out the souls of the king's worst slaves populate it because they were all massacred, and that there is evidence of a big magic weapon inside it.

[mephnick]

It's kind of a give and take between sandbox and story with a massive sliding scale between them. At one end there's heavy railroading (which some players like, it's not inherently bad), and at the other end is directionless sandbox. To make it more "sandboxy", you have to give up some control of your story. You say you know how the arcs and story will go, but in a sandbox-type campaign you can't guarantee that very well. In a sandbox it's more like you create a setting with a bunch of stories and then you see which ones your players gravitate towards and flesh them out as you go.

Now, you can heavily push certain story arcs with multiple hooks, and good players will likely go with it, but be flexible. If the players don't seem very excited about the "evil wizard in the dark tower" hook you've been pushing and really like the "demon monster in the dark woods" hook they randomly saw on a poster in the town square, just file the wizard away for later and flesh out the monster. Then, have your arcs progress regardless of whether the PCs interacted with it or not. Demon monster took 2 weeks to sort out? Well the evil wizard has since enacted his plan to enslave the riverside community near his tower. PCs hear about this and hey, they're probably back on the arc you want.

Every DM's style is a push and pull between the two opposites. Dip into each end in varying degrees to figure out what balance point suits you best.

[KorvinStarmast]

Random encounters.
Not all are combat.
Each encounter's resolution becomes an anchor in the world's reality.

We did a campaign like that for a year in 1e. It was a weird and neat way to do world building, and it helped that our DM was very good at improve/off the cuff decision making, and we had great fun.

Knowing when to be nice, and when to back away, and when to run like hell was part of the fun.

[mephnick]

Random encounters.
Not all are combat.
Each encounter's resolution becomes an anchor in the world's reality.

That's a good point I missed. I rolled a random encounter that just said "cultist + revenant". Because of my PCs actions and some improv that two word encounter turned into a 4 month long side quest about a secret cult targeting clerics with revenants who'd had their memories altered, like undead guided missiles. Now that's a thing in the world.

Random encounter tables get a bad rap sometimes, but if you use them to flesh out your setting and inspire you in the moment instead of just as a travel tax they can be extremely useful.

[Nupo]

Don't prep too much ahead of time. If you do, you will be reluctant to let the players deviate from it because it will mean all that hard work will be wasted. Prep one or two sessions ahead of time at most, and let it go where ever it wants to go. My best long running story arcs were never planed, they just happened. I see something happing and run with it. If the players go a different direction, I run with that. If you let it, a simple roll on a random encounter can turn into years of fun gaming.

[Naanomi]

Yeah being able to roll with the story and make it up as you go is really the best way to be 'sandboxy'. Have ideas... I knew I want to do a flooded dungeon; so when the party ends up hunting down a smuggler who cheated them, guess where he hides? But don't be married to anything

Also, it can be challenging to walk the line between 'real world' and 'death trap world'. In a sandboxy video game, you walk into the wrong cave and meet the dragon lord; no worries... reload; come back later. Neither reloading nor coming back later (well... and expecting the dragon to still be sitting there) work in tabletop; so you either need to have parties great at scouting and planning (and running); or still be prepared to shape the world with (mostly) level appropriate stuff as they wander around (unless they really screw up)

[Lonely Tylenol]

Look at the map of the campaign, look at the different places and ask yourself "What is there?" and add there whatever you imagine, also make the backstory of the PC relevant somehow, maybe the Farmer boy run into a Merchant that their parents use to sell the different products that they produce in the farm, etc...

This is my answer, but I want to add a little more detail to it:

Suppose you have four towns, named Northampton, Easton, Southborough, and Westchester. (We are not very creative with our names.) All four towns have problems.

Northampton is a mining town, but recently the mines were shuttered because the miners started disappearing once they dug deep enough, and now are refusing to work. In Easton, the person running the place has become increasingly paranoid, and is issuing out draconian sentences (like torture, imprisonment and death) for even minor crimes and dissent in Easton. Southborough is being raided by gnolls encroaching from over the border. Westchester has been plagued by rumors of a long-forgotten treasure in the ruins near town, which have brought adventurers from high and low places alike to find it.

These adventure hooks have different priorities that might dictate what the party goes for, even if they are all available at the same time. For example, if I'm going off "greatest good" principles, I might prioritize Southborough first, since it presents an existential threat to the town; that is followed by Easton, since lives are at stake there as well; then Northampton, where there is no immediate threat to life, only economic health; and finally Westchester, where the goal is mostly personal benefit. If I favor law and order above all, I may prioritize Southborough/Westchester, then Northampton, then Easton. If I'm motivated by personal gain, I'd go to Westchester and sniff out the treasure. And so on.

You may also introduce tiered adventures which allow for sustained interest in the continuity and long-term health of a sandbox game. For example, I've decided to go to Easton first, for whatever reason. It is there, after some investigating, scuffles with the local authorities, and a confrontation with the corrupt head of state themselves, that I discover that Easton's mayor has been corrupted by a hag coven. Specifically, a Night Hag has been possessing him in his dreams, and her Green Hag sisters have been harvesting souls from those he's sentenced to death! The party is compelled to find these hags and bring them to justice... But, they are only level 2, and the Hags are Challenge 5 (Green) and 7 (Night) in the coven. This introduces gates which forces the party to diversify their adventure: sure, the party could rush headlong into the coven and challenge the three, but they would be way too low level to have a meaningful chance to win. (In fact, maybe they try right away and fail, and the coven toys with them, humiliating them so the jags can revel in their defeat; this might lead to an escape adventure, followed by a second confrontation with the hags later on, now that the party has their own pride at stake, and the hags have been played up as the villains.) Gates need not always be level gates: for example, if the party started fighting the gnolls at Southborough, perhaps they realize that they are able to fight them back from the town's walls alone, but don't have the money and human resources necessary to wage an extended campaign against the gnolls.

You can also introduce time-sensitive elements into the campaign. Some or all of these things might be things that the party can take care of whenever, but some might get aggravated by inaction, or new things might just happen. For example, the party went from Easton to Westchester in order to find the rumored treasure and try to use it to defeat the hag coven, and they took their time in doing so (fighting monsters and helping other small-town folk with their lesser troubles along the way). Only the world didn't stop around them: while they were faffing about, Southborough fell to the gnolls, emboldened by a Fang of Yeenoghu sprouting up among their ranks, and now there is a full-on refugee crisis happening, with the survivors of the invasion fleeing to the other three towns. Northampton is short on food, since they haven't been able to trade their wares mining, and can't keep up with the influx of new people; Easton is in a power vacuum since their mayor fell, and they don't have a strong voice to direct the refugees; and Westchester is as chaotic as ever, since it wasn't built to accommodate all the adventurers that came in the first place, and this just aggravates things. Meanwhile, the gnolls, still unimpeded, are spilling northward, representing a greater threat with each day they remain unchecked. This guarantees that, somehow, at some point, this specific plot element needs to be dealt with by the party, even if the others don't; if they don't deal with it, it will continue to get stronger, and provide more of a threat to the people (which allows it to upscale with the party's rising level and abilities, as well).

This way, you can have an open world that has a certain feeling of verisimilitude that still has pressing plot elements and hooks, so the party always has something to do, if they don't pursue something entirely different on their own.

[Knaight]

This sounds like a weird question.

I've got an idea as to how I'd like my campaign to go as far as story arcs and campaign stuff. What I haven't figured out is how to make my game more of a sandbox open style of play. I don't like the idea of railroading the campaign. Any ideas?

That bolded part outright conflicts with a sandbox style. To get a bonafide sandbox it has to go entirely, to get something more sandboxy a mix could work - if the time between the arcs is lengthened you could theoretically have a sandbox style there, although if that doesn't then change the arcs after it it gets undercut a bit. You could implement a mission structure, where the group template is that the PCs are all members of an organization doing the work of the organization, but between missions they can go do whatever, where the missions are placed remotely enough to the rest of play that the integrity of the sandbox portion is stronger. Still, these styles don't mix easily.

[Sigreid]

That bolded part outright conflicts with a sandbox style. To get a bonafide sandbox it has to go entirely, to get something more sandboxy a mix could work - if the time between the arcs is lengthened you could theoretically have a sandbox style there, although if that doesn't then change the arcs after it it gets undercut a bit. You could implement a mission structure, where the group template is that the PCs are all members of an organization doing the work of the organization, but between missions they can go do whatever, where the missions are placed remotely enough to the rest of play that the integrity of the sandbox portion is stronger. Still, these styles don't mix easily.

Another option is to have ideas for a couple of possible arcs and develop one or more of them based on the players actions and interests.

[Rotsu]

What I do for things like this, is i spend a few hours thinking of who might want to deal with the problems, and then write them down, in detail. I usually come out with 20+ backstories and ask my players to pick one. Each one has a personal arc within the larger story, a purpose, connections, and the why. The players get choices, and if they don't like any of them, and this is super rare, I ask what they want, and write up a close sketch of what they want. Never exactly what they want, so its a challenge, but close enough that its comfortable and easy. Been doing this for a while (5+ years) and it works well.
Good luck!
Rotsu

[Hrugner]

It sounds like you want a fake sandbox. Generate your world and let the players go wherever they like. Lay a generous amount of plot hooks and make sure they always tie back into your intended story arc. Wherever players go, have them meet the characters you'd intended them to meet during your plot and build these character's out of the types of people available where the party encounters them. Have a few scenes socked away and use them when conditions are close enough for them to work. If the players abandon some activity, have a recurring rival complete the activity in their stead and win major prestige.

And there you go, a pretend sandbox where the players feel free but don't really have much say in how the game plays out.

[McNinja]

This sounds like a weird question.

I've got an idea as to how I'd like my campaign to go as far as story arcs and campaign stuff. What I haven't figured out is how to make my game more of a sandbox open style of play. I don't like the idea of railroading the campaign. Any ideas?Map out the entire area the campaign will be set in. It's easier if it's a smaller area, but I gave myself a ton of work by setting my campaign on an entire continent and letting the players loose (although they still like being given direction). It's half planing what towns and cities are where, half making it up as you go.

Best tip: have the DMG ready and open to the town/village creation guide.

[Sir cryosin]

Make a map of your world or the spot were the party will be traveling in. Put a few markers on it like rumor of Forgotten ruins, abandoned mine of methril. Then when the party is in town. Have them hear of rumors of different things that are happening. Look up Matt colvillle about west marches type of campaign on YouTube he explaines it better.

[GoodbyeSoberDay]

What's that Eisenhower line? "Plans are useless, but planning is indispensable."

Consider that story arc you have in your mind. Sketch out the plans, motivations, means, etc. for the major players (NPCs). Imagine how it would go without the PCs interfering - this is like writing a (potentially depressing) story. Then allow the PCs to turn everything on its head. It's almost guaranteed not to go the way the NPC planned it, but the NPC must have a plan in order for the PCs to change it. It's better if curious PCs can figure out what's going on, as well. One of the major problems of some sandbox games is having no informed direction to take (check the recent Kingmaker thread in the 3.5 forum).

To make it more of a complete sandbox, there should be several of these arcs going on at once, and they should interact with each other in myriad ways. Consider Lonely Tylenol's informative, extensive example, and then add even more complications and interactions based on motivations, plans, and means you sketched out previously. For instance, maybe the Night Hags don't like the Gnoll threat, and are willing to bribe the PCs to make them go away.

[Beelzebubba]

The only way to truly make a 'sandbox' is to have a lot of prepared material beforehand that is geographically situated in the world, have enough material to handle what the characters are after this session + a bit more 'around the edges' in case they surprise you, and the ability to wing it if you get surprised.

It also helps to have a bunch of resources on tap in general, i.e. random encounter tables tuned to the area, stats and maps for general situations they might encounter (such as a small town, oasis, whatever) that you can just throw down.

It also helps to have a library of interesting things at your disposal - i.e. instead of a random encounter being on a straight road in a forest, create an interesting land feature, like a set of exposed sandstone caves with the encounter creatures inside keeping watch, with a dense thicket of woods leading up to a gorge on the backside - so they are forced to make decisions. Depending on the parties' actions, they may sneak by the front with the risk of getting caught, avoid it entirely by the back with the risk of falling, go in guns blazing, etcetera.

I think the 'sandbox' is an older game style that had a lot of support in AD&D and has fallen away in the recent editions. It's a lot of work.

[Doug Lampert]

Don't prep too much ahead of time. If you do, you will be reluctant to let the players deviate from it because it will mean all that hard work will be wasted. Prep one or two sessions ahead of time at most, and let it go where ever it wants to go. My best long running story arcs were never planed, they just happened. I see something happing and run with it. If the players go a different direction, I run with that. If you let it, a simple roll on a random encounter can turn into years of fun gaming.

Have a bunch of possible adventures and hooks, last time I ran sandbox I tried to have at least 5 for every session, typically 3 they might reasonably do (maybe one job offer, one investigation, and one thing related to the "main plot"), at least one other thing that would be a complete pushover, and at least one other thing that would obviously steamroll them.

Information gathering skills get some extra use in determining which is which.

Make sure your characters are aware of at least a couple of the above, and at the end of any arc ask your players what they'll be doing next time.

Make sure your players are aware that not every hook is a good hook for them. That you'll be deliberately including stuff that isn't a great idea. (And then GaryStu the mighty wizard with the odd runes "DMPC" glowing on his forehead dealt with all the problems, again, while his hired guards watched, and got paid one whole silver piece for their trouble and for getting their gear destroyed by side effects of the enemy's attack.)

You want to do detailed prep on what they're doing, not what you think they'll be doing.

[Nupo]

You want to do detailed prep on what they're doing, not what you think they'll be doing.Well said.

I will also add, no matter how much you prepare, you will end up having to "wing it" from time to time. The more you do it the better you will get at it. If done well the players will never know you didn't have all this planned ahead of time. One thing that will help a lot is take notes during play. Minor details you come up with on the spot can be fleshed out later, before the next session.

[CaptainSarathai]

The thing is, do you really want a sandbox, or just to avoid railroading the PCs?

There are two ways that I generally see "sandbox" campaigns turn out:


This is the weakest sandbox. Players venture from town-to-town, following leads and rumors, and tackling threats as they arise. There is no real connecting story beyond "we came, we saw, we conquered"
This is the West Marches campaign, unfortunately. Yes, the overarching "story" is that the adventurers are exploring uncharted wilds and must fill in the map, however, ultimately it was still "go here, fight thing."

You could run such a campaign from a Guild, or other central hub where the players get their rumors.
I've found that these campaigns tend to be boring and short-lived, because you eventually run out of new encounter types, and the players never get invested in a story.


Sometimes, the players will find a story and latch onto it, or it makes sense to have an ongoing story spawn out of the myriad things they have accomplished.
DMs seem to think that a sandbox will include lots of walking from town to town without giving a reason to. In the last "sandbox" DMed to me, our entire party decided to call a single town home, and defend/liberate it from an evil king. We went from a "quest of the week" format, to at least a prolonged adventure. The DM had a choice - either let us defeat the evil king with ease, or design some kind of longer path to victory. He chose the long path, and over several levels we warred against the king and eventually defeated him. But then we decided that we would be kings. Suddenly we were essentially playing the Kingmaker campaign from Pathfinder.

Players will latch onto something - defeating an evil rival or defending a town, collecting certain relics or fulfilling a specific prophecy, and suddenly the game revolves around a central idea and all roads must lead to a conclusion which satisfies the players.

The trick to "not railroading" is very simple - ask the players what they want to do. That is the the first and most important step. I've written my method for campaign design several times, but here are my tips:

1. Session Zero-One
Intro the setting, and then put the players right into the adventure before anyone rolls a character.
"The 5 of you are standing at the mouth of a large cavern - a single set of rickety wooden stairs leads downwards into the dark beyond."
Ask them why they have come here. Who sent them? What is their goal? You can help them to come up with something, and guide them, but the decision should ultimately go to them. This way, they are already guaranteed to be interested in this story, because they're writing it.
Then they roll their characters. Since they already determined why they are working together, there's a far better chance of having characters with entwined backstories and a tighter party. Less likely to have the "mysterious loner" types.

2. Time Management
End the session when they leave the dungeon (or at some earlier point if they run out of time).
The session ends after they determine their next course of action, but before they can act on that step. This way, if they say that they are going to take their newfound treasure back to the Guild, you have a chance to write the Guild, it's Master, the town where it is located, and perhaps a few small encounters on the road. If you had written all of this before, then they would have been forced to be working for a guild, and forced to go there after the dungeon, or else your planning would be wasted.

This is how you always want to end a session. Know what they're doing next, and plan accordingly.
If for some reason you find yourself with spare time and can't/won't end early, then you need methods to pad for time.
One trick I've learned is to give them inaccurate maps. For example, rather than a grid or a map with a specific scale, give them a stylized "picture map," with no sense of distance. This way, you can make a trip take as long as you need. They want to head to a nearby town? It's a 2 day hike, and they'll need to camp a night.
Boom, instant chance for plenty of random encounters.
This is where you work in your throwaway, "filler" monster of the week stuff. Use wild animals wherever possible. If you have bandits attack them, for example, there's a chance that they'll want to find out where the bandit camp is and destroy them for good. That's fine, but now you need to write a bandit camp. If it's just a Bugbear, or a non-combat encounter of some kind - it's less likely that they'll be distracted from their current course.
Burn some time in combat and then end the session just outside their destination - giving you time to write it for next week.

3. All Roads Lead Home
If the party ever reaches a point in the story where they need a new quest, then you need to present multiple hooks. A single hook is obvious railroading, but multiple hooks gives the illusion of choice. I say "illusion" because all hooks should ultimately lead to the same bit of planning. A good example is tracking a Goblin hideout. You can seed in multiple hooks:
-- The local innkeeper can't sell any ale because his shipment was robbed.
-- A distraught mother's child has gone missing
-- The town drunk was roaming the wilderness north of town, and heard strange voices in the trees.

Don't specifically tell the players "there are goblins rumored up in the north woods," because then they'll know to attribute all of these things to goblins. But individually, they might go and investigate one of these hooks, counting the others perhaps as clues. In any case, they always stumble onto the Goblin den. Perfect, because now you only need to write one dungeon.

[Nupo]

For example, rather than a grid or a map with a specific scale, give them a stylized "picture map," with no sense of distance.
Kind of like this one that I gave my players a few weeks ago. At the end of a dungeon the hobgoblin chieftain threw a piece of paper into the fireplace. One of them said they grab it, and I gave them this.

http://www.americanfalconry.com/Map.jpg

[gfishfunk]

With few exceptions, I read over this and just said 'wrong, wrong wrong.' EDIT: That is a bit harsh - wrong for creating sandbox but right for other modes of play. This is mostly because I think most posts present valid ways of doing things that generally do not lead to sandbox. Information and backstory do not lead to sandbox. Build-it-as-you-go does not lead to sandbox. Sandbox occurs when you have sand AND the players play with the sand. First you have to establish to the players that there is no right answer or correct way of going about things. If they reject that premise (outright or implicitly), you will not have sandbox.

First: in order to have sandbox, you need to have sand that the players recognize as sand. Individual bits of things that they can use and understand to build their own trajectories.

Second: It won't happen on session 1. If you are lucky, it will happen on session 5. Sand has to be scooped in, bit by bit. Players cannot handle too much information at once.

Third: You don't need detailed anything. You just need enough elements to see what your players want to do. The players are making this up as they go along, so can you. At worst, you need need session notes and previously established facts. At best, you need a quick mind.

Practical considerations: Go into session 1 with 1 thing happening. It does not need to be open ended, but it needs to introduce the PCs to some NPCs and the world. Session 1 or 2 should wrap it up, and introduce other things that are happening. Paying attention to what the players respond to, Session 3 - 5 should establish Big Key NPCs that inhabit the world, figures that can play key parts in the story. By the end of Session 5, if you are lucky and are dealing with players that are proactive and not solely reactive, you can have a sandbox.

Session 1 should provide: 1 sworn enemy group - one that can be counted on to resist and hate the PCs no matter what. 1 reliable ally - one that can be used as a resource for the PCs and as a source of information. Session 2 - 5 should provide information about other factions, and those factions should be align-able, should have a feel for the type of goals that they will have, and should have a style unique to themselves.

Impractical considerations: Most sandbox experiences are awful because a DM waits for the players to do something, and the players don't know what is in the world to interact with. My last 'sandbox open world' game (as a player) had us enter a town and the DM asked 'what do you want to do?' We had no idea. The game just started. Not only do we not know the world, we also really don't know the other PCs. We really don't yet know our own. Throwing hooks doesn't quite help. Providing a quick, forced sessions 1 and 2 help by providing context.

[Lonely Tylenol]

Second: It won't happen on session 1. If you are lucky, it will happen on session 5. Sand has to be scooped in, bit by bit. Players cannot handle too much information at once.

This (and the stuff further down which relates to it) is indispensable. I've yet to run or play a successful open world where the first sessions were open-world. In all of the more successful games, the first few sessions were railroaded, time-sensitive, and plot-rich, and gave some lasting impression of what critical world elements would be like after those sessions.


With few exceptions, I read over this and just said 'wrong, wrong wrong.' I'm not going to pick fights about particular points, I just did not see a lot that I agree with.

This is ill-advised. If you can't give specific, detailed and constructive criticism that at least challenges the views you disagree with, then you definitely shouldn't offer vague, dismissive commentary as your alternative. The rest of your post was fine; this helps nobody, fits in nowhere, and could just as easily been left on the cutting-room floor.

[GreyBlack]

Define "sandbox."

I'm serious here, there's so many variations on sandbox that there's no way to adequately answer they question. Do you want an open world feeling for your players you explore? Do you want a feeling of interconnectedness and versimilitude of adventure? Do you just want a place for your players to muck around in and be murderhobos? There's many answers depending on what you want.

[ShikomeKidoMi]

More sandboxy? Mostly, don't rely on things being fixed.

Here's a couple tricks: Improv and pre-made encounters that can be used almost anywhere the players go, but which you pretend were always going to be in that location.

Throw out plot hooks and if no one grabs them, be ready to come up with new hooks (and plots, for that matter).

[Cybren]

So, years ago the line editor for GURPS, Sean Punch, did a talk at a con about running long term campaigns, and a lot of it was about being open and sandboxy. A podcast had recorded it, but they have long since seemed to not be around anymore (at the very least, a few years ago when someone on the GURPS forums shared their copy of the mp3 with me, they weren't around anymore, maybe they've since returned). So, if you have two hours to kiil, I find there's a lot of good advice in here
https://www.dropbox.com/s/bh9yd1kv0pt6azc/RC-LONGTERM-P1.mp3?dl=0

[Asmotherion]

Make it revolve around creating a new comunity. Perhaps some war refugees the party saved and esquorted need a new home? Perhaps some slaves from the underdark that were abducted 50 years prior will find their hometown destroyed and need to rebuilt it from scratch. Find some interesting hookpoints to introduce to the players the idea that they are expected to change their environment as they play.

[gfishfunk]

This is ill-advised. If you can't give specific, detailed and constructive criticism that at least challenges the views you disagree with, then you definitely shouldn't offer vague, dismissive commentary as your alternative. The rest of your post was fine; this helps nobody, fits in nowhere, and could just as easily been left on the cutting-room floor.

Acknowledged. I edited it and still don't get into too many details, but now I provide some reasons why I disagree.

[Nupo]

Impractical considerations: Most sandbox experiences are awful because a DM waits for the players to do something, and the players don't know what is in the world to interact with. My last 'sandbox open world' game (as a player) had us enter a town and the DM asked 'what do you want to do?' We had no idea. The game just started. Not only do we not know the world, we also really don't know the other PCs. We really don't yet know our own. Throwing hooks doesn't quite help. Providing a quick, forced sessions 1 and 2 help by providing context.
It's the DM's job to create the setting, and if the setting is peaceful, dull and uninteresting the campaign will also be peaceful, dull and uninteresting. If on the other hand the DM creates a vibrant seting filled with dramma, intrigue and exciting things happening, and does a good job of detailing this to all the players at the start you will have a campaign filled with dramma, intrigue and exciting things happening. Even if the first thing the DM does after relating all this to them is say, "what do you want to do?" You don't have to force anything, even in the first session, if you give them a rich environment that has interesting things going on.

[StorytellerHero]

Consider making the passage of TIME matter in the world of your campaign.

While the PCs are adventuring or resting, the world around them is not frozen and things can happen in other places.

To keep it within reasonable control, when you prepare notes or brainstorm for your campaign, you could focus on the objectives of one or two major antagonists and think about what they are up to and where while the PCs are doing their thing, and how the actions on both sides might interact over time.

[Sigreid]

A really good place to start is ask each of the players their characters hopes, dreams and fears. Toss your hooks from there.

[KorvinStarmast]

Consider making the passage of TIME matter in the world of your campaign.

While the PCs are adventuring or resting, the world around them is not frozen and things can happen in other places.

To keep it within reasonable control, when you prepare notes or brainstorm for your campaign, you could focus on the objectives of one or two major antagonists and think about what they are up to and where while the PCs are doing their thing, and how the actions on both sides might interact over time.
This is a good point, and a way to get them interested is to now and again have songs, rumors, or events of the larger world splash over into where they are, but they don't have to play/fight. They can choose to avoid or participate.

I'd like to add on something that made the random encounter thing we did work: our DM would sometimes use an encounter to create a trade route (on his map) and then create two towns at the end of that trade route, which we either did not didn't learn about, or get to, based on what we did next. He'd sketch them out on a 3.5 card(he showed me the index card box at the end of the year, which had begun only with an old silver dragon's lair in the mountains that we never ended up going to) and it was jammed with stuff.

The best mission for our cleric (not me, I was a magic user in this one) was brokering a peace deal between a Halfling village community and the wild elves in the adjoining forest.
(Our first adventure into the forest was an ass kicking on us at the hands of the wild elves. They were, in old Greyhawk style, wild elves and we just avoided TPK.)
That encounter had happened due to a random roll as we pursued a wereboar toward his den. (Never did catch him...)

[mephnick]

This (and the stuff further down which relates to it) is indispensable. I've yet to run or play a successful open world where the first sessions were open-world. In all of the more successful games, the first few sessions were railroaded, time-sensitive, and plot-rich, and gave some lasting impression of what critical world elements would be like after those sessions.

This is a good point that segues into another point for general DMing. It's really useful as a DM to look at how other media handles the things you want to do in your campaign. You want to run a sandbox game? Look at what the best open world video games these days do. How many of them just say "OK, Go."? Not many. Breath of the Wild, the last couple Elder Scrolls games, the Fallout series, the Witcher 3, Red Dead Redemption...they all start with a narrative railroad to get the player into the setting, explain the mechanics and provide some instant motivation. You need all three of these things at the beginning of a D&D sandbox as well for a better chance of it running smoothly. The first session or so should introduce characters, explain mechanics (travel system? downtime? injury system? chases?) and provide some obvious hooks that the players can take or leave. If you have a main hook you want to have running in the background, don't shroud it in mystery, throw it right out there in the first session. Doesn't mean beating that plot element is currently possible, but at least players know it's there.

[gfishfunk]

Agreed with last eight or so posts. Open world and sandbox doesn't mean no plot, it's more like plot buffet. Have stuff, but don't force participation in a faction.

[Garresh]

Create content you don't intend to use, and keep it handy. Create lore that is not story relevant, and can only be found by players who dig. Let them go off the rails when they want, and reward them with strange dungeons that weren't even planned, or sudden surprise plot hooks they didn't expect.

Basically, just make lots of crap you don't intend to use, and pull it out when they go exploring.

[Nupo]

Also have lots of NPC's handy. I use full character sheets for them. Just make them up as if you were creating a character. Make a variety of classes and levels. They don't have to be fleshed out with lots of details, just the basics. Details can be added later that are appropriate for the situation. They will come in handy.

I like to have lots of little dungeons ready that can be dropped in anywhere. You might need to change some details, especially with the entrance, to make it fit with the current spot, but having the basics down will be very helpful. The longer you DM the more of these you will have on hand, and therefore the more resources at your disposal. They don't have to be great big detailed novel like things, just a page or two. Here is an example (http://www.americanfalconry.com/Dungeon%20-%20Ship.pdf) of one I worked up for a recent adventure. It's not as generic as what I'm talking about, that can be dropped in anywhere, but it will give you an idea of the level of detail I'm talking about. Actually I spent more time on this one than I usually do, most are pretty rough. Fine details are filled in on the fly.