Any advise for a game master who is mostly improvisational?

[Ameraaaaaa]

Ya see i am possibly doing a 3rd attempt on a non one shot game master. But it's also the 1st time I'm game mastering in my new improvisational style outside some one shots with family.

2 paragraphs of context.

Before I'd try to come up with a plot beforehand but with some improvisation during the game. The 1st time i kinda rail roaded the players by accident. The 2ed time had no rail roading iirc but had many small issues relating to feeling pressured to please the players perfectly. Both were online.

But recently i started doing a few one shots with family as i said beforehand and they were completely improvised and they were amazing. I also did some other one shots as a pc instead and a attempt of a campaign with a published adventure in the osr style (to experiment) which my lil bro didn't want to play after his character died.

Context over.

So i decided to try to run a game of deviant the renegades mostly improvised with the exception of the conspiracies which I'll build after the characters finish with their characters.

So any advice for improving my improvisational skills for a campaign rather then a one shot.

If it helps I'm gonna be playing deviant the renegades.

[HumanFighter]

Your GMing style seems to be about the same as mine :smallsmile:
As a GM I improvise a lot (I prefer GMing over being a player)
There's been times where I had a whole dungeon and story planned out, but due to my players wanting to do other things, that all never happened (facepalm) maybe I didn't give them enough incentive? Idk.
With improvising, I find it can slide either way, going towards cringey failure to rapid success in any given moment. The game and players can be unpredictable (because of dice-rolling) and u just gotta learn to roll with the punches.
I find one thing that helps me as an improv GM is to keep the goal of the campaign in mind, and gently try to move the players towards it. Give out a plot hook to the players that they promptly ignore and do other sh*t? Let them know the consequences of that. But not like in a meta-sense where u tell your players OOC: "guys, if u don't save the baron from the villain now, the kingdom will surely fall into chaos" dont do that. Instead, show, don't tell. Describe to them that maybe a peasant uprising is happening now that the baron is gone and unless the players do something there may be a new person on the ruling seat, and they may not be as kind or merciful as the previous baron was.

Also another thing that I find helps with making improv GMing faster and more efficient is random tables. Random tables for encounters (in wilderness or in dungeon), random tables for loot, etc. Just make sure those tables are balanced somewhat so it doesn't get stupidly out of hand with one roll.
But random tables aren't everything. Instead there can be a "yes" or "no" answer and you roll a dice to determine which it is. This should usually be 50/50, but sometimes an outcome might be more likely than another, and I'll leave you to calculate that at your discretion. For example, your thief player wants to break into some lesser noble's house to steal a precious artifact, but the noble himself is a powerful mage and a dangerous opponent. He has house guards, but if he were home this would make the mission a lot more difficult. So, roll a dice to determine if he is home or not when the break-in occurs.

[Glimbur]

The problem I have when improvising is I do not remember everything from session to session. Take 5 - 10 minutes after the session to make sure your personal notes are in order, and start each session by having a (random) player give a quick refresher on the last session. This can be very illuminating for everyone.

[Batcathat]

The problem I have when improvising is I do not remember everything from session to session. Take 5 - 10 minutes after the session to make sure your personal notes are in order, and start each session by having a (random) player give a quick refresher on the last session. This can be very illuminating for everyone.

Yeah, I was about to suggest this as well. When I started GMing I improvised pretty much everything, including much of the setting (I was 11 or 12 and had zero experience role-playing, I'm not even sure I had more than a vague idea there even was such a thing as pre-written content) and while it had its upsides (I still improvise pretty heavily, even if I've also learned the advantages of planning) remembering stuff was definitely one of the challenges.

[Ameraaaaaa]

Your GMing style seems to be about the same as mine :smallsmile:
As a GM I improvise a lot (I prefer GMing over being a player)
There's been times where I had a whole dungeon and story planned out, but due to my players wanting to do other things, that all never happened (facepalm) maybe I didn't give them enough incentive? Idk.
With improvising, I find it can slide either way, going towards cringey failure to rapid success in any given moment. The game and players can be unpredictable (because of dice-rolling) and u just gotta learn to roll with the punches.
I find one thing that helps me as an improv GM is to keep the goal of the campaign in mind, and gently try to move the players towards it. Give out a plot hook to the players that they promptly ignore and do other sh*t? Let them know the consequences of that. But not like in a meta-sense where u tell your players OOC: "guys, if u don't save the baron from the villain now, the kingdom will surely fall into chaos" dont do that. Instead, show, don't tell. Describe to them that maybe a peasant uprising is happening now that the baron is gone and unless the players do something there may be a new person on the ruling seat, and they may not be as kind or merciful as the previous baron was.

Also another thing that I find helps with making improv GMing faster and more efficient is random tables. Random tables for encounters (in wilderness or in dungeon), random tables for loot, etc. Just make sure those tables are balanced somewhat so it doesn't get stupidly out of hand with one roll.
But random tables aren't everything. Instead there can be a "yes" or "no" answer and you roll a dice to determine which it is. This should usually be 50/50, but sometimes an outcome might be more likely than another, and I'll leave you to calculate that at your discretion. For example, your thief player wants to break into some lesser noble's house to steal a precious artifact, but the noble himself is a powerful mage and a dangerous opponent. He has house guards, but if he were home this would make the mission a lot more difficult. So, roll a dice to determine if he is home or not when the break-in occurs.

I love all of the advice here! Tho i don't really have a goal for the campaign yet. My plan is to let the players set their own goals then prepare for those goals somewhat with some room for improvisation.


The problem I have when improvising is I do not remember everything from session to session. Take 5 - 10 minutes after the session to make sure your personal notes are in order, and start each session by having a (random) player give a quick refresher on the last session. This can be very illuminating for everyone.

Well we aren't doing sessions it's pbp. But it is true that it's a good idea to take notes.

[HidesHisEyes]

I haven’t heard of Deviant the Renegades so I can only give general advice, but my number 1 piece of advice is to read this article:

https://thealexandrian.net/wordpress/4147/roleplaying-games/dont-prep-plots

And this one:

https://thealexandrian.net/wordpress/7949/roleplaying-games/node-based-scenario-design-part-1-the-plotted-approach

In fact read them twice. In fact, just treat that whole website as a sacred text!

I’d add that this kind of improvised play is MUCH easier, for me at least, with a lighter game system. Again, don’t know the game you mentioned so I don’t know if it fits the bill, but I’d say it’s best to use something lighter than D&D 5E, especially when it comes to enemy stats and encounter balance. My game of choice is Dungeon World, but most OSR games (and many other rules-lite games) are good too.

Finally, I’d point out that improvised doesn’t mean zero prep. You can do as much prep as you like, you just can’t prep actual events. You can’t prep a plot.

In any case good luck with it! It makes me happy every time I see someone say they’re running an improv-heavy RPG. I hope it goes well and your players love it.

[Ameraaaaaa]

I haven’t heard of Deviant the Renegades so I can only give general advice, but my number 1 piece of advice is to read this article:

https://thealexandrian.net/wordpress/4147/roleplaying-games/dont-prep-plots

And this one:

https://thealexandrian.net/wordpress/7949/roleplaying-games/node-based-scenario-design-part-1-the-plotted-approach

In fact read them twice. In fact, just treat that whole website as a sacred text!

I’d add that this kind of improvised play is MUCH easier, for me at least, with a lighter game system. Again, don’t know the game you mentioned so I don’t know if it fits the bill, but I’d say it’s best to use something lighter than D&D 5E, especially when it comes to enemy stats and encounter balance. My game of choice is Dungeon World, but most OSR games (and many other rules-lite games) are good too.

Finally, I’d point out that improvised doesn’t mean zero prep. You can do as much prep as you like, you just can’t prep actual events. You can’t prep a plot.

In any case good luck with it! It makes me happy every time I see someone say they’re running an improv-heavy RPG. I hope it goes well and your players love it.

Thanks! I'll make it my duty to read the sacred texts!

Edit btw the system I'm using is part of the chronicles of darkness. So not rules light necessarily but oh well. I'm loving my players excitement!

[Misereor]

So any advice for improving my improvisational skills for a campaign rather then a one shot.


For a free form campaign, have a body of material memorized.
Obviously this is a lot easier if you already created your own campaign setting, and becomes ever easier as your world grows.

Have a few side treks ready to go, for a change of pace.
There will be times when people get tired of sandboxing and just want to be pointed at a situation for them to solve.

Have a mental list of cool, but superficial ideas that you can quickly insert into a variety of situations.
Adding detail and the occasional twist makes for greater player immersion.

Know the rules.
Nothing worse as a player than having to wait half an hour while your GM looks up monster stats and rules.
That happens alot when your GM is not used to anything but using the stat boxes from published adventures.
Obviously you can't learn all rules and stats by heart, so just have a general grasp of how things work, so that the game can flow naturally.

[kyoryu]

I run highly improvisational, basically this: https://bookofhanz.com/#how-i-gm-fate-core

The key I think is to have story questions - not mysteries, but "will this happen, or will that happen?" And make sure that those questions are relevant. Make sure there are appropriate stakes to the players too - either preventing bad things or getting good things.

Secondly, assume that any plan the players come up with is reasonable. Okay, sure, sometimes this goes too far, but you should really at least look for the way that it's reasonable.

So, now, when the players want to do a thing, ask "how could this go well? How could this go poorly?" If the answer isn't obvious for how it could go poorly, start looking at who might oppose it, or why it might be inherently difficult. This is almost always easier if you think of the place where this is occurring.

If there's only one answer, you don't really need to play it out. If you can answer both of those, you ahve something worth playing through. Go ahead and use your system of choice to figure out which of those things happens.

Now, often times you have to figure out how much info/whatever to give people for a success. I usually try to plan for a certain number of scenes to play out a single question. Then make sure that each scene can get them at least that portion of the way there - if not, just gloss over it. If they do something that ends up giving them more progress? Great. Just make sure you give them at least the "right" amount of progress.

Okay, so to wrap it all together:

1) Figure out what the big question of the game is. "Will the PCs get the Treasure of Awesomeness, or will <rivals get it/it sinks and is lost forever, whatever>". "Will the PCs manage to defend their village, or will it be overrun by raiders?" etc.

2) Figure out how the PCs plan on dealing with this. For a particular plan, I often try to give three requirements for what they're doing. "We want to depose the Duke!" "Okay, cool. You'll need to either kill him or find a way to get him to step down, you'll need to get someone in place (maybe yourselves) to take over, and you'll need to get sufficient support from the other nobles to make it stable." Each of these things can spawn other requirements, too. And of course they're subject to player creativity to get around the requirements. For each of these, try to figure out what happens if the PCs don't fully succeed - either they fail outright, partially succeed, or succeed with some kind of cost.

3) Let the PCs decide which they're going to tackle first, and how. For that scene, figure out what could go well or poorly. Play it out.

4) Repeat #3 until you've met all the requirements, and either succeeded, failed, or (more often) had some success but with some side effects/consequences/drawbacks. Like sure you depose the Duke, but you didn't get all the support you needed, so now you're going to have to deal with a rebel faction in the future.

For encounters, figure out the types of things that they might encounter, and pull out a bunch of them in advance. Then worry about putting together encounters that make sense based off of that (presuming you're not running a game that's light enough to generate opponents on the fly)

[Ameraaaaaa]

I run highly improvisational, basically this: https://bookofhanz.com/#how-i-gm-fate-core

The key I think is to have story questions - not mysteries, but "will this happen, or will that happen?" And make sure that those questions are relevant. Make sure there are appropriate stakes to the players too - either preventing bad things or getting good things.

Secondly, assume that any plan the players come up with is reasonable. Okay, sure, sometimes this goes too far, but you should really at least look for the way that it's reasonable.

So, now, when the players want to do a thing, ask "how could this go well? How could this go poorly?" If the answer isn't obvious for how it could go poorly, start looking at who might oppose it, or why it might be inherently difficult. This is almost always easier if you think of the place where this is occurring.

If there's only one answer, you don't really need to play it out. If you can answer both of those, you ahve something worth playing through. Go ahead and use your system of choice to figure out which of those things happens.

Now, often times you have to figure out how much info/whatever to give people for a success. I usually try to plan for a certain number of scenes to play out a single question. Then make sure that each scene can get them at least that portion of the way there - if not, just gloss over it. If they do something that ends up giving them more progress? Great. Just make sure you give them at least the "right" amount of progress.

Okay, so to wrap it all together:

1) Figure out what the big question of the game is. "Will the PCs get the Treasure of Awesomeness, or will <rivals get it/it sinks and is lost forever, whatever>". "Will the PCs manage to defend their village, or will it be overrun by raiders?" etc.

2) Figure out how the PCs plan on dealing with this. For a particular plan, I often try to give three requirements for what they're doing. "We want to depose the Duke!" "Okay, cool. You'll need to either kill him or find a way to get him to step down, you'll need to get someone in place (maybe yourselves) to take over, and you'll need to get sufficient support from the other nobles to make it stable." Each of these things can spawn other requirements, too. And of course they're subject to player creativity to get around the requirements. For each of these, try to figure out what happens if the PCs don't fully succeed - either they fail outright, partially succeed, or succeed with some kind of cost.

3) Let the PCs decide which they're going to tackle first, and how. For that scene, figure out what could go well or poorly. Play it out.

4) Repeat #3 until you've met all the requirements, and either succeeded, failed, or (more often) had some success but with some side effects/consequences/drawbacks. Like sure you depose the Duke, but you didn't get all the support you needed, so now you're going to have to deal with a rebel faction in the future.

For encounters, figure out the types of things that they might encounter, and pull out a bunch of them in advance. Then worry about putting together encounters that make sense based off of that (presuming you're not running a game that's light enough to generate opponents on the fly)

Cool cool. I didn't plan for a big question tho. I kinda was just doing the game like how i gm for my family which is "Hey! Here's a setting off the top of my head. What do you want to do?"

Though maybe i should ask the players what would be their goals are for the campaign to centre that big game wide question around.

Btw overall very good advice that I'll be implementing. Just explaining that basically i prepared nothing. Not sure if that's too improvisational or something.

[Stonehead]

Prepare a list of names for people/places, even if you don't prepare the people themselves. Most people have a hard time coming up with convincing names off the top of their head, and nothing slows down a game quite like "...and you come to the town of, uhh, hang on, let me look up a random name generator online".

[Ameraaaaaa]

Prepare a list of names for people/places, even if you don't prepare the people themselves. Most people have a hard time coming up with convincing names off the top of their head, and nothing slows down a game quite like "...and you come to the town of, uhh, hang on, let me look up a random name generator online".

Good advice. Tho in this case less of an issue in this case since it's an online play by post game.

[kyoryu]

Cool cool. I didn't plan for a big question tho. I kinda was just doing the game like how i gm for my family which is "Hey! Here's a setting off the top of my head. What do you want to do?"

Though maybe i should ask the players what would be their goals are for the campaign to centre that big game wide question around.

Btw overall very good advice that I'll be implementing. Just explaining that basically i prepared nothing. Not sure if that's too improvisational or something.

"What do you want to do?" is a great springboard for the "big question". Take what they say, and then reframe it as a question about whether or not they succeed at it.

In most cases, try to attach an NPC to the antagonism.

"Oh, we want to start a bakery"

So, now the question is will they? What's stopping them? Rivals? The difficulties of starting a business? The red tape of the area that they're in? All of the above? Figure out what the main opposition is, and go to town. "Will they successfully start their business, or will they be driven out of business by the evil Rollie Croissant, de facto head of the Baker's Guild?" "Will they successfully start their bakery, or will they be stymied by the stuffy bureaucrat McBlergh, who wants to maintain his control over the businesses in the town?" (or you combine the two in some way, even better, but that might be more than a one-shot).

You don't have to come up with anything in advance, really. You can come up with the big question, or they can (and you just rephrase it). Utlimately, for this story, 95% of your job is to take what they want to do, and tell them why it's hard. Then they come up with solutions and you figure out if they succeed or not.

[HidesHisEyes]

Cool cool. I didn't plan for a big question tho. I kinda was just doing the game like how i gm for my family which is "Hey! Here's a setting off the top of my head. What do you want to do?"

Though maybe i should ask the players what would be their goals are for the campaign to centre that big game wide question around.

Btw overall very good advice that I'll be implementing. Just explaining that basically i prepared nothing. Not sure if that's too improvisational or something.

Yeah Kyoryu’s advice is excellent. With regard to the dramatic question, I find a good way to handle this is to play a bit and let the dramatic question emerge through play. Run one session with a simple, constrained scenario that forces the players to act in some way. They’ve all been captured by someone, or they’ve already agreed to some quest or mission and have just arrived at the place. Make it immediate. Run the scenario in an improvised fashion, don’t railroad, and *something* interesting will emerge. That gives you the dramatic question for the next session and potentially the whole campaign. (The Dungeon World way of doing this is that you turn up to your first session with nothing at all prepared, do character creation and in the process ask the players a bunch of questions about the world and their characters, and improvise that first scenario from their answers. Then take it from there).

One more tip: if you haven’t, look into the idea of “countdown clocks”. For any antagonist or ongoing situation, you draw yourself a little circle divided into about four to twelve segments, and write down (briefly) what will happen when the clock runs out. When the players screw up, waste time or neglect to act on that situation, shade in a segment. When it’s full, the thing happens. I prefer this approach to a detailed timeline because a) it doesn’t require you to actually track in-game time and b) it’s easier to amend on the fly in response to the game’s events - if the players render a clock irrelevant (which they can and will) then you can just throw it away.

[Stonehead]

Good advice. Tho in this case less of an issue in this case since it's an online play by post game.

That does take a lot of the edge off.

The only other thing I can think of is don't be afraid to abandon threads that don't interest you or the players. Be it npcs, storylines, mysteries, whatever, the best strategy for improv-heavy games is to throw everything at the wall, and drop everything that doesn't stick. If the players find a dwarf with a big white beard, and no one's interested, maybe they'll like the dwarf with a curly mustache, and if they don't like him, maybe they'll like the one with a cowboy hat.

[kyoryu]

One more tip: if you haven’t, look into the idea of “countdown clocks”. For any antagonist or ongoing situation, you draw yourself a little circle divided into about four to twelve segments, and write down (briefly) what will happen when the clock runs out. When the players screw up, waste time or neglect to act on that situation, shade in a segment. When it’s full, the thing happens. I prefer this approach to a detailed timeline because a) it doesn’t require you to actually track in-game time and b) it’s easier to amend on the fly in response to the game’s events - if the players render a clock irrelevant (which they can and will) then you can just throw it away.

One thing with clocks is that filling in the segments should be visible in the game world. Not necessarily all of them, but at least a good amount of them. I'd probably say at least 4 of the clock segments should be correlated to things the players can see.

Let's say that the players decide they want to overthrow the religious order or start a cult or something. Interesting external "advances" could be:

1. Occasional rumors of this cult
2. Cult symbols are seen periodically
3. The cult is denounced, and various people are heard making proclamations (legal/religious) against it
4. Cult members go about openly
5. Official banning of the cult happens - members are prosecuted and have to go into hiding
6. Warrants are issued for the PCs
7. Various levels of mercenaries are sent for the PCs
8. Elite mercenaries (especially some kid of spooky ones) are sent for the PCs

You don't have to use all of these, but they give a tangible sign of how things are getting more tense and advancing in that plot.

[HidesHisEyes]

That does take a lot of the edge off.

The only other thing I can think of is don't be afraid to abandon threads that don't interest you or the players. Be it npcs, storylines, mysteries, whatever, the best strategy for improv-heavy games is to throw everything at the wall, and drop everything that doesn't stick. If the players find a dwarf with a big white beard, and no one's interested, maybe they'll like the dwarf with a curly mustache, and if they don't like him, maybe they'll like the one with a cowboy hat.

Definitely second this. And don’t get hung up on things that seem super important from an in-universe perspective dropping away because they weren’t interesting to the actual people playing the game. If they hear rumours in session one of a cult trying to bring about the apocalypse, and by session 10 it’s never been mentioned again, that might seem a bit weird and unsatisfying - but I promise it’s better than having too many plates spinning and not getting an interesting narrative out of any of them. If anyone even remembers that cult just say “oh yeah they just tore themselves apart through infighting”.


One thing with clocks is that filling in the segments should be visible in the game world. Not necessarily all of them, but at least a good amount of them. I'd probably say at least 4 of the clock segments should be correlated to things the players can see.

Let's say that the players decide they want to overthrow the religious order or start a cult or something. Interesting external "advances" could be:

1. Occasional rumors of this cult
2. Cult symbols are seen periodically
3. The cult is denounced, and various people are heard making proclamations (legal/religious) against it
4. Cult members go about openly
5. Official banning of the cult happens - members are prosecuted and have to go into hiding
6. Warrants are issued for the PCs
7. Various levels of mercenaries are sent for the PCs
8. Elite mercenaries (especially some kid of spooky ones) are sent for the PCs

You don't have to use all of these, but they give a tangible sign of how things are getting more tense and advancing in that plot.

Yeah definitely. I don’t actually put the visible signs of a clock in sequence, I just jot down a few things the PCs might notice or encounter for each clock, deploy them as needed, and update them when the situations seems to warrant it.

Also, in the kind of game where clocks are most commonly used (like Apocalypse World and Blades in the Dark) I think it’s common practice to have the actual literal clock right there on the table where the players can see it. That might not be to your taste if you’re really into immersion, but it’s an option.

[kyoryu]

Yeah definitely. I don’t actually put the visible signs of a clock in sequence, I just jot down a few things the PCs might notice or encounter for each clock, deploy them as needed, and update them when the situations seems to warrant it.

Also, in the kind of game where clocks are most commonly used (like Apocalypse World and Blades in the Dark) I think it’s common practice to have the actual literal clock right there on the table where the players can see it. That might not be to your taste if you’re really into immersion, but it’s an option.

Yeah, I like attaching them to movements of the clock, but that's really based on DW. And the big "failure" I see in the DW fronts advancing is people writing fronts in such a way that their advancement is based on things the players don't see. Attaching visible changes to the world is my way, for me, ensuring that I don't fall into that trap.

(Really, Dungeon World fronts are an amazing bit of game prep tech)

[Easy e]

Just always be asking, "What are you going to do?" The players can carry it.

[Stonehead]

Definitely second this. And don’t get hung up on things that seem super important from an in-universe perspective dropping away because they weren’t interesting to the actual people playing the game. If they hear rumours in session one of a cult trying to bring about the apocalypse, and by session 10 it’s never been mentioned again, that might seem a bit weird and unsatisfying - but I promise it’s better than having too many plates spinning and not getting an interesting narrative out of any of them. If anyone even remembers that cult just say “oh yeah they just tore themselves apart through infighting”.

Another fun idea is that the players aren't the only problem solvers in the world. Every once in a while, it can be a good idea to introduce a rival adventuring party, and have them take down the cult.

[HidesHisEyes]

Yeah, I like attaching them to movements of the clock, but that's really based on DW. And the big "failure" I see in the DW fronts advancing is people writing fronts in such a way that their advancement is based on things the players don't see. Attaching visible changes to the world is my way, for me, ensuring that I don't fall into that trap.

(Really, Dungeon World fronts are an amazing bit of game prep tech)

Right. Dungeon World is my favourite TTRPG but in the end I didn’t get on with fronts. I always found the situation changed so often that I was rewriting whole lists of threats and grim portents after every session. So I settled on using clocks because they feel more modular and disposable. But that’s obviously down to differences in personal style. Honestly as long as you’ve got *some* way of organising content that works for you, it’s all good.


Another fun idea is that the players aren't the only problem solvers in the world. Every once in a while, it can be a good idea to introduce a rival adventuring party, and have them take down the cult.

Yeah this is great fun. Just throw em in there and wait for the players to start a vendetta against them!

[Grod_The_Giant]

Grab a piece of paper and take your own notes during the session. I can't tell you how many times I've made something up on the fly only to forget vital details the next time it comes up.

[sktarq]

A few things I have found works well for me is seriously sandbox games with little/no prebound adventures...or anything adventurelike in their histories once they are done....we call them escapades.

Have a bounded space. Know roughly how wide your players will range. get their meta game agreement on this...
Then build up lots of stuff to just have around. What is X town like? Use a mix of basic sketching an a couple deep highly specific things for flavor (especially for yourself) and potential first contact points.

But mostly figure out who/what the power players are. and what they are up to. This allows you to figure out roughly what will will be going on in any given region, industry, noble house etc. So if you players wander in you can see who has power, who wants power, what weakness exist in the system (ease of food/ore/communication disruption/raiding vulnerability) in ways that allow you to build up something believable very quickly. It doesn't really matter what your players DO you just flick down a quick list and see who would know, care, and whose reaction is worth noting/could add to the story. Also it will help you quickly get a hand on who may be helping or becoming foes of the party so that can come in later in a natural way.

Also this helps to keep the world alive with STUFF HAPPENING if the players don't do stuff. If you give players choices of doing X quest or Y quest have the other one resolve (maybe partially, maybe completely) this will reinforce player agency and the idea that their choices matter. Also is great source for rivals (either those hurt by the PC's doing other things perhaps , or other adventuring parties that did do those things and are getting acclaim) This also helps in driving PC/player motivation. Without a given adventure or plot driving them forward players may tend toward extreme caution in moving forward (this is a holdever from games where the world is basically static and just waiting for the PC's to move). another problems the players can have is a lack of direction and motivation...without a plot etc to tell them what to do they can do anything but have no real way of saying any given choice is better than another and so end up just waffling and not engaging. Feeling like lots of stuff has a countdown helps with this; it pushes the party to commit to trying to do stuff.

also try to make sure the PC's are engaged with the world. Lots of players make the classic outsider type. Hell "Hero" means outsider in some definitions (Read Joseph Campbell for example-great for all sorts of DM's in general IMO). But this can make it hard for players to really engage with a world that doesn't force them to. Without a plot they are told (even told metagame via subtle social norms of the table) to follow it is easy for lots of PC's to just "nope" out of stuff. So try and forge social bonds between the PC's and world around them. Mentors, patrons, old friends, old teachers, family whatever. Conduits of connections to the world. They can also be ways to feed your players adventure opportunities, exposition, etc. Now lots of players don't want such connections because they see such stuff as a weakness, the BBEG can't kidnap your sister if your family is all tragically dead type stuff. The best solution I've found is to make those connections a source of power. You've pissed off the guards at town X? well if you lean on your uncle who is friendly with them he may be able to get you forgiven somehow, maybe money maybe you doing a favor for the right guy etc. Or such people have connections to information, items, etc that is hard otherwise to get. Make such connections a source of strength. All this is because a highly impov driven game is about the PC being proactive and doing while the DM reacts more than directs.

Again most of this should just be power players (including leaders of criminal factions, major monsters, etc) and a second pass sketch of most places. but some basic flavor can be helpful (what are the building styles like, what to people eat, what treasure do they value, what threatens them) in order to give you a good descriptive base to work with for an area, as it is easy to have things become very generic when you are improvising (an inn is an inn is an inn, everything is Tudor or Roman styled until described otherwise, every guard/orc/etc sounds/is equipped/attacks the same). What you also don't want to do is over-commit to any of these things. Again it is sketch of the world not a full gazetteer. You don't need to know the layout of any of the power players homes, more than a couple lines about personalities etc until your players show interest and start wandering that way.

The point is to have enough that no matter where the players wander, or what the decide to do you have some starting points to think about how the world will react, what they will find there, what the consequences are of the PC actions are, and why this place/person etc is different than what they would have found if they made different decisions. Then it is a matter of you looking at what your prep work says may be relevant and deciding what helps make an entertaining fun table. It is about having a framework of thinking so that non-obvious but logical and entertaining options come to your mind, and about adding the depth in world building, choice development options, etc that many people find to be the weakness of a lot of improvisational/sandbox type gaming.

Finally I use WTF/IDK dice. When your players do something you have no real clue on. When it isn't based on a skill or anything but on, luck, basically. Take a basic common die (say a D20 in D&D or a D10 if playing WW or whatever and just ask yourself mentally how that turns out. I like exploding critical high+lows personally. High is good low is bad.
example: They take a random alley during a chase? I have no clue what is down that alley. Roll a die. High may have a clear shot to run, stuff to toss behind you to slow pursuers, Critical high may have a open back door to someplace you can loose them while a critical low may have a dead end fence that can trap or at least needs to be climbed over or something.
It is all about getting your imagination something to build a full reaction out of, many times it will toss out things that end up being the weird wild an wonderful that add flavor to the game. It helps drive variation and specificity which can be harder to come up with when flying by seat of your pants.

Again this is all stuff I found works for me, you do you and your mileage may vary but these are some things I do to counter the things that I have disliked about games I thought were railroady or improv-generic.

[HidesHisEyes]

A few things I have found works well for me is seriously sandbox games with little/no prebound adventures...or anything adventurelike in their histories once they are done....we call them escapades.
...

Just sticking my head in to heartily endorse this post. All really excellent advice. I’d emphasise the idea of your prep being a “framework” even more, and add that anything in your prep can be altered or thrown out on the fly if you feel confident doing so. It really is just there to make your job easier at the table.

And the IDK die (I know it as the “die of fate”, John Harper’s term) is a really awesome tool.