If players have a better idea than I have. I always go with it, but only after they give me a few monutes to write down a few notes during which they talk in-character.

I took improvisation to a whole new level one session when I let the players work out a plan for the task I was setting them when I wasn’t in the room. I gave them all the necessary information for the task, in this case it was breaking a prisoner out so I gave them maps and a few hints about what sort of equipment they would need. And I also told them I would be handing out XP for originality.

I thought I could anticipate the players, I was wrong.
.-= Katana Geldar´s last blog ..There’s one at every table… =-.

Trying to come up with improv-ed results can be difficult and I’ll admit I sometimes get stumped. I’m going to keep your revised list of techniques with me for my next game session to jog my creativity .

When I get really stumped, I have a fall-back: I use a tarot deck. I’ll surreptitiously draw a card behind my GM screen and I’ll improvise an outcome based on that card’s image. Heck, the result may not even remotely resemble the card — I just use it to jog my creativity. Truthfully, any stack of pictures can do the same thing, it doesn’t need to be tarot cards. I just happen to collect tarot decks, so there’s usually one in easy reach. But I’ve done it from Everway story cards, art history text books, photography magazines … anything rich in visual images will do in a pinch for me.
.-= Cheri Arbuckle´s last blog ..Wormy’s Back!!! =-.

This is typical of wanting to define something, in the world nothing is black or white and an all encompassing definition would take legal document to provide. Obviously the moment you try to put a simple description to something you leave to interpretation the things in the grey areas. Obviously poor choice of word like here can lead to assumption not intended.

The main point was that improvisation and sandbox-game are two different thing that could or not co-exist.

As for sharing narrative. I don’t see it affecting whether the game is sandbox or not. I would definitely expect it to be used more often by improvisation heavy GM then those with heavy prep as you’ll have a hard time to be fully prepared for it, but not exclusively.

Hope it clarifies what I meant.

@Tommi: Players are as likely to build worlds as the GM, but their contribution may be less oblivious. Often what my players say or do will shape my vision of the world and I’ll implement new elements to it based on their indirect input.

@Colmarr: The Primal/Within experiment is actually the first time I’ve built a game world as a team. Yan has the super extra bonus of building it both from the inside and out.

I’m curious to see if we’ll stay in this world over the next season.

I’m not sure this is always the case, and the current Primal/Within campaign is a good example. You are a player and Phil is the DM, yet it’s clear that you have been a strong force in designing the world.

Another example I encountered recently is “open skill challenges”. When our party got lost in Thunderspire Labyrinth, our DM simply said to us “How do you get out?” In doing so, he gave us narrative control that included creating the problems we faced (caverns and locked gates) and the solution to how we solved them (the “Jump” power and Athletics checks).

Admittedly, the DM will likely always bear the greater portial of world-building, but it doesn’t have to be (and often isn’t) completely black and white.

As for the shared narrative/creation with player, that is something else. What I meant by the GM provide the world is that he will tell you what happen when you do this or that. If you prefer he provides the physics of the world. The world in itself could be from a source book, create by a friend, yourself or a group. It has no real significance whether the game is an improvised non-sandbox game or else.

Like I said that is the way I see things and is in no way the meaning of life and the universe. We all know it’s 42.

The paradigm of GM taking the lead without having prepared much anything is little explored. Do people actually play like it? I have no idea.

Also: Players create world in play in the same way that GM does, but usually to lesser extent. They may do it by suggesting things or by simply adding them. I had little idea about what space ships looked like before a player applied pencil to paper. Now we have a map and all. Player creating world.
.-= Tommi´s last blog ..… but I’d rather be roleplaying =-.

Please pardon by plugging away:

http://eric.maziade.com/post/2008/12/22/The-Rules-of-Sharing-Narrative-Control-%28and-Improv%29

NB: What I construed as “sharing narrative control” back then is less about sharing the narration than it is about sharing story building.

I use two techniques for this. The first one is the Little Cliffhanger. Right before the dead time begins, do something that makes your players incredibly nervous. If the players are all on edge and chatting with each other, it’s not dead time. Either have something relatively minor go wrong, have the bad guys do something really clever, or if all else fails, just roll a bunch of dice and say, “Uh oh” (That last works especially well if your players know you have a large collection of random tables for weird things – the City District table, the Building Contents Table, the Wacky Street Encounter Table, the Magic Curse Table, and so on). Of course, if you actually have a fairly benign table for whatever you are doing, a roll on it will often do the trick too.

Second, make good use of Table Breaks. Any time you are doing a bit of serious improvisation, like your “What’s in this random house?” scenario, get up from the table and take a break while you do the brainstorming. Use the restroom. Grab a beer or a soda. Refill a bowl of chips. Sharpen a pencil. Whatever. When you get back, you’ll have your answer; with any luck, your players will still be mulling over your Little Cliffhanger; and it’ll look like you knew what you were doing the whole time.

@Wyatt: You GMing style is pretty much like my own. It’s improvisation heavy with some snippet of preparation.

I’ve got typical encounter plan that they could find in the region (mostly as David mentioned to avoid flow breaks of creating an encounter on the spot) a few line describing the major places of interest and that’s it. I’ve played 3 sessions in my campaign with 2 page of notes. The battle map is draws on the spot on my large vinyl map usually by the players with me giving them some general guide line, while I get the figurine and make some last minutes adjustment to my planned encounter to feel more in tune with the current situation.

@Graham: I suppose what it all comes down to is that improv is a method and style of running a game, while sandbox is a game type. They really define two different things, which may or may not overlap.

That’s pretty much my thoughts on the subject.

@Oddsey: Thanks for the definitions! I also agree that player ‘jumping the rails’ could shift a campaign from DM-driven to player-driven. In fact I’m sure some structure-light campaigns alternate between DM-Driven to Player-driven quite often.

@Yan: Ptolus is indeed the ultimate published Sandbox setting (it’s also 700 pages long) and would serve as a great tool for DMs who want to test such types of game but are daunted by world building. Although, as you say, there are intermediate game types where you prep only as much as you expect players to explore in a given session and improvise anything you didn’t plan for.

@Eric: Lol! Yeah, a lot of us rely on your tool #3. I have to try #2 soon, as I’ve seen in in the book you gave me. As for Laws’ process, I’ll print it out on a sheet of paper and keep it nearby when next I need to improv…

@Wyatt: I think you are one of the DMs of the second half I spoke about in my post. Like Yan, you’ve made improvisation your main way of running a game. While you may not t always feel confident about it. I think that prepping some modular pieces of adventures (Tiles, Stats, plot ideas) is a great way to help you piece together an adventure rapidly.

@David: The issue of Continuity is very real. While I wouldn’t sweat using the same name for a NPC, I would use the slip to fuel something funny in the game, I would be worried of making stuff up that completely contradicts what I did in an earlier session. That’s one of the reasons why I write game reports.

In terms of avoiding dead time, so far my technique of having all my pre-drawn battlemaps (I have dozens) and game stats all done has allowed us to play without breaking the flow. So yeah, that’s also a very good point.

@Tommi: I really think that the ‘say Yes’ cardinal rule comes to the forefront of an improvised game. Saying no closes doors and stifles the GM’s creativity for sure.

As for doing impro work in a stats heavy/mini-driven game like the last 2 instances of D&D, I would say that it’s possible but needs a bigger toolbox, a bit like David and Wyatt mention (pre-drawn maps, pre-made stat blocks, etc)

Thanks for the comments so far.

What about games where you don’t provide the world, maybe only a starting situation, and let the players steer?

Actually, regardless of how much is prepared to begin, you would still be providing the world to the players. Just not all at once, and some of it will be pulled out of your ass, but it’s still provided by the GM.
.-= Graham´s last blog ..Damn you, Dave! You and your… logic… =-.

1. In case of player character and risk, roll the dice (or say yes to the player).

2. Go to the direction of most uncertainty; do whatever gives others such choices that you don’t know what they will decide.

3. Do the obvious thing. It is generally not obvious to the others, especially when the obvious actions of many players create very nonobvious outcomes.

4. When in doubt, think of what might happen and roll the dice.

5. Ask ideas from the players. Really. Just go ahead and ask.

Applying them in the given order should give decent results. I think that I don’t quite do that, but it is something of an approximation.

A set of rules that helps in impro would be useful; 3rd and 3th edition of D&D put pretty much focus on interesting combat encounters, which require some planning or lots of skill to pull off without it, so they are not the best systems out there for improed play.

Yan; you wrote:

Sandbox game = let the player steer the story only provide world
Non sandbox game = steer story.

What about games where you don’t provide the world, maybe only a starting situation, and let the players steer?
.-= Tommi´s last blog ..… but I’d rather be roleplaying =-.

I find there are certain skills necessary to run an improv game, and thinking fast is only one of them. Whilst these tips are very good regarding thinking fast they do not deal with some of the other problems you may run into, then main two for me being:

1) Continuity. Every now and then your brain will play a trick on you and you will use a name you’ve already given to something else, or you will put an NPC where he isnt supposed to be. There are several ways of dealing with this, but the best bet is to try keep notes on things as you make them up so this sort of thing doesnt happen too much.

2) Flow-breaks. If you have to stop for 10 minutes to draw a map, find monster stats and total up XP budgets (especially prevalent in 4e) then your players are likely to get bored and the attraction of an improv game (it can go anywhere do anything) wears thin quite quickly. To avoid situations like this encourage your PCs to aid you in combat prep. Have someone do math, someone else draw the map while you rummage through books/PC folders etc. to find the relevant information, doing this prevents your players becoming bored during combat-setup.
Another way of avoiding this is having pre-made combat maps that you can drop in where relevant, and a couple of pages with “random encounter” stats on for the same level as your PCs, but if you start to do this players will begin to question how improvised their improv game is.

Latest Blog Post: Before 1st Level, WIP

I don’t plan a session, I plan a flowchart of the major objectives of a quest or adventure. I have in mind some ideas of how players will move from A to B, and I even have some ideas of, if I’m bored enough, slapping them towards B with a large trout. But I don’t have a straight line of “players will go here and do this”. I LOOK like I do, because whenever a player goes anywhere, there’s something there for them. But in reality that something is basically pulled out of a hat. Usually.
.-= Wyatt´s last blog ..A Punishment Ill Fit (III) =-.

1) Throw something (some would call it a red herring) at the players and listen to them figure it out for inspiration
2) Come up with my best ideas and use the third one
3) Freeze, stutter and curl into a ball of shame.

I love Law’s ideas and I should have a post-it on my DM screen.
.-= Eric Maziade´s last blog ..Meta != Crunch =-.

Sandbox game = let the player steer the story only provide world
Non sandbox game = steer story.

Improvisation DM: Create on the spot.
Scripting DM: Create in advance trying to cover all possibility.

A Sandbox game without improvisation is the equivalent of running a game in Ptolus… I mean you’ve got 300+ pages of material, you can just say “Ok, you’re in a Ptolus inn what do you do?” and flip page from there on…

An all improv sandbox game would be something like “You are in… a bar! in the city of … Port Swala which is a coastal city in… Lala land. What do you do?” You create world and story as you go.

A non sandbox game you often star the campaign in media res or in bar when somebody burst through the door looking for you. The PC are hooked right from the start and lead from hook to hook.

Whether this is improvised or written is just a matter of whether you prepared all the scene in advance or just create as you go.

I think that no GM are purely any combination of these. As we’ll use tool from the other approach to fill the weakness of our favored one.

Good point about the DM-driven improv games.

I suppose what it all comes down to is that improv is a method and style of running a game, while sandbox is a game type. They really define two different things, which may or may not overlap.
.-= Graham´s last blog ..Damn you, Dave! You and your… logic… =-.

Thus, there’s strong overlap with improv-heavy games, but it would also be possible (easy, even) to a run a game that was both entirely improv’d and entirely DM driven. Start with “ninjas attack” or equivalent, move on to “the ninjas leave a mysterious clue to their origins!” and roll from there. It’s not sandbox because the players are in a primarily reactive role.

But that’s just for a given value of “sandbox.” And, heck, since I’m defining the term in a way that describes player behavior/social contract stuff as much as it does DM style, a game could move from “non-sandbox” to “sandbox” pretty easily, particularly an improv game. The players just have to decide they don’t care about the ninjas and go looking for a caravan to guard or something.
.-= Oddysey´s last blog ..Sort of a Death Frost Doom Review =-.