The other day I posted a link to the blog that I wrote about my first experience with Dungeon World to the Google community and in response I was asked about my methods for improvisational GMing. As my response got longer it seemed like it might be easier just to blog about it.
First, let me say that I don't claim to be any kind of expert on the subject. I'm absolutely certain that there are others who do it better than I do. Nor is it the only way that I GM. For me it's all situational. If I have the time to do prep, I usually do, and my games are generally better for it. Right now however I have work and school and four kids at home. That kind of time is precious, and very rare.
That said, even when I have plenty of time I've never been into super heavy prep. I think that games run best when they incorporate the ideas and plot seeds woven by the players, and too much prep makes that harder to weave in. In general my prep revolves around world building and finding ways to fit the character in to what is going on around them, to mix their ideas with my own.
All of that aside, that's not what this post is supposed to be about. It's about how I improv without any prep in general, what techniques I used in my Dungeon World one-shot in particular, and what kind of tools the game gave me that I used to generate setting in play specifically.
Let's start with general tips.
The very first thing I like to know about a game is: Who is the bad guy, what do they want, and what will they do to get it. I ask myself these three questions first any time I start thinking about a game, regardless of how much prep time I have. If I have those three questions answered, I have what I need to play. From here, I simply play the bad guy like my PCs might run their characters. They don't get screen time of course, but I work from the assumption the bad guys are going to get what they want unless the PCs intervene, and then rely on the players to come up with a plan for how they intend to spoil things. Once that happens the bad guy reacts to the new circumstances, and the cycle begins anew until the PCs eventually win. That's very simplified of course, but it works very well.
Of course, for that to help you already need to have some sort of inspiration.
And by inspiration I mean, of course, a willingness to steal good ideas. Bonus points for stealing them from the players, but any source of media will do. In my actual play I mentioned that the government was called the Federation, that it had outlawed magic, and implied that it would be a major long-term threat within the setting. Every bit of that was ripped whole cloth out of the Heritage of Shannara series by Terry Brooks, which I am in the process of re-reading. I needed a villain, and it was what came into my mind so I grabbed it. You don't have to be that obvious and direct, but there's no rule saying you can't be. A long time ago I ran a two year campaign of DnD 2nd edition where the players had to gather pieces of a mysterious artifact that would expel the demons who had been invading their world that was inspired by a cartoon called The Pirates of Darkwater and no one ever noticed. The key is to make a choice quickly, move on, and be willing to adjust it later as needed. As long as the game keeps moving few people are going to second guess where your inspiration comes from so long as you don't use obviously stolen names.
As for making choices quickly and moving on, that can be the real hard part. For that I recommend that you find some excuse to give yourself a breather when you need it. Time only passes when everyone is sitting there staring at you, waiting for you to say something. Very few groups need much of an excuse to take a short break, get a drink, and tell a few bad jokes, and so long as you're not in 'game time', no one even notices if 5 or 10 minutes pass. Give yourself what time you need, move away from the table and out of range of whatever questions people might be asking. Allow yourself a quiet moment to just think without being responsible for anyone else's entertainment. I like to grab a drink and step outside for a little fresh air, but sometimes it's easier to make a quick trip to the bathroom if you don't want anyone following you or trying to talk to you (which you really don't if you need to think). Worst case you can always just tell them why you need a second and excuse yourself. Players get it that you sometimes need to process. The real key here is to do it sooner rather than later. Don't wait until you're stumped or you run the risk of creative block from frustration. Take a break right after characters are all together or right after the initial action to put at least the outline of what you think is going to happen together in your mind.
Next, think of a couple of generic openings. The kind of things you see in movies that aren't really related to the story at hand, but introduce the protagonist. James Bond movies are great at this. A lot of times the movies start at the end of his last assignment, just to show off what a badass he is. Personally I like to start with action. This serves a huge number of functions: It helps the group to learn the system, introduces all of the characters, let's them stat to gel as a team from the very beginning, and shows you as the GM how they interact with one another. Also, it buys you time. For most systems an opening combat is low enough impact that you can start to link character histories, and think about who that bad guy should be. Do they look like they are enjoying beating these guys up? Are they making up good reasons why they're in this fight? Ask them. If they are, then maybe these bad guys are related to the larger plot. If they're not, move on to something that hooks them better and just treat this one as a warm up.
Speaking of asking them, don't try and do it all yourself. Got a detail you don't know or care about? Ask someone. I recommend a single player, rather than the group at large. Point to somebody and say "Why are these guys trying to kill you?" This is not a time for discussion. If they hesitate, or don't seem to have something quickly, point to someone else. If it's weak, have another player expand on it. "We stole something from them," should be followed up with: "What did you steal, and why?" Spread it around the table though, the game shouldn't be any one player's story, not even your own. If it is a player asking a question ("Does the Federation have somewhere they could keep a prisoner in town?") let them answer it themselves ("I don't know, you tell me."). They're obviously the ones most interested in the answer.
Finally, keep notes. I'm all about running games with little or no prep beforehand, but you really need to write things down afterwards. It doesn't have to be anything big or extensive, but at least enough to give you an idea of what happened and any important names that you should be remembering. It's ok to come up with something on the fly and just throw it out there, but sooner or later it needs to be integrated into the overall plot or explained away. Not every random encounter with a goblin horde needs to become a major storyline, but by explaining it later in a way that makes sense ("Hey, did you hear that someone took out those goblins that have been raiding the King's Highway? Seems like they finally messed with the wrong group, maybe trade will start flowing again.") adds to the sense that the world is lived in, that what the players do matters, and ties up loose ends. For me personally I think this is what makes playing improv games work in the long term. Their biggest weakness is that they can feel random, like a monster-of-the-week. By taking a few minutes after the game is over to remember important things and integrate them in you r mind you can avoid the worst of that.
Ok, I promised to get into some specifics, but it's getting late and this has gotten really long. I'll have to wrap up how all of this fits into Dungeon World tomorrow.