You’ve prepped, plotted, and planned. You have character sheets from the players, printouts from the rulebook, and everyone found a spot on the calendar that works. Now, your players are sitting around the table, dice in hand, and are looking expectantly towards your end of the table. What do you do?

I wouldn’t go so far to say that running your game is easier than prepping for it, but it is a completely different set of skills. Many of those skills, like using the game’s rules and putting yourself in the headspace of a character, apply equally to all players, whether they’re the GM or not. Others, like taking notes and tracking what’s going on in the setting, look the same whether they’re happening during the session or in prep time before. There is one skill, though, that is both admired and dreaded in equal measure: improv.

You can’t run a game without improv. Not a good one, at least. Even a pre-written module can’t cover every eventuality that comes from a player’s head, though they can try. Similarly, the act of diminishing or negating player action to push them back towards what you have planned is called railroading, and it’s always bad. So, if you want to be a good GM, and create memorable characters, stories, and sessions, you’re going to have to make stuff up.

How much you actually improv depends on your personal style, and the more you prep, the more information you have at your disposal for the session. The problem I always see with this is twofold: First, the amount of prep you actually use during the session is quite low, so even if you’re a master illusionist, constantly reshuffling your prewritten NPCs and encounters, actually prepping everything out ahead of time is an exhausting amount of work. Second, your hours of prep absolutely do not make a better, more involved game than my random generator plus saying yes to everything. There are two types of GMs, those who prep to avoid improv and those who improv to avoid prep, and I’m absolutely the second.

As was made clear in the previous articles, my mode of prep is to build a scaffold (what I called the problem space) that you will use during the session to support the narrative as it comes into being. This means understanding what’s being presented during the session, yes, but also means understanding how much control you have over the events that take place therein. A traditional role-playing game is a conversation between players, the GM, and the dice, with each having equal say. With that in mind, let’s take a look at how best to set the stage and run a session.

Setting the stage

Actually running an RPG is a bit like pushing a boulder down a hill. You’re going to have to do a bit of work to get it rolling, but after that you’re just running alongside making sure it doesn’t hit anything. When you’re starting a session, there’s two important things you need to do before you can really get going. First, you (or one of your players) need to provide a recap of what happened last time. The important points here are the rough events that took place, and anything that was particularly of importance to one of the players. This shouldn’t take more than a minute, even for four hour sessions. If one of your players says you forgot something, let them butt in; they have just told you that there’s an event from last time they think is important, and you’re going to want to loop it back in.

If it’s the first session of a game, then ‘recap’ the session zero by having everyone introduce their characters and going over how the characters know each other. This part may bleed into the next part, as unlike a true recap you’re not entirely going over things that have already happened. Whether it’s the first session or not, though, that ‘recap’ is putting everyone on the same level of knowledge for starting the first scene of the game.

Generally the first scene is going to take place in a familiar location, or at least a neutral location. This is an opportunity not just to begin shifting everyone to thinking in character, but also to take care of the logistics of how much time has passed since last session, have characters rested or healed, and anything else that didn’t have to happen on screen. It’s best not to do the bookkeeping onscreen, but at least note that it has happened so everyone is up to speed. Then, noting the recap, simply ask what the characters should do next. As I noted in the last article, you’re probably going to want a nudge available to you if everyone is just looking at each other. Similarly, if you’re running a sandbox game or are otherwise at a point of resolution where there’s no immediate problem, you should have one up your sleeve. That was part of your session prep, most likely. While you can have some friendly NPC find the characters and deliver your quest du jour, it works just as well (better, honestly) to give the players a few reminders and let them decide how to go find the plot themselves.

In the event that the last session ended on a cliffhanger, you’re in luck! You start the session where you noted the ending last time, and quickly recap the immediate problem the characters are facing. From there, it’s off to the races.

Running the session

Every RPG adventure boils down to characters moving through an environment overcoming obstacles to accomplish a task. This is why Fate Core works with defining only four tasks: Attack, Defend, Overcome, and Create an Advantage. The first two define interactions with character obstacles, while the second two define interactions with environmental obstacles. The players provide the characters, the task is generated mutually through what’s happened in the game so far, and you, the GM, provide the environment and the obstacles. Where this all gets complicated is in what happens after you’ve provided those things, as that’s the part you can’t prep for. The biggest curveballs players throw to their GMs are based on how they choose to overcome obstacles. This starts small, where a character throws an NPC out a window, both ending the combat and forcing the GM to consider what happens in the street below when they hit the ground. It then gets bigger, as characters on the move through a town on the way to the presumed destination of the entire chapter of the campaign decide instead that they want to stay there, changing the entire storyline. GMs must improvise when they have curveballs thrown at them, and it will happen all the time. This is why you prep your problem space.

Let’s go back to the first example of throwing an NPC out the window. If you prepped your problem space, even if the room where the fight took place wasn’t a preset location, you should know what kind of neighborhood the characters are in, what floor the room is on, and maybe some basic things about the NPC, at least whether or not they were a significant character or just some low-level mook. Now, after the dice hit the table and that NPC is getting defenestrated, there should be two thoughts in your mind: First, what is the next logical thing that happens? Did your description of the room cover whether or not the window leads out to the street or to an alleyway? How high up are you? Does your system have falling damage (like it or not, the rules are here to help keep things consistent)? Second, what’s the most interesting thing that could happen? Is it genuinely interesting to have the guy land in a dumpster and survive, or does that just feel like it’s making a cool combat maneuver toothless? If he lands among a throng of bystanders (clearly interesting), is that consistent with how you set the scene? Instead of actively considering these questions every time a player does something I don’t expect (which is all the time), I generally tend towards choosing the outcome that a) fully lets the player get the win they were expecting, and b) makes things more complicated than they were before. Some people may look at a) and think I’m being generous, but the dice, not you, should be telling the player when they aren’t effective. If you feel the need to mitigate player ideas, use a system with partial successes, not GM fiat.

The dice and the rules should be your friends. That’s not to say you’ll end up using a system completely rules-as-written, but if you’re bothering to use a system with more detail and more crunch you should be making that work for you. Consider encumbrance. Classically a rule that’s ignored in D&D, many systems use alternative methods for encumbrance to make it usable and interesting. The existence of encumbrance, though, in something like D&D can still help you and pose a challenge. Imagine your characters making it to the bottom of a dungeon and finding the treasure in the form of several hundred pounds of statuary. Now imagine that you’re also making use of the wandering monster rules, and getting out is going to be as interesting as getting in (and you didn’t have to prep any more). I’m not trying to say you ‘need’ more or more detailed rules, but when you have them they can help put more and more interesting constraints on a situation.

Situations are ultimately what you’re going to prep. When you put a simple task in a difficult environment it can provide hours of fun, and it’s going to be up to you to provide that environment. In Electric Bastionland Chris McDowall states that GMs should be generous with detail; that goes for any sort of game, not just Electric Bastionland or other OSR systems. ‘Gotcha’ moments are rarely fun for anyone besides the GM, and they’re rarely as fun as you think they’ll be. On the other hand, dawning comprehension of what’s going on can make for great tension; in an ideal world the players will have to think on their feet as much if not more than you. Play the environment consistently and let the dice tell you when things get screwy, and you’ll be able to run most sessions with ease.

---

Let me give a longer form example; this is the finale session of a game I ran using Interface Zero, a cyberpunk setting for Savage Worlds. The campaign involved shenanigans with religion, with an initial run-in with Scientology turning into the pursuit of a religious extremist trying to consolidate power in the North American Coalition (or NAC). Said extremist had made it back to his private home in Delaware, where he had scheduled an informal lunch meeting with Max Bell, a powerful CEO in the setting. The campaign had started in Delaware (as a tie-in to our group’s in-person trips there), so I started the session at the same bar the campaign started at, where one of the PCs was a part-time bartender. They were told the antagonist had made it back to his home, and they knew the address. They were told the CEO was flying in. They were also told that if the meeting took place undisturbed, it could mean very bad things. That was all.

I prepped one location in detail, the antagonist’s house. It was in a corporate neighborhood with high security, the lot was walled, and there was a security detail on site. I noted in Google Maps the nearest municipal airport, just so I’d know where the CEO was coming from. The very first question I was asked in the session was what limo companies were nearby. Curveball one. I made up a limo company, and our hacker broke in to try and see if they were sending a black car to the airport around the time our CEO was set to arrive. Instead of artificially prolonging the situation with another limo company, I said yes. They then, as they had already broken into the (not very secure) local limo company system, looked up what driver was set to drive the limo, called him, and told him to take the day off, paying his rate from their personal funds. Then one of the players dressed up as a valet and went to the limo company office, explaining there had been a mixup. Long story short, the team got the limo, and went to pick up the CEO. After he got into the limo, one of the characters popped out and put a gun to his head. Instead of immediately calling security, the CEO gave the character thirty seconds to make a proposition. Roll persuasion, successful proposition, the CEO agreed to let the team get onto the grounds with the limo. After that, instead of directly engaging the security team, one of the characters opted to disguise herself as the antagonist’s wife. Curveball two. She rolled fairly well, but I made the judgment that she’d fool the security guards but not the antagonist himself, who would clearly recognize his wife. This worked, albeit temporarily; the wife wasn’t supposed to be there and everyone knew it. Still, it bought the team enough time and leeway from the guards to  drive the limo through the house’s large decorative plate glass window (because why not) and for the team’s heavy to pop out of the trunk, sock the antagonist once in the face, and knock him out. It was a great synthesis of a simple problem (stop a meeting), a lateral solution (steal the limo that was supposed to be onsite for the meeting), and enough prep to allow me to say yes to every crazy idea one of my players came up with.

---

During the game, a GM has two jobs. First, they are referee, they apply and arbitrate the rules. Second, they are the graphical engine of the game. For an RPG to feel truly transcendent, you need to do things real graphical engines can’t do; you need to make stuff up for which there are no art assets. Ultimately, this comes down to two things. First, do your prep and go into the game feeling like you understand how your setting behaves. That way, when you make stuff up it will be (or at least seem) consistent and believable. Second, do the most interesting thing that makes sense at a given moment. Let the dice tell you how well or poorly everything goes, but once that basic fact is established, make it interesting. Interesting doesn’t mean outlandish or wacky…keep your players engaged, and never be afraid to loop back in on what’s already happened.

So that’s How the Wonk GMs. My games have shenanigans, and it probably makes sense why I tend to like systems like Powered by the Apocalypse that give you a lot of wiggle room but also a lot of structure. I know my style may not be to taste for everyone, but I think there’s one takeaway that anyone, no matter how much prep or how much improv you want, can use. Treat the world your game takes place in as real. Understand how it works, understand why it works. Whether it’s physics, tropes, or simply the ‘rule of cool’ every setting has underlying rules that hold it together. The more you get the world, be that a classic D&D setting or Night City or a Federation Starship in the Gamma Quadrant, the more easily you’ll be able to build your campaign around you. And at the end of the day, building our stories is all any of us are actually doing.

Like what Cannibal Halfling Gaming is doing and want to help us bring games and gamers together? First, you can follow me @LevelOneWonk on Twitter for RPG commentary, relevant retweets, and maybe some rambling. You can also find our Discord channel and drop in to chat with our authors and get every new post as it comes out. You can travel to DriveThruRPG through one of our fine and elegantly-crafted links, which generates credit that lets us get more games to work with! Finally, you can support us directly on Patreon, which lets us cover costs, pay our contributors, and save up for projects. Thanks for reading!