Recent discussion of con games and one-shots leads me to believe this is a subject that a lot of people grapple with. I've had some experience doing con games at our local university game club, so I thought I'd share some of the things that worked for me. I think it's a distinct skill from GMing an ongoing game, so it merits a distinct post.
I'm going to write here mostly from the perspective of con games, but I think most of it will also apply to the single-game experience at your own table.
The obvious problem of con games is that you are working on a very small budget of time. You may be used to a leisurely session of character creation, maybe even an extra one setting up the world, and then unspooling a dozen or so episodes of a game over six or ten months, but you need to forget all that when you sit down for a single-setting experience. Everything has to be distilled down. Finding which ingredients are the most important ones, and making sure that you include those, is half the battle.
The first ingredient, and the one many people either forget or do a chintzy job of, is social. The GM walks into a group that is very likely people that they aren't familiar with, in a con, and even if they know the people they likely don't all game together every week. You need to create at least the beginnings of a group social contract, starting with the most basic thing: introductions. The GM needs to introduce themselves and make people feel welcome, shaking hands and letting them know they are welcome at the table. It's a great idea to take a few minutes going around the table getting to know everyone, learning names, asking what games they like and what their gaming background and expectations are. (Writing down names or providing some version of name tags is very useful for GMs like me that have trouble remembering names on the fly.) That may seem like an unnecessary thing, but knowing where your players are coming from makes all the difference when you're teaching them a new game and guiding them through a quick session.
Sometimes I like to bring snacks to the table. People always warm up to a game when they have an opportunity to break bread (or eat Hallowe'en candy) together.
Take a moment to talk about what kind of games you run at your table, the stuff you like to see in a game, and what the con game you're about to run will be like. What kind of tone are you going for? Make sure people know. If you're using a particular setting, explain the basic assumptions of that world. Keep it brief. I once had a GM take half an hour out of a four-hour con session explaining the background of the Shadowrun universe. That's twenty-nine minutes too many.
Get people excited about the game. Talk about the things that are cool about it, and if you have a rulebook full of awesome artwork, make sure you pass it around. If people have played the game before, find out what their experiences were like. Make sure they understand if your game will differ from that experience.
The next thing is characters. Most of my suggestions here are going to be broad ones, but just this once, I'm going to say something very specific: UNDER NO CIRCUMSTANCES SHOULD A PLAYER COMING TO A ONE-SESSION GAME EVER, EVER HAVE TO MAKE A CHARACTER.
Was that clear? Let me repeat it: NEVER. EVER. HAVE. PLAYERS. MAKE. UP. CHARACTERS. BEFORE. A. CON. GAME. OR. ONE. SHOT. For myself, I think games that demand this sort of thing as an integral part of play are no-go for one shots, unless set-up is as concise as FIASCO. You simply do not have the time it takes to educate players about making informed choices in character creation on a four-hour clock. You don't. Don't do that. It's not fun to sit there for an hour or two making up a character before you get to roll dice and play. Players don't want that, and you look like a lazy asshole who hasn't put in the effort to make the game run smoothly.
I like to present players with characters that are about 4/5 finished. Players like to personalize characters, so it makes a lot of sense to let them make a few choices (especially a name - players are very thorny about people choosing names for them) before play begins. That gives them a chance to make a personal investment in their character, and draws their attention to the information on the character sheet and how it works. Some games - I'm thinking specifically of stuff like APOCALYPSE WORLD and its variants - are already set up for this sort of thing, with players just required to make a few choices and go. That's fine, but you need to be able to help players expedite their choices, or you could still be waiting half an hour if someone wants to think things through very carefully and maximize their choices. Making suggestions relevant to the character or allowing the choice to be made later, during play, can speed this along, so it's important that the GM be familiar with what choices are available.
A lot of games struggle with a proper opening. I like to keep it simple. Present the players with a straight-forward situation where they understand what they're expected to do, and they have the opportunity to try out some of the most important rules in the game in a hands-on way. For example, in my Ghostbusters con game, the very first scene had the players busting some ghosts in a hockey arena. This was a low-pressure situation where they weren't likely to be seriously hurt, which sucks going into the rest of a four hour game, but it let me teach them how to attack, defend, and leverage the scene to their advantage. It's also important to make sure this first scene strikes the right note, tonally - if I had made that first scene scary, it sets the stage for a scary game to follow. If I make it goofy physical comedy, they know that this is going to be a light, fun adventure.
The introductory scene needs to be relatively self-contained and tight, but after that you need to give the players the freedom to play with their characters a bit. Build in a little space to play the characters and find out what their relationship to each other is like. The characters inform all the most interesting parts of what follows, of course, so if the players don't get this opportunity to do this kind of exploring, the adventure is going to have a flat, generic feeling. Let them make significant decisions as soon as possible, and let as much of the adventure as possible flow from that. For example, in a FIREFLY con game, I had the crew of the ship hired (conscripted is more accurate) to transport a load of convicts to a mining colony. I offered one of the players a Fate point if they happened to know one of the convicts, and let the player decide what their relationship was. It could have been a friendly relationship, a rivalry, an enemy, even a family member. What the player decided changed the whole rest of that adventure. That kind of narrative power gets players' attention. Another adventure, I set up that there was a terrible secret buried in a cargo container under a load of ore, and let the players say what it was. They decided it was a body. That really set things moving.
You need to remember not to plan too much in a con game, just enough to keep things moving. Often, it's enough to have a structured beginning and the opportunity to make some interesting choices, then simply watch them play out. I used to plan pages and pages of formal adventure for these events, but some of my best games have been based on handwritten notes scribbled on a single piece of loose-leaf paper.
Teach the game as you play. Don't provide a big data-dump of rules at the beginning. Most handouts go unread. Just explain things as necessary, and help players make the best choices and uses of their abilities. If a rule isn't entirely necessary to play, you should feel free to skip it. Con games are about the broad strokes, and providing a fun, compact experience. Nobody is grading you on your rules mastery or fidelity to rules-as-written.
Players want to use the time to explore the characters, so they want in-character moments, and they also want to try out any nifty abilities they might have. Give them moments that their special talents can shine. If you've got a doctor, you need to have someone who needs medical attention. If you have gunfighters, there had better be a gunfight. Watch what the players are doing, and play to that. When a player is angling their character in a certain direction, give them what they need. A hookup artist needs people to make passes at (and possibly be slapped). A Companion needs situations where she can be shown to be elegant, noble, respected. A roughneck wants to get their hands dirty.
Push the characters. Challenge them. Find out what they believe, and then give them the opportunity to prove it. Make sure that anything they brought to the table gets played out, to the best of your ability. If this is the only chance the players get to walk in these characters' shoes, they should feel like it is a full and rewarding one.
Remember to ask questions of the players, and try to see the game through their eyes. Being open and transparent about how things work, what you're thinking, and what you expect of them means they will have the best possible game experience. No one wants to struggle because they don't understand something, or are operating under a false assumption. Again, you're teaching this game, and you need to make sure you don't assume the players have knowledge just because you've got it.
This is going to worry some of you, but it's honestly not that big a deal. Remember how you've got four hours for a con slot? Well, that's not exactly right. You've got less than that. Figure about 5-10 minutes for introductions, 5-10 minutes for character talk, 5-10 minutes of unproductive goofing around, and one 15-minute break, and you've probably got closer to 3 hours for actual play. This is not a problem, and it shouldn't feel like a deadline creeping up on you, smothering your fun. What it means is you need to keep your eyes on the time. Have a watch, a phone, or a clock somewhere you can easily check it... frequently. Don't let early set-up talk go on too long, or side conversations that slow down the game. Know when your break is coming, and play toward that. Try to have a big reveal just before the break, to charge everyone up so they come back to the table raring to go, possibly with their hands full of chocolate bars and sugary drinks. That's energy you need. If you have a big bad, or some other kind of threat, this is when they show their hand. The rest of the session is about how the players deal with them. And since players need some time to scheme, some time to act, and of course some time to roll the dice, you need to budget this into your 195 minutes.
Remember that in most systems I'm familiar with, if you run a combat, it will take much more time than you like.
Don't worry about filling the four hours completely. If you fill three and a half hours and everyone has a good time, don't stretch things out - finish and let people spend the rest of the time talking and enjoying their time together. They may also need a break between sessions, if they're playing in a game soon afterward, to eat or pee or both (hopefully not at the same time).
I'm sure there is stuff I'm forgetting at the moment, but I hope that's a good starting point for any readers who might be struggling with con games. Please feel free to ask questions, if you have them.