Strategies for Cold Opens of Impromptu Games

Also Applicable for Getting Any Game Going

In this post, I’m going to write about a strategy I’ve used to great effect when starting an RPG 📖 session with a new batch of characters. I’ve found this effective for one-shot games, in which players make their character in the moment and you play for that evening. It is also a useful technique for kick starting a session in which you’ve lost the momentum from previous sessions.

Let’s walk through a few examples.

Example One

You and a group of friends agree that you want to play a one-shot. Everyone’s going to roll up a quick character. Maybe it’s Dungeon World (DW 📖), Old School Essentials (OSE 📖), or Stars without Number: Revised Edition (SWN 📖). I’ll pick Dungeon World.

Everyone grabs a playbook, fills it out, then you get to the character bonds. I’m going to roll up the first bond using my Random Bonds Generator for Dungeon World. While bonds are something that might not be in other games, you could still use the random table.

The first player rolls up “I will convince insert character name here about running away with an angered friendship.” The first player names the second player’s character.

And here’s the first step: Ask the second player to expand on being named in this bond. Who’s the friend that angered them? Have the second player name that friend. Maybe ask another question.

Then, and here’s the second step: Ask the third player’s character about their thoughts regarding that bond. Maybe ask a loaded question: “Why do you think their anger justified?” or unjustified? Or “When did you learn about this?” or “Where?”

And while it may not be evident, encourage the players to ask questions that they’re curious about or name their involvement. Though, as a GM 📖 , halt the questions when you feel like you’ve got enough adventure material to now throw at the characters.

I then move on to the next bond. Either from someone that has yet to engage in the conversation. All the while, I’m sketching out a relationship map between characters and bonds. And holding the answers as the creative constraints for kicking off a game.

When I run Dungeon World, I always love it when there’s a Thief. The bond “blank and I have a con running” is paydirt. The Thief names an accomplice, the accomplice defines the con, and an observer gives commentary on that state of the con.

The goal of this approach is to begin binding these newly created characters to a common cause. You’ve not yet “placed” the characters in a location but are beginning to tease out the situation that will certainly become the starting location/scene for your session.

Example Two

Let’s say you ran a game that fizzled out awhile ago. You all decided you want to pick it back up. Instead of trying to pick up where you left off, consider advancing the time a bit, but asking the characters about the events that transpired just after that last session.

Let’s say your characters were in the middle of a dungeon, and due to scheduling constraints you had to keep postponing, and eventually lost momentum. However, you’ve found time again in your schedules, and decide to give it a go.

Instead of picking up in the dungeon, start asking questions to re-frame the campaign and start up again.

“What’s something you lost as you fled from the mummy in the dungeon?“ In this case, tie in a powerful creature in the dungeon. Let them know about the opposition.

With that answer, ask the second player “What was your role in that loss?” And then the third player, “And what are your thoughts about that whole situation?”

“When you returned from the dungeon, what were you surprised to have learned?”

That question might be more shared world-building than you’re comfortable with. If so consider a harder framing question: “When you returned from the dungeon, the Darkcloaks had descended on the town, why do you think they’re there?”

Again, ask questions, take the answers and then ask another player a question that builds on that first question.

Also ask “The loot you brough back from the dungeons running out, what have you frivolously spent that money on?”


Ask a few framing open-ended questions; questions that require more than a simple yes or no answer. Then ask follow-up questions and as the adventure takes form, dive into your game. Don’t let one player answer too many questions, instead involve all players with answer questions. The goal is establishing a shared vision of the initial adventure premise.

While running that game, when in doubt, look back to the notes you took regarding those framing questions. Keep drawing information and insights from those questions and answers.

I’ve used the above approach for the last 6 or more years.


For those that have read Mel White and Bill White’s Low-Prep Adventure Design, I highly recommend it. It’s similar approach to what I outlined.