Sam Feldstein's Notebook

Improvised D&D

Problem: Roleplaying sessions take a long time, so they happen rarely, if ever.

Solution: Total improvization. Minimize prep, maximize fun.

A lot of the broad principles I'm going to discuss here are already known to the roleplaying world. So I'm not laboring under the illusion that I'm inveinting anything. So while I'm going to start with some broad principles and implications, what I'm mostly interested in here if figuring what an improvised game would look like, and how it would actuallly work. How do you run encounters if you haven't look at monster stats ahead of time? How do you start a quest if you haven't alreayd come up with an antagonist? How do you invent NPCs on the fly?

As with all things I write, I'm writing this first and foremost for me. I'm writing this because I'd like to figure this out. The creative potential of total improvisation excites me, and I think it has the potential to be just as fun and satisfrying a way to play as a planned campaign.

Collaborative worldbuilding

The most significant implication of total improvisation is that the players will be just as involved in worldbuilding as the GM. Since the GM hasn't planned anything, it'll be up to everyone to put their creative caps on and start thinking outside the box.

For example. Let's say there's a tavern brawl that the players may or may not have started. And let's say one of the players was on the second floor when the brawl started. They come out onto the landing, assess the situation, and decide that the best way to join the fray is to leap off the bannister, grab the chandelier, swing across and land on the other side of the room.

Well, nobody knew there was a chandelier there before the player said it. So now the GM has to decide whether or not accept the player's creative proposal. I think in this case the proposal is harmless enough the for GM to accept outright.

And now that there's a chandelier, we're all thinking about the tavern a little differently. Maybe it's not a dive so much as a fancy-schmancy house of pleasure. Or maybe it is a dive, but it's owned by a retired bandit chief who pillaged the chandelier from a fancy-schmancy house of pleasure. And maybe that retired bandit chief isn't going to be all that happy about a group of strangers starting a brawl and destroying his tavern. And maybe this campaign was in need of an antagonist anyway...

That cause and effect progression is what makes storytelling feel magical and organic.

What I love about this idea is that instead of laboring over a ton of prep, the GM can truly just focus on being a referee.

Problems

The main problem with a totally improvised game is that anything technially goes. But realistically, anything doesn't go, because that would mean any player could just so happen to have a bazooka in his pocket, or the GM could oblierate the party with a very convenient sinkhole whenver she felt like it.

At the end of the day, a lot of this is going to come down to the players' good faith and the GM's judgement. But as long as we have those as a bedrock, I think this will be a ton of fun.

Whereas “Yes, and…” requires you to accept any premise from another player and then build upon it, “Yes, But…” allows you to accept the player’s premise, but then subvert it to support previously established continuity. This is a great tool for players who are collaboratively building a world together through play, rather than playing in a world that their DM created for them.

For example, Player A might say, “Player B, your character is a mercenary, right? I heard that there’s a fortress full of unscrupulous mercenaries in the desert of Anauroch. Did you come from there?”

If Player B hasn’t created a fully fleshed out origin story for their character, and likes Player A’s idea, B might say, “Yes! That’s right, my character did come from Fort Doom in Anauroch.” Now, Player B likes the idea of coming from a fortress, but they had the idea in their head that their character was a noble-hearted mercenary who only offered their services to compassionate people. B might then say, “But, I left a long time ago. The mercenaries in Fort Doom and I didn’t see eye to eye.”

James Haeck

Where to begin?

The way a lot of improv groups do it is to start with a prompt from the audience. In lieu of an audience, D&D groups could source prompts from generators, LLMs, or tables.

Or another way you could do it is just begin playing and see what comes of it. Put the characters in a tavern and have them introduce themselves. Let the conversation flow from there until it touches something that would make a good quest.

PCs

If they find them through improvisation, you'd have to figure out how to find your stats and abilities. Alternatively, after the first session, you could build the character based on how it went. If you're character kept making poor decisions, maybe he has a low Intellgence, etc.

Alternaitvely, you could roll stats on the fly. The first time an ability becomes relevant, the player rolls for it and that becomes their ability score forever. The usual rules will apply (no intellgence below 6 or whatever, etc.).

Spells could work in a similar way. PC spellcasters may pick them at will until they've reached their allowed number of prepared spells. After that, they are stuck with those splles until they are allowed to change them per game rules.

NPCs

For unique NPCs, you could base them on a monster or roll for their stats. Similar to the improvised PCs method above, roll for ability scores and skills as needed. Use d10 for skills, where 10 is zero. Use a d20 for ability scores. This may of course result in some highly underpowered or overpowered NPCs, but fun lies in the extremes.

This also results in somewhat of a gooey framework wherein characters aren't so bound by the rulebook. If a characters winds up with and Charisma score of 8 but a +10 Persuasion skill, you will be forced to find a creative solution to this problem. To use our example, you'd have to conclude that such a character's ability to sway people can't possibly be based on charm. So okay, maybe he has a penchant for blackmail. Maybe, like Jack Sparrow, people find him kind of repulsive, but he has a way of combinging leverage with circumolcution such that he confuses people and they wind up giving him what he wants without even realizes it. Or maybe he wines and throws trantrums .

You can see roleplaying as putting a character over yourself, and putting yourself into a character. If you we did away with skills, and the outcome is determined by how well the character actually persuades the NPC. The numbers are there to settle any disputes.

That's called creativity. This is the heart of D&D and all roleplaying games, in my opinion. The rules are there when we need them, but we are not to let them do the thinking for us.


The first time a character uses an ability, roll for the score using the D&D method: Roll four six-sided dice and sum the highest three. The total is the character's ability score.

For unique NPCs, you could base them on a monster or roll for their stats. Similar to the improvised PCs method above, roll for ability scores and skills as needed. Use d10 for skills, where 10 is zero. Use a d20 for ability scores. This may of course result in some highly underpowered or overpowered NPCs, but fun lies in the extremes.

This also results in somewhat of a gooey framework wherein characters aren't so bound by the rulebook. If a characters winds up with and Charisma score of 8 but a +10 Persuasion skill, you will be forced to find a creative solution to this problem. To use our example, you'd have to conclude that such a character's ability to sway people can't possibly be based on charm. So okay, maybe he has a penchant for blackmail. Maybe, like Jack Sparrow, people find him kind of repulsive, but he has a way of combinging leverage with circumolcution such that he confuses people and they wind up giving him what he wants without even realizes it. Or maybe he wines and throws trantrums .

You can see roleplaying as putting a character over yourself, and putting yourself into a character. If you we did away with skills, and the outcome is determined by how well the character actually persuades the NPC. The numbers are there to settle any disputes.

That's called creativity. This is the heart of D&D and all roleplaying games, in my opinion. The rules are there when we need them, but we are not to let them do the thinking for us.

Quests

Begin with a prompt. There are any number of generators to help with this, but I like this Perchance Adventure Prompts.

Wing it. In the spirit of true improv, just start playing and see what comes of it. A lot of improv sessions begin with a conversation. You could do the same. Put the characters in a tavern and have them introduce themselves. Let the conversation flow from there until it touches something you think would make a good quest.

Encounters

There are probably two ways to handle improvised encounters. The first is with an encounter generator like Goblinist.

You could also do this on your own, if you have a good enough working knowledge of monsters. You'd have to be careful to be fair about it. But if you've played enough dnd, you probably have a good enough feel for it that this should be possible. It would be just like any other internalized skill.

Once you have your monsters, you can add them to the D&D Beyond Encounter Builder and manage the encounter that way.

Notes

Further reading

Generators