A one-shot adventure has to do everything a campaign does, from hook to rising action to climax to resolution, in a single 3–4 hour sitting, with characters the players met an hour ago. The constraint is the craft, because every scene has to earn its time. Below is the structure that fits the clock, the hooks that get players moving in minutes, and a node-based layout borrowed from Justin Alexander that makes even a one-shot non-linear without adding prep.
How do you structure a one-shot?
Build a complete three-beat arc of setup, conflict, and resolution around exactly one core problem, and front-load the exposition so play starts inside the first fifteen minutes. One hostage to rescue, one ritual to stop, one mystery to solve. Subplots are a campaign luxury. In a one-shot, every thread either converges on the climax or becomes a thread you’re leaving hanging at midnight.
The 4-hour pacing template
| Block | Time | Contents | Cut here if running long? |
|---|---|---|---|
| Open | 0:00–0:20 | Hook delivered, stakes stated, party moving | Never — protect this |
| First scene | 0:20–1:05 | First obstacle or investigation; teaches the adventure’s logic | Trim, don’t cut |
| Middle scenes | 1:05–2:35 | Two scenes of escalation, choice, and discovery | YES — this is your flex zone |
| Pivot | 2:35–2:50 | The revelation that reframes the climax | Never |
| Climax | 2:50–3:40 | Final confrontation; all threads converge | Never — start it on time even if scenes remain |
| Resolution | 3:40–4:00 | Outcomes, epilogue beats per character | Compress to 10 min if needed |
The operating rule here is that the middle is the only expendable part. When you’re behind at the two-hour mark, cut a middle scene whole rather than rushing everything. Players never miss a scene they didn’t know existed, but they always feel a rushed climax. Movable objectives help too, and the session-timing tools in our Lazy DM toolkit guide were built for exactly this problem.
Hooks that start the adventure in one sentence
The fastest hook formula is Sly Flourish’s Simplest Quest Model: an NPC wants you to do something at a location. Make the want urgent and concrete, like a body to recover before nightfall or a ransom due at dawn, and deliver it in the opening scene rather than after a tavern crawl. A one-shot has no time for hook-shopping. The players already agreed to the premise when they sat down, so let the characters agree fast too.
You can strengthen any hook by attaching one of three pressures: a deadline (the ritual completes at moonrise), a personal stake written into a pregen’s backstory, or a visible consequence (refugees fleeing the thing you’re about to fight). Need the NPC fast? Our generators will hand you one with a name and a quirk in seconds.
A hook that worked at my table
One hook that worked really well for me was starting the players right at the problem instead of making them go find it. I had the party arrive at a small settlement where people were already scared, something had already happened, and the NPC basically said, “We do not have until morning.” That got everyone moving fast because there wasn’t really anything to debate — the danger was visible, the clock was ticking, and within maybe ten or fifteen minutes we were already in the first real scene.
Node-based design: non-linear structure without extra prep
Linear one-shots are fragile. If players refuse scene two, scenes three through five collapse with it. Node-based design, articulated by Justin Alexander at The Alexandrian, fixes this by prepping the scenario as nodes (locations, NPCs, encounters) connected by leads, which are the clues pointing from one node to another, instead of a fixed sequence. Players choose which lead to follow. You prep the pieces once and any order works.
For a one-shot, a five-node layout is plenty:
- Start node, where the hook lands, seeded with leads to nodes 2 and 3
- Two investigation nodes, each revealing part of the truth and a lead onward, reachable in either order
- Pivot node, the revelation, with both investigation paths leading here
- Climax node, unlocked by the pivot
Alexander’s safety mechanism, the Three Clue Rule, says every conclusion players must reach needs three separate clues pointing at it, because they will miss one and misread another. In a one-shot you can compress it down to this: every node should hold at least two leads forward, so no single failed check dead-ends the session. The structure also pairs naturally with cutting for time, since dropping an investigation node just means consolidating its leads into the other one.
My node prep in practice
For node prep, I keep it pretty simple. I do not want a giant flowchart that turns into homework, so I usually think in terms of a few locations, what clue or problem lives there, and where that clue can point next. Even just sketching it out as boxes and arrows is enough. The big advantage is that when the players go somewhere “out of order,” I am not scrambling — I already know what that place reveals, and I can just let the session bend around their choices instead of trying to drag them back onto my original path.
Encounters and endings
Budget combat honestly. A meaningful 5e fight eats 45–60 minutes of real time, so a 4-hour one-shot holds two of them, maybe three. Make each one mechanically distinct, say one fight with environmental stakes, one social or stealth set piece, and one climax, rather than three rooms of hit-point attrition. Design the climax with multiple resolution verbs while you’re at it. It can be fought, sure, but it should also be possible to bargain with it, trick it, or escape it, so whatever the table has become by hour three still has a path to an ending.
End with epilogues. Thirty seconds per player, asking “three months later, what is your character doing?”, turns a one-shot from a session into a story, and it tends to be the part players quote back to you later. If you’re building in the horror genre specifically, our horror one-shot guide adapts this whole structure for dread and pacing.
One-shot prep checklist
- One core problem, statable in a sentence
- Hook with a deadline, stake, or visible consequence
- Five nodes, two leads out of each
- Pacing template marked with your flex scenes
- Two to three mechanically distinct encounters
- A climax with at least three resolution verbs
- Epilogue question ready for each player