Heist sessions are some of the highest-energy nights a tabletop group will ever have — when they work. When they don’t work, they collapse in a way that’s specific to heists: three hours of players drawing maps, arguing about insertion vectors, and never rolling a die. The score never happens. Everyone goes home tired.
This article is for the GM who’s about to run a heist — in any system — and wants it to actually pop. I’ll cover the failure mode (the planning trap), the structural fix that Forged in the Dark games pioneered, and the four things that have to be true at the table for a heist to actually feel like a heist.
The Planning Trap
Heist movies feel like heists because the audience watches the plan come together in a montage. The crew checks blueprints, picks locks, casses guards, lifts a keycard, and then we cut to the job. We never sit through twenty minutes of the crew arguing about whether to enter through the loading dock or the roof.
Tabletop RPGs default to the opposite. The GM presents the building, the players want to be smart, and the next two hours become a planning meeting. Every contingency gets debated. Every variable gets a roll, or worse, doesn’t — because someone says “before we go in, let me first…” and the loop begins again. The session ends with the crew still outside the target, demoralized, and the GM exhausted.
The planning trap is structural. It’s not the players being slow. It’s a system that rewards planning and punishes execution.
Traditional RPGs make planning safe (no consequences for thinking) and execution dangerous (every roll has stakes). So the rational player optimizes by planning forever. The fix isn’t to ban discussion — it’s to make planning less safe and execution less binary.
The Flashback Fix
The single most important mechanical innovation in heist RPGs is the flashback. It’s the load-bearing column for every modern heist game and it works like this:
Skip the planning. Cut straight to the score. When a player hits a problem — a locked door, an unconscious guard they need a keycard from, a sniper on the roof — they say “earlier, I…” and describe a flashback to the prep they did. That flashback either auto-succeeds or gets a roll right then, with stakes attached.
In Blades in the Dark, flashbacks cost stress (the meta-currency for pushing past limits). The cost is proportional to the audacity of the flashback. Free keycard? 0 stress. “Actually, I planted explosives in the vault last week”? 2 stress. The economy keeps players from auto-winning while still letting the score move.
This single change rewires the whole conversation. Planning doesn’t happen at the table — it happens inside the score, exactly when it’s relevant, with stakes attached. The pace stays high. The fiction stays cinematic. The crew gets to be cool.
If you take one thing from this article: your heist game needs a flashback mechanic. If your system doesn’t have one, bolt it on. Players bank a small pool of “prep tokens” (3-5 per score), each one spends to retroactively make a piece of prep true. Make it auto-succeed at small stakes and require a roll at large ones. Run a session that way and you’ll see the difference inside twenty minutes.
Designing the Score
A heist score has four moving parts. If any one is missing, the table feels it.
▸ The Target
Specific. Nameable. “The Bishop’s personal ledger” beats “some documents from the church.” The target gives the players something to hold in their head. If your players can’t describe what they’re after in one sentence, you haven’t designed it tightly enough yet.
▸ The Obstacle
Not a list of guards. A specific, animated threat. “Captain Mara Voss runs night security and she does not miss” beats “there are twelve guards.” Name the antagonist. Let the crew learn her name through prep. The score is now a duel with a person, not a logistics problem.
▸ The Complication
Something that turns this from a heist into a story. The vault is being moved that night. The Bishop is hosting a wedding. A rival crew is hitting the same building at the same time. Complications are not added during the score — they’re built in before play begins. They guarantee the crew won’t just execute a clean plan.
▸ The Cost
What it costs the crew if they pull it off. Heat from the city. A betrayed contact. A debt to the faction that helped them. A heist that costs nothing is a video game level. A heist that costs the right thing is a story.
Build all four before the players sit down. Write each on an index card. The rest of the score emerges from play.
Running the Heat: Escalation Clocks
The other mechanical innovation that heist RPGs gave the rest of the hobby is the progress clock — a visible, fillable circle representing some force that’s opposing or advancing the score. Heat is the most famous flavor: how aware the authorities are of the crew, ticked up whenever the crew is loud, sloppy, or visible.
Clocks do two things at once. They give the GM a way to track threat without secret notes (the players can see the heat rising). And they give the crew a measurable thing to push against. “Heat is at 4, one more incident and the Bluecoats kick the door in” is a much sharper pressure than “the guards are getting suspicious.”
How to use clocks if your system doesn’t have them: draw a circle on the table. Divide it into 4, 6, or 8 wedges. Name it. Fill a wedge when something happens that should fill it. Tell the players what happens when it’s full.
Three clocks I run every heist:
- ▸ Heat — how aware the cops/regime/authorities are.
- ▸ The Score Itself — progress toward the target. Fills with successful actions.
- ▸ The Complication — whatever’s making this heist harder than a normal one. Fills on a schedule. When it fills, the complication actually happens.
Three visible clocks turn every action into a tradeoff. Want to take this faster? Heat fills. Want to be quiet? The score clock barely ticks. The crew sees the math and makes interesting choices instead of optimizing on a single axis.
Crew Composition: Specialists Beat Adventurers
One reason D&D parties struggle with heists is the adventurer model. A fighter, a wizard, a rogue, and a cleric is a combat unit, not a crew. A crew is a Driver, a Hacker, a Hitter, a Mastermind — people defined by their role in the job, not their fantasy archetype.
When you build characters for a heist campaign, ask each player two questions:
- ▸ What’s the one thing your character does that no one else in the crew can?
- ▸ What does your character refuse to do?
The first question forces specialization. The second creates faultlines — moments when the crew has to make a choice about a thing the specialist won’t touch. Both produce drama. A crew where every character can pick a lock and shoot a gun is a crew where every score feels the same.
Systems Actually Built for This
If you’re running heists in a system that doesn’t support them, you can bolt on flashbacks and clocks and a complication budget — and that works. But it’s worth knowing the systems where this is the default mode.
Blades in the Dark
The originator. Crew-based industrial-fantasy heists with the cleanest expression of flashbacks, clocks, and downtime in the form. If you’ve never run a heist game, this is the system to learn it on.
Scum and Villainy
Same engine, applied to a sci-fi space crew. The job structure is identical but the missions feel like Firefly episodes instead of Peaky Blinders.
Leverage RPG
Built directly on the TV show. Has its own flashback-equivalent (“the plan”) and crew-role mechanics. Best for groups who like their heists comedic, character-forward, and structured around righting wrongs against corporate antagonists.
Omega Project: Infamy
My own game (mentioning it here because it’s in this category). Forged-in-the-Dark heist game set in a surveillance-state dystopia called Amerius. Flashbacks, Heat clocks, and Vice are core; the twist is a public-credit system (S3) that tracks every operative and constrains the crew’s next move. On Amazon as the Hooligan Kit.
The Practical Checklist
If you’re running a heist next session, here’s the four-bullet pre-flight:
- ▸ Write the target, the obstacle (named antagonist), the complication, and the cost on index cards before the session.
- ▸ Set up three visible clocks: Heat, Score Progress, Complication. Draw them on the table.
- ▸ Tell players they can flashback. Give them a budget (5 prep tokens, or 9 stress, or whatever your system supports). Make this rule out loud at the top of the session.
- ▸ Cut directly into the score after one paragraph of fictional setup. No planning phase. The crew is already at the target. Begin.
The crew will be uncomfortable for ten minutes. Then they’ll figure out the flashback economy. Then the table will accelerate. Then someone will say “wait, this is actually fun” and you’ll have done it.
Further Reading
- ▸ Every Forged in the Dark Game Worth Playing in 2026 — the full landscape of systems built on this engine.
- ▸ Take the Operative Quiz — two minutes, find out what kind of crew member you’d run.
- ▸ About Ragesix Games — the design influences behind OPI.