Designing the Heist

Your game has combat, and combat works. Combat has rounds and actions and hit points going down and all these lovely interlocking gears that create tension and demand decisions.

But then your players want to sneak into a building, or maybe seduce an NPC.

And you go: “Uh. Mmkay. Roll Stealth.” Or, “What’s your Charisma?”

The trad game approach is one roll, with a binary outcome, pass/fail. So the party does thirty minutes of planning, and it’s over in five seconds of resolution. The heist that took all that table time to conceive lives or dies in a single roll. With swingy D20s this is often quite dissatisfying.

The problem is “how do I make a scene out of something that isn’t a fight?” We want to give a negotiation or an infiltration the same mechanical weight, the same dramatic shape, the same narrative rising tension that combat gets for free.

4E was polarizing, let’s get that out of the way. But, among other things, D&D 4th Edition tried to solve this. 4E called them Skill Challenges. My game CRAWL:CORPSEC tries to solve this too, and I call them Ops. There are different solves for the same problem. They arrive at almost opposite answers. Why they are so different is, I think, kinda interesting.

Skill Challenges work like this: the GM sets a target (like, 6 successes before 3 failures). Players take turns picking skills, justifying why that skill applies, and rolling. Successes accumulate and failures accumulate. When either bucket fills up, the scene resolves for better or worse.

Pros: It’s clean. It’s learnable in five seconds. It’s transparent to the players, because you always know the score. “We need two more successes, we can only fail once.”

And so… it ends up feeling like a spelling bee. Or whatever, a game of parchesi. Something like that.

Ok, fine, not always. Sometimes it sings. But often, very often, what happens is: players stop thinking about the heist and start thinking about which skill gives them the best modifier. “Can I use Athletics here? I have a +9.” The mechanical optimization impulse (which, as I’ve written before, is a real and valid play drive) eats the fiction alive.

One issue is that 4E Skill Challenges are accumulative. You’re filling two buckets and each roll gets you closer to success or failure, and you know it. The emotional arc is momentum: “We’re winning! We’re almost there!” That’s… heroic fantasy, which is fine for D&D. The heroes are supposed to build momentum.

But that ain’t how infiltrations feel. It’s not how cons or grifts feel. It’s not how any operation against a responsive system feels.

When you’re sneaking through a facility, you don’t feel like you’re accumulating victories. You feel like you’re spending something. Your luck, your window, your time budget, your preparation, your margin of error. Every second you’re inside, the probability of getting caught ticks upward. The emotional arc isn’t momentum. It’s erosion.

That’s the genre expectation talking. (Genre expectations again. Always genre expectations. The sign out front says what the ride should feel like.)

The fancy game designer term for what 4E is doing with Skill Challenges is dissociative mechanics. The bookkeeping (six successes, three failures, tally marks on scrap paper) is dissociated from the fiction. Nothing in the game world corresponds to “we have four successes.” There’s no in-the-fiction reality to the bucket filling up. It’s pure metagame infrastructure. The players are aware, at all times, that they’re playing a minigame that’s been overlaid on the fiction like an acetate transparency sheet. “We need two more successes” said nobody in the dungeon. The characters don’t know what “successes” are. Worse, nobody at the table is thinking in story terms when they say it. More like spreadsheet terms.

This is ironic (sorta). Skill Challenges are supposed to give non-combat scenes mechanical weight and texture. But the mechanic draws players out of the fiction. The more transparent (or, if you prefer, intrusive) the system, the more it becomes the thing you’re engaging with. You’re not heisting, so much as rolling to fill up buckets.

Trad games had no answer for extended non-combat scenes. “Make a Stealth check,” and the rest was in the whitespace of the game design. “Rulings not rules.” 4E had a different answer. Then there’s Blades in the Dark. It solves this quite differently, and BitD is doing something intentional and nifty, but also isn’t quite what I want.

BitD uses Clocks, which are just segmented circles that fill in as situations develop. Really, they’re just a countdown bar. The “infiltration” clock might have six segments. Each complication, each failed roll, each alarm fills a segment. When the clock fills up, the consequence triggers. Players see the clock, and they see it filling. This makes visible (and hopefully palpable) the fiction escalating in tension.

BitD also has Position and Effect (Controlled, Risky, Desperate) as an explicit, mandatory conversation and negotiation between the GM and the player before each roll. The GM says “this is Risky” and you can push for better Effect or accept worse Position. It’s negotiated, transparent and collaborative.

This is better than 4E. Much better, IMHO. Because the Position/Effect conversation is about the fiction (you’re discussing what your character is actually doing and what the risks actually are) rather than about which skill gives you the best modifier. Plus, Clocks are visual, intuitive, and they communicate narrative pressure fluently.

But… and this is taste, not criticism… BitD’s transparency is still transparent. It’s still dissociated. The players see the clock and they can count the segments. They know exactly how much runway they have, and they can optimize around it. “The clock is at 4/6, so I’ll take the Desperate action now because we’ve only got two ticks left.” That’s smart play. But it’s also metagame reasoning wearing a narrative hat.

Also, the Position and Effect labels are spoken aloud at the table. The GM says “Risky.” The player hears a game term, not a fictional state. It’s more diegetic than 4E’s success/failure tally, but it’s still a mechanical label that exists outside the story. The characters don’t know they’re in a “Risky” position. They know someone’s checking badges more carefully. There’s a gap between the label and the experience, and the label is the thing that reaches the player’s ears.

BitD made me think. What if you hid the clock? What if Position was never a word spoken at the table?

Ops in CORPSEC inverts the accounting. Instead of filling a bucket of successes like 4E, you’re draining a pool of uncertain operational margin. You start with an Ops Pool number (based on a skill roll), and the GM secretly chews away at it every time you do something risky, fail a check, trigger a response, or just take too long.

When it hits zero, you’re blown.

The GM doesn’t tell you the number. They don’t tell you anything mechanical. They never say “you’re at Risky position” or “your Pool is low.” They say: “The guards are relaxed, joking about last night.” Or: “Radios crackle with your description.” The Position labels (Icy, Controlled, Risky, Desperate) are GM-facing tools, used as a cheat sheet for calibrating narration. The players never hear them. All the players get is the fiction.

This does a couple of things please me.

It forces the GM to narrate rather than adjudicate. In a 4E Skill Challenge, the GM’s job is bookkeeping: mark the success, mark the failure, announce the tally. In BitD, the GM announces a label: “Risky.” In an Op, the GM has no mechanical shorthand to fall back on. They can’t say “Risky.” They can’t say “your Pool is low.” They have to describe a world that’s reacting to you: “The facility releases more recon drones.” “Your contact’s smile just faded.” “The terminal accepted your credentials but took two seconds longer than it should have.” The mechanic is invisible to the players. All they experience is fiction that’s getting worse. The GM can’t phone it in. The position labels on their side of the screen are just a calibration tool for how bad “worse” feels.

It gives players control over pacing without giving them control over outcome. In 4E, the pace is fixed. It’s a 6-before-3 challenge, and that’s how long it takes. In Ops, the player(s) decide how fast you burn. Go loud and your Pool evaporates. Play slow and methodical and you buy yourself room. And here’s my favorite bit: you can divert Effort from a successful check back into your Pool. You can trade immediate effectiveness for operational longevity. That’s a real choice. That’s a choice that matters. That’s a decision where knowing when to pull the lever is the “game skill,” the human Player skill (as opposed to the character skill… which I’ve written about before.)

It models adaptive opposition. A maglock doesn’t get harder because you’ve been picking it for a while. But the research facility absolutely does. Cameras track anomalies. Security rotations adjust. Pressure sensors learn footstep patterns. Even without true AI, a complex security system’s interconnected responses create something that feels adversarial. The Op mechanic encodes this: the depletion die gets bigger when you do risky things, when you escalate to violence, when you leave bodies. The adversary-system learns. 4E Skill Challenges don’t model this at all. The DC is the DC, whether it’s round one or round six.

It scales consequences to timing. Failing a check at Icy position is a warning sign. Failing at Desperate is a catastrophe. The same action (picking a lock, telling a lie) has dramatically different stakes depending on when you attempt it. Tension escalates naturally, without the GM having to artificially ratchet anything.

However, there are a couple things 4E and BitD do that CORPSEC Ops don’t do.

“Get X successes before Y failures” takes five seconds to explain. Even BitD’s “the clock is filling” is immediately intuitive. You can see the pizza slices filling in. Ops requires understanding Pool, Position, depletion dice, the Effort-to-Pool diversion, scaled consequences, and the GM’s hidden tracking. It’s a more sophisticated machine and it demands more from everyone. CORPSEC is not a mall food court game. (Can you tell?)

But there are gaps in my design of Ops. This one bugs me: 4E Skill Challenges are collaborative by construction. Everyone can roll, everyone can contribute, and the party succeeds or fails together. BitD handles this well, too. Anyone can jump in with an assist, a setup action, or a flashback. For CORPSEC, Ops can spotlight a single character… the hacker during a Tech Op, the face during a Social Op… while the rest of the party sits there. I haven’t solved this elegantly yet. The assistance rules apply, sure, and a good GM will cut between operators, but the mechanic itself doesn’t force ensemble engagement the way Skill Challenges or BitD’s teamwork actions do. That’s a gap. Or a feature. I’m not sure yet.

4E directly, mechanically enforces using different skills. You can only use the same skill X times, or secondary skills unlock bonuses. This encourages “creative” problem-solving: “How does my Religion skill help with this chase?” Ops doesn’t do this. You could theoretically roll Cyber twelve times in a row. The variety comes from fiction and GM judgment, not mechanical structure. That’s fine when the GM is strong, but may be problematic when they’re not. (As a side note, I put “creative” in quotation marks above, because in practice this “creativity” ends up looking like, “Ok, what’s my highest skill bonus?”)

BitD’s Position/Effect negotiation is pretty admirable design. Before a roll, there’s a negotiation about what’s at stake, what’s possible, and what the risks are. That conversation creates shared understanding and establishes stakes. It gives the player informed agency over an approach. Arguably it leaks too much player metagame perspective into the fiction, but… I digress. Ops doesn’t have this explicit back-and-forth. The GM tracks the Pool behind the screen. Players act; the world reacts; they find out how bad it is through narration. That’s diegetically coherent, but it gives the player less metagame agency over risk framing. BitD lets you lean into Desperate on purpose, because the XP reward is there. In Ops, Desperate just means you’re about to get caught.

These three systems sit on a spectrum of diegetic commitment.

4E Skill Challenges are dissociative, full stop. The mechanic exists entirely outside the fiction. Successes and failures are metagame objects. The players engage with the system, and the GM pastes fiction over the result.

I’d put it this way: BitD Clocks are semi-diegetic; medium-rare dissociated. The Position/Effect conversation is about the fiction. Clocks are a visual metaphor for narrative pressure. Labels are spoken aloud as game terms, and the clock is visible, countable, optimizable. As a player, you’re going to look at them. So, there’s a seam between the mechanic and the story.

Ops is fully diegetic. The players never interact with a mechanical abstraction. They interact with fiction that’s being generated by a mechanical abstraction they can’t see. The seam is hidden behind the GM’s narration. There is no “game term” that reaches the player’s ears, just the cluster-f*** getting worse.

Here we see a spectrum of design choices. BitD’s transparency is a feature for its genre: scrappy criminals who are meta-aware of their position, who deliberately take crazy risks for XP. 4E’s transparency is a feature for its genre too: superheroic adventurers striving towards success. CORPSEC Ops’ opacity is a feature for CORPSEC’s genre: professionals who don’t know how close they are to getting caught, who can only read the room, read the body language, read the two-second delay on the terminal, and make a gut call.

Genre drives mechanics. Expectations are prior to mechanics.

It’s worth acknowledging that all three systems live or die on the same thing: GM narration.

If a 4E GM just announces “success, failure, success, you win,” the Skill Challenge is a two-day-old tuna sandwich. If a BitD GM says “Risky” without painting a picture of why, the Position/Effect conversation is just as dead. If I just describe the same corridor twice, the Op is dead. Every one of these systems is a hook for the GM to hang fiction on. The mechanic generates the pressure, and the GM generates the experience.

But, Ops requires the GM to narrate, because the hidden Pool means narration is the only channel of information the players have. You can’t communicate “the system is catching up to you” without describing what that looks like in the fiction. 4E lets you get lazy. “Mark a success.” Done. Next. BitD is in between. The labels do work, but a good GM will narrate through them rather than with them.

Whether obligatory narration is a feature or a bug depends entirely on how much you trust the people at the table. I’m designing for GMs who like this modality. Which is, again, a niche decision. Mall food courts don’t push anyone. That’s why they’re popular. Hello, 5E.

But I am determined. Dogged. I keep making design decisions which are coherent, but slide towards a niche. I know. It’s a curated dinner party instead of the mall food court. Every time I make one of these niche-y design choices, I can hear the ghost of 5E whispering: “Sure Steele, sure. But I have millions of players.”

But your heist is dissociative.

Try CORPSEC instead.