Playing the Discworld RPG: As Full of Life as an Old Cheese

Some settings feel like they ought to be RPGs. The Discworld is one of them. It’s already about people blundering through plots they don’t fully understand, improvising wildly while the universe throws Consequences at them in the name of narrative satisfaction. That’s a Tuesday night game, right there.

Like a lot of nerds of a certain vintage, I read Pratchett at exactly the wrong age for developing a healthy relationship with authority. His books happily rewired my sense of humour and ethics: be kind, be sceptical, and whenever possible, make the pun. So a Discworld RPG is not just another licensed game to me; it’s poking at the part of my brain that still quotes Vimes under stress.

We’re also in a very specific cultural moment: late-stage licensed RPG glut, where everything from prestige TV dramas to 1980s toy lines is getting a “cinematic roleplaying experience.” I wrote recently about that pile-up and how tie-in games often struggle to justify their existence beyond “you like the thing, yes?” Discworld: Adventures in Ankh-Morpork has a harder job than most. It’s stepping into a beloved, finished body of work, with a fandom who can tell when the tone is wrong by the second adjective. It can’t just be “D&D, but with funny hats.” It has to feel like Pratchett.

So: does it?

The Narrativium Engine: Rules in a World That Runs on Story

Discworld: Adventures in Ankh-Morpork is built around the Narrativium system, a lightweight, story-first engine whose core rule is basically: the story comes first; the mechanics help you get out of the way of it.

Characters are defined by Traits: freeform phrases that capture who you are and what you can get away with: background, species, organisation, niche, quirks, and whatever nonsense you accumulate along the way. “Built like the proverbial shithouse,” “Knows everyone on Treacle Mine Road,” “Designated distraction”, all of these are Traits, and the book encourages you to stretch language until it squeaks.

Mechanically, everything revolves around tests. When the outcome of an action is interesting and uncertain, you:

1. The GM sets the scene.
2. You describe what you’re doing and justify it with a single Trait.
3. The GM decides which Outcome die you roll: d4, d6, d10 or d12, based on how appropriate the Trait is and how much experience you plausibly have.
4. You roll that die.
5. The GM rolls the Narrativium die, always a d8, representing the story pushing back.
6. Dice are compared:
   • If your Outcome die is higher, you succeed.
   • If the Narrativium die is higher, you fail and face Consequences.
   • If they’re equal, you still succeed, but also get Consequences, i.e. the world pushes back.

Degrees of success and failure (the margin between the dice) are used to flavour how well things go. A 12 vs 1 is a triumph; a 6 vs 5 is scraping by with your dignity in tatters.

The real fun is in Consequences. The book leans into them hard: failed tests spawn new Traits like Stinks like Foul Ole Ron or A laughing stock amongst thieves, which then come back later to either hinder or (with a bit of verbal contortion) help you. This is very much “fiction first, comedy second, then mechanics,” in that order.

Luck, the Lady, and Failing Forward

Each character starts a story with 4 Luck, a shared currency that measures how long the Lady is willing to humour your nonsense. Luck can be spent to:

• Help another character on a test (you spend 1 Luck to roll your own Outcome die and potentially replace theirs, better or worse).

• Fuel magic (more on that in a second).

• Avoid Consequences from a failed test, by narrating how you get improbably lucky.

Once you’re out of Luck, the world starts hitting back much harder. The GM is encouraged to award Luck for good roleplay, ridiculous plans, or moments that feel very Discworld, but with a warning not to give out so much that the stakes vanish. In my one-shot (more on that below), the 4 per character was already a lot and I didn’t need to give any more out.

You’re also reminded, repeatedly, that this is a fail-forward game. The GM is told not to block progress with a failed roll, but to make failure about timing, collateral damage, social fallout, or interesting new problems. “Find the fun in failure” could be printed on the GM screen.

In practice, that combination, swingy dice vs a fixed d8, plus a finite Luck pool, makes the system feel more unforgiving than the light tone implies. We’ll come back to that when I talk about Hogswatch.

Group Tests, Million-to-One Chances, and Other Discworld Bits

The game supports group tests, where everyone rolls their own Outcome die against the same Narrativium roll. If only one success is needed, you compare the highest Outcome die to the d8; if one failure would doom the attempt, you compare the lowest. Everyone else’s results add colour – extra successes, extra mishaps.

There’s also a Million-to-One Chance rule: when a plan is truly audacious and feels like the sort of stunt that should only work in fiction, the GM can just say, “Fine, it works,” possibly with a little extra d10 flourish to see if the stars really align. The text is very clear that you shouldn’t let players farm these; as Pratchett taught us, you can’t cheat narrative causality without it noticing.

On top of that, the GM gets tools like re-rolling the Narrativium die using an NPC Trait (the Librarian’s wrath, say) or awarding new Traits mid-scene to reflect how the story has twisted.

Magic, Violence, and the Things Pratchett Didn’t Linger On

There are no traditional combat rules. You don’t get a hit point track or initiative order; a brawl in the Drum, a chase down the Shades, and a shouting match with Vetinari are all handled with the same test-and-Consequence structure. The book is explicit: violence exists but is rarely the point, and detail is kept deliberately fuzzy in the Pratchett style.

Magic likewise isn’t a shopping list of spells. Instead, the text has a section on “Magic and Mechanics” that emphasises limits, side effects, and cost. Magic almost always costs Luck and tends to produce its own special brand of magical Consequences, even on a success. It’s closer to “nudge the story and accept that reality may nudge back harder” than “I cast Fireball, again.”

If you’ve spent time in trad systems, this can feel alarmingly vague at first. If you’re happy to improvise and lean on tone, it’s exactly the right level of structure.

Building a Party in Ankh-Morpork

Character and party creation starts with the organisation you all belong to: City Watch, Unseen University, The Times, Beggars’ Guild, Seamstresses’ Guild, Fools, “Visitors to Ankh-Morpork,” and so on. Each organisation comes with a short pitch about the kinds of stories it attracts (heists, magical mishaps, domestic politics, petty mysteries) plus example goals.

You then create the party as a shared entity:

• Party Name (ideally a pune or play on words; the book provides many gloriously stupid suggestions).
• Party Past (how long you’ve worked together, how you got into trouble last time).
• Party Present (your slogan or immediate situation).
• Party Goal (what you’re trying to achieve this story).

The party itself is a Trait you can invoke, being the City Watch or a Times reporter tangibly matters in play, giving you authority, contacts, and excuses to poke your nose into trouble.

Individual characters then layer on:

• Background (what you did before, in Discworld terms).
• Species (human, dwarf, troll, etc.).
• Niche (what you bring to the group).
• Quirks (little personality hooks that inevitably become Traits when you least expect it).

There’s also the idea of Core moments: big character beats that, once they happen in play, crystallise into powerful Traits you can call on for dramatic spotlight scenes. The game assumes short-form stories but gives you rules for Returning Characters if you want to carry your watchmen, reporters, or wizards forward between adventures.

In terms of GM support, there’s a solid “Running the Game” section that talks through scenes, safety tools, Pratchett’s way of handling grim material (off-screen, with humanity and humour), and practical advice on failing forward without making characters feel incompetent.

The City on the Page: Ankh-Morpork as a Play Space

The setting chapter focuses tightly on Ankh-Morpork, “brawling city of a hundred thousand souls” plus several hundred thousand people, set in the Century of the Anchovy, just before Going Postal and Thud!

You get:

• An overview of the city’s history and current “Now,” including guild politics, technological advances (clacks, print), and Vetinari’s delicate balancing act of keeping things just functional enough.
• Key locations: Pseudopolis Yard, Sator Square, the Patrician’s Palace, Pearl Dock, grim little pubs that would make even the Mended Drum look classy.
• A chunky “People and Places” section full of NPC write-ups: Vimes, Vetinari, Death, the Watch, The Times staff, the various guild trolls, dwarfs and undesirables you’d hope to bump into.
• Three full adventures at the back (including Little Women, which does some impressively careful work around dwarf gender politics and bigotry in a way that feels very Pratchett).

It’s all written in a tone that sits comfortably close to the novels: footnotes, daft asides, and the sense that the authors like the setting and understand what it’s for. It’s not a huge world book. It’s “Adventures in Ankh-Morpork,” not “On the Whole of Discworld”, but for play purposes, that focus is a strength. You have enough fuel to run watch procedurals, guild heists, newsroom farces, and Hogswatch specials without having to reread half the series. There’s also a bunch of additional supplements in the Kickstarter package which I’ll be honest I haven’t gotten to reading yet.

Hogswatch at the Table: The Night Belief Went Missing

I actually put the game on the table for a Hogswatch one-shot, using a scenario I wrote called The Night Belief Went Missing. The pitch: someone is siphoning belief through stolen children’s teeth and a dampening signal on the clacks, and if the city doesn’t get its festive nonsense up to minimum standards by midnight, Hogswatch starts to come apart.

The scenario was structured as three scenes:

1. Vanishing wish lists at the Post Office grotto.
2. Teeth, towers, and The Times in a little investigative middle.
3. A big finale in Sator Square or the Opera House, where the players stage a pageant or reality-heist to kick belief back into shape.

None of these locations actually worked out once the players had devised their troupe:The Layt Shyft of the Streyt Sweypers Guild. But they gave me a skeleton framework to at least try to make some plot magick happen.

My party, including Dusty the sweeper (a golem), Florance (a dwarf), Lightyr (a boy-turned-into-a-candlestick), and Cobblestone (a troll), immediately got into the spirit of things by naming the first tavern the “Baa Bar” and having Dusty damage the door frame just by entering. Inside, they found a gloomy crowd and a despondent urchin; attempts to rouse the bar with improvised street-sweeping songs went down so badly they were thrown out into the snow. Perfect Discworld failure: the right idea, utterly wrong execution.

In Sator Square, Dusty took offence at the Auditors “stealing” people’s possessions, in reality, an officious type with a clipboard trying to get the stalls properly tidied and logged. Some gentle table persuasion and a decent roll later, Dusty and gang ran off the auditor, and the “Belief meter” I was tracking in the background ticked up.

The finale saw Dusty doing some perfunctory sweeping in a warehouse while Florance analysed the giant ledger contraption the Tooth Fairy Organisation was running, Lightyr tried (and mostly failed) to set the paper feed on fire, and Cobblestone discovered a crate of stamps. Which to a professional stamp licker was ambrosia.

The actual victory condition ended up being Dusty shoving a handful of small rocks (previously vomited by Cobblestone – long story) into the machinery whilst Florance broke her foot kicking the metal contraption. The rocks jammed the works, broke the clever little extraction scheme, and Hogswatch was saved by mild vandalism and geology.

It was mad, fun, and very Discworld. The players went big on characterisation, weird plans, and punning. At one point I wondered if they might side with the Tooth Fairy Organisation but they talked themselves round to stopping them isntead. Each of the characters got to do their thing. As the player’s later recalled: “Lightyr got to commit arson. Best.” and “Yes, got to smack someone with a broom, always a good ending.”

The system absolutely supported that palette-cleanser energy: one roll per interesting beat, consequences that were funny rather than punitive, and a sense that the world would carry on wobbling even if the dice hated you.

And the dice did hate them. The core mechanic proved surprisingly unforgiving: in straight-up opposed rolls against the Narrativium d8, they lost more than they won, even when rolling d6s and d10s. Luck got spent early to dodge more unpleasant Consequences, which meant by the last act they were eating every bit of fallout the story could dish out.

I ended up leaning heavily into the book’s advice: fail forward, make it funny, keep them competent. A failure at the Baa Bar didn’t end the investigation; it just meant they left smelling of beer and humiliation. A nudge from the Death of Rats got them back on the story. Pushing back auditors in Sator Square didn’t “beat” the conspiracy, but it nudged Belief and set up later NPC reactions. The system feels best when you’re comfortable treating the dice as a vibe checker rather than a strict pass/fail gate.

As a seasonal palate cleanser, it worked beautifully.

Comparisons

Versus QuestWorlds

Mechanically, Discworld’s Narrativium system feels like a cousin of QuestWorlds (review forthcoming): both are narrative-first engines built around opposed rolls, freeform abilities/Traits, and degrees of success that feed directly into how you narrate the story.

Where QuestWorlds is a generic framework you can skin for almost any genre, Discworld is unapologetically bespoke. It assumes:

• You’re in a city where narrative causality is a character.
• Failure is often funnier than success.
• Social and moral Consequences are more interesting than physical damage.

QuestWorlds gives you more formal structure around contests, extended conflicts, and how benefits/Consequences carry forward mechanically. It also expects the GM to do a lot of interpretive work but gives more scaffolding for “serious” stakes and long campaigns. Discworld’s rules are looser, more conversational, more willing to shrug and say “if it makes a good story, do it.”

You’re going to pick Discworld if you want your Pratchett baked in the book and like the idea of Traits as evolving comedic baggage, not just abilities. It’s also well suited to short arcs (two or three sessions) where long-term growth isn’t really a concern.

QuestWorlds is a toolbox which you could use for this kind of thing if you want more procedural clarity with explicit rules for building climactic sequences. Or perhaps the jokey style of Discworld is fun to read but you prefer a rule system that is easier to navigate at the table.

In a way, Discworld feels like “QuestWorlds for people who specifically want to play weird, grounded, messy comedies about institutions and belief in Pratchett’s city.” If you’re one of those people, that’s worth the price of admittance.

Versus Liminal (and Other Urban Fantasy Ensembles)

On the thematic side, the game that Discworld reminded me of most was Liminal, which I talked about in my “gothic feelings” piece. Liminal is modern British urban fantasy: ordinary-ish people dealing with the supernatural in a country that is mostly just trying to get to work on time. It’s melancholy where Discworld is sardonic, but both are ultimately about people in a city negotiating with powers bigger than they are.

Liminal uses a simple 2d6+skill system with Conditions and faction clocks; Discworld uses the polyhedral Outcome vs d8 structure. Both encourage ensembles, both put a lot of weight on NPCs and locations, and both are at their best when players lean into the emotional and social texture of the setting.

Discworld is satire rather than wistful mystery and good at telling short, punchy stories rather than slow burn building of a long-term web of debts and favours.

Liminal is for when you’re looking for something less comedic and less specific but still British and urban. I never thought of trying to be funny with a Liminal scenario but there is a certain British sardonic humour that could be brought to bare here. It’s also fair to say Pratchett did a range of styles and genres in this world, particularly later on, that expand beyond the pun-tasy norms of his earlier works. A Discworld story using Liminal is not as daft as it may sound at first! maybe…

So… Who Is This For?

Discworld: Adventures in Ankh-Morpork is, unsurprisingly, for Discworld fans first. That sounds obvious, but I think it matters. This isn’t a generic “comedy fantasy” system you can easily re-skin; it is absolutely marinated in Pratchett, in the writing style, the advice, the adventure prompts, the organisations, the NPCs.

If your group:

• Knows the books and loves them,
• Enjoys riffing in-character,
• Is happy with story-first, mechanically light, sometimes swingy play,
• And wants 2–3 session arcs or one-shots that feel like lost TV episodes,

then this game is the one. As my Hogswatch one-shot proved, it’s a lovely palette cleanser: everyone gets to loosen up, do ridiculous things with a broom, and still somehow save the season.

I could see using this for one-off forays to loosen me up from the grim fantasy or cosmic terror I enjoy most of the time. It’s also nice to explore Pratchett’s world anew. Sadly, there are no more of his books to read for me.

It’s also a good exercise for players who need to find the fun in failure and as a GM to really practice fail-forward, “consequences, not blockages.”

But, and this is important, I don’t see this becoming a regular campaign workhorse for me.

The core system is clever and on-theme, but that opposed dX vs d8 structure plus finite Luck makes it feel harsher in practice than you might expect. My Hogswatch group lost more tests than they won, and while I could always push things forward, it required a fair bit of GM energy to keep failures feeling like funny twists rather than a long slow grind of “you try, you fail, you pay in Consequences.” The text does warn you about this and gives tools, but it’s still a thing.

It’s also very tightly scoped: Ankh-Morpork, in a particular Now, with specific institutions. If I want a longer-running character drama with similarly light mechanics, I’m more likely to reach for QuestWorlds or something in that space and build my own setting around the players.

And honestly, in an age where we are absolutely drowning in good games, this one doesn’t force its way to the top of the pile unless you or your group have that strong Discworld affection

So: if you’re a Discworld reader with a table of similarly afflicted friends, this is absolutely worth having on the shelf for occasional, joyous visits to Ankh-Morpork. If you’re mostly here for systems, and the licence doesn’t tug at anything in your soul, you can safely admire it from afar and spend your money elsewhere.

For my part, I’ll keep it handy for the next time Hogswatch rolls around and I need an excuse to write “Baa Bar” on a post-it and let someone save the city by putting the wrong rocks in the right machine.