Deep Dive: Shattered City

Before These Unprecedented Times, I used to hit the gym three times a week. Working with a personal trainer, I got into weightlifting later in life. It was fun to hit new numbers! It was fulfilling to feel strong. But lifting late in life meant I got injured. Take just a little too long off, screw up my form a tiny bit, and boom. There goes my back, or my shoulder, or a knee.

Running UFO PressShattered City, designed by Douglas Mota and Mina McJanda, these past several months has been the RPG equivalent to my late-life injury-risking weightlifting program. It finally broke me, but along the way it was great to hit new numbers. It was fulfilling to feel accomplished. To feel strong.

I wrote a bit about Shattered City when we first started. Twenty sessions in, this is my final deep dive. And just in time: Mina tweeted about the hardbacks finally making themselves into the world recently.

Shattered City

PDF available at DriveThruRPG (affiliate link)


To recap: players in Shattered City play Houses, big important factions shaping the future of the eponymous City, Montara. In the game’s fully realized form, one player (not the GM) instead plays a Guild, representing the occupying force oppressing the city. We went all-Houses, and it was fine. But I’ll bet a motivated Guild player would be a ton of fun.

In our game, we had the Crystal Crowned, weirdo vampire-elf alien nobility who were the original masters of Montara; the Verdant Priesthood, supportive plant-people with neat plant-magicks; and the Rebel Underground, Montaran patriots pushing back against the occupiers. We decided on the Volarees as our Guild, fallen nobles looking to make a name for themselves by turning Montara into a powerhouse.

Players also play a hero of their house. We started with a Warlock for the Crystal Crowned, lots of funky powers related to the setting’s ambient magical energy; a Warden for the Verdant Priesthood, a relationship broker and peacemaker; and a Martyr for the Rebel Underground, because who doesn’t love a rebel martyr?

Gameplay was so broad and involved that there’s really no point in trying to retell any of it. They fought giant monsters, explored the nearby countryside, and eventually overthrew the Volarees. And that was just in the first “cycle.” The game is played in big strokes like that, with little vignettes within the bigger sprawl of history.

Behold! A delicious, completely bonkers r-map of our final game-state.

Since we played online, I produced a lot of documentation about our game. I built an elaborate character keeper, wrote a setting bible, used Miro to maintain the relationship map above, and so on. It might not all make sense but if you want a clear breakdown of what all happens in the course of a campaign, check out my Current Situation document.

There Is So Much (Maybe Too Much)

There are so many ways to combine houses and playbooks in this game. So much stuff. Maybe too much.

Players choose from eight houses, five guilds, and fifteen playbooks (helpfully broken up into bold, civic and uncanny themes). The Houses won’t change much but players are expected, even encouraged, to hold onto their heroes lightly and regularly introduce new playbooks. Advancement is left entirely to the players, who can just decide to advance as they play through the faction roles they take on in the course of their lives: leader, traitor, outsider, agent, rebel, prophet and agent. Once you’ve taken four roles (and advanced four times), the character either retires or becomes a Champion (which is kind of its own house).

Oh how could I forget? There are also, formally, five species in play. Those also impact play, both mechanically and culturally. Several of the playbooks are also de facto species, like the Awoken (a self-aware robot from the ancient technomagical past) and Chimera (genetically engineered byproduct of a recent war), or the Verdant Priesthood (all of whom are part plant) and the Architects of the Living City (space-manipulating aliens from another dimension). Players looking to optimize a build are going to get quickly frustrated. Not everyone is happy to run whut they brung.

Then there are the move sets. So many move sets. There are hero moves, for when the characters are doing normal roleplaying stuff. When you pull back and play more abstractly, there’s a whole other set: house moves. And to execute the biggest-picture stuff, starting and ending whole cycles, you’ve got story moves. The moves take up three tabs on that character keeper I mentioned up there.

It’s a lot to keep in one’s head. Occasionally, the moves or their attached economies will affect across scales. If the players aren’t fully invested in the excitement (challenge?) of mastering perhaps the most complex PbtA game ever conceived, it’s gonna be a rough ride for everyone.

The Moves Are Alien But Productive

Let’s talk about these sets of “common moves.” Being several designs into the Legacy concept (Legacy and then Legacy 2nd Edition, Free From The Yoke, and now Shattered City), the move sets have gotten — and I swear this is not pejorative! — weird. You need to set aside your opinions about “correct” PbtA doctrine to really evaluate them.

The takeaway all my players shared in debrief was that it was hard for them to engage with Shattered City‘s moves in an organic way. Moves don’t easily trigger from the fiction as it happens. Moves don’t snowball into other moves like one might expect. The game wants players to engage very intentionally, even boardgame-ily! But the moves’ fictional triggers are unintuitive and the outcomes that emerge from those triggers are highly abstract. My players settled on looking at mechanical outcomes they wanted, and reverse-engineering the move(s) needed to get there. That’s not necessarily a new thing to PbtA-style play — players regularly evaluate outcomes and steer the fiction toward moves — but it’s an inescapable feature here.

Let’s take a couple examples from the Hero move set, Unearth and Compromise.

Unearth reads like this:

You might expect a move like Unearth to produce outcomes like Read a Sitch (Apocalypse World), Discern Realities (Dungeon World), or even the “extend your senses” option in Let It Out (Urban Shadows). Those are all about nailing down situational details right here and now. I think there’s some equivalent to that effect in nearly any PbtA style game I can name. That’s not what happens with Unearth.

Unearth is an invitation for the player to co-develop setting details. None of the options in the bullet list are situational details, though. I mean, sure: in typical PbtA fashion, anything that isn’t a move can just be narrated. Whatever. But my players were looking for more practical affordances and I, as GM, wanted more guidance. To be sure, Unearth produces super interesting content! So it’s not a “bad” move. It’s an alien move. Alien, but productive.

Compromise is another in that category. You’d think you were getting something like Intimidate/Manipulate (Impulse Drive), or Turn Someone On/Shut Someone Down (Monsterhearts) or even Propose a Deal (Cartel). But, like much of Shattered City, the “social move” operates at a whole other level. Here’s how it reads:

Structurally I really like this. It’s how the game’s Assault move works as well: you earn objectives and suffer reactions. But you know what? In almost every case, it was very hard to fit the objectives (other than Enlist and Earn) to the fiction, and the Reactions all felt wildly out of scale. The move felt vital — there’s a lot of interpersonal work in the game — but Compromise almost never produced the types of outcomes we were hoping for. Not every Compromise outcome looks like something in the (immediate) fiction.

The House moves are also quite abstract. In fairness they need to be, given they’re driving the big-picture fiction of the game. Mostly House moves don’t have fictional triggers and don’t act like conventional “moves” other than by writing the rules in Move language (weak and strong hits, pick lists, etc.). One terrific innovation in Shattered City over its predecessors is the addition of “decrees,” an economy that lets the GM limit how long folks spend in up-high mode, and drive the action back down into hero-level roleplaying. So mostly the triggers are worded like “when you spend a Decree…”. Which is mechanical, not fictional. It’s fine, we adapted. That feature is not universally true: there’s a “react to bad things” move called Endure, kind of a House-scale Defy Danger (Dungeon World), that we literally never thought to use in twenty sessions because it didn’t trigger the same way.

The most alien bit of the House moves, in the end, is that there’s no simple move to achieve most of the game-state stuff you’d want. Like, there’s no specific move to “earn Influence” over another House. You have to come at it sideways, see, either by dropping into a Flashpoint and taking Influence as a Compromise outcome (see above), or come at it sideways via some other move. I wrote an entire reference document about this for all the economies in the game. From my GM seat, I rather liked this indirectness. But the players hated the uncertainty and reverse-engineering necessary to pursue their boardgame-y factional goals.

A Couple (Almost) Fatal Flaws

Matters of unmet moves expectations and overwhelming player options aside, there are a couple places the game for-real nearly fell apart for us.

Manifest is a Mess

The big one is a move/concept called Manifest. In Shattered City, literally everyone in the game can tap into ambient magic called Qoam. Very broadly, Qoam is telekinesis, tuned to specific aspects depending on the focus (Potent, Swift, Wide etc.) mastered by the character. That ability is addressed by Manifest, but it also shows up indirectly in other surprising places: the move used to take down gigantic Titans, the Warlock playbook, the Firebrand playbook (which switches out telekinesis for fire!), and so on.

Each player participating in a Manifest move can add their own focus to the move, and doubles the strength of “one person applying force.” Ostensibly these foci create very elaborate effects and creations. There’s an appendix at the end of the book, easy to miss, that offers sample effects at various combinations of foci. But these effects don’t reflect the “double human strength” guidelines. Like, you can create a magical automaton with three foci! That just isn’t covered by Manifest or its supplemental text.

I suspect Manifest is meant to be eyeballed, both by the players and the GM. Like, the player(s) suggests a magical effect and the GM says what foci are needed. That’s cool, it’s “a conversation,” but I feel like the GM doesn’t get enough guidance to know whether something is, you know, three-ish foci or five-ish foci or even if a particular effect is possible under the “telekinesis” umbrella. It’s setting-important and very, very loose. I think if you went into the game with the GM very clearly telling everyone “imma eyeball this,” and clarify that every Manifest is going to be “a conversation,” it might be okay.

Missing Misses

Second (almost) fatal flaw: On the mathy side, it’s just too darned easy to roll hits. And given how large-scale every move is written, you don’t see a lot of rolls in a session. Fewer rolls + rare misses = a feature to route around.

Players help on Hero moves by handing over extra dice. I normally like the simplicity of Advantage-style help, but where I’ve seen it (most recently in Impulse Drive) it’s usually limited to one extra die. Not here! We had three players, therefore four dice in many cases. You do need a Covenant (relationship) with the person you’re helping, which can be fun after a Covenant is broken and the players hustle to create a new Covenant with an eye toward future help. Assuming they can help, it’s easy to assume everyone will help every time if possible. The odds of missing on four dice is quite small — around 9% according to, compared to 18%ish on three dice.

At the House move scale it’s even more pronounced: Allied Houses can provide two help dice! Two Allies (Allied is a game-state term of art) helping on a roll means six dice. They will never, ever miss. There’s a cute bit where the lowest die in the pool means the helping House might’ve fucked things up between the Houses, triggering additional complications, but as far as actually giving the GM an opportunity to make a Reaction (GM moves in this game) from a miss? Doesn’t really happen. Of course the GM can still make Reactions under all the typical PbtA triggers, like “when they look to you to say something” or “they hand you a golden opportunity.” But the House Moves are so abstract that it’s tough to embed the moves’ outcomes in the fiction. Golden opportunities and “waiting for the GM to say something” don’t come up in an organic way.

To be clear: we played twenty sessions. Nothing even close to “fatal” about that, right? It meant we had to master reverse-engineering our desires, working backward from desired outcomes to the string of moves that would get us there. That’s a tough vibe for folks accustomed to tighter interplay between the immediate fiction and move triggers. It’s tactical, abstract, boardgame-y.

Big Little Picture

As with the rest of the Legacy branch of PbtA, the secret of the split between hero vignettes and big-picture systems is that it tricks everyone into co-GMing. This is Shattered City’s best magic trick, and it really is magical when it fires off.

By asking the players to advocate for their factions, everyone brings more energy to their struggle than just “wouldn’t it be cool if you and I had a fight?” The players are quite invested! But almost every move demands the players come together to work through the consequences of their big-picture actions, frequently without GM intervention.

The same thing happens in the Flashpoints, where we play characters in a conventional RPG-y way. Many of the moves ask the players to create their own truths and answers (see Unearth, above). This can be a surprising cognitive load for players. More than once I’d offer ideas for moves that ask the players to provide answers. There is something weirdly different between discovery (revealing what I came up with) and invention (sharing what they come up with). Mostly my players wanted to discover what complications and through-lines I had up my sleeve.

Setting aside the aesthetics or fatigue of collaboration, the play product itself — the actual outcomes, the resulting fiction, the through-lines — is, honestly, exceptional. It was a hell of a run, our twenty sessions. Nothing else quite like it.

The Killer App(s)

There are a few other standout bits of design that stuck with me.

  • Covenants are a super interesting replacement of History (Hx) type mechanisms. Covenants are ranked (from 1-3) and characterized via a pair of keywords, each keyword provided by half the relationship. If you get additional Covenants with the same character, it’ll have its own keyword pair and rank. Sometimes the rank matters and sometimes the number of Covenants matters! But it’s sharp in so many ways. The big one is that the players remember why they had a Covenant in the first place. But also, sometimes you Test the Covenant, and on a good enough roll the players can narrate a new positive memory about that relationship. That may sound frivolous but it produced several genuinely moving moments in play.
  • The grid of political relationships between factions is terrific. Factions can hold Influence over each other (which may not be reciprocal), they can be allied, they can be in conflict, or they can hold dominion over another faction. These are all mechanically meaningful, and produce a ton of interesting context.
  • Assault is maybe my favorite one-roll combat resolution minigame in all of PbtA. On a hit you pick a tactical goal and a couple complications, and on a strong hit you pick two goals, a “better” goal list opens up, and you only face one complication. It played out great every single time we used it. The trigger includes an invocation of “appropriate weaponry,” which meant we were able to use Assault to play out an angry speech to whip up a mob at one point.
  • The centrality of the mechanized map, constantly evolving and growing, is very powerful. I’m a big believer in having central information on a table, and Shattered City mandates the map. Every “strategic asset” is on the map somewhere (and therefore can be targeted with City moves), new locations get added all the time via move outcomes, it’s great. We used Miro to build our map, and being able to zoom in on Montara for individual buildings and neighborhoods, and then zoom out for forests and mountain ranges elsewhere, was very cool. Not sure how we’d pull that off at a real table.
  • The Roles attached to every hero playbook. There is so much design packed into this bit of the game! Every Role is really a little mini-kickoff to their personal arc, either a thing they need to do to enable the role, or a thing that has happened that gives the GM stuff to frame around. When those interact with the players’ House choice, it’s pretty magical. So what’s that all add up to? 8 houses x 15 playbooks x 6 roles per playbook = 720 kickers? That’s kind of bananas.

In Conclusion

In the end, I think what broke me was that it was hard to care about the characters.

That’s a structural feature. The game purposefully breaks up big-picture and personal play. Players are incentivized to take, finish and move on from their Roles as the only path toward advancement. And they only get to advance four times, then it’s on to another hero of their House.

It’s hard to be a fan of an abstract organization. And while our shared history was pretty exciting, it’s hard for me to invest in “history” as its own end product, Microscope style. I was a little skeptical when we set out that the Houses are the real stars of the game, and it turned out they are. Which is not, it turns out, what we cared about the most.

I could definitely see fighting the game’s incentives and playing Shattered City in a slower, more expansive way. Setting aside advancement until you’ve fulfilled your role in good faith, rather than the one-or-so session vignette per spotlight character we went with. Treating the House moves in a more intentional co-GM way, perhaps. Like most long-form games, though, it’s hard to see myself getting back to them to apply all these best practices.

Shattered City is perhaps the most elaborate, difficult PbtA game anyone’s ever attempted. But if you’re willing to put in the work, maybe you’ll get 20 (or more!) sessions out of it too. It’s got nearly limitless space to expand a campaign into. But you have to be strong.

If you’re looking for additional thoughts on long-form play, I wrote about that in the recent past!

6 thoughts on “Deep Dive: Shattered City”

  1. I have read your deep dive with strong mixed feelings over a long time. I was the lead designer of the game until Jay/Mina abruptly took over my manuscript and forced untested changes that nearly broke the game, as you reviewed. And I can prove it. It was a lawful act from her, but highly unethical and ultimately clumsy.

    Shattered City was a huge effort from my part, and I think you would have enjoyed the original version much more.

    1. Oh my gosh, I would love to have seen it. I hope you put something of your own out! We really loved what we got despite the wrinkles.

Leave a Reply