Cataphracts Design Diary #2

Cataphracts commanders: there is no actionable intelligence in this post. Read on.

Since my first design diary entry, quite a few people have reached out to me asking about how to run their own. It’s really exciting! It’s very gratifying to see so many people excited and interested in running their own games based on my silly little wargame. That in mind, I thought this time I’d focus more on my development process, and explain a bit how I made the stuff that turned into the game that now is Cataphracts.

I wrote the rules that became Cataphract in more or less a daze, a flurry of ideas and development without much pause. I don’t necessarily recommend this—I stalled out hard on other work—but these kinds of game design benders can, in some circumstances, prove rewarding. I encourage you to tweak and modify these rules for your own campaign.

As I wrote the rules, I began work on the map and the broader campaign. I can’t share my detailed map or campaign files here (because my commanders read these posts), so instead I’ll just work through a small example, a vertical slice of a larger campaign.

First: choose a setting, a time and place to pattern your campaign on. In general, I recommend the “historical setting with serial numbers filed off” approach. It gives your players recognizable elements to work with, but frees you from the research necessary for a historical campaign, and allows you to scrub off some of the ickier parts of the history. I’m going to stick with the same one I use for my game—an Eastern Mediterranean / Black Sea type of place in the high-to-late middle ages, 13th–14th centuries, right around the Fourth or Fifth Crusade—because I’ve already done a fair bit of reading about this era and it’s pretty familiar. You should choose a period and area of history that interests you! One of the perks of premodern settings is that a lot of the basic technology of warfare is broadly similar, so these rules should scale relatively cleanly from the Parthians to the Safavids.

Let’s start with the map. My original Cataphracts map came from my friend Jason Ripplinger (they had a spare topographical map from a few years back they made for fun, since they’re that kind of guy), but you can honestly use almost any map, it doesn’t really matter—just make sure the map you choose includes rivers. Here’s one I made earlier.

Blank topographic map

Afterwards, chuck a hex grid over it—I just print one off on hexgridgenerator. The hex grid matters mostly in terms of scale: my normal Cataphracts map is about 50 hexes (flat-topped) by 60 hexes, but roughly 40% of those are in water. We’re opting smaller for this demo map, and it’s wider than it is tall, so I’m setting it at about 40 hexes wide by about 20 tall (ish. I’m not measuring all that carefully). At 6 miles a hex, that puts this at something in the ballpark of 200 miles east-west and 100 miles north-south.

With a hex grid

Next comes cities and towns (not yet fortresses). There’s no one formula for determining where humans build their settlements, but in general low, wet, flat land is easiest for farming, rivers and coasts offer access to trade, and distance provides a measure of safety and control. I would tentatively recommend only a handful of cities (1–2 per faction), but plenty of small towns.

Remember that, like a lot of maps, the measurement of space is often more critical and applicable as the measurement of time. Armies march two hexes a day, three at a forced pace; two towns fifty miles apart take four days to travel between, or three at a forced pace (forced-march cavalry only take a day and a half or so). Remember, too, how fast messengers travel—at about 8 hexes per day, a messenger could leave one of those towns in the morning and reach the other by evening. The map you draw strongly influences the flow of the campaign, and the complexities of maneuvering troops—distance becomes time, and time becomes information lag. The further commanders are apart, the harder it is to coordinate.

Here’s what we’ve got:

Cities are squares, towns circles

You can see I’m not sticking to my own guidelines totally, because neither do people. Even ancient cities were built in unusual places, and well, cities in strange places are more fun. But generally the trend holds: the lowlands and coasts are thickly settled, the hills and mountains prove thinner.

Now, the roads. Roads connect settlements, and want to stay on level terrain. On topographic maps, they often loosely trace elevation gradients. Also, bridges are far more common in settlements than they are just out in the countryside, so keep your bridges confined mostly to settlements.

Armies almost always march on the road. As Devereaux describes, while maps often look like a big flat expanse of open land, they’re really more akin to a web of connections and pathways with walls or empty space between. (If you’re an RPG nerd, this is a hexcrawl, but in reality most armies end up treating it more like a pointcrawl.) Commanders will spend quite a bit of time figuring out how to best get from A to B, and the roads form the bulk of the setting, the playing field, with which they must contend. Those two towns fifty miles apart as the crow flies can end up far further if the road between them wends and winds.

Here’s our map with the roads.

Roads!

As with most things on this map, there’s probably a more-historical way to fill these roads in, but what we’ve got here feels good enough. It’s got some roads that feel very efficient and logical and some that feel sort of odd and arbitrary. All settlements are connected to somewhere else, but there are sometimes gaps between major settlements’ connections.

Next: fortresses! Think like a commander, or a jealous and fearful king: where are the enemies going to march from? Where do you need to stick a castle to keep watch? We haven’t started drawing faction lines yet (though perhaps you’re already starting to guess where they’ll fall), but historical territory lines often shift, and fortresses designed to keep one foe out often end up repurposed against another.

Generally, this means crossroads, chokepoints, and positions that allow you to see a long ways. Sometimes, two fortresses end up getting built pretty close to one another—one to watch the other—even if they later end up on the same side.

You want a lot of fortresses. At least as many as cities and towns combined, if not significantly more. Even a small army garrisoned in a fortress can cause a ton of trouble for passing armies: settlements are the goal, fortresses are the obstacles. (If you study the history, there should probably be even more.)

Here’s our map with fortresses added:

In theory, any single point could be the site of a decisive battle

You can see they sit at high points, choke points, and other key locations—prime real estate for commanders looking to force their opponents into a prolonged and costly campaign.

Right now, the map is more or less done in terms of topography, roads, and strongholds. There are two things left to do: settlement level and faction turf. Settlement level is effectively just a measure of population, or perhaps population and wealth put together. (If you’re a Crusader Kings nerd, consider that development level increases supply limits—we’re rolling both into one number here, but it’s the same idea.)

For settlement scores, I stick to 20-point increments, ranging from 100 (100 persons per square mile, quite densely settled) to 20 (20 persons per square mile, quite thin). If you’re into the deep verisimilitudinous worldbuilding, these settlement scores plug nicely into donjon’s Medieval Demographics Calculator.

I don’t have an official way to map settlement scores. Regions around cities should be quite dense with outlying suburbs, mountainous hinterlands should be very thin, and settlements follow rivers and roads. Otherwise it’s more or less a matter of eyeballing it.

Here’s our same map with settlement scores:

Probably should’ve picked a different shade of blue than the ocean lol

You can see how the settlements follow relatively reliable metrics: dense along the big coastal port cities and rich valley towns, thinner as you get deeper into the wilds, and then quite empty way up into the mountains.

While I don’t show my players this map of settlement scores, wise commanders quickly learn to intuit where to find more and less population. Because settlement levels determine how many supplies an army can forage at once, this information can swing the tide of a campaign: foraging around a city can support an army for weeks or even months, while foraging in the outer wilds rarely lasts for more than week or two (depending on the size of the army and so on, naturally). Armies march on their stomachs, and campaigns live and die based on their ability to keep the troops fed.

Now, faction lines! Since this is a smaller map for a smaller game, I’m going to add just three factions, rather than the full five I have in my current Cataphracts game. With six cities, that gives each faction two cities, a pretty reasonable number.

In general, I encourage you to draw weirder faction lines than you might expect. While historical states and polities were almost always contiguous, drawing unusual lines encourages your players to get funkier with their movements and diplomacy.

Here’s what I’ve got:

I would say Purple probably has the best position, but also likely the smallest starting army

Not terribly abnormal, but it does put each faction in a position where they need to start making decisions quickly about where and how they want to march. Once commanders commit, the flow of the game will get pulled to one region or another and things will get spicy fast.

If this were a map we were using for an active game, now would be the time for two rather tedious tasks: naming all of these strongholds, then doing the math to count up how many soldiers each faction starts with. I’m not going to do that here, since it’s kind of a pain, but that’s the next step.

(Eagle-eyed readers will notice that I’m skipping a possible step that gets mentioned in the core rules: regions. While I drew region lines onto my original map, I’ve quickly found they don’t actually matter, and are extremely fiddly to communicate to players. I plan to edit them out when I update the rules—in the meantime, you can just substitute anytime regions are mentioned with “hexes closer to a chosen stronghold than any other,” with priority going to larger strongholds.)

In terms of starting detachment types—how many cavalry, heavy detachments, wagons, and skirmishers—I mostly just eyeballed it. Strongholds (and thus factions) that seem richer, with more cities and denser population and an older history, got more heavy detachments; factions with more wilderness and hinterlands got more skirmishers and non-heavy cavalry.

I would highly recommend you start each faction with all available soldiers raised, skipping over the weeks and months it takes to levy all the troops and get them organized. Start each faction lead with all their soldiers in one place in one big block, ready to go.

At this point, you might be saying “Sam! These factions don’t seem very equal? Doesn’t this give one side an unfair advantage?” To which my answer is… yes, basically. Historically speaking, wars fought between equals are quite uncommon, especially perfect equals. You can do some work to even the lines out (factions with more troops get fewer heavies, etc.), but in general the factions are likely going to be uneven. Besides, being the bigger faction isn’t necessarily an advantage: it means more borders to defend, more soldiers to feed, and more commanders to organize. In my Cataphracts game, how many troops a factions started with has had effectively zero observable effect on both a faction’s success on campaign or their players’ enjoyment.

At some point in this process, you’ll also need to draw a player-facing, somewhat-inaccurate diegetic map (it’s the main image in my first post). Mine includes the names of all cities and towns, but not fortresses; its dimensions are close, but not perfect.

Now, let’s talk about player materials.

Each commander gets three documents: a faction sheet, a commander sheet, and an army sheet. A faction sheet is shared across all commanders in a faction, and includes a brief summary of the faction’s history and culture, their special rules, a quick overview of the other factions, and some tables of names and titles for commanders. Here’s an example:

High Duchy of Sarai-by-the-Sea

Three hundred years ago, Saint-King Mikhail the Mighty sailed to these shores, converted to the faith of the Orthodoxy, conquered the giants dwelling along the water, and proclaimed his own kingdom. Since then, the kingdom’s borders have shrunken somewhat (not least due to the Mikhailovichi fighting amongst themselves), but the lords of Sarai-by-the-Sea still trace their lineage back to the holy Saint-King.

As descendants of people from beyond the sea, the Saraians place great value in shipwrighting, warring along the coasts and rivers. From the others in Kyrenia, they learned horsemanship, poetry, and alchemy, offering their craft in ships and their deep knowledge of magic in exchange.

For the past seventy-five years, the High Duchy—as the kingdom became, in time—has been a subject of the Prince-Bishopric of the Ikosion, a large and prosperous theocracy along the southern coasts of Kyrenia. A proud people, the Saraians chafe beneath such subjugation, but scattered rebellions over the years have always failed.

Now, with the sudden and suspicious death of Grand Duke Vasily, the realm is thrown into chaos and fury. Why tolerate the Ikosenes’ rule? Why not rise once more, an independent kingdom for Saraians alone? Why not withhold the ancient territory set down by the Saint-King?

Fast, easy, lots of historical analogues to draw, not more than a few paragraphs. Brevity is essential here, since players must read and understand quite a bit before getting to play.

Here’s some sample special rules:

Saraian Special Rules and Detachments
Embarking and disembarking from ships takes half a day. Ships travel 6 miles per hour faster.

Knight-Wardens: Knight-Wardens are heavy infantry. When defending a stronghold, they count quadruple for the purposes of determining numerical advantage.

Marines: Marines are skirmishers. An army of exclusively marines can undergo a forced march to travel along a river as if they were in a ship (at a regular pace), even without a ship.

Quick, breezy, and operational-level. It’s very easy to sucked into the minutiae of battle, but resist that urge: the juice of Cataphracts is not in the fine detail of tactics (go play any other wargame, if you like that), but rather the logistics of moving thousands of soldiers in synchrony.

I’m going to skip the tables of names and titles here along with other factions’ descriptions, since I assume you know how to make those—they’re much the same as they would be in any other tabletop RPG book.

That’s faction sheets, which all commanders in a faction have access to. A commander sheet looks something like this:

Grand Duke Zdravko Vasilyevich “the Tall” Razumihin

You were born to high nobility, first-born son of the duke, prepared for greatness from birth. Raised in the royal courts of Simurghzal, you’ve always been the fastest horseman, the canniest sailor, and the most handsome of men. Yes, your people were subjugated by the foreigners of the Prince-Bishopric of the Ikosion, but until this year that mattered little: you had your life, and they had theirs. That all changed when your father, Grand Duke Vasily the Harpist, died, poisoned at your banquet feast by Prince-Bishop Eulalia’s underhanded dogs. Now, your people howl for blood, and they look to you for leadership—and vengeance.

Stats
Age: 26 (October 11th, 1299)
Faith: Orthodox
Traits:

• Beloved. Your army gains +1 resting morale.
  
• Stubborn. Your army does not lose morale on defeat in battle.

Objectives
As determined by yourself, your family, and the Boyars’ Duma:

1. Free the High Duchy from the rule of the Prince-Bishopric of the Ikosion.
   
2. Reclaim the Saint-King’s Hawkfeather Crown, stored in Karkota’s reliquary vaults.
   
3. If possible, kill Prince-Bishop Eulalia.

(This is a lead commander, so they start with a couple of bonus traits, a nickname, and standing objectives. Fresh subordinate commanders get only the normal amount of traits, receive their objectives from their superiors, and must earn their nicknames.)

Like the faction sheet, the goal here is to give players enough to understand who they are in brief, then get them playing fast. Don’t get bogged down in the nitty-gritty of history—the social complexities that emerge over the course of play are going to be far more exciting and compelling than any backstory.

Finally, the army sheet: a spreadsheet, with space to track supplies, morale, and detachments. These don’t translate neatly to markdown text, so I’m just going to link them: here’s a sample army sheet for Grand Duke Zdravko (probably after forking off some of his troops to subcommanders), and here’s a blank template army sheet for you to steal.

While this seems a lot, I realize, this is basically everything Cataphracts is on a material level. A mildly-complicated map, a bunch of documents and spreadsheets, and then a discord server with many individual channels. For a game that sometimes feels like it’s going to overwhelm me with its sheer size, it’s actually quite lightweight.

Just to recap:

1. Take a map and layer hexes over it.
2. Add rivers, cities & towns, roads, and fortresses.
3. Paint on the settlement levels.
4. Draw faction lines.
5. Write up faction sheets. Name important locations.
6. Do a bunch of arithmetic to figure out how many troops each faction gets. At some point in here, draw the diegetic map that all players see.
7. Write the lead commanders’ sheets and draw up their army sheets.
8. Find some players and put together a discord server.

Okay! That’s design diary entry #2. I hope that this sheds some light on my process, especially if you’re interested in running your own Cataphracts game. Next time, I think we’ll talk about the day-to-day running of the campaign itself from my side as referee. Thanks for reading.

May 15, 2025 cataphracts design session report

Cataphracts Design Diary #1

Cataphracts commanders: there is no actionable intelligence in this post. Read on.

About two months ago, I reread several series on military historian Bret Devereaux’s blog, ACOUP: analyses of Helm’s Deep and Minas Tirith, breakdowns of pre-modern command and pre-modern logistics, and, of course, a post simply titled “How Fast Do Armies Move?”. I’m a fan of Devereaux’s—he writes in that delicious space of really knowing his history yet also with the understanding he’s writing for a bunch of D&D dorks who care about the fiddly minutiae of, say, the efficacy of maille vs. plate armor against English longbows. The fun, gritty little details.

In military history, as Devereaux explains, there are basically three levels of decision-making: strategy (why you fight, the goals of the war), operations (how you get your soldiers to battle, how armies move), and tactics (how you win battles). It struck me, as Devereaux often alludes to, that there are a lot of games about strategy (Civilization, Diplomacy, the world map of Total War) and innumerable games about tactics (any given medieval strategy title or tabletop wargame), but very few about operations. Almost no games, as far as I’m aware, are interested in, say, the logistics of feeding an army, or communication structures between commanders in the field. Games usually simplify or skim over all those fine details of running a war, or just skip them entirely.

So, I said to myself, why not try to make an operations wargame? Embrace the logistics, the gritty details, the fuzziness of the fog of war?

Thus was born Cataphract, the written ruleset, and Cataphracts, the game I’m currently running. They have the same name because I made both of them, but—as I’ll discuss—if you run your own campaign, it’ll probably be different. (Really, this one should be called something like “the Voreia Campaign” or “the End of Emperor Michael’s Peace.”)

Almost immediately, things started happening that I have never seen before in any game, anywhere.

The pitch is pretty simple: it’s an asynchronous play-by-post real-time wargame, set in a pseudo-Black Sea region circa 1300. That means no sessions—instead, my players write what they want to do in a discord channel, and I keep track of their orders on my map. We track things in real time, so that means if an army takes two weeks to march from one place to the next, well, I’ll see you in two weeks. Messengers are just “a guy on a series of horses,” so they, too, are tracked in real time—a letter sent from one stronghold to another 150 miles away takes, at minimum, about three days to arrive. While I allow a little bit of “rubber-banding” to keep things moving (especially with players in multiple timezones), we generally stick to this real time element quite strictly. No teleporting armies, no instantaneous messages.

The actual written rules are straightforward: 6-mile hexes, a basic 2d6+mods roll for battles, pretty simple rules for special units, and a few other odds and ends. You can read these rules quickly, and they’ll feel familiar if you’ve played a more simulation-y tabletop RPG before (or, you know, Mount & Blade).

The other twist I added is command structures: each of the five factions (more on them in a minute) starts with a single commander leading one huge army. Whenever a commander wants, they can bring in a new commander (and thus a new player), hand them a chunk of their army, and send them off. Once that happens, those commanders are broken into their own text channels, and cannot communicate except through messengers. (The one exception is when two commanders are in the same place—then, they get a channel together where they can talk freely.) Those commanders can then appoint their own subordinate commanders, and so on. I started Cataphracts with five commanders—at time of writing, I have twenty-three.

I’ll talk more about the IRL logistics of running a play-by-post with twenty-three players in a future post, but the short version is this: I set up a channel on a reasonably-popular RPG discord server I’m on, then each commander gets their own thread, using discord’s thread feature. Each commander then gets a little doc with their character writeup and a sheet with their army numbers. They write messages to me, I reply and notify them as events occur, and I keep track of everything on a big spreadsheet and a running Photoshop map file. Critically, commanders do not get to see the detailed world map: they have a public diegetic map that’s low-detail and not totally accurate (see below), but only get the detailed hex map with all the place names of regions they explore. Even then, they don’t get positions of armies; no commander has perfect information.

The inaccurate, in-universe map of Voreia (the setting of my Cataphracts game), which all commanders see.

That brings us to factions. The five factions in Cataphracts each get a short little writeup, a couple of special unit types (each of which is no more than a sentence or two), and a table of names and titles—only a page or two per faction. The five factions are:

• The Novan Empire. Think late-game Byzantines, somewhere between the Palaiologans and Angeloi. Their lead commander is not the (young, snot-nosed, sniveling) emperor, but rather a legate, a military veteran.
• The Divine Eparchy of Sakarziya. Basically, “What if the Orthodox Church had Papal States, and what if those Papal States were in the Caucasus Mountains?” As with many archbishops and emperors, the exact power split between the two is a little unclear, but the Novans and Sakarziyans are not at war.
• The Zlanic Principalities. Pre-Golden Horde Kyivan Rus’, more or less. Formerly a client of the Novan Empire, the Zlans declared independence (and thus war), but still follow the Orthodox faith.
• The Third Valgar Kingdom. Take the Second Bulgarian Empire, make them Bogomilist, and scooch them forward a bit in time. Heretical subjects of Nova, now turned to rebellion against them and the Sakarziyans alike.
• The Kalkhanate. Almost exactly just the Ilkhanate (lol)—pseudo-Turcopersian horse lords bent on conquest, with a leader recently converted to the local faith to win extra support.

Each of those five started with a single commander and a big block of soldiers, somewhere between ten and twenty thousand apiece. There’s a possible version of this game that starts earlier, with each of those five lead commanders raising their armies and bringing them together and so on, but I opted to skip that—lead commanders started with all available troops raised, ready to march, what would amount to something like 2–3 months after the war’s official declaration.

I wrote the first five commanders—a legate, an archbishop, a grand princess, a king, and a khan—but each subsequent commander is randomly generated. New subordinate commanders get a relationship to their superior commander (often, but not always, family), their age, and a few traits—older commanders get more traits but are worse in hand-to-hand combat (a relatively uncommon occurrence). These basic tables have resulted in some remarkable dynamics and relationships: one faction is entirely composed of family members of a dead former king; another involves a secret parent-child relationship; another still is composed of no family at all, just work colleagues trying to get along. My players love these relationships. While Cataphracts could basically be played entirely in clipped sentences and spreadsheet math, so far, every single one of my twenty-three has, to a greater or lesser extent, gotten into character and spent a lot of extra time and effort in their letters and messages to each other.

(Worth mentioning there are also a handful of other commanders, ones that don’t strictly exist in any one faction: monastic orders of knights, mercenary companies, peasant revolutionaries, one or two others. I’ll talk about them more in a future post.)

For me as a game designer, though, what’s been more compelling is the complications that arise from the logistics. Commanders know only what their scouts can see (within a fifteen-ish mile range) and what they receive in letters. The real-time tracking continually creates a powerful, sometimes overwhelming level of “information lag,” where operations succeed or fail not based on the strength of the armies in question but rather on commanders simply knowing (or, more likely, not knowing) what’s actually happening in the field.

Here’s a (mostly) hypothetical example: Red Commander marches into Blue territory, and reaches Blue Stronghold A. The Blue NPCs in Stronghold A write a frantic letter to Blue Lead Commander, saying “Oh God they’re here! Send help!” Before that letter reaches Blue Lead Commander, Red Commander captures Blue Stronghold A in an assault, and begins marching to Blue Stronghold B. Upon receiving the letter from Stronghold A, Blue Lead Commander brings in a subordinate (Blue Subcommander) and says “The Reds are at Stronghold A! Go stop them!” Blue Subcommander marches, and unexpectedly runs into Red Commander at Stronghold B. Blue Subcommander sends a letter to their lead saying “Oh God! Now they’re here at B! Send help!” at which point Red Commander again assaults Stronghold B and defeats Blue Subcommander. Blue Subcommander retreats to Blue Stronghold C, only a short ways ahead of Red Commander.

At this point, what does Blue Lead Commander know? They know that Red Commander is at Blue Stronghold A, and they know that Blue Subcommander is somewhere between them and Stronghold A. That’s it. If Blue Lead Commander is at Stronghold C (they might not be, but bear with me), the first time they hear of both Stronghold A and B’s fall to Red is when Blue Subcommander arrives in person fleeing Red Commander. (It is, quite literally, the Two Generals Problem.)

This kind of thing happens constantly. Almost every week, one faction or another has made blunders or errors based on a lack of clear information. Fancy strategy and complex maneuvers have consistently failed simply because coordination is extremely slow and difficult. The most successful factions thus far are not the ones that have fancy strategy or even especially powerful units, but instead the ones that keep their plans dead simple.

Sometimes, it goes even further: from time to time, a commander will get unlucky and, for one reason or another, they’ll send two or three letters to another commander and get no response. When this happens, the usual response is panic: they freeze and hold position for days or weeks, they pick a random direction and march away hoping to escape and/or find their targets, they turn around and march straight back to talk in-person, or otherwise freak out and start behaving irrationally. Because information is so limited, commanders are constantly under pressure, a high level of background uncertainty. To really move with confidence is a very rare occurrence: most of the time, commanders only really trust what they can see with their own two eyes (or rather, their own scouts’ many eyes)—everything else is suspect.

One of my friends, a lead commander, said it best: “Sam, I feel like the extreme difficulty of this game—and thus its possible brilliance—is that I never have any idea what the fuck is going on.”

These constant logistics requirements also add further wrinkles of complexity to the diplomacy. While a lot of Cataphracts plays along similar lines as Diplomacy, Crusader Kings, or Model U.N., the intense fog of war and need for resupply (combined with unpredictable elements like weather) muddies the waters. While I expected this to result in backstabbing (a very common occurrence in Model U.N. and large-scale megagames), it’s actually done the opposite—subordinate commanders are loyal, almost fanatically so, to their lead commanders. Players are desperate for direction. While hostile factions do engage in a measure of deceit and trickery in their letters to foreign powers, inside each faction the various commanders have thus far been extremely loyal and honest.

In megagames and larps I’ve played in (not to mention smaller-scale board games and videogames), players betray each other for fun, to make things interesting. In Cataphracts, players don’t dare betray one another because they need each other.

Factional alliances, likewise, have thus far reached an intriguing conclusion: no faction has yet broken a deliberate alliance with another (no betrayals), but multiple factions have already failed in their alliances. That is, they agree upon a time and place to meet and thus crush their enemies, but along the way, one or both factions just… don’t show. And not even intentionally! In the alliances I’ve seen so far, more or less all commanders intended to honor the alliance and stick to the plan, but failed to do so. The weather slows them down, they need to stop and resupply, their superiors redirect them, they miss a critical message and show up in the wrong place—this kind of thing happens all the time.

Again and again, operational issues are the linchpin of the game. Commanders succeed and fail based on their battles, sure, but more than that, they succeed and fail based on their ability to coordinate, stay organized, and stay aware of their surroundings. The simplest elements—knowing where the enemy is, knowing where your allies are, knowing how long it will take to get from A to B—become essential to victory. In other games, this information is usually obvious; in Cataphracts, operational intelligence is a rare and valuable resource, more precious than treasure or soldiers.

I have so much more to talk about (the barriers between the written rules and my own referee style! time, scale, and domain play! the logistics of running for twenty-three players! simulation as a tool for education and research!) but I’ll save them for a future post. Thanks for reading—I hope you enjoyed this little snippet of an ongoing project. Hopefully, there will be more design diary entries in the future.

April 23, 2025 cataphracts design session report

Doing a Cool Move

My roommate ty asks an important question of each game he plays. The question first emerged, I believe, during hours-long sessions of mid- to heavy-weight board games (Oath, Twilight Imperium, the big crunchy League of Legends board game, several unusual David Sirlin titles), many of which were a struggle in one way or another. The question is this:

“How hard is it to do a cool move in this game?”

Let’s consider this question in some common games:

• In chess—a game I am not at all an expert in, but have spent a lot of time watching the chess hustlers play in Union Square—a “cool move” is mostly about making your opponent miss a key observation. Chess players start to lose when they fail to notice an important play, often many moves in advance. They blunder a knight, accidentally land in check out of nowhere, or make suboptimal trades. Thus, a player’s cool move is to see this possibility in advance, slip past their opponent, and pounce on the weakness. Very cool.
• In tag, a cool move is about physically moving. You duck or twist past your friend who’s it, you make a daring leap across the playground stucture, you make some mad sprint over long distance to escape. It’s cool because you’re physically doing it. Awesome.
• In poker, a game I also don’t know very much about (but I have been reading Frank Lantz’s book, which often discusses poker), a cool move is always about bluffing. It’s about fooling and deceiving your opponents, one way or another, to get them to buy into your con. You read them, you hook them and reel them in, and then you take their money. So exciting and compelling there’s a whole James Bond climax that’s just Bond and Mads Mikkelsen playing poker, both of whom make some very cool moves.
• In Titanfall 2, a cool move is about precise timing, reflexes, 3D spatial maneuvering, and understanding how to exploit a videogame’s character controller. You bunnyhop, wall-run into a double jump, achieve max speed, then whip around two-hundred-and-seventy degrees with the grappling hook and land a cross-map sniper rifle headshot, all in mere seconds. It’s difficult to master, but beginners start trying cool moves very quickly, and learn the good ones as they play.
• In a Dark Souls speedrun, a cool move is about extreme manual dexterity and combat mastery intermixed with exploiting glitches and quirks in the game, plus a total understanding of the environments and levels design. Every speedrun gets strange quickly, but runners make bizarre and fascinating moves as they play. Jumping off a would-be lethal cliff to survive, glitch the camera at an impossible angle to convince the game you’re dead, and thus clip through boss gates? Very cool.

For a move in a game to be cool, it has to be unexpected or unpredictable in some way—surprise is an important element of a cool move. Whoa! Wild! Check it out! I exclaim to my group chat as I post the latest trickshot compilation. We love to be surprised and amazed by what we can do in a game.

A cool move is also powerful. A cool move demonstrates mastery, understanding, or talent, even on the small scale. A cool move to a total beginner is different than a cool move made by a highly experienced player—what starts as cool becomes normal, and the cool horizon expands further. When you start playing fighting games, a quarter-turn seems like some impossible task; for players climbing the ladder, the question stops being if you can do the move but rather when you deploy each of your many cool moves (including the humble quarter-turn).

Fresh and unexpected action combined with a high degree of skill together demand an open playing field, a relatively broad state space. For every cool move, there needs to be a larger field of uncool moves: regular moves, boring moves, average moves. Perhaps to a new player, every move of the experienced player might look cool, but for the experienced players there always needs to be that horizon to chase, new levels of unexpected prowess to seek.

Many games tighten this horizon in order to raise the floor. When I play, say, Assassin’s Creed, my basic attacks turn into long, flashily-animated combo chains as my dude rips through the mobs, all just by me pressing X, X, X over and over again. It looks cool, sure, but it doesn’t feel cool. I know anybody who plays Assassin’s Creed could make the same moves, since it takes next to no talent. That level of “guaranteed cool” often reduces skill ceilings as well: it just doesn’t have much depth to its combat, movement, or stealth, because the systems aren’t very open-ended. The developers need players to hit a minimum “threshold of cool” no matter what, so the field must be constrained.

When you play, say, Dark Souls—a videogame with superficially similar combat to Assassin’s Creed, but in truth far complex and more satisfying—you learn all the mistakes you make as you play. Same goes for Titanfall, or chess, or most other games with this freedom and depth. When you play the latest open-world action-adventure AAA map game, you learn primarily only how few mistakes are even possible. Cool moves demand the existence of uncool moves, and designers often seem eager to remove every possible uncool move.

Many crunchier systemic games fall into a related problem, which is having a single obvious move you should use every time, the optimal choice in every case. You run into this in roguelikes or certain deckbuilders: a flush deck in Balatro is a lot of fun the first few times, but it gets old quick. You can deliberately avoid flushes and still succeed, sure, but having such an obvious, easily-repeated, usually-optimal move gets stale quick.

Tabletop RPGs often run afoul of this issue. If I play a rogue, in say, Dungeons & Dragons 5e, I should try to land my bonus damage sneak attack every turn. Rogues don’t get any per-day abilities (an intriguing design choice that I support, for what it’s worth), so it’s just sneak attack, sneak attack, sneak attack. This gets old fast, and Wizards of the Coast knows it—one of the (good!) changes made in 5.5e was to, at high levels, give rogues alternative options for sneak attack: they reduce the damage total, but allow tripping attacks, blinding attacks, and so on. This helps rogues stay less bored, in theory, because now they have more to do.

You can see this philosophy pushed to its outer limits in something like Pathfinder. Pathfinder has innumerable options available to its players—ancestry, background, class, spells, pets, feats of a hundred varieties, dozens of weapon tags, and so on. More options, more choices, more cool moves, right? Surely, Pathfinder players must always feel like they’re doing more cool moves than their less-optionful compatriots at other tables, right? More, more, more?

In a word: no. This blind maximalism seems obviously wrong to me.

From my perspective, there are two basic ways to do a cool move in a tabletop RPG, and the difference hinges on Markus Montola’s third rule of roleplaying, the Character Rule. To recap, Montola describes roleplaying as following three rules: the World Rule, the Power Rule, and the Character Rule. They state, respectively:

1. Role-playing is an interactive process of defining and re-defining the state, properties and contents of an imaginary game world.

2. The power to define the game world is allocated to participants of the game. The participants recognize the existence of this power hierarchy.

3. Player-participants define the game world through personified character constructs, conforming to the state, properties and contents of the game world.

In brief, roleplaying is the act of imagining a world and how it changes, together with your friends, through only the characters you play in that world. World, power, character.

If you read the histories (Playing at the World and Shared Fantasy, mainly) most of the early wargamers that played early RPGs and proto-RPGs followed these rules to a greater or lesser extent. Most tables had one or more referees in charge of the world, and players had one character they controlled; outside of their characters, players couldn’t affect the world. Classic D&D.

This is still the model for most RPG tables, and many (most?) rulebooks simply assume it to be the case. By Montola’s three rules, much of what these players do at the table is roleplaying—except for, of course, the jokes, Cheetos, rules arguments, chatting, and so on. But when they play, these players define a world, together, through their characters.

But! at many tables and according to many rulesets, you often break this third rule, the Character Rule. That is, players regularly define the world outside their characters. In Blades in the Dark, this looks like a player (rather than a character) spending Stress points to declare a flashback—that there’s a gun under the table, or that the window was left unlocked, or even that the night watchman and the gentleman’s valet are secretly in love and having an affair. That particular diegetic Blades scoundrel didn’t do anything to trigger that Stress spend: it occurs purely at the player level.

In Wanderhome, this gets pushed further, with players writing all of the world. Wanderhome players often take inspiration and suggestion from their characters, but are never confined in their ability to affect the imaginary world by what their character can do. These rulesets often lean towards a more “storytelling” sort of model, at the far end of which lie collaborative writing exercises (which range from “Hey, let’s dream up a new short story collection together” to something like Microscope)—and improv.

By Montola’s definitions, improv is not roleplaying because improv actors often define the world outside the limits of what their characters can do. If we’re doing a scene and I say “Boy, it’s cold in here,” and you say “What did you expect, living at McMurdo Research Station?” now we, as characters, are in Antarctica. By contrast, if we’re playing Mothership and I say “Boy, it’s cold in here,” and you mention McMurdo, the Warden would very likely say “No, actually, it’s because you’re on the deserted ice moon Galicon-9-3, inhabited by heatsucking slugs.” Improv actors define their world through dialogue, the words the actors say, but the characters themselves are not diegetically changing the world.

Collaborative writing likewise does not qualify as roleplaying for much the same reason. Authors possess tremendous ability to alter the world outside the characters they write, and do so more or less constantly.

Despite Montola’s definitions (if you can imagine), many RPG books and tables embrace these methods. Accordingly, many people who claim to be playing a tabletop roleplaying game are, in fact, doing some other activity, at least some of the time. You can imagine this as a simple little scale: Microscope defaults to being out-of-character and dips into scenes only on occasion (mostly not roleplaying), while Blades usually follows the Character Rule and only steps out of the characters for a few beats at a time, like when you spend Stress (mostly roleplaying).

At all tables, much of what constitutes “a session” is not roleplaying. Those Cheetos, jokes, scheduling arguments, rules arguments, safety tool invocations, smoke breaks, and so on—all are not roleplaying. Even just when your GM asks you “Hey, you’re from the Novan Empire, right? Do you think your family has land in the Western Themes?” and you say “Yeah, probably, my background says I’m from Skiathos, so I think that makes sense”—in that moment, as you step outside your character to change the world, you stop roleplaying.

Thus, we return to doing a cool move. In these more collaborative, “writers’ room”-style games, a cool move is rarely something that your character does. Declaring your guy in Masks to be the most awesome forcefield-invincibility-triple-multiclass badass doesn’t mean very much because someone else can just say “Yeah, Dr. Deltron’s laser fries you anyway, and your armor begins to peel away, revealing your old childhood scars.” That ability to override the imaginary world via authorial fiat makes powers and features mostly useless from a goal-achievement perspective. Nearly all RPG rulesets about collaborative storytelling thus are not about solving problems in the world, and so the cool move must change.

Because the goal in these collaborative storytelling games is not “To win” (slay the dragon, solve the mystery, etc.) but rather “To tell an engaging and meaningful story, together,” a cool move is much more about how it contributes to the story, or the world, or whatever else you care to write about together. The cool move in Microscope might well be to say “Oh, wait, what if each of the psionic cyborgs had like, an astral umbilical cord situation, and sometimes those cords got cut and the cyborgs were severed from the Hive Mind and became independent? But it was like, an existential crisis each time?”—a move that solves no problems for the characters in its redefinition (indeed, it probably only creates more), only redefines the world in an exciting way.

Does this meet our former criteria of a cool move—that is, to be unexpected and effective? Yes, I think. If your goal as a table is to tell an engaging story, dreaming up exciting and dramatic narrative elements is both unexpected and effective. If other players guess what you’re about to say, well, that’s much less cool; likewise, if it’s just kind of some long diatribe that doesn’t add anything, that too is rather less than cool. It’s easy to spot a cool move in a collaborative story because we, as humans who consume media, regularly see so many uncool stories.

But let’s consider the other route—the second way to do a cool move in an RPG. In collaborative storytelling, it’s cool to have exciting ideas. In what we might call “strict” Montolan roleplaying, exciting ideas are filtered through a very specific lens: the lens of character. If my idea can’t be enacted through my character, it’s maybe fun to consider but worthless for the purpose of the question at hand. It’s fun to imagine, say, “What if goblins had hot air balloons?” but that idea alone doesn’t solve the problem of your characters needing to climb a tall cliff. In these kinds of strict roleplaying games, the cool move often revolves around coming up with an exciting or clever solution to a problem within the narrow scope of what’s possible for your characters to do.

A cool move in a post-OSR game looks like, say, using the engine block from your party vehicle to run electricity through the toxic oil seeping out of the facility, thus shocking all the techno-goblins inside. Or surveiling the vice president’s wayward child for scandalous kompromat to use as leverage in gaining access to a CIA field base. Or even just jamming a pair of sabers through the handles of a door to hold off the Imperial soldiers trying to break in. Whatever. The point is, a cool move involves using the in-world resources and options available to the characters in an unexpected, powerful way. The options available to characters and the options available to the players are, for the most part, one and the same.

This brings us to the question of rules. Earlier, I mentioned state spaces, the field of possible game states within a rules system. Are these various cool moves “in the rules?” On paper, no: most OSR and post-OSR rulesets are quite spare and lightweight, with only a few (if any) powers, abilities, or features available to player characters. But in practice, these moves do fall in the rules, because, well, it’s not like using a car battery to electrocute goblins is cheating. These odd sort of rules-outside-the-rules are a result of the different frames of play in an RPG. Old-school-ish play usually relies far more heavily on the rules of the diegetic frame—the rules of the imaginary world—rather than endogenous rules of the written ruleset. Car batteries, the conductivity of toxic oil, techno-goblins dying from excessive electricity: none of those exist in the endogenous written rules, but all clearly exist in the diegetic.

By emphasizing the diegetic rules over the endogenous, OSR and post-OSR play can get away with minimalism in their written endogenous rulesets, because they just don’t matter that much. (By contrast, this kind of play gets tremendous use out of detailed adventures, especially ones that do the legwork of fleshing out an imaginary world in detail.) The cool moves in a post-OSR game are about exploiting some unexpected—often only implied—element of the imaginary world to achieve a character’s goals. The open-endedness necessary for surprise comes from the near-endless breadth and depth of the simulated world, rather than the vast complexity of the written ruleset. (This is sometimes referred to as “tactical infinity”, but it holds true even outside of tactics.) In these kinds of simulation-heavy games, I do a cool move by employing the features of the imaginary world in clever ways.

Okay, so, two basic kinds of RPGs, two basic kinds of cool moves. Collaborative storytelling games, where cool moves are about enhancing and building the story and world, and world-simulation games, where cool moves are about wielding the contents of that world. Very different kinds of play, very different kinds of cool moves, but we can see a path forward in each.

What about RPGs without many cool moves? I get how to do a cool move in A Land Once Magic or The Vanilla Game, what kind of games make it hard to do a cool move? There are two common answers I see: constrained rulesets and confused rulesets.

Let’s start with the latter. Imagine you’re playing an RPG that, on its face, is about problem-solving—slaying dragons, political scheming, mystery-solving, and so on. You find a secret chest hidden in a wall and crack it open, excited to see what secrets lie within. But then, the rules say something like “Now you decide: what lies within the chest?” and whatever you say, it becomes true. Where we were previously all happily following the Character Rule, suddenly some rule inserts itself and asks you to break it, to step out of character.

As a player, this confuses me. If I’m trying my hardest to win using all the tools available to me—like I would in a post-OSR kind of campaign, or, you know, most other kinds of games—then I should obviously say “The chest contains a +3 sword of dragonslaying!” or “The chest contains signed documents proving the Vice President sold cocaine!” If we’re trying to win, I want to win. But, there’s a strong cultural tendency to avoid such blatantly favorable world-defining. It’s seen as crass, brutish, powergaming; it’s seen as exploitative, somehow.

I’m sure you’ve seen the opposite, too: you’re having fun playing out a collaborative soap opera of melodrama and big emotions, but then suddenly some rules-switch flips, and you step into grid combat. Where before making self-sabotaging decisions was part of the fun, now play becomes all about damage optimization and number-crunching, with poor tactical decisions being heavily penalized. That, too, is confusing.

When I play an RPG (or RPG-like game), I want to know upfront: is this a storytelling kind of game, or a problem-solving kind of game? The rulesets that try to blend both often feel like they pick up the worst of both worlds, demanding players switch between two very different sorts of minds or risk spoiling the whole affair. These overlapping rulesets make it very hard for me to do a cool move: I can no longer trust that my melodramatic self-sabotage won’t shaft us later, nor can I trust that my clever rope-beartrap-soap trap won’t be undone by the suddenly rules conjuring “an unexpected downside, hidden cost, or heavy price” into existence. It’s confusing! Why do we keep releasing these kinds of rulesets? I can’t tell what kind of game we’re trying to play, and that makes it hard to do cool moves!

These kinds of half-storytelling RPGs are but the first kind of RPG I see that makes it hard to do a cool move. The second is what we might call the overly-constrained RPG ruleset. These are the rulesets that insist on every action falling into their written rules system—that the only state space that matters is the state space explicitly defined by the rules. If it isn’t in the endogenous rules, it isn’t allowed.

This was super common in the ’90s and ’00s (and still fairly popular today), where you saw ever-larger lists of skills, items, and options to try to cover every possibility of the imaginary space (an impossible task), like the endless glut of 3.5e splatbooks. These rulesets become a “dictionary” of possible moves: anything outside that dictionary doesn’t exist. I don’t think its coincidental that many of these overly-constrained RPGs emphasize tactical grid combat—far easier to say “this is what you can do” when space is so neatly and rigidly defined.

These constrained rulesets make it difficult to do a cool move because every option must be designed by the book’s writers. Videogames get away with this because they can do far more math in the background, running physics engines and procedural generators and so on—on paper, Diablo and D&D have similar basic rules, but the sheer quantity of math being done to keep Diablo running far outstrips anything a human could do. Things can happen in a videogame that the player doesn’t quite understand and it’s fine; in an RPG, if the players don’t understand how a class feature or spell works, the game screeches to a halt. Practically zero videogames explain the full extent of their internal mechanisms because those mechanisms are so complicated they’d just confuse players.

RPG writers can’t afford to create rules of such complexity. Without players, RPGs do nothing, so an RPG ruleset must be simple to understand. But, all this rules simplicity creates a problem: how do you do a cool moves if the state space of play is very constrained? If the designer must write out every possible move, and the players must be able to read and understand every possible move, where do we get cool moves?

The typical answer for these kinds of traditional rulesets is breadth. They bloat so quickly because in order to maintain momentum, there always need to be new options available to players. If I show up to the table with an official splatbook in hand, ready to play some harebrained new class, for a few sessions, the fact that no other player knows what my character can do makes it seem cool. It’s hard to grapple and comprehend with vast quantities of options, so even simple-but-unknown actions can maintain some of that cool factor.

Eventually, though, this breadth runs out. If feeling cool relies on research, your players will hunt down every option and combination and use it to build ever-more-optimized characters. As a game designer, you must constantly supply new options to stay ahead of the diehard fanatics—who, unfortunately, are also your main market. At the table, this kind of emphasis on character-building transforms play into essentially score-tallying: how well did you build your character? Did you maximize every value, optimize every choice? Did you spend enough time doing research? Sure, you might occasionally stumble upon a cool synergy in your build, but a cool character build does not make for a cool move in play, live, at the table together.

I see both of these—the ruleset that demands all actions stay within the predefined endogenous bounds, and the ruleset that constantly breaks the Character Rule while asking players to problem-solve—all the time. They account for the vast majority of unfun sessions I’ve played, sessions where it felt like I was either being pulled in two directions at once or else utterly locked into a grind. I want to feel cool playing an RPG, and to do that, I need clarity of purpose, and I need freedom. Without these, play loses its spark.

I described two main methods for doing cool moves in an RPG (or RPG-like game): collaboratively storytelling and redefining the world for extra creative drama, and wielding an ironclad imaginary world in clever exploits. Critically, neither a videogame nor a board game can recreate these—both are too constrained, too closed of systems. And yet, because so much of RPG design culture cribs from these other, more popular mediums, I keep seeing designers trying to imitate the closed systems, and it always fails eventually.

So, the next time I sit down to play your game, I’m going to ask you: how hard is it to do a cool move in this game?

March 19, 2025 design game studies writing

Ten Tangible Tips For Editing Your RPG Manuscript

1. Remove instances of “to be” verbs: be, am, is, are, was, were, being, and been. Instead of “The goblin is dancing,” write “The goblin dances.” In most cases, this also includes the added benefit of changing the passive voice into the active voice. In cases like “The goblin is wet,” instead say “Water drips down the goblin,” or simply “The wet goblin…”

2. Remove instances of “to have” verbs: have, has, had, and having. Instead of “The guards have the keys,” write “The guards keep the keys on their belts,” or even simply “The guards hold the keys.”

3. You can probably remove instances of “any,” “every,” and “all.” Instead of “The cursed flame melts all metals near it,” write “The cursed flame melts metal near it.” Instead of “Every day, the townsfolk eat 1d6 rations,” write “Each day, the townsfolk eat 1d6 rations.” Instead of “Whenever the clock strikes twelve…” write “When the clock strikes twelve…” The first case—any and all—are just unnecessary. In the second case, “each” is more precise than “every.” In the third, “when” does the same work as “whenever” with less space.

4. Along the same lines, avoid using “Anyone who gets…” or “Anyone who does…” Instead of “The explosion burns anyone near the door,” write “The explosion burns those near the door.” In general, those works wonders for these kinds of if/then statements.

5. End sentences on the strongest, most evocative word. Instead of “A ghoul stands in the yard,” write “In the yard stands a ghoul.” Ghoul is a far more exciting and grabbing word than yard, so put it last. When you can, end your paragraphs on the strongest sentence (which ends on the strongest word).

6. Conditionals belong in the present tense. Instead of “When the water reaches the ceiling, the glass will break,” write “When the water reaches the ceiling, the glass breaks.” Almost every instance of the word will can be cut from your manuscript, changing the future tense for the present.

7. Avoid expletive constructions. Instead of “There is a red vulture on the roof,” write “A red vulture sits on the roof.” In general, be careful of there is when it appears.

8. Important game quantities (enemies, loot, precise distances) are written with numerals, but all other quantities are written out. Instead of “Six orcs attack intruders,” write “6 orcs attack intruders.” Instead of “The village contains about 15 houses,” write “The village contains about fifteen houses.”

9. You don’t need that bolded keyword: if it’s an important rules term, capitalize it, but otherwise leave it alone. Instead of “The dungeon contains 6 threats,” write “The dungeon contains 6 Threats.” Almost all manuscripts full of bolded words (or words in smallcaps, italics, etc.) are worse for it.

10. Paragraphs are your friend, as are paragraph breaks. You don’t need fancy parenthetical notation, you don’t need complicated nested bullets, you just need to write sentences. When you need to connect multiple ideas or have too many sentences, put multiple paragraphs together using regular old paragraph breaks.

11. Bonus tip: use the New York 1d6 for dice values. Instead of “d6 goblins appear,” write “1d6 goblins appear.” As you reach higher multiples of dice, the notation maintains consistency throughout.

If thinking about editing like this inspires fear and dread, I am available for hire as an editor.

September 13, 2024 writing

What Does a Work Not Need?

Ten years ago, Steven Soderbergh released an unusual cut of Steven Spielberg’s Raiders of the Lost Ark: he put the entire film in black and white, then removed all the audio and replaced it with Trent Reznor’s soundtrack to The Social Network.

I watched this cut of Raiders for the first time only recently, and it struck me. It’s like watching the most exciting 1940s action movie ever made set to music that sounds like it belongs in Cyberpunk. I’ve seen Raiders before, years ago, and I know the basic plot beats, but it hit me, again and again, how tense and fun and engaging the movie remains even without the detail of spoken words or Williams’ iconic soundtrack. Even the dialogue-heavy scenes, where Dr. Jones talks with the feds or Belloq flirtily captures Marion, are still dramatic and gripping. It was stunning, shocking, to realize I could watch and fall in love with a movie that seemingly lacked so much of what, in other films, felt essential.

Before I watched this Soderbergh cut of Raiders, if you asked me what components made a film, I would’ve listed the usual things: acting, directing, lighting, cinematography, sound, costumes, set design, etc. Now, though, I feel like my third eye has cracked open just a slit, and I can see lots of other elements, things I’d never noticed before until everything else was taken away—blocking, staging, choreography, composition, props, cuts, pans, silhouettes, and more that I lack the film vocabulary to describe.

The next few hours and days were spent googling film terminology and rewatching old “Every Frame a Painting” videos with new eyes. I started to grapple less with Soderbergh’s Raiders cut in its own right and more with the academic-artistic process the film pushed me through. It’s rare for me (perhaps embarassingly) to find a work that really makes me re-evaluate my understanding of a medium, and even rarer to find a work that does that through the process of subtraction. Soderbergh’s cut is, basically, just Raiders. Sure, he removed the colors and solors and tacked on the Reznor soundtrack, but the resulting piece feels like less than the original. In reducing the work—the form, even—down, Soderbergh illuminates all of the artistry in the piece. And, indeed, reveals all the artistry that other films sometimes lack. Do you think you could watch a Marvel flick in black and white with no dialogue and still understand what was happening, let alone enjoy the experience?

In short, by removing many elements I took as necessary, Soderbergh’s cut of Raiders revealed all the depth and meaning I had previously missed. By cutting away elements I typically fixate on, Soderbergh illuminated new dimensions of what was possible.

And now, a small confession:

This post is not about Steven Soderbergh’s cut of Raiders of the Lost Ark. This post is about The Isle.

The Isle is a 2022 tabletop RPG adventure written by Luke Gearing with editing by Jared Sinclair and graphic design by Micah Anderson, published by Spear Witch. Its stats are for The Vanilla Game, but it’s easily converted to other rulesets.

The headline: The Isle completely changed the way I thought about RPG adventures. It completely changed the way I thought about writing in RPGs. It completely changed the way I thought about the work I want to make and the projects I want to release. If I had to point at a single project as the thing that made me most change my views on making RPG books as a craft and medium, it would be The Isle.

As a work, The Isle is an enigma. On the one hand, it’s basically just another dungeon. On the other, it’s unlike any adventure I’ve ever read before. When I reached the last page, alone in my room the night before my session, I literally gasped aloud from shock and surprise. The confusing thing about The Isle, the almost paradoxical nature of it, is that it doesn’t actually do anything new, really. Rather, it just strips out everything I thought was necessary, and the remaining result made me rethink what an adventure could be.

So, what doesn’t The Isle have? Well, for one thing, it doesn’t have an overview. There’s the briefest of introductions, and then it launches straight in; I had no idea what I was in for when I started reading. It doesn’t have illustrations or major graphic design elements; Anderson’s work is sparse and simple, and their cartography is equally minimal. It doesn’t have fancy typography or text effects (Cormorant body text; Neue Haas Grotesk headers; stat blocks and details done in italicized newlines; all in black and white). It doesn’t have sidebars, advice, “how to use” sections, or other explicit author-to-GM asides. And barring that very short introduction, it doesn’t have any text that isn’t a location or description of things in a location.

To put it bluntly, The Isle is basically just a bunch of room descriptions. And yet, reading the book was a continual process of repeatedly asking myself “Why did they [not] do it this way?” thinking about it for a minute, and realizing “Oh, of course.”

Why doesn’t it have an overview? Oh, of course, to make reading the book a process of discovery. I had to read the book to know what was in it so I could run it; I wanted to read the book because I was desperate to know what was on the next page and in the next room.

Why are the rooms in sometimes slightly odd orders on the map? Oh, of course, so that the most impactful, most exciting, most jaw-dropping reveals come last. Whenever I finished a floor, I wanted to go back and read the whole thing again just to pick up all the hints and connections I realized I missed.

Why isn’t there a sliced-up minimap on each page? Oh, of course, because these map files are included so you can print them out next to the book and write all over them. This is a text meant to be used at the table, and it’s vastly easier to just have another page with the map on it.

Why do all these rooms have little preview descriptions as their first lines? Oh, of course, because they’re what you can detect from another room, so when you need to scan through from one room to the next they’re easy to find and read. This is a piece of design tech that, while it requires careful coordination and consistency across the map, makes huge strides in making the dungeon a more connected place.

Why is there no bold text, no keywords, no summaries? Oh, of course, because this is a book to be read and not a reference manual to be skimmed. This is also why, I suspect, the rooms have keyed numbers but not names: if you want to know what’s in a room, you have to read it.

This goes on and on. There are dozens of minor technical choices in the writing, editing, and presentation that depart from the standard of contemporary adventures in favor of a blunter, more direct, more impactful methods.

The Isle was, I think, the first adventure I’ve read where I felt like I actually had—and yet also wanted—to read. If you want to know what the dungeon holds, there’s no other way to find out than to simply read the text and learn. In so many RPG books, the text of the book opens with a pitch (“It’s a cursed vampire’s castle in a haunted land!”) followed by explication of that pitch over the course of dozens or hundreds of pages. Sometimes the keyed locations of an adventure elaborate on that pitch in exciting, unexpected ways, but so often it’s simply more words detailing what you, as a reader and sometimes as a player, basically already knew. That the summary exists at the front encourages readers to skim the main text; it gives writers a sense that their writing won’t be read. The Isle is not like that. It plays its cards close to the chest; if you want to sniff out its secrets, you have to read the words on the page. It’s an adventure that’s impossible to wing your way through without reading it first, but at the same time, after reading the adventure you have everything you need.

And this is without touching on the words themselves, really. The imaginary dungeon described in The Isle is a text in its own right: it has embedded histories, recurring motifs, foreshadowed mysteries, overarching themes, and ideas about the world and human condition it wants to communicate. Dark, horrific, disturbing ideas, often, but ones that shine throughout. Gearing is a veteran RPG writer, and understands better than anyone how to seed key concepts early, hook the reader (and player in), then pay it all off by the end.

One of the most striking moments running the dungeon (after reading the book) was when my players made a realization—a sneaky hidden lore thing—that I myself had not. I was the one narrating the rooms they saw and the creatures and objects within, but I’d missed the secret connection between a couple different areas. As a reader, I had access to more or less all the same information they did as players, moreso even, but I hadn’t made the same realizations. By not revealing its secrets directly, by not explaining itself, The Isle places the GM in the unusual and delightful position of sometimes being more or less equally informed as the players. At the same time, when my players went to execute on that information by exploiting an enemy’s vulnerability they garnered from that lore, I found that the module had already written that vulnerability into the monster’s statblock. Despite not knowing all the secrets of the module as GM, by simply following the text I provided my players with an exciting game.

This is exceptional writing and design. It removes a layer of interpretation, of noise, between the author of the dungeon and the players at the table. Once I realized that I didn’t need to know everything, that I didn’t need to worry about patching holes or filling in gaps, I could simply relax. Gearing’s writing (and Sinclair’s editing) is such that, once I read the book, my work was done. I needed to adjudicate as referee, of course, but I never felt pressured to guide them along some intended story or deliberate outcome. Some adventures feel like they’ll fall apart if I don’t hold them together, but The Isle is the opposite—I don’t think I could break it if I tried.

The Isle is unique in that is meant to be read and meant to be played, but little else. Many RPG books seem obsessed with GM tools, resources, options, generators, and other such things. They serve up a glut of raw stuff with the expectation that a GM will pick and choose their favorites for a given table or session. By packing every conceivable feature into a book, there’s an expectation that a given GM is getting more value. These are books designed as compendiums for out-of-session; bits and pieces to use when making a campaign, but not actually seeing live use behind the screen at the table. The Isle isn’t like that. While you certainly could rip out its monsters, items, and other content, it’s not designed to do so. There are no tables, no generators, no lists of isolated content to slot elsewhere. Rather, it’s a very specifically designed adventure to be read, cover to cover by a GM, and then used behind the screen as you play a session.

Ironically, or perhaps fittingly, I think this actually enhances the degree of impact that Gearing has over play. Because The Isle is just a series of rooms, of content, it means that players moving through it necessarily engage with his work. As a complete imaginary space, the Isle—the literal island—defines the possibilities of the game being played vastly moreso than any ruleset or series of touchstones could hope to.

People love to talk about the importance of vibes in an RPG book, but I think The Isle makes the opposite case: if your technical foundations are solid, if the writing is clear and the content exciting, then you don’t need to know the vibe at all. The GM just needs to read, and the players just need to play. Will their vibes match the adventure? Will their character smoothly fit into the world Gearing made? Will they get the “intended experience?” No, of course not, but that’s the point—Gearing’s monsters and locations and setting simply exist. What you do with them, as players and GM alike, is up to you; nowhere in the text does the trite RPG declaration of “The Isle is a game about [X]…” appear. The Isle isn’t a game, and it never claims to be one, and yet rests confident in the strength of its content. If you want to play in the pseudo-historical Viking-horror mode that most of The Isle exists in, you certainly can—but you never have to, and Gearing makes no claims regarding the best way to play.

Herein lies the contradiction of game design in RPGs and the adventure: the author of the adventure presents the world, the contents of the space the players move through, and thus many of its diegetic rules. At the same time, the adventure makes no assertions over what players want, or how they might go about fulfilling those desires, or indeed even the means available to players to achieve those goals. A typical RPG book offers a thousand options and variations, none of which need be followed, but in their width and breadth ensure that every table more-or-less follows in the designer’s vast footsteps. The Isle—and adventures in its style—forcibly inserts a spare few elements into the imaginary world that demand action and reaction by the players, but offers nothing beyond. No player can avoid Gearing’s influence when playing through The Isle, but Gearing’s work never escapes unmarked: no two tables will play through The Isle in the same way, and none will play like Gearing. By presenting direct authority over the imaginary world, he imposes more than a ruleset ever could; by presenting nothing else, Gearing ensures the players need never respect his wishes.

This seeming-paradox, this ultimate enforcement of the diegetic world-rules combined with total abdication of the exogenous and endogenous game-rules, presents a solution to a question that’s vexed me for years: how can I be a game designer without being an authoritarian? That is, how can I make games and game-like things without needing to enforce my whims and will at every turn? Most RPG scenes are filled to bursting with designers very eager to dictate their exact intended experience, but that mode of design has long left me cold.

The answer that The Isle opened my eyes to is the adventure. Adventures present a world, but no constraints. Good adventures present a playground to romp through, but have no intentions. In the medium of imaginary worlds that RPGs play with, the adventure is not the game but rather the toy. Players can do whatever they please with a toy; in all likelihood, they’ll invent new games to play with it that I never dreamed of.

But adventures are also full of game design! Really classic game design, even—building a dungeon that’s fun and exciting and engaging to navigate through uses every skill of what we call “game design.” So much of what goes into RPG books comprises less active game design than a kind of meta-instructional game design: there are myriad books that tell a GM how to play and run a campaign or session, but far fewer that present the content needed for those same sessions. The Isle contains no explicit advice, yet presents dozens of extremely deliberate examples of stellar design in action, ranging from the information NPCs possess to the floorplan of the dungeon levels to the symbolism of recovered magic items. The Isle is not a book about good game design but instead is a book of good game design.

The Isle contains almost none of what I expect an RPG book to have, yet remains among the most playable RPG books I possess. Gearing, Sinclair, and Anderson stripped out the shiny components that sell an RPG book and instead present merely the most basic elements of their medium: words, maps, and the imaginary world.

What if every movie, stripped of color and sound, looked as good as Raiders?

September 3, 2024 design review writing

In Praise of Legwork

A question: you ever try to run a campaign with a city in it? Not just a town or settlement or fort, but a proper city, one with an actual street map and whole neighborhoods and thousands of people living within the walls? I certainly have, many times. It’s hard! It’s really hard.

I live in New York—one of the biggest cities in the world—and the complexity of the streets and the depth of the city itself is just staggering. Every time I think I understand a neighborhood, or even a couple streets, I find new nooks and crannies tucked away. I often wonder, when my players reach a city, how I can best replicate that feeling. How can I create that particular impression of unfolding layers of people and places, where there’s always something new to discover just around the corner?

The answer, I think, lies in City State of the Invincible Overlord, the ’70s-era Judges Guild city module. It depends a little which version you read, but more or less all City States share a common feature: huge lists of NPCs, in their hundreds and thousands. Crack open the PDF, and seemingly-endless quantities of people leap out at you, all sorted by their main locations—shops, apartments, temples, and so on—which in turn get sorted by the street and neighborhood.

On some level, City State overwhelms. Its ancient graphic design and information ordering certainly doesn’t help. But at the same time, once I started digging into the book, I found myself getting acclimated. There’s a learning curve to the text, but once I started climbing the curve, I realized the huge breadth of information and content available. Here’s an example, from By-Water Road (“PROB 60% Run Off Road by Horse Racing Pages”):

Three example locations from the 1977 Revised version of City State

The horizontal line of numbers is the stat block: class, alignment, level, hitpoints, armor class, save level, strength, intelligence, wisdom, constitution, dexterity, charisma, and weapon. Rhino Rudigore, the innkeeper of the Blue Dolphin Inn, is a neutral level 2 fighter—not particularly powerful in combat but quite physically fit.

Below his statblock, the book details the bartender, Koris Brightips, also with her compressed statblock (which you’ll notice, uncomfortably, includes an extra keyword at the start: “FEM”) and the note that she “sings.” From there, it mentions the Blue Dolphin’s customers—freemen, sailors, and nobles (with stat blocks)—the contents and dangers of his chest, the prices of the food sold (snakes fried in bear fat!), and gambling rates. I admit I’m not quite sure what “Legend of the Flying Citadel: Storm Giants Castle in Harridan Gap” is; I’d wager it’s a reference to another Judges Guild module.

Rhino, Koris, and the Blue Dolphin form just one entry of literal hundreds that the book includes. All this info is, of course, very geared for the particular sort of slash-and-burn ’70s adventuring style, but even sans stat blocks, these entries contain loads of exciting information. Just look at the next two entries: Jolly Naben, the (chaotic evil) smith that works behind the Dolphin, is close to broke and thus probably happy to sell you his “unicorn,” but also possesses valuable information about the Dwarven Mines; Squeaky Werter, the local racketeer, is the Thane of a Senator and thus mostly immune to prosecution, and so accordingly is loaded to the gills. The text could use a sharp editor’s pass, of course, but this is good stuff, ripe for adventure, all contained in just a few lines. Across an entire city’s worth, the book brings enough content to run a dozen campaigns, each unlike the other.

Which brings me to my original point: City State of the Invincible Overlord is, as far as I’ve ever seen, the closest thing to a city module that’s actually ready to run. GMs need to get through a lot of reading, obviously, and the book is more than a little painful to our modern eyes, but it does present a whole city. Bob Bledsaw and Bill Owen, the two writers, really did just take the time and effort to detail more or less every street there was. I’m sure it took a breathtaking amount of time and effort on their part, and for all its flaws, I’ve never seen anything quite as large or as detailed as City State.

This is what I call legwork: the long, slow, detailed, sometimes painful process of going through to figure out and then write down all the important details of everything that’s supposed to exist in the imaginary world of the game.

Consider a starker example, using John Harper’s Blades in the Dark (of which I’ve played dozens and dozens of sessions over the years). In the original Blades, Harper included a map of Doskvol, the city in which Blades is set; here’s a lightly-interactive digital version of Harper’s map. It describes about a dozen districts, each of which includes a few key locations and the basic outline of the major streets.

A few years after the release of Blades, Tim Denee released his “Detailed Doskvol Street Maps” with Harper’s blessing. Denee’s maps are huge: they include detailed maps of the streets, almost all of which Denee named, along with the canals, bridges, and rooftop crossings of each of the (also now named) neighborhoods within each larger district. A tremendous amount of effort went into these maps, but the difference at the table proves immediately noticeable. The ability to say “Oh yeah, our gang runs Fishmarket, with our headquarters located on the roof of Pike Court between Sharktooth Alley and Roe Lane” and have those be real places on the map feels immensely cool and exciting. (In my old campaign, players would war with other gangs and redraw turf lines according to the canals and streets—“Who gets tolls on the Moon Bridge?” was a hotly-disputed question.)

You can, of course, just name the streets and neighborhoods in Blades yourself, without these detailed maps. The option always remains there. But that takes time and effort, the kind of time and effort that I as a GM maybe can’t or won’t spend—the legwork. I was more than happy to spend the $10 that Denee charges just to avoid doing all that legwork myself.

Legwork is hard. It’s basically all the boring or fiddly or tiring parts of writing up an imaginary place. It’s pretty fun and breezy to say “Ooh, what if there was a city that was, like, on an island in a lake full of blood, and the blood came from the monster under the city, and the sewers, like, turned into its veins?” It’s even pretty straightforward to start coming up with names of important places and people in that city—the Fanged Gate, the Sanguine King, the Crimson Keepers—but sooner or later the going gets hard. Sewers that turn into veins is an exciting idea, but… what’s the map of those sewers look like? Which streets have manhole covers, and which buildings’ basements include access? For that matter, what exactly do the streets look like? And when the Crimson Keepers carve off chunks of the regenerating beast below, where do they store it, and how much do they carve per day? Once the text hits the table, cool concepts demand legwork, and legwork is a lot harder than dreaming up mere ideas.

In many cases, GMs who (quite reasonably) don’t want to do all this legwork just decide to wing it. I employed this strategy for a long time when it came to cities—draw a simple overview map, write down a dozen possible street names, then just slot them into place as needed when they came up during the session. This works basically fine, it’s enough to run a session or two (particularly if the players don’t spend long in the city), but it starts to fall apart when players start needing persistent details. And winging it really doesn’t work for dungeons, which, as more zoomed-in locations, demand a higher level of detail on a room-by-room basis. Unless you’re an expert improviser, sooner or later you’ll find yourself wanting the details that only legwork provides.

I come now to my sharper point: legwork is so hard that sometimes it feels like RPG writers and designers aren’t willing to do it at all.

A short list of some things that don’t take much legwork to make: rules; lore; character options; generators; “inspirational” material; and faction descriptions. By contrast, a short list of things that do take quite a lot of legwork: hexcrawls; dungeons; cities; and all the NPCs that actually go in a faction. There’s obviously some gray area between these—monsters, items, spells, and so on—but take a quick glance through DriveThru or itch or Kickstarter and tell me which you see more of. Is it the stuff that takes a lot of legwork to write, or only a little?

Once I saw this pattern, this avoidance of doing the legwork, I started noticing it everywhere. In OSR-land, depthcrawls and other generators provide shortcuts to making huge spaces (megadungeons, labyrinths, etc) without needing to do all the legwork in between. In storygame-land, the onetime rallying cries of “don’t write a story” and “play to find out” have become code for “just make the players (or GM) come up with everything themselves.” In trad-land, established companies and studios churn out books full of extraneous systems and new character-building options instead of adventures. All of these, one way or another, allow writers to skip doing the legwork.

I talked about all this a little with some friends, and they brought up an important question: is this actually an issue? What’s the problem with having lots of other non-legwork stuff on hand and just winging it? Is legwork actually all that useful?

The short answer to this question—and many others—is that a writer can afford to take the time and effort to come up with better content than somebody improvising it live at the table.

This isn’t true in every single case all the time (there are many great improvisers and even more bad writers), but in general, particularly when it comes to adventures, writers have the ability to work at a breadth and depth that GMs who wing it simply can’t. Yes, GMs can tweak things to enhance them for their specific table in a way that writers never can, but a GM can’t improvise a whole dungeon, let alone an entire campaign. When the imaginary world of the game brings a real breadth and depth, when there truly is something over the next hill and the hill after that and the hill after that, new dimensions of play emerge. Many campaigns and books claim to be a sandbox, but unless the world proves rich and detailed across the board, can players truly explore and do what they like?

Legwork proves its value to games even in cases that may appear less than obvious. Take, for example, solo games—a genre and form arguably about players doing the legwork. One of the most enduring and popular solo RPG books, Tim Hutchings’s Thousand Year Old Vampire, stands apart from many others in that it includes hundreds of prompts. And not just a random list of prompts, or prompts on any topic Hutchings could think of, but carefully-crafted prompts that nest and loop and feed into each other. The rules themselves of 1kYOV are simple enough; the juice, the reason that I and many others keep coming back to it, lies in its prompts—the legwork. If Hutchings hadn’t taken the time to write a couple hundred pretty good prompts, sure, we’d have a cool pitch and some neat vibes, but actually playing the game would be far more difficult and far less engaging.

Here lies the key message, the axiom I want to impart: the more legwork the writer provides, the less GMs and players need to do themselves.

It’s for this reason that I largely stopped buying RPG books that aren’t adventures. I can come up systems and vibes and inspiration myself, that stuff’s all exciting, but what I don’t like coming up with is legwork. In fact, I’m very happy to modify and adjust my own systems, or even my own campaigns, to accommodate a bunch of good legwork from another designer—like Denee’s maps. My Doskvol looked a bit different before, but suddenly acquiring an entire of Doskvol vastly more complete than my own made for an excellent reason to change those previously-established details.

My recurring example here, the city, proves only one case. Many recurring popular daydreams really can only be solved through lots of long, slow, fairly draining writing.

Ever wondered about a campaign of rusty space truckers who constantly need to swap and repair parts of their ship? You can of course just improvise it, or abstract it, but to really run that space trucker campaign, you need lists of hundreds or thousands of parts that interlock with each other—legwork.

Ever mused on a system of magic with different magical languages and mysterious components and complex rituals? You can do it on the fly and use a bunch of system math to simulate, but to really make that magic work you just need hundreds of texts, components, materials, and ritual processes—legwork.

Ever daydreamed about a real, proper heist, with guards and disguises and alarms and everything? You can retrofit and improvise it (like Blades does), but to really run a heist, you need guard patrol routes and workers’ weaknesses and camera locations and power lines and a million other tiny fiddly details—legwork.

The answer to my original question—how do you make a city feel like a city?—seems to have only one real answer: make a city, a big one, that’s broad and deep and includes all the pieces. There is, as far as I can tell, no other substitute.

Historically, we designers typically use clever math, rules, and out-of-character solutions to run these kinds of games. These work, sure, but they aren’t the real thing. To truly get down in the weeds of an imaginary world and still have it maintain consistency and excitement, you need a ton of legwork in the background to support your moment-to-moment play. You need the ironclad backing of a great deal of content, ready to hit the table without further preparation or improvisation. To play in a fantasy city, we needed City State of the Invincible Overlord; to play other dreams, we need more legwork.

June 19, 2024 design running the game writing