Architectural History for Gamers, 1: why build city walls?

Restored walls of the Ark (fortress), Bukhara, Uzbekistan

I’ve previously tried to sneak some architectural history into the gaming space with my maps of classic dungeons and D&D tourism series. This is a lot more direct – a distillation of some lectures I’ve given on topics around cities – why they happen, why they take the specific forms that they do, what the essence and function of various city structures might be. I’m putting them here because I think they might be useful for worldbuilding – if you think about the why of a thing, it informs the what and how.

So. Why build a city? Or, if you’re in a Jane Jacobs mood, why do cities form?

…..ok, before I go any further I should warn you that a lot of what follows is fairy stories. Any time anyone tries to tell you about what people were thinking or why they did things, you have to be skeptical. Look at the news – people can’t tell you why they’re doing things right now. Imagine how unreliable our information is about other countries, or times from which we have no eye-witnesses to interview. Imagine how much of what we “know” about prehistoric peoples is just made up – stuff that “makes sense” to us. And yet. When we’re telling the history of cities, we keep going back to our place of maximum ignorance – trying to explain how they first came to be.

The usual answer is that they’re either defensive formations – a group of families huddling together to dissuade raiders – or overgrown homesteads, where layabout children can wait for jobs to become available. Once you’ve hunted and gathered more than you can use this week, it’s nice to be able to trade some of it, maybe turn it into something unique that other homesteads will want to exchange for their surplus. With surplus and trade, you wind up with routes and hubs – centers of interchange, where it’s useful to hang about for a bit, waiting for other traders to show up and buy your stuff. Therefore roads and crossroads and market squares, eventually a network of hubs, each with their own identity and stored trade goods/treasures. And eventually you get surplus labor, i.e. bored people who invent stuff, and then the city becomes a center for novelties. And eventually those layabout children get jobs designing games or some other frippery. Lewis Mumford called the city “a point of maximum concentration for the power and culture of a community.” Jacobs said it inevitably becomes the center of invention, because of what Spiro Kostof called “a certain energizing crowding of people:” the fizz of ideas being exchanged in the public square, the availability of surplus materials, concentrated together where they could be recombined.

So here’s the ancient ruins of Khirokitia, on Cyprus, inhabited roughly 9000 to 6000 years ago. A concentration of round houses built along a road (or “great wall” – honestly either interpretation can be supported, all we have is the foundations – but it looks like it links them together). Note, the “wall” around the whole thing is an archaeologist’s intervention – as far as we know, the community had no built defenses beyond the walls of the houses.

It might have looked something like this, only without the bombard damage on the big house, which conveniently shows that it has 2 storeys inside.

And here’s more-or-less contemporary Çatalhöyük, in Turkey – a city of unprecedented density (over 2000 people!), continuously occupied for over 7000 years.

Don’t pay too much attention to the captions – archaeologist James Mellaart was convinced everything was either a shrine or a proto-Roman courtyard. Now we know the “courtyards” were rubbish tips.

2 things to note:
1. the houses are so tightly packed together that they’ve gone square instead of circular… which means one wall can work for two houses!
2. there are no roads between the houses at all.

Why no roads? Because people got into their houses – and from house to house – via the flat roofs. Which is an ingenious form of defense (against wild animals, other communities)… if you don’t mind the cost of making every roof in the place load-bearing, strong enough to have all your neighbours traipsing across it at any moment.

…also, all those ladders. I’m glad it’s not my job to carry big pots of water into the middle of all that.

Didn’t you say this post was going to be about walls? Yeah, it seems walling in communities was actually not that popular during the Neolithic – there were plenty of other ways to defend against the sorts of threats people faced in the years while they were inventing pottery. High castellated walls in particular seem to only become a thing once there are more effective missile weapons than spears.

But they still show some characteristics that walled towns would share:
1. they’re compact. People have clustered together and been accepted by their neighbours, and they’re clearly different from whatever is outside.
2. they share communal costs. Maintaining houses, common structures, and defense are all group responsibilities.
3. they show a clear inside/outside boundary, between domestic stuff that is close to hand for raising children and reproducing society and the stuff that doesn’t fit, that people don’t want in the collective front yard, which all gets ejected from the centre (or buried in a midden). This also happens in (unwalled) Kayapo villages in the Amazon rainforest in Brazil:

I can’t quite believe I’m retreading this old 19th century conceit of forest peoples being “primitive,” best considered alongside the ancient world, but this example does fit the theory neatly. This photo, of course, is of a totally museumified village, so you have to take the anthropologists’ word for its representativeness.

Here’s the anthro theory of the Kayapo village – that it’s a set of social categories expressed as a concentric spatial order. The community gathers in the centre (ritual hut and dancing space, shown in red, the Kayapo colour of sociality/engagement), they have individual family homes around that (in the unmarked white zone), and outside that ring of houses is the dirty space (in the Kayapo taboo/ignore color, black), before you get to the forest.

No physical wall is expressed but there’s a clear inside/outside, and what goes outside is… the same for the Kayapo as for ancient Romans – lepers, outcasts, tanners, burials, catacombs (OK, the Kayapo don’t build catacombs, but they are all outside the city wall in Rome).

So the theory of the city wall that’s common in architectural history and anthropology is that it expresses something like this:
– a social boundary thats separates the local community from the undifferentiated space of the outside
– a liminal zone where stuff can be ejected, therefore an area of non-stewardship or non-responsibility
– a controlling device that keeps threats (animals, raiders) out and productive citizens in.
The wall unites the city – it is usually its single biggest public work. Building city walls shows up repeatedly in ancient Greek foundation myths, as the critical moment of founding. Our word “urban” refers to the walls (urbs) of the city. Vitruvius’s ideal city is walled, and Aristotle’s definition of civilization (itself a word derived from civitas – the city) is a situation where a woman will be safe, because her scream will be heard from the city walls (and therefore any city that’s too big for the scream to carry must be too big to be properly civilized).

OK so, some principles.

1. walls define the community

instead of showing you a load of Greek stuff I’m going to China. These are Tulous – multi-family community houses – built by the Hakka minority in Yongding, in the mountainous southwestern part of Fujian.

Tulous are houses and towns. Surrounded by a defensive ring wall, they’re made up of several separate houses (segments, if you fancy a trip down the history of anthropology) built together. Communal activities happen in the middle courtyard, private ones in the individual houses’ courts. Property is shared communally within the tulou: the whole community is thought of as an extended family.

(it is BTW rather unfortunate that Disney’s live action Mulan decided to put the nationally-identified saviour of China in a Hakka tulou, given how the Chinese have recently been trying to erase the Hakka and incorporate them in the national “harmony”)

So the whole ring is the whole community, and the segment (one house) is a cake slice of it:

With living space on the outer rim and work space closer to the middle. The next tulou over is another, separate community, or was, before the Han ostracized and museumized them.

2. as a corollory, walls express inequality

Once you have specialization and division of labor, you get hierarchies and status differences and separate sub-communities and administration. In the classic/prototypical Islamic city this leads to a social division in the city – a high town of nobles and administrators is walled off from the low town of workers and producers. The walls may be practical or ceremonial – either way, their primary role is communicative.

Bukhara Ark, containing the emir’s palace and vizier’s offices. When first constructed, this wall was surrounded by commoners’ houses, but the Soviets tore them all down to leave it marooned in a level plain.

The wall of the Ark is in blue, the city wall encircling pre-Russian-invasion Bukhara in pink.

Aside from protecting the rulers from the citizenry, walled subdivisions might also maintain boundaries around foreigners’ quarters (usually outside the locals’ wall, rather than walled in, although the caravanserai/fondaco is a special walled building type for enclosing foreign merchants), royal hunting parks/gardens/harems, and family compounds, as in those hilltop walled towns of Tuscany that everyone loves to cite as birthplaces of rationalist humanism:

Tiny San Gimignano still has 14 giant towers, each guarding a separate noble family compound that at one stage completely screwed the town’s street plan and traffic flow. UNESCO offers an overview from the viewpoint of preservation, Medina Lasansky has a whole book about how the city you see today is Mussolini’s reinvention.

When the Ottomans conquered Constantinople they put up a load of mosque-hospital-bath complexes, free for the use of all the city’s Muslims, to show the citizens the benefits of converting to Islam and being good Ottoman subjects. But they walled them off, so that the city’s Jews and Christians, who didn’t have the right to enter, would be both excluded and intrigued. The walls became a symbol of social hierarchy.

The Suleimaniye kulliye in Istanbul, walled off from the rest of the city even though it’s supposed to serve the public.

3. walls have gates, and pirates and emperors love a choke point

Control over the boundary = control over community membership. And if you can control who goes in and out through a narrow choke point, then you can have power over them every time they want to cross that boundary.

diagram of a city: the community is orange, the outside black.

So the second purpose of building a wall is to restrict traffic to a single point where you can extort money from it.

The choke points of Europe’s waterways are a wargamer’s or toll-taker’s wet dream. If you want to get from Mariupol to the Atlantic you have to cross the Kerch Strait, the Bosphorus and Sea of Marmara, the Dardanelles, the Marsala gap, and the strait of Gibraltar, and every one of those chokes has been used as a toll gate.

High, thick walls have a propaganda effect – look how safe you are in here – but the everyday point of the wall is toll-taking.

House for weighing imported goods to assess customs duties, Amsterdam. In this case the wharf is the gate.

One of the oldest types of neighbourhood associations in Europe is the Gate Defense Association – a communal organization that collects all the tolls and then redistributes…. some of them… to the deserving poor and public works like… repairing the church or the association’s HQ.

gates of the Ark, Bukhara

4. walls state what you are prepared to defend

The Mongols tore all the city walls down when their empire was expanding, then put them back up when it started contracting: if you control the countryside, you don’t need them. If you need them, then you only really control the city. That impulse, to add or subtract walls from cities, has played out many times in Europe and Central Asia. For an emperor, walled cities are potential sources of rebellion, fortresses of power for over-mighty subjects. The license to crenellate your manor house is a barometer of state power.

If you charge customs on goods that enter the city wall and also refuse to defend the parts of the city that fall outside your wall, you get faubourgs – commercial districts that operate (comparatively) freely from state intervention. And of course if the city extends too far beyond the walls, you have to build new walls (funded by new toll-gates).

For a legendary unplanned example, check out Baghdad as it stood under the wise and pious Harun al-Rashid:

Baghdad was planned and built from scratch by Caliph al Mansur and his vizier Khalid ibn Barmak to be the Caliphal seat, the perfect city of peace, round as the world because it represented the world – or at least the world of Islam, the abode of peace. There were ideas about being able to shelter the whole population of the realm within its encircling walls. And it had four gates pointing to the four corners of the Islamic domain and/or four greatest other cities in the Islamic world – a fortress, a symbol, and (at last) an administrative center capable of governing the vast and unruly Caliphate. It was, by several orders of magnitude, the biggest single building work conducted under Islam. al Mansur’s brother had recently been propelled into the Caliphate by a bloody civil war, his reign depended on slave soldiers and propaganda. Baghdad was to prove that there was substance to the new world he promised. The specs were outlandish – millions of tons of baked and glazed bricks, bronze, copper, and gold domes, viziers’ offices and barracks and mosques and great parade grounds, and the whole thing was to be enclosed in a stout, circular wall. There was just one problem: to build such an orderly monster, you had to import hundreds of thousands of workers, who needed somewhere to live, and markets for food, and stores for materials, and brick factories….. And so before the Round City could be raised, there first had to be built a sprawling, unplanned, supposedly temporary city for the workers. And of course today there is no trace of the Round City, but Baghdad the sprawl remains. And that’s why the Arabian Nights is full of stories in which Harun (al Mansur’s son) disguises himself to walk outside the palace among the common Baghdadis and discovers terrible injustices, which he demands Ja’far (Khalid’s grandson) put right.

Successful cities outgrow their walls and successful territorial states have progressively less and less use for them – a bloody great wall in the middle of your city plays havoc with traffic flow (by design).

Paris in the 19th century, showing concentric walls showing several eras of the city’s expansion. Before Haussmann’s road-building program

The traditional thing to do (in London, Paris, Vienna, and Moscow, at least) is to tear the walls down and turn them into unusually broad streets, maybe with a line of trees or a whole park to show how important they are – the origin of the Parisian boulevard (bulwark) and Vienna’s Ringstrasse (ring-road – because the walls form a ring around the old city). And I know it sounds contradictory, to first build a giant wall and then flatten it again, but trust me, that’s easier than displacing all the people who would’ve built houses all over that prime urban land if the wall hadn’t been there.

The Ringstrasse is now one of Vienna’s top attractions, as well as the means for visiting the others. Not bad when you consider it used to be a rubbish tip.

Paris’s last wall, erected in the 19th century by Thiers, became the Paris ringroad – which is not leafy or park-like, but with its notorious traffic jams and noise it could be considered a defense of a sort.

Next: how to defend your walls, or: fortifications through the ages.

XCOM in the mirror – Phoenix Point

In the previous posts in this series, I’ve praised XCOM for its focus, clarity, and elegance. It gives you the right amount of information at the right time to allow you to make informed decisions. It aims for simplicity, winding up at a level of complexity that allows for good tactics. In those parts of the game where it’s less clear, it’s also less successful.

Those posts are all retrospectives – written after I’d had a chance to absorb and think about XCOM’s lessons, not as I was experiencing them.

This post is different: it’s about Phoenix Point, another game made by some of the same developers, which shows how you can get it wrong, where XCOM gets it right. And it’s written as I’m in the middle of trying to get to grips with it.

Phoenix Point is what happens to XCOM when its enthusiasts want more of everything. More enemies, more complex variants on those enemies, more research, bases, faction diplomacy, decisions, action points per round, weapons… and most of all, more ways to get into a position from which you cannot win. At the same time, less information up front – more interface ambiguity, more traps hidden in subsystems, which slow you down in your race against the enemies, more forcing the player to take leaps in the dark and then backtrack later. The overall result is more confusion, more paralysis, more stress (of the “I think I unwittingly did something wrong last week and now I’m screwed” variety), and less fun.

Geoscape view looks familiar! Each of those little circles with logos inside is a site I’ve explored. Blue and pink logos are bases of different factions, whom I can befriend or raid or rescue from disease/alien incursions, which come out of the red areas.

On the tactical level, where XCOM has 2 action points, for MOVE and SHOOT, PP has 4, so you can have actions that take 1, 2, 3 or 4 points… the result of which is that your heavy gunner and sniper, who have 3-point guns, are slowed to a crawl, having only 1 point left for movement. Where XCOM has a lot of one-shot kills, PP routinely takes 3 characters’ actions to take down one enemy soldier… the result of which is that combat is much slower and less predictable, tactically – because most of an enemy squad will escape any ambush you set, or turn the tables on you by failing to die and instead swarming you. And where XCOM gives you infinite reloads, PP wants you to gather up ammo from the battlefield (or, by default, leave it there when you end a mission… so you get into a situation where you’ve probably killed all the enemies but maybe not, but you don’t want to Do The Thing to declare victory because there might be some valuable bullets lying around in the opposite corner of the map). And where XCOM has a sometimes-enraging %-to-hit counter, PP has a touchy-feely “how much of this circular sight does your enemy occupy” interface, which takes multiple clicks to access but maybe appeals to sniping enthusiasts but… those guys are probably playing Call of Duty instead.

In brief: XCOM = quick and deadly, PP = slow, fiddly, and less tactical.

Character Advancement is both less clear and less varied than XCOM 2. There are 3 classes and everyone can dual-class at level 4, so your squad of 6 people certainly contains duplicates – it winds up being a sort of class-and-a-half system. Why would you do this? The main reason I’ve found so far is to overcome class restrictions on equipment. Snipers can’t normally get the Heavy’s jump jets, which is the only practical way to get to many of the sniping platforms in levels. Dual-class, like Omar, below, and you can jump to a sniper’s perch… and leave your short-range Heavy gun at home. Update: turns out more classes unlock later in the game but they have, like, funny heads? Or mutato-cyborgo bits? I have no idea, and they won’t ever explain. The fun will be in finding out.

Ammo management is a huge part of the game, but there’s no clear interface for reloading a weapon. And there are like 6 different minor variations on every basic weapon because…………..? I guess to make ammo management harder. So that’s why you end up with a giant Storage bin on your loadout screen. For 2 dozen ammo types. Update: and there will be more kinds of very similar weapons as you advance. Also, make sure you never fire that stripey cannon shown in the picture below except in direst extremis, because you’re never getting any more ammo for it.

But where the game really makes strides is the strategic layer. This was kind of rudimentary in XCOM. Now it’s…. a lot more complicated.

This video, ostensibly offering tips for the strategic game, actually provides a really thorough list of its design problems. First, it says the way to win the game is to strike before the aliens get to show you much of the content (which, OSR sensibilities aside, seems like a funny way to spend your development dollars). Friendly relations with other human factions are vital, so you should murder the first few humans you meet for their stuff, because the bad will that generates will only grow more expensive later (mixed messages, anyone? What sort of character are you playing, anyway?). And you should avoid researching anything that you can get by any other means – the video explicitly calls most of the research options “traps:” wasteful sinks for time and resources. Avoid developing your bases, building aircraft (steal them instead – but early on, before factions get properly mad at you for doing so), and try to win faster than both the aliens and the human factions’ mutually-assured destruction countdown. In other words, absolutely do not try to explore the things that the game puts in front of you. Along the way, there are several sub-games you should exploit without getting attached to them: a gold-farming trade game, diplomacy that’s kinda doomed, and a flying Godzilla, impossible to attack, that destroys the friends you first invest in.

Do I sound bitter? Well, I’ve only sunk 20 hours into it, and I’ve already realized that I should probably restart rather than sinking another 20 into a lost cause. I bet that puts me ahead of the average player.

The experience has, however, reinforced one thing for me: the importance of managing the player’s cognitive capacity, when learning a new game. Dribbling out information, letting the player grok the basics before getting fancy with special conditions. Some folks resent XCOM’s hand-holding tutorials, which callously kill off your soldiers to show you that soldiers will die and throw grenades at you to show you how dangerous they are and so on. When you’ve gone through XCOM’s 0-level funnel-like first mission, you think “that was really basic, I totally could’ve handled more options” ….but XCOM is smart to start with that funnel, allowing you to figure out movement and cover without thinking about the affordances of different character classes.

PP offers a very unattractive alternative, where you’re assumed to know how XCOM works, therefore you’re ready to have your expectations subverted. It throws you right into its XCOM++ world with all the options open from the start and… it’s overwhelming. Hundreds of (trap) research topics; pages and pages of (mostly probably irrelevant) lore about factions, your own history, the aliens; half a dozen separate base/aircraft/weapon/manufacturing systems to optimize, much of which, apparently, not worth the bother. And laced through it all, a basic failure to explain. To take one, stupid example, look at this interface element:

a yellow, dotted circle around your aircraft. It shrinks as you move toward a goal.

I bet you think it represents the range the plane can go without refueling at a friendly base, right? I was pretty scared even to approach that edge. I thought I might have to always have enough fuel aboard to make it back home. No. The plane can auto-refuel (if, in fact, that’s what it means at all) at any abandoned graveyard/alien base/curious swamp-grotto you’ve explored. You can tell it to go somewhere far outside that circle and the circle will just keep re-setting to max every time you pass over an icon. It took me 20 hours to find that out by accident. There is no manual where the dotted circle is explained.

If only the game dribbled that stuff out a bit, maybe its complexity would be a virtue, not a vice. So my positive lesson from this not-so-positive experience is, an empire-building game/earth-girdling campaign begins with a single step. Is there exploration? Let the players explore a small map (maybe an island or cave) first, revealing the next, bigger step when the first one is internalized. Is there redundant research? Let the players do some useful research first and then give them clues about what redundancy looks like, before having them lose a game to find out… There is never a point in the game where you should abandon this design principle: any time you’re about to radically expand the scope, even late in a campaign, add a little training, where the players can figure out the new possibilities before they have to depend on them. And for god’s sake, teach the players about any new thing with a good, reliable example, before betraying their hopes with a treacherous one.

Update: this is reinforced by the fighter plane minigame, which unlocks around hour 40. If you want to know why that ends the game (for me, at least) read below. If you don’t need any more reasons to avoid, then avoid this too.

——— actual whining below ————

LATE ADDENDUM: I get obsessive about video games – in general I avoid picking them up because I find it so hard to put them down again, but Phoenix Point has offered me an off-ramp after 60 hours of play, and I’m taking it. I mention it here because I feel there’s a lesson about RNGs and the need to place some limits on them.

PP has a small but real chance of just throwing a giant boss monster into a level full of mooks. Specifically, a super-beefed-up version of the Acheron into a game where the ordinary Acheron has just appeared. The Acheron is itself a Swiss Army knife of disasters – it introduces a new mechanic that directly erodes your soldiers’ abilities to resist it, it summons 1d4 mooks a round (which means it might double the monster count of a level in 2 rounds before you even know it’s there), it jumps, giving it more and different mobility compared with other monsters, and it’s a plain old giant bag of hit points and damage-dealing. So randomly throwing one into a level is wildly destabilizing, like making the level 6 times harder, where the regular monsters are supposed to offer a “balanced” challenge. It’s like having “red dragon” on a wandering monster table for a first level dungeon and then locking the players in one room with the dragon so they can’t run away.

OK, fine, reload the level, it won’t have the Boss, continue playing. If you allow savescumming, I guess this is just the game scumming back. But… what if you have a similar random avalanche waiting to kill the player, that gets set at the beginning of the game and that only goes off, say, 40 hours into play? 40 hours is a lot – maybe the player could plan around it? Maybe it’s not that big a deal if they know it’s coming?

Yeah so the story is, if you buy the complete PP now, it comes with all the later DLCs auto-installed. And one of those DLCs has a flying Godzilla that randomly wrecks everyone you’re trying to trade with. It sets you up for a whole arc where some people will get wrecked but then you’re in another arms race to learn how to resist it and eventually you can fight back. And to get you started and give you some hope, the DLC also straight up gives you a fighter plane to fight it with, at exactly the same time the Godzilla appears. It’s all scripted – hey, here’s your fighter… oh noes a Godzilla, go fight it. Except my fighter plane arrived in Africa and my Godzilla’s in Brazil, and the plane doesn’t have the range to cross the ocean in between, so it can only get there via the north pole, and only if I’ve built a daisy chain of bases where it can refuel. It would take, say, 20 more hours to navigate a path for it to come and fight. During which, I think, the only way I could resist the Godzilla at all is by researching and building another fighter plane.

So I tested that idea – made a new savegame, invested a couple of hours, tried to arm up a new fighter (knowing it would bankrupt me, prevent me from pursuing any of the other half-dozen life-saving technologies… but that’s actually a typical feeling for the XCOM strategic game so); for science, I decided to sacrifice my morning. And… the fighter minigame has no documentation. I don’t know if it’s not working because I’ve failed to do something, or because I can’t find the fire button, or what. Maybe I just gave my fighter some missiles to carry rather than a missile launcher (a thing you can do with infantry, and no the game doesn’t warn you that anything’s wrong). Maybe I need a new pilot or gunner class that the game didn’t tell me about. Maybe it’s an actual bug. I dunno, the interface has nothing to say.

So, that’s my off ramp. Thanks, PP.

Learning from XCOM, 5: the bit that would be better as a TTRPG

Ok, so I’ve been singing the praises of XCOM here for a while, but it’s not flawless by any means. A friend tried the game, based on my posts here, and did not like it at all, principally because of the various doom counters that it uses to make you feel like each mission has larger consequences than just success/failure in the moment.

My experience was that when they work, the game’s various hurry-ups put you in this interesting double-bind where:
– you need the toys that the strategic game unlocks in order to survive the next tactical mission,
but
– you also urgently need another tactical mission to raise money for the strategic game.
So you’re constantly worried about just surviving to the end of the month. A worry exacerbated by soldiers needing to heal, your second-best squad being some levels behind your first, etc.

My friend’s experience was that the strategic game was a hectoring cacophony of doom counters and impossible threats and constant interruptions preventing any sort of coherent thought. It was especially galling that the aliens continually did stuff that was deliberately designed to be far out of the player’s reach, so there was no chance to save all the days or all the people (the aliens also occasionally go on spite murder rampages, which are supposed to discredit your resistance – and yes, it’s upsetting to have a counter of civilians you’re supposed to save, and see them being picked off before you can even get to them. It’s storytelling – these aliens are despicable murderers – but it’s also kind of an anti-game moment – are you not playing this in order to prevent exactly that?).
My friend did not appreciate the feeling that the game was sadistically making them feel stressed and powerless, watching things spiral out of control.

There’s a possible macho response to this – that that’s the difference between surfing and drowning – but my friend is an experienced game player. And anyway, if you signed up for a squad tactics game, that doesn’t mean you signed up to surf a noisy wave of countdowns. Which seems like a generally applicable thing to say about games: if you have a game that does one thing well, it’s probably best not to require your players to also do some other thing, in order to experience it. Multiple flavours can combine to taste even better together, but if you put peanuts in my chocolate then I will spit the chocolate out and consider it wasted.

And after some consideration, I think the issues are really intrinsic to XCOM being a pre-programmed video game, where the strategic/campaign level would be better handled – more responsive, more free, more explorable – as a TTRPG.

The strategic level of the game has a lot in common with the boardgame Pandemic Legacy – and XCOM2 especially has a lot on common with Pandemic Legacy 2 (hereafter PL2).

It consists of a map of the world, on which things go wrong every month and you get a chance to place workers to fix some of them. Like PL2, you don’t have access to the whole map to begin with: you only get to know about stuff that’s happening around your contact network, so as your contact network expands (through communication tokens with locals, and radio masts), so does your understanding of the plot.

Here we are, fairly late in the XCOM2 supplement War of the Chosen, having made contact with 3 other resistance factions.

The first and greatest weakness of the XCOM implementation of this strategic layer is, it’s essentially passive. It’s a “race against time” but your method of racing is largely… waiting for missions to pop up. And, while you wait, hoping that not too much goes wrong before you can research plasma rifles. In a TTRPG you could handwave the waiting and invite the players to take on proactive missions – spy on the aliens, find caches of tech, rescue experts, and raid supply convoys. Then the players would always be busy with something when a surprise mission lead comes in. But that’s… a whole other game to program.

That waiting weakness leads to the other great problem XCOM digs for itself: it has to artificially generate urgency, in the form of a race against a Secret Alien Scheme (represented by those red and black squares at the top of the screen – as the aliens progress on their scheme, the squares turn red – that’s the primary doom counter). XCOM has a nice tech tree (for which it never shows you a map), a set of resource-gathering dependencies (always expand your network first, that gives you money for little things like medicine and weapon powerups) and a story, told in missions you access in series, via research. But if it weren’t for the hectoring doom counters you could just take a leisurely stroll through all that content, unlock all the branches of the tech tree, and level your soldiers up to max. It’s only the doom counter that makes your choices, about what to prioritize and what to defer, difficult or stressful ones. Because now you have to anxiously wonder if it’s better to spend time doing an autopsy on the new bugsquid or trying to figure out more efficient radio, whether you must expand east toward the alien base you’ve heard about, or south toward a nest of potential friends. And whichever thing you choose, XCOM will force your stress level a bit by dangling things just out of your reach.

But why is this a weakness? Aren’t races and tough decisions… cool?

Well, first because XCOM doesn’t give you any information about the likely outcome of your tough decisions – it’s really keen on the story of you unlocking secrets so… it’s a long series of leaps of faith in the dark, rather than the kinds of carefully considered risks that the tactical game excels at. And second, because the way XCOM’s race works is – boom, the RNG pops up a mission:

(some missions are monthly regulars, some are triggered by you hitting certain milestones, some are on some other mysterious schedule).

and you can either go fight it in tactical mode or ignore it – which is the same as fighting it and failing. If you win, you might set the aliens’ doom counter back, buying time to see what happens when you choose “nest of friends” and ignore “alien base lead” on the next decision fork. But if you lose/ignore the mission, the doom counter clicks forward. Which is, obviously, a positive feedback loop or death spiral – failure means you have less time. And that death spiral is intensified because:
(a) you don’t get whatever rewards the mission could have handed out,
(b) your reputation with the people funding you gets worse, so they may give you less funding,
(c) during the tactical mission, some of your soldiers probably got killed or injured, which means you lose their xp progress, their readiness for the next mission, and any unique equipment they had.

And (d) you trigger a flurry of “we can’t afford to lose like this, the aliens will win” NPC voiceovers, which are surprisingly stressful and effective at making you feel bad. If the aliens’ doom counter gets all the way up to 12, BTW, you get a literal doom clock:

and scare windows, just in case you weren’t feeling the failure.

Notably, XCOM never tells you that you’re doing well. Or, if it does, it does it with a broad wink that says “hubris? Pride before a fall? You know what comes next.” (skip to 51:30)

So this “race” is really just a challenge to:
(a) accept and win all the missions that pop up
and
(b) guess the right order in which to develop your base, soldiers, and informant networks.

In contrast, Pandemic Legacy has a negative feedback loop – if you’re not doing well, it gives you more resources. And in PL2, where you’re uncovering the mysterious map as you go, you know that you don’t know where you should prioritize expanding, so you take action and hope for the best. And, critically, although your decisions change the way PL2 plays, they don’t lock you out of winning or put you on a long, torturous path to assured destruction in several turns’ time, which is what XCOM explicitly threatens to do, even though that threatened destruction might really be farther away than they suggest.

So, XCOM and PL both have arbitrary cutoff victory/failure conditions. XCOM’s aliens win if their doom counter goes off, the player wins if they reach a critical research point and definitively sabotage the Secret Scheme, at which point the story beats have set you up to believe in eucatastrophe. PL is limited to one year (12-24 game sessions), at the end of which you might lose outright but are much more likely to score a limited, weird victory by stopping some of the diseases or limiting the damage they do. Anyway, you’re guaranteed to at least understand what’s going on and why.

BTW Phoenix Point, an off-brand XCOMalike made by the XCOM developers, doubles down on the Pandemic elements of the game – now instead of aliens it’s a virus that turns your fellow humans into aliens, some of them distinctly copyright-infringing.

And it makes a point of telling you, at the end of the tutorial:

But imagine if you were playing a TTRPG and all you needed to provide, for the players to go looking for trouble and plot heists, was:
1) a map with some targets on it*: e.g.:
– mystery lab,
– supply depot,
– places with frequent street battles that point to resistance cells (actually, these are all things that XCOM2 tells you are going on, but the only interaction it gives you is “sit and scan this for 6 days and get 50 supplies or an engineer.”)
2) some random tables of mission ingredients and complications, e.g.:
– destroy the transmitter
– put false info in the dead drop
– person you’re supposed to retrieve is unwilling
– ambush
Then you wouldn’t necessarily need to provide definitive end conditions – or you could have them emerge through play. Think Game of Thrones’s endless political roil, out of which emerges a win/lose dyad of zombies or Targaryen world-empress. For any long campaign, the players would find out enough about the world to potentially set their own win conditions, if they even want that.

…I’m actually (still) implementing a thing like this, called Countercolonial Heistcrawl, so I’m not going to anticipate that by writing a reduced version of it now, but… in the spirit of this blog post series, here are some principles, learned from XCOM, for making such a thing. Note – I am not talking specifically about making an alien invasion/human resistance game, for which this video covers the bases of XCOM’s story very neatly, but rather any warbox game where you have a strategic overlay (or “campaign”) that sets goals for a series of tactical missions (or “dungeons,” if you like).

Principle 1: the game continues as long as the players want to play

Or, “don’t impose a limiting win condition at the start.” Players are apt to give up if they lose all their followers but… Babur, the first Mughal Emperor, was just an idiot with a horse and a sword when he set out to restore his ancestor Genghis Khan’s empire. His son Humayun lost it all, being reduced to a horse and a sword, and got it all back (with a surprise twist ending!). Neither one accepted that their game was over. In the history of (actual anti-colonial) resistance movements, it’s very rare for anyone to definitively be done away with and never heard from again. Instead they mostly just lose ground – resources, reach, personnel. There are cases, in fact, where every individual person in a movement dies, or the leader dies, and a couple of generations later a new resistance movement forms around their memory and ideas. From a game-running perspective all of this can just be counted as the same player – if they get knocked right down, they get a negative feedback boost to keep them in the game. So then the stakes of the game are not elimination but capability. If their factions have levels (of reach, influence, military might), those can be drained but the faction doesn’t die unless it chooses to.

If the game is getting stale, by all means change things up, maybe introduce a story that can definitively take out an enemy – but you don’t have to Ragnarok the whole setting: any single definitive change can act as a campaign keystone, leaving room for a sequel with somewhat different rules.

Principle 2: information is not just power, it’s also interactability

XCOM’s tactical layer is a superb model of clarity in information – you know your chance to hit, you know what damage you’ll do, you know where the enemies are and you quickly learn what they can do… and the result is you, as a player, can interact in a sophisticated way with the game. XCOM’s strategic layer is a blind graveyard of opportunities that you stumble over – or cheat your way through by reading other players’ stumblings. You have no idea what to build because the game doesn’t tell you what anything does; you can’t guess at the likely outcomes of research projects; and there are myriad easter eggs in the design that would get you to difficult destinations… if the game told you how they worked (skullmining, I’m looking at you). I get that the designers want the aliens and their tech to be mysterious. I get that they want you to feel joy in discovery. But. The practical result is that you can’t play these parts of the game strategically, you’re just surprised at every turn.

So don’t do that. Instead, think about what you really want to be a surprise reveal, and let everything else be meaningful information that the players can use. Tell them how strong they are (give their faction a level rating, if you like) and what they can take on. Tell them what is rumoured to exist in the wilderness (and make the rumours mostly true), the likely outcomes of research, what their options are for expanding and what those choices mean, even what the next obvious step is, trusting that there will be enough surprises and mysteries and missed connections to make their busy times interesting. It’s when they can choose between two obvious steps that they will surprise you with a creative third option that you hadn’t considered, not when they’re paralyzed by a dizzying load of Decisions on Zero Information.

And when you really want a surprise, tease it – if the player anticipates it, half expects it, they’ll be interested and alert when it finally shows up. XCOM2 War of the Chosen has these three super-enemy nemeses – The Chosen – who can unexpectedly show up and destabilize any mission. Each one gets a theatrical entrance, each one taunts you and laughs villainously, and each one is announced with a set of strengths and weaknesses when it first shows up. So these Chosen, out of all the game’s elements, show real design thought regarding their lines of clarity and unknowability. They’re unpredictable but still interactable. They’re a surprise you can try to plan for, an extra enemy you dread. They make you wish you’d planned better, they don’t make you throw your hands up in the air and say “how was I supposed to know about this?” That’s a good surprise.

Principle 3: the world is made of agent units (targets), and those are dungeons

As players explore/trade for info/spy/listen to the radio, the map gets populated with targets. These targets are elements or agents of the enemy’s network of power – that is, their organization is broken down into modules (or, from another perspective, players can cleverly identify weak points) where attacks can make a difference. So a target is a railway signals box or a warehouse or a general or a genetics lab or some bit of enemy infrastructure that they depend on, that you can disable or appropriate, the loss of which has knock-on effects for the enemy’s operations. Ideally, you draw up an actual network of dependencies, tying these points together. Attacking targets is played out in tactical mode.

Per Principle 2, recon can tell the players the threat level of (almost) any target, so the players can compare it with their own to get an idea of whether they can hit it, take it down, take it over, or leave it for now as a long-term goal, for when they reach level 5.

XCOM really only has 3 types of missions:
1. Kill all the enemies
2. Sneak into an enemy installation and steal/destroy something
3. Get a resource safely across the map to an exit point
and that already allows for a lot of different sorts of challenges, that are not really all dungeons. But you could have a lot more –
– show someone something (especially: show a potential ally the enemy behaving badly… by getting them to chase you).
– plant/construct something in the enemy’s operations – this could be installing a bug, or planting evidence to destroy trust between enemy agents…
– pull enemies into an ambush
….

Defeating targets gives the players rewards and puts a hole in the enemy’s network. The enemy’s network heals with time, but if you shoot it full of enough holes fast enough, it may drop entire capabilities/departments.
Rewards could be:
– information (new targets),
– expertise (new tech/toys),
– personnel (new soldiers),
– resources… etc.
If you merely destroy a target, you might get a subsidiary reward (local resistance contacts, defectors, info etc). But if you take over a target of equal or higher level to your own group (extract useful resources/abilities from it), you level up.

Principle 4: separate the game ingredients by function and give them a face

This is a thing XCOM does very well – your research subgame is a scientist with a personality. Your make new toys subgame is an engineer with a personality. It makes the bits of game easier to find and understand – this is a new technology, therefore research, therefore go talk to the scientist character. That is a thing we can make multiple copies of, therefore it’s engineering.

So given that a warbox depends on information, give the sources of that information personalities – and assign different ones to streams of different reliability, so the game can set up a language for establishing the level of confidence that the players should invest int them: game mechanics/advice comes from Judi Densch, the enemy informant is Andy Serkis, and the supplier of rumours is Steve Buscemi.

Curiously, Phoenix Point doesn’t do this. So maybe there is a need for posts like this one, full of obvious advice, after all.

* actually, it looks like Phoenix Point might adopt exactly this sort of structure for its strategic layer. I’ve just started playing it and… there’s a bunch of targets to investigate around my base in the upper Amazon, including some supply-scavenging missions. Just when I thought this series was over. OK fine, I shall play more and report back with an addendum.

Maps of classic dungeons 4: the rats in the walls of the Opera Garnier

Alexis Dornier: Stilt Studio. If you like that, see also his Birdhouses resort in Bali.

Part of the myth of modern architecture is that it’s all function. But that’s hardly ever true – and even when it is, somehow the expression of that function tends to come across in an ostentatious, theatrical way. 

Arkwright’s cotton mill, Cromford, 1771

Take this cotton mill building, for instance, often forwarded by British people as the birthplace of the industrial revolution. It’s supposed to be purely, practically functional. You can tell because it’s unadorned – no fancy pillars or porticoes or stepping of the facade architecture – and it’s sometimes cited as an important moment in the development of the new clean modern movement in architecture; if you’re wondering why the industrial age abandoned ornament just when it could mass-produce it, the relative cheapness of these buildings is your answer.

And yet even Arkwright’s mill was theatricalized, as a place of work, with this entrance gate, reminiscent of a medieval castle barbican or a prison. The gate announces that within these walls is the place for mechanical labor.    

Architecture can’t really help communicating. Splendor – which Arkwright’s mill self-consciously eschewed – also has a function – if it didn’t, people wouldn’t spend so much on it. It tells a building’s inhabitants how to feel and behave, what to pay attention to and, conversely, what to ignore. An institution’s buildings are its visible presence, its self-portraits, and theatrical settings – showcases for its power, arenas for its executives’ performances. 

Foster & Partners: Apple store, Macau

Corporations love gigantic foyers and atriums, that both dwarf the visitor and offer an interior world to get lost in. Hotels and cruise ships have grand social gathering spaces – dining rooms and ballrooms – where they can make a spectacle of their guests. And temples have whole complex schemes of teaching spaces that make their creeds visible. Theaters of theology.

Anatomy Theater, University of Leiden

..in short, theaters are everywhere. And theaters are dungeons and dungeons are theaters: they frequently contain self-consciously theatrical storytelling spaces – the old raised dais at the end of the room, the sacrificial altar, the monster pit, the pentagram in the middle of the floor. Also, fairly often, these theatrical parts of the dungeon are left as set dressing – whoever made the temple, they’re not here any more and instead it’s giant rats or opportunistic hobgoblin mercenaries, staining the tapestries.

But a theater is a terrible thing to waste – especially one dedicated to fiction. The lies people tell you can be so informative about what’s really on their minds.

Ok, so, some theory.

Theaters (arenas, ritual spaces) dramatize the functions of society.
Justice is theatrical – prisoners and executions are displayed in public to show an example to the rest of the court.
Power is theatrical – both to attract the court’s attention and to set a standard that enforces humility among the servile classes.
And rituals/displays show a society its own structure – as well as offering moments when that structure can be suspended, changed, and re-asserted. Weddings, promotions, judgments, declarations are all status changes for members of the community – moments of structural vulnerability, where stuff is up for grabs, where everyone has to update their hierarchies.
(Victor Turner wrote the classic book on this – the whole thing is interesting)

And/So theatrical rituals offer moments of crisis where PCs can insert themselves to maximum effect – Bond and John Carter routinely escape right at the moment of their spectacular, exemplary punishment, when all eyes are fixed on them – the intended display of the tyrant’s power instead becomes their spectacular humiliation. Years of careful plotting are undone in a way they wouldn’t be, if the whole thing hadn’t been so damn public. And when the heroes escape, they don’t just illustrate the contingency/risk that gives the ritual its power, they also tell us about the holes in their captors’ moral architecture: Carter can slip out of his spectacular starvation chamber through the same tunnel the tyrant uses to enact his sadistic urges – the tyrant is defeated by his own vices, embodied in his theater of cruelty.

But it’s also worth thinking about how a theatre divides its space into the scene – what you are required to look at, to understand the drama – and the obscene – both that which is hidden behind a curtain because it’s too upsetting for the public, and that which is off the stage: the stuff that is necessary to the production but which the audience has to ignore – the lights, flies, and rude mechanicals that make the magic happen. And, fairly often, simply people right there on the stage whom you must pretend are not there (kuroko, servants, boom operators).   

People fret about the obscene becoming visible, but it’s just as useful as a (open) secret place, from which one can observe without being observed. Bond gets inside (infests) Goldfinger’s model of Fort Knox (showing us how the villain’s plans are paper-thin) and learns all of Goldfinger’s plots from that blind – even the bit where Goldfinger plans to betray his mafia allies.

In The Hunchback of Notre Dame the spy gallery is the cathedral bell tower, which offers a commanding view of the bustling square and its centerpiece gallows. While the tyrannical archdeacon sets up a punishment theater for the burghers of Paris, to condemn Esmeralda for failing to return his love, Quasimodo uses his bellside perch to spy on the proceedings from above – a nice reversal of the panoptic power of the state. Also he improbably uses a bell rope to Tarzan down to her and snatch her away from the archdeacon’s own rope necklace, which is probably super Freudian or something but I digress.

gargoyles and saints, locked in an endless staring match

Getting up to that bell tower, by the way, involves entering a rabbit warren of passages – the private scurryways of a very public building, which the archdeacon only barely understands. If you’re intimate with gargoyles then you’re definitely not in a place intended for the worshipful visitors – or at least you weren’t before Victor Hugo’s novel made the bell tower tour one of the most popular parts of a tour around Paris’s Old Dame.

I can see the king’s bedroom from here

And then there’s the actual theater: a building devoted to fixing the audience’s gaze on only one place–the royal box, where Louis XIV and his court just might let their guard down, as they watched their entertainments. Or the stage, if Louis XIV himself (or his wife, or his mistress) were performing.

the Opera at Versailles in the mid 19th century, dressed for a rather staid performance by Queen Victoria

The quintessential theater, the Platonic ideal of the modern palace of illusions, if you like, is of course an imitation, rebuilt on a grander scale, of Louis’s court theater – Paris’s Opera Garnier. It was commissioned by the commoner emperor Napoleon III to form the centerpiece of his “theater state,” and had to be capable of accepting the bourgeoisie as well as dukes and visiting royalty.

The Opera Garnier, Paris’s other hunchback – the giant peaked-roof carbuncle spoiling the symmetry of the dome houses the flies – all the scenery-changing machinery that turns a stage into a cave into a forest or a mean peasant’s hut.

This is the Opera that’s now mostly famous for its Phantom… which was originally written off the back of the fame of the building itself. So it goes with the ravages of time.

I wish I could tell you that Victor Hugo wrote both The Hunchback of Notre Dame and The Phantom of the Opera, and that these two epics about disfigured men with yearning hearts hiding out in Paris landmarks were both scathing satires on the French society of their day – that from their secret vantage points they laid bare the hypocrisies of the powerful. Both do indeed theatricalize French society – Notre Dame offers a “god’s perspective” from which we can see the venal goings on of the city. The Opera Garnier, (at least when it was first built) was a factory for fantasies, both on the stage and in the auditorium, the perfect metaphor for the pasteboard Second Empire. And Hugo had plenty to say about Napoleon III as a pasteboard character – “Napoleon the little,” “the nocturnal strangler of liberty”, that would fit snugly into the whole Phantom thing. But unfortunately Phantom was written in 1910, 25 years after Hugo’s death, by melodramatic mystery scribbler Gaston Leroux and if there is, indeed, a common moral shared by the two books it’s something about the dangerous power of being a sexy woman in a city full of deformed horny romantically-inclined men. At best, they could be partnered with Cyrano de Bergerac to make a series on ugly monomaniacs of Paris, 1630-present.

But let’s forget the sad accidents of history for a moment and pretend that the two books and buildings really are in dialogue. Because they are, regardless of their mere authors’ intentions. The books’ obsession with vertical limits – Quasimodo’s heavenly tower and Erik the Phantom’s subterranean sewer-palace – seems to says something about the anatomy of Paris as a great beast impelled by carnal drives, looking to the stars, etc. The secrecy of the spaces suggests a Paris of labyrinths – which it is, from sewers and catacombs to “secret” entrances to the Louvre and other landmarks. To say nothing of the Law. And the physical structures of the buildings…

Here’s the thing. Notre Dame is a labyrinth, sure – from subterranean crypt to chorister’s gallery to belltowers, there’s plenty of places for ecclesiastics to get up to no good. A family of urchins could easily hide in its decorative niches, posing as statues when the organist shuffles by. And, like most cathedrals, it has a wooden skeleton, so it’s not just a giant pile of stone. But the Opera is another sort of labyrinth altogether – the sort of collection of hanging facades with big holes behind them that you could only build in the 19th century, with metal frame construction and a limitless fund of decorators. The grandiose rooms that make up the public front of the Opera

are surrounded by hundreds of other spaces, many of which can’t really be dignified by being called “rooms,” which threaten to engulf it.

detail of a longitudinal section model of the Opera Garnier, at the Musee d’Orsay, Paris

The famous auditorium is that rather small gilded space, 4 stories under the dome. The skyscraper next door is the stage and flies.

There are at least 3 buildings in that building: the gilded front of house, the actors’ backstage, and a third hidden set of service and maintenance spaces, which are not just tiny scurryholes but whole galleries and kitchens and staircases and staging areas – an architecture of supports and buttresses and food carts and stagehands and whispered orders, where the rooms are cut into by the backsides of domes and mock plaster pillars. These are the spaces that even the Phantom doesn’t talk about: room for an army of goblins – or of plumbers, seamstresses, and waiters. Like Manhattan or Disneyland, where the public thinks they’re standing on solid ground they’re really on the upper floor of a building that extends below their feet for several stories; a whole other city that leaks into their view only through steam blasts and the occasional odd clank.

The front/maintenance divide continues into modernist works, of course. Eero Saarinen’s masterwork design for Bell Labs (which was instrumental to creating this internet thing we’re using right now) was a featureless box on the outside, leading to a classic grandiose atrium inside

but the bit that made the internet work was between and around the famously pristine labs, in the service corridors and closets

Nowadays these cable tangles have got big enough to demand whole skyscrapers of their own, and the front of the theater is your cellphone screen.

the windowless block of a “hut hotel” (telecoms data exchange), suggestively dropped on Manhattan’s Avenue of the Strongest

OK, great. So what do you do with all this?

The first thing is obviously to imagine and then make use of all the ductwork, plumbing, and behind-the-scenes service spaces that your dungeon’s Death Cult Temple demands. Does your local post-apocalyptic warlord live in a palace that used to be a big public building? Then he can’t possibly know all the murder-holes that lurk behind the seemingly-solid walls. Have you always been bothered by the weirdness of classic D&D maps, of rooms connected by long noodle corridors? Now you know that those black “solid” spaces are just rusticated fronts stuck over service corridors that connect to other dungeons, maybe displaced half a floor or half a genre up or down from the version you fell into. How do goblins and ghouls get everywhere? In this swiss cheese environment, the occasional secret door seems like a jokey lampshade.

Do you guys still watch Die Hard? Then you know how grandiose ductwork can be. But just in case you missed Bldgblog’s amazing post on the topic, here’s a reminder. Describing the Israeli Defense Force’s invasion of Nablus, it talks about squads blasting their own passages through a basically hollow spatial labyrinth, which wasn’t previously suited to their movements, but which they turned into an architecture of continuous surprise:

“soldiers moved within the city across hundred-meter-long ‘overground-tunnels’ carved through a dense and contiguous urban fabric.” Their movements were thus almost entirely camouflaged, with troop movements hidden from above by virtue of always remaining inside buildings. 

Breach-charging through walls, floors and ceilings, they deliberately avoided the expected entry points – doors and windows – that the architecture offered them. From Lethal Theory:

In the IDF we now often use the term “to smooth out space” when we want to refer to operation in a space as if it had no borders. We try to produce the operational space in such a manner that borders do not affect us. Palestinian areas could indeed be thought of as “striated”, in the sense that they are enclosed by fences, walls, ditches, roads blocks and so on . . . We want to confront the “striated” space of traditional, old-fashioned military practice [the way most IDF units presently operate] with smoothness that allows for movement through space that crosses any borders and barriers. Rather than contain and organize our forces according to existing borders, we want to move through them.’

…which makes me wonder why D&D parties are typically so respectful of the frame of the dungeon. I mean, obviously the answer is because it’s set up that way, like a Skinner box, part of the invisible rulebook of dungeon delving, which also blocks out the rest of the world with its social consequences etc, to produce a murder-palace full of monsters and trophies, sure, but… if you’re also interested in dungeons as heists, then it makes sense to set them in heistable spaces. And to be aware that spectacles always turn back to face their viewer, who is also on display – the royal box at the opera allows royals to watch the audience and vice versa. The invading murderhobos, who think they’re being sneaky as they inch up their darkened 10′ corridor are being studied for their tactics and equipment – if the dungeon boss isn’t a complete idiot.

And if you’re interested in dungeons as heists then you can build dramatic cues right into the map. Let’s say you have a boss with personality – a Strahd or Acererak or Alp Arslan, and you’ve been thinking about the space for the final showdown. There are advantages to making it a theatrical space, beyond the merely melodramatic – it can be at the focal point of the dungeon, where everyone can see the fight. Its boundaries can be clearly marked, so the players can prepare themselves – and do their homework -before stepping into the arena. And of course they can’t step into that space without themselves becoming part of the spectacle – what are their stakes? Why should they be hesitant to grab that cup, defile that fane, take over that dark lord’s sepulchral domain? The theatrical is also where you find the thread that pulls the villain’s whole architectural scheme apart – the whip-carrying slaver that everyone loves to hate, the feast you can ruin to turn the court against the villain, the seemingly minor functionary who turns out to be critical to the alarm-raising system. If you’ve done your homework, you can make your grand entrance deliberately, confident that you have your own surprises waiting in the flies.

You didn’t think I’d write all this without mentioning Renzo Piano’s and Richard Rogers’s Centre Georges Pompidou, did you? The sneaky trick here is, they shifted all the vital infrastructure to the facade in order to leave uninterrupted space inside for a clean, pristine and reconfigurable gallery box, all the better to show whatever crazy ductwork modern artists can cook up.