RPG Design: Realistic Dungeons?

Posted by Rampant Coyote on December 7, 2015

A couple of weeks ago, we had a discussion in the comments section about using real-life castles and dungeons in RPGs. I thought I’d elaborate on my thoughts in a post.

Many years ago, with access to a university library, I researched tons of actual castles – mainly European – searching for maps and floor plans. I still have that folder laying around somewhere. Combined with an old and well-used purchase, Palladium’s The Compendium of Weapons, Armour, and Castles (one of my favorite game guidebooks ever), I not only had a whole bunch of ready-to-use maps for our dice-and-paper games, but I had a much better understanding of many of the whys and hows of castle design.

But there is a problem. Lots of problems. Realistically, these structures were built to prevent enemies from doing exactly what the players are doing. Assuming an organized, intelligent defense (which we almost never do in RPGs), the party would be dead before it breached the first wall, let alone penetrated the central tower / keep / donjon and could begin the room-to-room clearing so common in these games. And if they did get that far, things were even nastier.

See, everything in a castle was designed (at least originally) so that a tiny force of defenders could hold off an army of attackers. That was the point. The shields and weapons hanging on the walls weren’t for decorations. The rounded design was less vulnerable than square corners to siege weapons like trebuchets and cannons. The tiny, twisty halls are supposed to limit attackers to moving through in single-file. The entryway into the front was often designed to run parallel to one wall of the castle so that it exposed attackers to attack along the entire route. Even the spiral stairways were designed to impede the ability for a right-handed attacker to swing a weapon, while giving a defender much more freedom. And of course, there were things like murder holes, arrow slits, and so forth to maximize defender cover while exposing the invader to constant attack.

While this could be a really cool tower defense game idea, or a decent strategy game (and it has been), it’s a lot more problematic for role-playing games. Likewise, the best strategy against something like this would be to settle in for a long siege, if possible, and starve the defenders out. Not exactly fun from an RPG perspective (“The second month passes. Do you do anything differently?”).

We were (still are? It’s for Frayed Knights 3) using a floor-plan for Bran Castle as a baseline for one level. We ran into some problems based on real-world architecture, even as we took a lot of liberties with the floor plan. Halls and stairs and doorways are designed to be tight and confining in the real world, and are almost impossible to navigate on a 3D screen. Perhaps in a VR environment with a setup that can track your fine body movements as you duck, twist, and lean to go through the environment, that would work out okay (although without tactile feedback, maybe not), but in the simplified 3D environment and controls of modern video games, it just doesn’t work. Ditto for caves. Those can be hard to navigate in real life, and even harder for the player – let alone AI – to believably walk through in a game.

The best we can do is borrow some elements from real life to give it some of that authentic flavor, and then go from there. The trick, really, is to borrow the cool parts and challenges from reality, and turn it into something fun and dramatic for the players. That’s why many of the classic fantasy role-playing adventures actually happen in ruins or abandoned castles, with many of the defenses falling into ruin or enemies unable to take full advantage of what is left. Dealing with kobolds pouring boiling oil through murder holes once or twice might be an interesting challenge for adventurers; being subjected to a constant barrage of harassing attacks gets old.

Reality isn’t always fun… that’s why we have our escapist fantasies. We don’t really want “realism” – we want drama and action and story that feels authentic. Borrowing liberally from real life gives the games a better feeling of verisimilitude and helping it ring true to the players. And it can provide some cool ideas.

Filed Under: Design - Comments: 12 Comments to Read

WhineAboutGames said,

According to my fantasy novels, the proper RPG way to take down an inhabited castle is through guile and psychological warfare. Siege ’em, starve ’em, bribe someone on the inside to stir up trouble, infiltrate in the guise of a servant and investigate subtle ways to turn the castle’s defenders against each other.

‘Take down a castle from within by burning the soup and stirring up unrest’ is on my vague list of game design ideas that I’ll probably never get around to.
Maklak said,

Thanks for writing about this.

@AhineAboutGames
That’s standard stuff and you forgot bribery. You can do so much more with fantasy. Some of it is even why castles became obsolete in FK. Basically the way to solve problems is to use a combination of magic and dickery.
* Siege engines kinda still work. Walls are probably reinforced by magic (I’ve seen a spell to increase material hardness somewhere in DnD) and I guess defenders have healing magic and can summon water elementals. Therefore stones, Plague-infested corpses and molotov cocktails aren’t going to be fully effective, but they’re sure !FUN! options, especially in bulk. But how about using something even more !FUN! for ammo? Earth elementals, summoned monsters, golems, shock-activated magic traps (don’t have to be evocation), etc.
* Air superiority works, either with flying magi or griffin cavalery or whatever. Give them protection from arrows and have them use ranged weapons and/or bombs.
* Magic works, especially AoE spells, like stink cloud. Bonus points for using invisibility, flight and summoning magic. More bonus points for teching up into things like cannon-shooting golems, animated siege towers, mind-controlled viverns, contracting mercenaries from the lower planes, etc. Even more bonus points for weaponising ghosts who produce more ghosts from their victims.
* One of the most amazing DnD spells I’ve seen was 9th level evocation with a costly material component (but really a 10th level evocation + necromancy). It made an explosion that was increasing it’s range damage and duration depending on how many people it killed and I don’t think it cared about walls. It usually only took one of those targeted at a blob of defenders on the wall for a decisive victory.

Fantasy wars rarely progress to the point where these become accepted tactics, or if they do, typically only one side uses them, but I see no reason not to spice things up. At the very least have oversized stationary crossbows firing magical bolts that explode, get through damage resistance or whatever.

As a GM, I’d take Tucker Kobolds level of organised defence and resulting TPK as a given for a standard DnD party who aren’t paranoid and dickish in the right way. If I wanted to really screw them over, I’d make a complex filled with traps that actually work (such as closing off a remote porticulis and flooding the entire level so the party drowns). The treasure? Close to the entrence there was an unmarked 3m wide wall where the owner used passwall. There are no markings enywhere on the floor, because he was magically flying. No skill checks, no notes, no nothing. Checkmate.

Oh and Dragons should be night-impossible to kill with standard means. They’re too smart to fight you on your terms. They know when the party comes too close for comfort to their lair because they have alarm spells. They have a cult of Tucker Kobolds making traps and putting up resistance. They can circle-strafe the melee guys in the open, then teleport / rune of recall /fly away when low on health. Or maybe the dragon is ruling a city through proxies and the party finds itself hunted down with extreme prejudice by the army, city guard, temple guard, thieves guild and assasin guild at the same time.
Xenovore said,

As much Corvin as Bran. =)

And castle ruins… whole ‘nuther story.
Maklak said,

Mount and Blade has fairly realistic castles, but you have like 100+ guys to help you, which in tabletop RPGs is only possible with very simplified rules, like tabletop, savage world or horde rules. In a heroic RPG having so many allies never happens for some reason.

Part of it may be time constraints. It takes a long time to get through a single character’s turn in something like DnD and playing with a summoner is stretching it. To play with lots of NPCs a turn would have to be so streamlines as to take less than a minute per character.
T2.0 said,

These explanations about why castles were designed that way in the Middle Ages are very interesting 🙂 Thank you for sharing !

I have a hesitation, though, about the statement that realistic dungeons would be impossible to beat in an RPG unless you would settle in for a long siege.
In one of your previous article, you were talking about Gary Gigax – quote :
“Some veteran players (who had even played through the module before, back in the old days) came up to him later and asked what they had done wrong. His response (as I recall) basically came down to the fact that they were trying to “brute force” the dungeon.”
Therefore I wonder : are the brute force (destined to fail) and the siege (not fun) the only options against a dungeon ?
Rampant Coyote said,

Depends on the style. See what @WhineAboutGames and @Maklak said, above. Those would be very challenging to implement into a computer RPG though as a generalized system though.

But as a general rule, tight, confined spaces don’t work well with generalized collision and pathing, either. Again, nothing that’s impossible, it just adds a good deal of complexity to the development.
Maklak said,

> Those would be very challenging to implement into
> a computer RPG though as a generalized system though.
Yes, but game designers can and do cheat all the time 🙂

In a Pen and Paper game, to run battles, you’d need a horde system (which replaces a group of people with something similar to a swarm, with ramped up damage and different hp rules) or otherwise a tabletop system, like Warhammer. I’ve seen a DnD book somewhere which simulated war and there were sub-quests, such as “destroy supply wagons” or “take and hold that hill” for the adventurers. Basically the tradeoff between PnP and cRPG is “computer limits choices and freedom, but can handle numbers and bookkeeping a lot better”.

In a strategic computer game, things like these are done through unlocking technology / building tree, which unlock usits or otherwise diplomats who can bribe eneny units and castles into surrender. Or some king of simplified attrition mechanics.

In a cRPG, you’d have static (non-rocedural) NPCs and quests. It is easy to imagine, a small group or even one person could be smuggled inside (possibly before the siege), disguise as one of the defenders or a messanger from the lord, then talk defenders into infighting / using their resurces inefficiently or poison the only well or place some remotely-detonated magical orbs in strategic locations or report to superiors with Sending spell. It is not difficult to imagine having rogue/bard-ish quests like that and the havoc these could wreck.
T2.0 said,

@Rampant Coyote: Thank you for this instant reply 🙂
I admit that, as often, my approach is very theoretical – I speak from the player’s point of view and tend to ignore the budget constraints or the technical difficulties that a dev could encounter when trying to put the idea in practice.
Maklak said,

One more note on this is that if a location is inhabited by intelligent beings that need to eat, I expect to find the following (and I could write a lot more about each):
1) Room layout that actually makes sense, with multiple paths between locations and not just one snaking corridor. Even a cave will often have some circular paths.
2) Water supply.
3) What is their food supply? Farming is the most feasible option for anything big, but fishing, herding animals, hunting and gathering work to an extent. Raiding is fine for smallish groups.
4) Food storage and food preparation area. Enough food to last a winter for however many people are there. This depends on season, but in RPGs there never seems to be a reasonable amount of food, such as 1kg / (person * day).
5) Shithouse.
6) Workshop(s) for maintaining and manufacturing equipment.
7) Storage rooms, including main armoury, but mostly supplies, like tools, firewood, oil, spare clothes, raw materials, etc…
9) Sleeping area with beds or whatever.
10) Stables with fodder storage?
11) Main vault for valuables?
12) Laundry / bath / cleaning?
13) Buisinesses?
14) Chapel?
15) Are there women and children or just fighting men?

There are a lot of considerations that go into making an area believable. Sadly game designers seldom understand enough logistics to pull it off or even care about it, which breaks immersion for me as much as stupid layout does.
JT in PDX said,

Hello there –
I am desperately searching for high- (or at least higher-) resolution complete floor plans of Castle Bran, hopefully including a key. Can you share the ones you used, or their source?
Thank you if possible –
JT
Rampant Coyote said,

All I found was on the Internet. Since ours was only a fictional castle “inspired” by the real castles, we were able to take a few liberties. If Kevin has any better sources he’ll let you know.
JT in PDX said,

Thanks! Even an address “on the internet” to point to would be great if they’re reasonably hi-res images. All I’ve found are the pictures on the Lego Castle site… looks like exactly the same size as the thumbnail you posted here. I wrote that guy but he said he had no memory of where he got the plans.

Anyway thanks again for responding.
JT