Typology of Random Encounters: In the Dungeon

Most of my games this past weekend were driven by the random encounter. And, despite the image above, none of them were with bears. Although one was with a crocodile. That, along with general considerations about dungeon design, has me thinking about the types of random encounters you'll find in a large dungeon or megadungeon.

Type I: Faction Members. When you have a faction in the dungeon, it should be represented on the random encounter table, unless there's a good reason that its members stay put. So if there is a goblin tribe, a gang of bandits, or a cross-species evil cult, you should be able to run into them. This allows for PCs to figure out what else might be in the dungeon without knocking down every door, and provides opportunities for faction members who leave an encounter alive to tell their allies about the PCs' activities.

Type Ia: Factions without a Base. It's possible for there to be a faction in the dungeon that is an itinerant group. This should be a temporary condition; perhaps you have a group of cavemen who've wandered into the dungeon for shelter, or maybe a half-dozen goblins got away when the PCs beat their king. They will initially be "homeless" within the dungeon, but within a session or two, they may take over a cleared room.

Type II: Predatory Monsters. There are "big" monsters that have their own room in the dungeon, but logically will be wandering about looking for things to eat. It could be a pack of hell-hounds or the minotaur; whatever it is, it's not home. If the PCs kill a significant monster outside of its lair, of course, now they have to find its treasure before anyone else does.

Type III: Animals and Unintelligent Creatures. These are your Rodents Of Unusual Size, your skeletons and zombies, your oversized bugs and so forth. They are here on little more than instinct, and most are optional encounters one way or another. The PCs can throw some rations to distract the R.O.U.S., or have the cleric turn the skeletons, or just go around them and hope the reaction roll isn't an instant attack. (These encounters should always involve reaction rolls.) They don't have the same impact as faction monsters, although they may also come from another room. I particularly enjoy relatively innocuous animals like dungeon chickens (there is a coop in my dungeon).

Type IIIa: The Clean-Up Crew. Jellies, slimes, oozes, globs, glops, puddings, molds, blobs, and various other unwholesome growing things come along and eat dead stuff. Sometimes they convert the PCs into new dead stuff to eat. These come in particularly handy if the PCs are in an out-of-the-way corner and you roll a random encounter.

Type IV: Non-Player Characters. In the encounter tables for older editions of D&D, NPCs with classes and levels were regular entries. (OD&D used the level titles for this, which is great if you remember the level titles.) They are distinct from faction members; generally once somebody has at least 2 hit dice, you should think up a name and a background rather than leaving it at "bandit #479". Fighters and thieves may have hirelings; wizards may also have familiars; clerics are more likely to have acolytes. They may have a relationship to one or more factions, and in time may constitute a faction in themselves. It's also entirely possible that they will join the PCs, and if they do so, they could also betray the PCs. Sure it's possible for bandits to do that, but in my experience PCs get stabby once someone says "stand and deliver."

Type V: Wandering Oddities. Perhaps the most neglected category, strange and wondrous effects can sit on the random encounter table along with all kinds of monsters. This can be as simple as an unknown sound, seeing a figure disappear around a corner or down a hallway, optical illusions, or even minor magical effects. This is a good way to reinforce a theme throughout a dungeon or dungeon level without necessarily dedicating a room to it or hammering it home constantly. It could be a detail like a particular fungus that is growing in this level, or mini-features like having carvings left behind by the cult that is operating nearby. This makes the dungeon a richer and stranger place.

Each of these types feeds back into solid dungeon design, and helps tie together a living environment that has strange and wondrous features. Featuring multiple examples of each type should help fill out that encounter list without the exploration feeling like a slog.

Learning from Nature: the Mechanics of Traps

A fascinating article in the Atlantic describes the mechanism of the Venus flytrap as scientists have studied it. It's a great description of the Venus flytrap's mechanics, which through evolution shows how an unthinking plant can hijack the instinctive survival response of an animal. Its process is instructive in how we can think about trap design.

First, the trap is appealing. A fly comes along because it smells something sweet. By making itself attractive the flytrap uses the appearance of other plants around it as an offensive weapon. Since the fly can't tell immediately that it's a trap, it comes in to inspect. Only then is it doomed. This is a good principle for elaborate traps in D&D. PCs are always looking for treasure. Appearing to present a treasure is a good way to present a trap, but not the ideal way.

D&D players aren't, no matter what some referees may tell you, as dumb as flies. A trap that's simply too obvious, such as a chest of shimmering jewels and gold coins unguarded in the middle of a room, will be regarded with the utmost paranoia. It's obviously a trap. A really great baited trap is one that looks like hidden treasure. It's hidden in a secret compartment or panel that can be found by careful PCs. Such a subversion is cruel, but sometimes the best way to spring a trap on players is to make them work for it.

Second, the trap has error-checking. It only closes when two of its cilia are brushed. This is a wonderful principle for mechanical traps: the trap is not triggered on the first pass. The primary reason for this is to avoid false alarms, as described in the article. This is good and pragmatic. But with a D&D style trap, there is a further benefit where the party is drawn deeper into the trap before it springs. A simple pit or arrow trap will kill only one member of a party, but it is reasonable for trap designer to aim for multiple kills. And it may also be a way for monsters (or clever PCs) to avoid a trap by letting it reset after the first trigger.

There are two good ways to implement this principle with a D&D trap. One is to require a single trigger to be pressed more than once, such as a pressure plate that activates a two-step process. The first step cocks the arrow and the second fires it. The other is to have two separate triggers, one that starts the trap process and a second that finishes it. Either can work and both are nasty surprises.

Third, the trap imprisons without killing. This can be useful in a faction dungeon where the monsters might prefer to question a member of a rival group spying on their territory rather than killing them. The victims of such a trap become useful bait or can be traded in a prisoner exchange. This can lead to a tense period where the PCs have an opportunity to try to escape before the monsters who set the trap come to check it. Or the designer may have abandoned it and the trapped characters are stuck until a wandering monster comes along.

A trap that doesn't kill outright is extra fun if it affects only a single PC. The rest of the group simply sees the lead PC go missing and isn't sure exactly where they are, while the trapped character has to deal with their predicament on their own. It presents the immediate dilemma of how much effort to spend on saving the trapped character. And as noted about the factions above, such a PC may be a bargaining chip that stops the rest of the characters from barging in and killing a monster group.

Fourth and most brutal, the trap makes its victim kill itself. The fly's struggles against its captivity doom it. Standing still would be the best policy, but it goes against their instincts. This can be copied in straightforward ways, such as by having the character trapped so tight that they can't struggle to get out without impaling themselves on a spike, or puncturing a container of poison gas or acid. It could also be a question of physics, if the trap is suspended more than 10' above the ground. Even a simple quicksand type of trap, where struggling to extract yourself actually pulls you in further, does a great job with this principle.

But this can be used in more devious and subtle ways. Efforts to escape can let the players, who again are still smarter than flies, outthink themselves. Elements of the trap itself may be unstable or unsafe, or designed in a misleading way. What looks like it will open a door actually operates a hidden ballista, or opens a chamber above full of heavy rocks. You can go way too far in this direction and wind up in Grimtooth's Traps, but I think there's a lot of fun trap design short of that.

I love the concept of the Venus Flytrap. It's an elegant and simple life form, and does its work without even the simplistic thought patterns of the fly, but it does something very intricate and involved. Its principles can lend a lot to any trap in your dungeon, whether you pick one or all four.

Oh, and don't forget the simplest way to apply lessons from this article in your dungeon: giant Venus Flytraps.

The Tao of the Empty Room

The dungeon stocking table in OD&D results in 5 out of every 9 rooms being empty. (We're excluding rooms stocked by the referee.) These rooms have no monsters and no treasure.

I've written before that empty rooms are critical for spacing and timing in OD&D. The question that we have to ask, though, is how empty they should be.

Of course, some of these rooms will be trap rooms. Those are good, noble, and deserve their own separate post. I will make one side observation about traps here: pit traps in OD&D only trigger on 1 or 2 in 1d6. A pit trap that has PCs have passed over without it opening can be even more dangerous than one that is fresh, because the PCs think that nothing is there. But here we are mostly talking about the empty ones.

Sometimes rooms should be empty of anything at all. This is a useful reminder to the referee that the dungeon is a mythic underworld. Not every square inch of the underground needs a rationale. In logistical terms this also allows the referee to use rooms that only appear to be empty. When they turn out to contain traps or secret treasure, the reward is all the better.

James Maliszewski wrote about the mystique of the empty room. Every turn in a dungeon is a use of resources runs the a risk of wandering monsters. Timekeeping being important, such empty rooms become tense situations.

But not all empty rooms have to be devoid of everything. When you have members of a dungeon faction in an area, nearby empty rooms can create a more "lived-in" feeling. This could mean, for instance, that there is a second area where denizens spend their time. Monsters might even split their time between the area that the map key indicates for them and an ostensibly "emtpy" room. This also makes it possible for PCs finding the empty room first to get a clue of what is up ahead.

There could also be structural reasons for the room to be empty. A room noted as empty on the key might, for instance, have a large pool of standing water. Perhaps it is unsound, with cracks in walls or floor. Anyone who has been through a home inspection can think of many reasons a room for a room to want for occupants. Mold, mildew, even just something with an unbearable stench makes for a good excuse.

If your dungeon has tinkering monsters, the room could house a partial or failed construction project. A bunch of bricks might have been removed from the wall or ceiling. There may be the discarded makings of a fortification. This is a good excuse to give fodder for clever PCs to make attacks on monsters.

Another interesting trick is to use a room that is only mostly empty. Such a room might have architectural interest that provides hints to the history and nature of the dungeon. It can also have one of several physical features typical of underground areas. A fun one is to have a spot where it is possible to see or hear what is happening in an adjacent room. Perhaps there is a loose stone, a pipe, or a spot where the acoustics just work out.

Remnants make a great feature of the mostly empty room. A discarded wine skin or broken weapon implies that the PCs aren't the only adventurers who have been in the dungeon. Broken or torn and generally useless items are classic red herrings. They also present opportunities for the referee to place vermin and insects that don't merit a proper monster entry. Even non-poisonous insects can make for a great creepy underground experience. (Such vermin can later appear in monsters' stew-pots as a fun callback.)

The condition of empty rooms is one of the most useful ways to give out information about the dungeon. The condition of its floor can hold crucial information about what lives in the dungeon and where. One floor may be dusty and another well-trodden. A particular path may bear the marks of frequent traffic, especially when clawed feet walk across stone.

A fun variation is to feature graffiti on the walls. Scratched or painted writing and pictures are another excellent source of hints. For an extra twist, consider having graffiti on the ceiling to reward players who think of checking above their PCs' heads.

There is also the opportunity here to make the dungeon get weirder the deeper the PCs go in it. If the first few levels are a bit more quotidian, you can change things up with empty rooms that seem more Stygian and have stranger features. The random noise table from the 1e DMG is a great resource for unexplainable sounds. Likewise, air currents deeper underground and shifts in temperature or humidity become ominous.

One of the reasons I like the empty and mostly empty rooms so much is that they avoid the "fantasy IKEA" effect. Excessive rooms full of stored stuff takes away the mystique from the dungeon. I also don't like it because generally I prefer when the dungeon is well scavenged. A room that was once a dining room shouldn't have a table in good shape; it should either have rubble or nothing in it.

A critical fact about empty rooms is that, in a living dungeon, they don't need to stay empty forever. Putting monsters in a room that players have already explored and found empty is the best payoff for having empty rooms. A group of monsters that wasn't finished off the first time might change addresses, or a new monster might move in. Given the 5 of 9 ratio, there should be more rooms available than taken, so the dungeon can make a complete change through play. This is one of the factors that keeps drawing me to the megadungeon idea.

The title of this post implies its central idea: the Tao of empty rooms matters. Letting a dungeon room stay empty can be more rewarding than you'd think.

Mythic Underworld: Eurydice

Continuing with the theme of classical myth, Eurydice was a wood nymph (or possibly a demigoddess) who was married to Orpheus, who was the greatest musician and poet ever born. (He was also possibly a demigod; there was a lot of that stuff in classical myth.) She was bitten by a viper, and died.

Orpheus, then, followed Eurydice into the underworld realm of Hades, where he found Eurydice. His music was said to be so piteous that Hades and Persephone permitted him to bring Eurydice out of the underworld and back to the world of the living, but he was to walk in front of her and not turn back. But as soon as he reached the outside world, he turned around; Eurydice had not yet left the underworld, and was lost to him forever. He eventually was killed by women in a Dionysan fury. The story is one of the great tragedies of Greek myth.

In D&D terms, of course, he should have just had a friendly cleric cast Raise Dead.

We talked in the Persephone entry about people belonging to the underworld, so I won't repeat that part. I do think it's worth underlining how final death was in Greek mythology; there was no easy way back, and even the gods would intervene to keep their favorites alive (think of Athena and Odysseus). It is somewhat cheapened when any cleric with access to fifth level spells can pop dead Eurydice back.

The basic frame of the story, of going into the underworld on a perilous quest, is of course inescapable in D&D. It's the whole point. Getting back out safely is, of course, an exercise in ingenuity (and sometimes luck) for the player.

The landmark of the underworld in Ovid is the great river Styx. It was this river that was supposed to have given Achilles his power (and the vulnerable heel that he was dangled by), although it is notably not the river navigated by Charon; that was the river Acheron, although Ovid still mentions a ferryman. Still, the idea of a river bounding the underworld has a lot of potential. If you are using the mythic underworld concept in your dungeon, an underground river or a gate (Ovid mentions the gate at Taenarus) is a terrific way to convey it.

Having such a boundary makes it a definite choice for the characters to cross into the realms below, and allows you to incorporate areas such as the basement of a castle that is "just" an underground location and not part of the underworld proper. The rules that Philotomy discusses, such as doors remaining stuck or monsters wandering the corridors, only apply once you have crossed the Stygian border. There may be consequences or challenges for taking something out and into the broader world.

One aspect of the Eurydice story that I love is how the rulers and inhabitants of the underworld are moved by Orpheus's song. Ovid is quite explicit, describing how the famous inhabitants of the underworld such as Tantalus and Sisyphus stop their efforts as they hear it, and even the Erinyes (Furies) are brought to tears.

A genuine sense of aesthetics and taste in its denizens is generally under-utilized. Evil is often aesthetically "ugly" and unappealing, and that has overall been the trend in D&D. But it isn't necessary; the underworld can contain and appreciate beauty, even if in a dark and twisted way. You can read the "music tames the savage beast" into this, where particularly beautiful offerings might be useful for negotiating with underworld entities.

One other rabbit hole that can be gone down with Orpheus and Eurydice is mystery religion. Orpheus is closely linked with the Orphic Mysteries, a well-documented early form of the mystery cult. Initiates were taught rites and rituals that had a secret mystical meaning, and we have several surviving Orphic hymns. These are poems, attributed to Orpheus, that contained detailed information on the mythological world. This strikes me as a fertile religion for a fantasy world; its initiates form a secretive group, and it represents a way of thought that is different from modern rationalism in an interesting way.

We'll leave Orpheus looking back at Eurydice. For the curious, I also intend to look at Odysseus and Heracles within the realm of Greek myth, and getting into Egyptian and Norse myth. Any thoughts on other mythology to approach would be appreciated.

The Mythic Underworld: Persephone

This is the start of a new series inspired by the discussion several years ago by Philotomy on the Dungeon as a Mythic Underworld. To boot, I am going to be talking about ... the underworld in mythology, and its implication on Dungeons & Dragons and similar games where characters go underground to seek treasure. One convention: in these posts I'm going to use the Greek names for deities, even when the Roman names are better known.

In the short version: Persephone was a goddess of nature and flowering plants. She was abducted by Hades, god of the underworld, who burst forth from a rift in the earth. Her mother, Demeter, eventually convinced Zeus to send Hermes to retrieve her. However, she had eaten either four or six pomegranate seeds from a fruit offered to her by Hades, and the lord of the dead was able to claim that each year, she must spend an equal number of months as his queen in the underworld.

This was a transparent explanation for plant growth in winter, which is alternately described as either Demeter's sorrow for her daughter in the lost months, or the effect of Persephone's simple absence. But that's not really what I think is interesting in this myth. Rather it's the idea that things in the underworld can permanently change a visitor.

A literal interpretation of the Persephone myth takes us to an interesting sort of "special" in a dungeon chamber. A table is laid with exquisite food and drink, but drinking it places the characters under an enchantment specific to the underworld. For instance, they might fall under a powerful wizard's Geas, or any command or curse type of effect, and optionally the amount of food eaten might be proportional to the price taken. This is suitable for any food coming from the dungeon, or perhaps any kind of "garden" sub-area, et cetera.

Hewing still closer to the Persephone myth, one possible consequence is an impressment into service within the dungeon. A character might not be the "queen" of the underworld – Persephone was a goddess, after all – but they might have to help build out a new sublevel, or maintain and clean traps, or guard against intrusions by other parties, et cetera. Of course this has the potential to thoroughly hijack a campaign.

Moving further afield, there is a general idea that things of the underworld are not wholesome and safe, quite possibly in a sense that implies moral corruption or chaos. If you like, corruption based on, say, a powerful magical treasure bound to the underworld can follow a path similar to the physical corruption of wizards in Dungeon Crawl Classics, or it can stay closer to home with the kind of mental drawbacks that we usually find with artifacts. It's interesting to make the PCs also slowly become creatures of the dungeon, in the sense of Nietzsche:

He who fights with monsters might take care lest he thereby become a monster. And when you gaze long into an abyss the abyss also gazes into you. – Friedrich Nietzsche, Beyond Good and Evil

This sense of corruption can be signaled in a lot of different ways. Perhaps the coins, instead of showing good and wise kings, are transfigured into the faces of wicked tyrants; instead of portraying the lawful and virtuous gods they now portray leering demons. Metal implements such as iron spikes left too long in the dungeon might take on a sinister patina, or wooden ones growing dark and twisted where once it was straight and true.

As the PCs delve deeper, the dungeon can subtly make them its own. When they leave, perhaps, or if they are foolish enough to overnight underground, they can have vivid dreams of the underworld passages, ending just before their bloody death. A low Constitution PC might pick up a lingering cough or chill that marks them, or the PCs' faces might take on the cast of the dungeon's unhallowed darkness.

Of course, the end stage of this is the Innsmouth effect - where the PCs become fully monstrous. This should be preceded by both PC choices, and some kind of warning effects. Perhaps drinking from mysterious pools and fountains has a temporary helpful effect but slowly changes the characters; they might get infravision, or the ability to open stuck doors, without the normal hassles assumed by the D&D rules. But there needs to be a way to reverse or turn back from this course. It's worth thinking about having the PC "go monstrous" at certain intervals, like in lycanthropy, so they don't experience it all at once.

The story of Persephone was a myth that deeply affected me when I read it as a kid. I've always felt that the idea that you make one mistake and now belong to the underworld is one of the more chilling ideas in mythology. So naturally I see it as a key building block if you want to make your dungeons mythic.

More Dragons in the Dungeons

So there's this game you may have heard of called Dungeons & Dragons. It seems to have more dungeons than dragons, which is a shame.

Ancient and medieval depictions of dragons aren't of multi-story lizards breathing out gigantic fireballs. They are generally small, more slender reptiles, comparable in size to large horses, that seek out treasure hoards and guard them. It's a classic underworld trope to have dragons atop a burial mound or ancient tomb, which makes them a natural fit for dungeons.

Dragon inflation has been a constant of D&D, and it has slowly pushed the dragon out of the game, except at high levels. Dragons went up in hit points significantly in the first edition PHB, and much further in second edition, firmly ensconcing them as upper-echelon enemies. They have stayed that way ever since; a party will pretty much have to be 5th level or higher before even thinking about slaying a dragon.

OD&D dragons aren't unbeatable super-beasts. White dragons only have 5-7 hit dice, and are 25% likely to speak, and 60% likely to be sleeping when encountered. A Very Young white dragon only has 5-7 hit points and a similar score on its breath weapon; in my opinion that's not out of the question for an encounter on the first dungeon level. It's a logistical challenge, since the party will want to avoid being hit by the breath weapon, but not one that the party can never overcome. Even at 3 square inches (30 feet), the cone of a white dragon's breath will not necessarily hit multiple party members if they arrange themselves properly, using terrain obstacles or shapes to their advantage. If nothing else, they won't just assume they can kick in every door and slay the monsters.

All OD&D dragons are treasure type H. This is a hoard with literally piles of coins; there is a 75% chance of 10,000-60,000 GP. On average (and not counting magic items), a hoard should be in excess of 80,000 GP, with most of that coming in jewelry and gold pieces. Slaying a dragon and taking this much treasure would be a fine haul, although the reward stops being quite so dramatic after 5th level, when a dragon's hoard is no longer an automatic level gain. But up to level 5 or so, dragons are an efficient way to gain a level.

If you want to picture a hoard, let's say that gold pieces are about the size of a Krugerrand. 60,000 gold pieces would be about 50 gallons of space, which would be larger than the volume of a typical bathtub. This would have about 25 pieces of jewelry and/or 50 precious gems mixed in. Armored PCs could only take out about 1800 coins each; you'd probably want a few wheelbarrows. Transporting gold out of the dungeon should be almost as much of a challenge as slaying the dragon. It's impressive, but not cavernous rooms full of coins. The largest possible hoard would have 184,000 coins, which is in the area of a refrigerator.

By design, dragons should be a low to mid level monster: a hard challenge at level 1 or 2, but able to be handled with much more ease at levels 4 to 6.. If the referee is running OD&D by the book (as it were), it's a poor decision to take on a blue or red dragon. They are more likely to talk, and thus to parley, and less likely to be sleeping. And their treasure will be the same as a white dragon, which could be as much as 6 hit dice lower than its red cousins.

There's another reason to put dragons in your dungeon, though: factions. Dragons will naturally relate to the other denizens, although obviously this relationship may be more of a predator/prey one. But it will always make the dungeon more interesting if there is a dragon sitting on a big treasure hoard. Whether or not the dragon talks is a big factor in this, as a brute white dragon that picks on the goblin tribe would be very different from a crafty green dragon that has a group of gnolls or bandits doing its bidding.

The moral is simple: don't be afraid to add dragons to your dungeon. They are designed in OD&D to be things that you can handle but still have some "badass factor."

How did these people get into the dungeon?

OD&D volume 3's Monster Level tables include numerous human-types in the dungeon, mostly a mix of fighting-men, magic-users and Chaotic clerics. Of the humans in volume 2, only Bandits and Berserkers appear. The rest are listed by their level titles from volume 1, and run the gamut from Warriors (level 2 Fighting-Men) to Wizards (level 11 Magic-Users) and Evil High Priests (level 8 Chaotic clerics).

It's telling that the charts don't contain any Lawful clerics. If you run into priests, they are either Evil Priests, or Evil High Priests with a retinue of apprentices. In typical OD&D fashion, other possibilities include at least a couple man-types: Mermen (who are not necessarily fish-tailed) and cyborgs (who are not mentioned elsewhere).

Dungeons probably aren't suitable places to find dervishes, nomads, buccaneers, or pirates, but it is a bit odd that cavemen don't show up on the monster level table. Given everything else that's in the dungeons, why not a Neanderthal? After all, we call them cavemen because they tended to leave artifacts underground, which seems consistent with the underworld. Brigands are also missing, but they're really just Chaotic bandits.

Supplement I: Greyhawk changes things around. We find various levels of thieves at all levels, and encounter dwarves and elves only on the first monster level. We also get our only Lawful wanderers, as there are paladins roaming about these dungeons.

B/X D&D changes things around, by creating a whole category of random humans to wander around the underworld. Class-leveled characters still appear, but so do Traders, described as armed merchants carrying trade goods. This has potential that simply has never been fully exploited, in my opinion.

OD&D, though, has an interesting wrinkle: the Monster Determination and Level of Monster Matrix. Basically, the monster encountered on a given level might be from the same monster level, or from 1 level lower, or from up to 3 levels higher. The upshot of this is that dungeon level 3 is the only level where you can encounter all of the class-levelled types in OD&D. Level 4 (Evil Priests, Myrmidons, and Enchanters) is the most common Monster Level encountered there.

If we extrapolate logically from all this, the OD&D megadungeon suggests to us at least 3 human factions, and probably more. First, there are both bandits and berserkers, who are likely to appear on level 1 of the dungeon (both monster levels 1 and 2 appear most frequently on level 1). It's possible that one or both groups have a "base" on the first level of the dungeon, or that they are coming from a base camp outside the dungeon entrance. But either way, a classic OD&D megadungeon has bandits and berserkers nearby.

Beyond that there is at least one faction of leveled humans. Lords, Wizards and Evil High Priests – presumably the leaders of any such faction(s) – appear from level 3 and downward, so it is unlikely that the base of any faction would be on the first two levels. Evil Priests are most frequent on levels 3, 4, and 5, so one of these levels is likely to have a Chaotic temple, or at least some kind of idol or presence, in its depths.

Other factions are possible. Any faction with magic-users or fighting-men to it should have at least one outpost on level 3; otherwise, the Warriors, Conjurers and Theurgists in the monster level 2 table would have no "home base." But the faction's leader is increasingly likely to be on a deeper level, suggesting a faction that holds several areas in the dungeon.

This is an interesting challenge for dungeon design. Say that a faction has its "home base" on level 7 and its "outpost" on level 3. There should be a way for members of that faction to get from one to the other without fighting their way through. This may include truces with other factions, or a secret passageway, or a specific "safe" pathway that could also be found and exploited by the PCs.

Beyond these factions, it is of course possible that at least some of the humans are freelance. They may be a "rival NPC party," an idea that has been written about before. They might be traders; some accounts in Alarums & Excursions show that OD&D play did in fact include such oddities as wandering salesmen and religious proselytizers (think '70s Hare Krishna types) as dungeon encounters. Such a dungeon should have "neutral" spaces that are not simply about fighting all the time, and areas where at least some of its human inhabitants actually live.

Or, you know, they're just guys wandering around the dungeon. It doesn't have to make sense. But you can do a lot of interesting dungeon design if you extrapolate.

Red Herrings and Reaction Rolls (Plus Kickstarters)

A couple of brief observations from my last game this past Saturday.

Reaction Rolls: These can be funky things. One NPC gets extreme positive reaction rolls – 11s and 10s – and the other negative ones – 4s and 5s – and you wind up with one accompanying the party and the other getting killed (after using Charm on the hobbit). It's kind of an inspired mechanic.

Red Herrings: I keep placing things in my dungeon that appear promising but are ultimately not useful in terms of finding treasure. In play, this can turn into a massive time sink. Players want the good treasure, so an area that appears like it might have treasure but doesn't can lead to a lot of time spent with no serious experience point gain.

So here are half a dozen ideas to spice them up:

1: Add a timer to the area. Rising water, sinking ceiling, clanking noise getting louder, treasure that only exists during the current planetary alignment, etc. Give the players ample notice and such timers can make their exploration at least much quicker.

2: Red herrings with teeth. This works especially well in D&D if it's a big, slow monster. The danger is that the players might think it has treasure and stand their ground for a fight, but the next group will know better.

3: It's a trap! Exploring too closely leads to various cave-ins, pit traps, arrow traps, explosions and flattenings. Especially good if there is a warning prior to the red herring area. For a fun variant, trap them inside and make escape a challenge.

4: Just wandering by. Increase your wandering monster rate in the red herring area so that it becomes prohibitive. Again, this has the disadvantage that too many player groups will try to fight it out.

5: Tell them it's a red herring. Mark off a torch or a flask of oil, let the players know they found nothing after diligent search, and get on with the game.

6: Fuck it, it's a special. Sure, you had planned on this being a time sink and nothing more, but they've spent long enough on it already. Grab your copy of The Dungeon Alphabet and throw the players a bone (roll 1d8 for type).

So with that said, I want to highlight the Kickstarter campaigns I'm currently backing.

Primeval Thule 5e is, well, a 5e campaign setting. It sold me quickly by listing "Conan and Cthulhu" as its reference points. Howard and Lovecraft are still big reasons why I game, and I think having a 5e book that supports this style will be really useful. I like Sasquatch Studios's material for 5e (Lost Mine of Phandelver, Princes of the Apocalypse), so I'm happy supporting this.

Metamorphosis Alpha: Epsilon City is Jim Ward's big city level for the Starship Warden. Followers of this blog know I'm a fan of MA, and I'm extremely happy to see that Goodman Games is putting out another adventure for the system. There are noises about expanded material and a boxed set, but the Kickstarter isn't exactly rolling along despite a fairly rapid start, so folks should get their pledges in already!

Ernest Gary Gygax Jr.'s Marmoreal Tomb Campaign Starter is a dungeon module by Ernie Gygax and Benoist Poiré. This one is growing quickly, with the Gygax name and some good press driving it. Benoist is a superb mapper whose work can be seen in issue #3 of the AFS zine and issue #3 of Gygax Magazine; he is doing all of the project's maps. Ernie is the son of E. Gary Gygax, and is writing this as a module in the tradition he learned from his father, as well as other luminaries like Rob Kuntz, in Lake Geneva. It's a wonderful project and I felt I simply had to back it.

One final note. I've been quietly using the Arduin Grimoire's critical hit chart without telling my players. It is monstrous, but I understand why people made so many imitations. I doubt whether I'll continue using it, but I always reserve the right to have it in my back pocket.

A Missing Link: A Miniature Megadungeon

(A brief edit to note: This is officially my 300th post on this blog. Wow.)

Zach Howard at the Zenopus Archives has managed to surprise us again, this time showing the original cross-section of Holmes's sample dungeon. Take a look and come back, I'll be talking about the hand-drawn image a good bit.

This little diagram is a master class in the "megadungeon in miniature." It condenses into 3 levels most of the principal ideas that underlie the Gygaxian megadungeon. There are multiple entrances, multiple connections between levels, different types of terrain, sloping passages, and generally everything you'd want from a complex dungeon, all in three neat levels. None of which is to disregard Tom Wham's great Skull Mountain diagram, but Holmes manages to do a lot with a few levels.

We come to two entrances on separate hills. This creates an interesting choice right off the bat. There might be something low-level like goblins or kobolds guarding the hillside entrance, but the descent into the mine is obviously the deeper way down. It's a good idea to let the rumor tables give a hint that the mine shaft to level 2 is a way down to more difficult monsters than the hillside cave.

In level 1 we're about equidistant from the two ways down. The level below the ladder has one of the long, gradual slopes that Gygax was fond of, and a party may reasonably be surprised when they go up a level of stairs and are still on the second level. Holmes's original wandering monster tables, drawn from Supplement I: Greyhawk, might give us a good idea of what kinds of threats lurk in each of them. Given how many humans with levels there are, it either suggests that the dungeon has a significant human faction, or is actively plied by rival adventuring parties. Either choice is interesting.

I really love those two carved-out areas by the mine shaft. They just have a ton of potential for mischief. As soon as I saw them I was envisioning a nasty monster swooping out as the PCs try to go down (or up) the ladder and causing all kinds of havoc. Or they could be rooms that are rigged with nasty traps, maybe something explosive, or a simple arrow trap that happens to knock the PC a long way down to the dungeon floor below. Or one of them could have a treasure visible, but a monster or a trap nearby that will turn the PCs' avarice into their undoing.

The third level's relative size suggests it should be a nice, big, sprawling level with lots of rooms and interesting tricks. Then there's the cave, and I have to admit if it had a lake indicated it'd be exactly after my heart. As it stands, the cave feels like it should really be the lair of a dragon as the culmination of the whole adventure (and a justification for the game being "Dungeons & Dragons"). It would really make the whole thing a summation of D&D in three levels.

It doesn't have the domed city and the great stone skull of Wham's drawing, but Holmes's original sketch points to a dungeon that can be played in the three character levels suggested by his rule book, and still give you the megadungeon experience. There's no reason the weird and cool stuff we've talked about previously can't be on these three levels, and Wham's drawing really only gives us an extra two or three true "levels" (counting 2 and 2A, the entrance cave as "one up", and the cave separate from the third normal dungeon level). What I think using a dungeon like this does is gives just enough of the complex elements without going into "megadungeon fatigue" that modern games often run into. By the time the players are bored with runs into the same dungeon, they're finished.

From that perspective, this is the missing link between the megadungeon and the smaller types that came to dominate the scene after the mid-1970s. You could describe this in a relatively short module but have months of play material come out of it. The only published dungeon that really comes close to this is Caverns of Thracia, whose reputation should tell you how great of a dungeon I think this could be turned into.

How Not to Write an Adventure

Jason Paul McCartan at OSR Today wrote a short link to an article from a site called RPG Knights that alleges to give advice for how to design adventures. Unfortunately, it's really not. The advice given is a recapitulation of Freytag's pyramid (in a modified version slightly different from the above, where the rising action is temporarily interrupted), the dramatic structure you learn in middle school, without significant insight into how to make it into an RPG adventure.

This kind of adventure writing is lazy, bad and everything that should be avoided both by referees and by writers creating modules for RPGs. If you've already written the plot, the PCs aren't the protagonists; they are just along for the ride. And that sucks.

In a well run roleplaying game, the elements of Freytag's pyramid (exposition, rising action, climax, falling action and denouement) arise organically out of player choice. Plotting them in advance prevents this from happening; if your climax requires that a certain character be in a certain place at a certain time, well, the best-laid schemes o' mice an' men gang aft agley*.

An RPG adventure needs something completely different. It's a type of writing that is totally different from writing a screenplay or short story, since the referee is not an author and doesn't know what the protagonists will do. So it should be no surprise that the elements needed are totally different as well.

The main responsibility of the adventure is that it becomes plot when PCs are exposed to it. This requires it to have potential conflict, or the seeds of conflict, within it. This doesn't need to be anything fancy; it's just another way of saying there should be monsters and/or NPCs standing between the PCs and what they want. A dungeon will often do this literally, for instance by having the quintessential orc and pie. If the PCs decide they want pie, that instantly transforms into conflict between the PCs and the orc. Nothing fancy is required, and it can be as detailed or simple as the referee prefers.

Conflict can be between factions, or between NPCs, or simply with the PCs. The more complex your potential conflicts are, the more ways that adding PCs can make the plot go pear-shaped. What is critical is that nothing ever be indispensable. There can be no NPC who can't be killed, no monster that must get away from a fight, nothing that the PCs need to find or know or do that will stop the adventure cold.

Everything else, really, is optional. A dungeon with monster and treasure keys is a baseline for a solid adventure. But there are a few different elements that help a good adventure.

• Background. This can be revealed through exposition, items, and dressing. The real shame of a lot of professional adventure writing is that it has extensive background that is not revealed to the PCs organically through the elements in the module.
• Methods of discovery. Ways to reveal background and information about the world and their enemies to the PCs are helpful. This can be through books, talkative NPCs, visions of the distant past, or many  other strange and odd ways to show the world to the PCs. Rumor tables are a classic method for revelation and point up the key fact that they are not necessarily reliable
• Physical obstacles. Sticking a chasm between the PCs and a goal, or making an adventure location particularly dangerous to approach, are good ways to add to the conflict without reference to more NPCs or monsters. Traps, of course, are a personal favorite.
• Dynamic world elements. A good adventure has elements, usually random, that can happen throughout the adventure so that it is not static. For instance, a random encounter table indicates that events outside the PCs' adventures are happening, and it is not necessarily a good idea to respond to all of them. Other examples include timed changes to the setting, such as the Swedish Army that will be coming soon in Better Than Any Man.

Again – none of this relates directly to plot, and if the players want, no story other than "PCs go in, get gold and leave" needs to be told in the game. Each of these points can be covered whether the adventure is a good dungeon crawl or a solid city adventure where a sword is never drawn and a spell never uttered. What's important is that the adventure be open-ended and have several potential forks, because no plan ever survives contact with the enemy, and no plot survives contact with the PCs.

* The English for this line from Robert Burns's Scots poem "To a Mouse" is usually given as "The best-laid plans of mice and men often go awry."

Looking at a Real "Megadungeon"

Hat tip to Dave Younce on G+ for sharing this link.

BLDGBLOG: Nato's Underground Roman Super-Quarry

This is a terrific article that looks at a massive megadungeon abandoned Roman quarry / NATO bunker beneath the Netherlands and Belgium. You go into one nation and come out in another. There is a brilliant description of what it was like to navigate the underground tunnels:

Most corners were roughly 90 degrees, but only roughly. Going through the caves was an exercise in left and right turns every 50 feet or so. Navigation was helped by street names. Unlike in the USA, where streets are numbered on a sort of grid pattern, these were zigzag streets. My office on Main Street and J Street, so if I got lost I would just keep walking until I came to either Main or J, and join it. If I went the wrong way, eventually the street would peter out either at the perimeter or a T-junction, and you would just turn round and go back the other way.

This massive complex was built from limestone, and was soft enough to cut with a chainsaw; the bunker was a "black hole" where everything that was brought in was thrown out in one of the underground landfill sites.

The BLDGBLOG link doesn't show much of what this place looks like, but it provides links to some photosets that do. Here they are in a clean format that doesn't block right-clicking:

Talk Urbex - N.A.T.O Quarry
flickr (Behind The Signs) - NATO Quarry
28 Days Later forum - N.A.T.O Quarry, France.
28 Days Later forum - Nato Quarry, Paris Suburbs May 2011
Urban Ghosts - Urban Explorers Discover Corroding Military Vehicles in Abandoned Subterranean Bunker

It's fascinating how wide these corridors are; many are 20' or wider, but even the narrower pathways are anything but claustrophobic. You could easily have full, programmed encounters at any point in the dungeon, regardless of whether it was an "office" (cleared out room) or not. The stairs down are dizzying. And when you find a pile of rocks like this, it only raises the question of what is lurking beneath it.

One of my favorite parts in the quote above is that there were "street names" in this labyrinth. It just seems like a very organic way to tackle a labyrinth; of course, in a fantasy world these "names" might not be written in a comprehensible language for the PCs, but rather a series of sigils or unreadable characters that the players have to figure out are in fact used by the denizens to navigate. And the idea that "streets" are not straight makes sense if there is any defensive value to the layout, as anyone who has driven around Washington, DC can tell you.

The manmade-but-not-worked nature of the stone also has a lot of potential. Since it is relatively soft, it's a great material for laying in secret doors or cutting out niches in floors, walls and ceilings to hide treasure. Tunnels – the kind beloved of kobolds or giant rats – are pretty easy to dig, and PCs can even do some small-scale mining. And I love the idea of ladder rungs right in the stone.

Of course this is just one example of a large structure built underground, but it has some really great ideas for what things look like when people build things beneath the earth.

The Monumental Chamber

The size of the Son Doong cave, in Vietnam near the border with Laos, is massive. Its largest chamber is 5 kilometers long and 150 meters wide. It has massive stalagmites 70 meters tall. Almost 500' wide and over 16,000' deep is almost unfathomable. But it makes me think about huge underground areas. (Technically Son Doong is not all underground; a hole has burst through its roof and there are plants growing in parts of it.)

There are two parts to what I'm going to define as a monumental chamber. First: a person standing at one point and using a torch cannot see the edges of the area. A torch occupies a circle 30' in radius, or about 2827 square feet. A 60'x60' room is the smallest such monumental area; at 3600 square feet it has around 1.5 times the square footage of an average house. The second criterion is that the monumental room has multiple entries in the dungeon key. A room that is huge but empty doesn't count.

A monumental chamber is a gripping idea to me, mainly because of its sheer scale. Frequently they will be much more than the average 10' high, and as often as not will have some support, whether in columns or through a stable structure like a dome. Vast caverns can shift elevation even within a single room; a manmade version of such might have several platforms linked by stairs. The monumental room is open to pedestals, megaliths, gaping chasms, cliffs, lakes, rivers, waterfalls – any of the geographical features you've always wanted to use in a dungeon but can't fit without making the room humongous.

Inhabitants, too, need not be single. Sure, you could fit an entire orc tribe inside of a sufficiently large chamber. But a truly large room could be home to multiple entities, particularly if it has niches, alcoves or other defensible areas. Perhaps humanoids come out to harvest mushrooms, or crystals, or some metal, while other creatures live in a secluded area. This is particularly potent if the chamber is very high and has multiple effective levels. Bats and stirges, among other types of flying things, are prone to live in the upper reaches. One idea I've always relished is giant bees with a hive up at the top of a chamber at least 50' high. Lower levels, a sub-floor or a body of water all create other natural divisions for multiple encounter types. The sheer size can encourage a whole miniature ecology.

Huge rooms can house multiple wonders as well. Natural features can be spectacular, as can megalithic sculpture and large-scale architectural designs. At the same time, you have some interesting options. If you've ever been to the Metropolitan Museum of Art in New York City, you may have seen the Temple of Dendur – an Egyptian temple inside the museum. There's likewise a small Japanese temple in the Philadelphia Museum of Art. Massive caverns can logically have multiple sub-caverns. A monumental chamber can be a dungeon room with other rooms inside it.

The monumental chamber is also a great way to link several different levels together. Entrances can be at multiple heights; a chute from one level, a landing leading down from another, a third on the same level as the floor, or a fourth with a chute or stairs leading down to a deeper level. They make great "crossways" of the dungeon where wandering monsters can come from a variety of original dungeon levels as well.

I feel that I'd be remiss if I didn't mention that dragons are naturals in such areas. A truly massive hoard needs a massive room to fit in. But generally I think that, especially in megadungeons, it pays to think big. (No, bigger than that.)

OK, so it's a little wet ...

The galerias romanas (Roman galleries) in Lisbon, Portugal are one of the harder architectural marvels to get into. A crack in one of the galerias means that the underground ruins are accumulating groundwater, and most of the year, they are flooded. Every year, the ruins are painstakingly drained (the process takes a full month) and visitors are allowed to enter for about 3 days. The rest of the year, the water is allowed to stay, at about 3m depth (10').

Which is a fun way to transition to talking about water in dungeons. Homeowners know all about the tendencies of underground spaces to accumulate water; we've invented sump pumps, French drains and all sorts of systems to reverse this tendency. But in dungeons it's a fascinating obstacle that I've thought about for a while now.

The first thing you have to decide about water is whether it is flowing or stagnant. If it is flowing underground, there must be some source: a river, spring or aquifer that is feeding the body, and a corresponding outlet. If there were no outlet, the body of water would continuously rise. (This is certainly a possibility: an extremely slow rising body of water, fed at a rate of well less than half an inch per year.)

Stagnant water has been deposited by some source and is not continuously refreshed. This is extremely dangerous to humans, since stagnant water is a breeding ground for disease and bacteria. Contact with a pool of such water should require some kind of appropriate disease checks; suitable rules exist in Blackmoor, the Dungeon Master's Guide and Lamentations of the Flame Princess. Drinking it is even more dangerous.

After that the next decision is depth. Water becomes a different type of obstacle depending on the depth. A foot deep permits more motion than two; three might be navigable for humans but is swimming territory for halflings. Deeper than that becomes a serious problem, if not totally impossible to cross. Water, of course, ruins most of the things adventurers carry; waterlogged boots can get mildew or mold, and generally become useless, while soaked torches or spellbooks are useless. Swords and armor will rust if not properly cared for. And swimming while in armor is basically impossible. Watch the video if you don't believe me – heck, even if you do believe me, it's a great experimental video.

Other factors can also be interesting: temperature, current, or brackishness. For instance, clear, cool water is very different from hot mineral spring water. In a dungeon near the sea, some water may be salty and undrinkable. And it could be very dangerous if the water has a strong current, pulling adventurers along with it as they try to navigate it. Varying any of these can create unique environments and pose certain challenges.

Now that we've put water into the dungeon, comes the payoff: all the water monsters in D&D are fair game. From sharks to sahuagin, from ixitxachitl to aboleths, and any kind of new monsters based on fish or whales or anything else you like, you've now got a reason to use these in dungeons. A long-standing version might have a blind cave version of the classic lake monster archetype, or vicious schools of piranha, or pretty much any cave swimmers you might imagine. It's a rich environment for monsters that aren't typical. And let's not forget about amphibians, arthropods and the like.

Of course where there are monsters, there is also treasure. This can be a bit tough to figure out. In chemically pure water, basically, copper, silver and gold should be relatively unaffected by corrosion. (In fact, copper coins in fresh water should tend to reduce the bacterial load.) The problem is that water is not pure and contains other minerals and compounds that will interact with metal. It could also complicate things if the coins are not pure metal. Seawater will corrode copper and silver coins, but leave gold fairly intact. Only acidity, introduced by the breakdown of wood or the general condition of sandy soil, would pose much of a danger to gold coins. The most dangerous thing would be an untreated wooden chest containing them.

Most gems will not be water soluble and should survive pretty well in tact. But iron is toast even if it gets minor exposure, as are most books or scrolls (leather parchment will be ruined and the ink will dissolve), and pretty much any wooden objects that aren't covered with tar, pitch or some other sealant. Magic items are the referee's call; it would be interesting to find a shining steel blade in a rotting leather scabbard and with a ruined handle, but the blade and tang in tact. Potions may or may not leak, which could have very interesting results – say, a wizard's laboratory where several potions have mixed together to form ... well, pretty much anything you want.

Water complicates everything very nicely. Spells will be interesting, from Sleep (does it turn into an effective Drown?) to Lightning Bolt and others. A Fireball might boil some of it. And of course there are many interesting spells and items that only work under water.

Given all these possibilities, it's a disappointment that there aren't more modules and books dealing with flooded environments. It's definitely something I keep intending to explore. I hope this inspires other referees to do the same.

Witches, Lovers and Pigs

In addition the charisma score is usable to decide things such as whether a witch capturing a player will turn him into a swine or keep him enchanted as a lover.

This is an interesting allusion. Gygax is clearly referencing Homer's Odyssey, where the enchantress Circe uses her magic to turn the companions of Odysseus into pigs and then keeps Odysseus as her lover. He stays on her island for a year, then finally frees himself and sails home to Ithaca. Holmes changed this to a frog, perhaps thinking that The Frog Prince would be more familiar to the young audience of his Basic Set than The Odyssey.

Lately I've been thinking about the idea of various sorts of magic-using types in D&D. Generally, these prompt the objection that a single Sleep spell can beat an entire party of low-level PCs. And while this is very true, they remain a standard part of the encounter tables, and in my opinion provide rich encounter potential. It helps to think of the Sleep spell as similar to the NPC having a powerful weapon and requiring negotiation instead of being dealt with in the "standard" way.

One particular role I like for magic-users is that of a dungeon "neutral." I think these are generally underused: creatures and personalities of the dungeon that aren't necessarily harmful, nor helpful; depending upon the PCs' actions, they can be taken in either direction. Magic-users are great for this, in a whole variety of ways. In the early levels of a dungeon, they can be conduits for spells and items that PCs won't necessarily have access to; a 5th level magic-user in the second dungeon level, for instance, can do all variety of divination that the PCs don't have access to yet. And the referee can make the cost of such assistance whatever they like, allowing them to set up all sorts of missions into the dungeon.

The second role is that of a nemesis. An enemy witch or magic-user is a great foe, in no small part because of the threat of polymorph and charm spells. While I am generally leery of the "boss fight" concept from video games, they do make good leaders of dungeon factions. The built-in fragility of magic-using types in D&D means that they will not be there for long drag-out fights, but rather short, desperate struggles as the PCs try to disable them. In the mean time, their abilities make the PCs' lives very interesting, since their comings and goings might be monitored, illusions cast, or many other magical tricks used to deter them.

The third concept is something I've mentioned before: giving NPC magic-users totally sui generis abilities. Circe, after all, was somewhere between a witch and a minor goddess, depending on what interpretation you choose. They can be the source of magic abilities unseen in the books, or even magical items that are unique without being artifact-level objects. (I'm thinking here of the magic sword forged by a witch from lightning bolts in The King of Elfland's Daughter.)

At heart, this dilemma — lover or pig — is an interesting one to place PCs into. It requires care that it doesn't become a railroading tool, but embracing both possibilities opens wide the options for magic-using NPCs as a major force for the kind of emergent dungeon roleplaying that I find really fascinating.

Keying the Corridors

If you bought Dungeon Crawl #3, you know that there are several places located in the corridors where I've keyed the traps. I'm thinking more and more that this is a philosophy I want to use and see more of in published modules.

I've found in writing adventures for Dungeon Crawl that numbering rooms tends to lead to a "one thing per room" mentality to dungeons. You can only get so much detail into a single entry before you cause your reader's eyes to glaze over; but at the same time, there needs to be more in the dungeon than answers to "what is behind this door / in this room?"

A big part of this is the idea that there are interesting things in the corridors. After all, the stereotypical old school dungeon has 10' wide corridors; my office is narrower than that. A rat's nest that is in a room could just as we'll be in a corridor; a hallway can have niches filled with statues, or hollow areas where treasure or traps are hidden. Creatures, traps and puzzles in the corridors make for a more tactically rich environment.

A lot of creature types are well suited to corridor encounters. For instance, the "cleanup crew" - slimes, oozes, jellies - make a lot of sense in corridors. Rats and other creatures of unusual size, which I've talked about in the past as much more compact than the space they're usually given, are another good choice. Non-intelligent undead make as much sense in a corridor as anywhere else; why would a lumbering zombie prefer a room? Generally opportunistic creatures should logically be in hallways.

For mobile threats, one interesting wrinkle is to make their presence a numbers game. If there is a giant rat nest in a niche of the corridor, perhaps there is only a 2 in 6 chance that the rats are there. They can go in the wandering monster list for the rest of the time.

Obviously the corridor encounters won't have non-hidden, unguarded treasure, but the possibility of hidden treasures is real. A single gem in the base of a statue can raise the stakes of a dungeon permanently, with PCs vigilantly checking every possible location for treasure, at least until they find a trapped one, and of course pit raps, arrow traps etc are natural corridor hazards.

The intrigue is in the ability of PCs to avoid encounters. These should not be in places that can't be circumvented; instead they give certain paths a higher difficulty cost. Perhaps the players find a shortcut but there are shriekers, or green slime, or another obvious challenge. Or perhaps they found an empty giant weasel nest going down a hallway; do they risk counting on it not being home on their way back?

So you'll see more of this from me in the future. I'm curious what others have done with this and what you think of the idea.

Disasters in the Underworld

Historically, we as humans have spent a lot of time digging holes in the ground. Throughout that history, we've had those holes collapse or blow us to holy hell. Coal miners have been known to bring canaries underground so that the birds would die from carbon monoxide poison while the miners still had a chance to escape alive. Even in modern industrial mines, a blowout can still lead to fatalities and miners spending days or weeks trapped underground with limited food and no sunlight. In Centralia, Pennsylvania, a coal mine fire started in 1962, is still burning in 2013 and will continue to burn unabated for 250 years.

In Dungeons & Dragons there's so much we can take out of this. The point I want to reach is in making underground disasters a challenge and not an annoyance or a simple declaration of "rocks fall, everyone dies."

The first type is to take a disaster that happened in the past. It becomes dungeon dressing: this is a collapsed hallway. Perhaps it conceals treasure, or a short cut, or a danger so awful that someone tried to bury it beneath the earth. Or maybe it's just a cave-in. There could be any number of horrors lurking in the earth, both giant insects and various tunneling monsters like the bullette. A wholly or partially collapsed dungeon area is fair game for being inherently dangerous; if it collapses, the existing fall should have served as adequate warning.

Other kinds of disasters than simple collapses could be interesting. Fire is one very dangerous thing; at low levels, a party will tend by its nature to have torches and/or lanterns. In general, there are plenty of flammable gases underground. A methane leak 50 years ago could make the air (or, as we know from modern gas mining, the water) flammable. Areas in an old mine or similar area might have the chance of causing a dangerous flare-off. On the other side you have possibilities like the Centralia mine, where there is literally a long-standing fire underground.

The second type of disaster is the one that only exists in potential. The most obvious of these is the inherently unstable area, where the local stone is weak and prone to shift catastrophically. The legend of the "tommyknockers" is drawn from the knocking sounds made by wood mine supports as they buckle and fail in advance of a cave-in. This is generally a feature that can make dungeons very dangerous, but players are likely to find it grossly unfair if the referee simply has a cave-in kill the PCs, however naturalistic that may be.

Such disasters can be foreshadowed in various ways. Noise is one; queer rumblings deep in the earth could be a dragon or balrog, but they could also be the signals that the earth is waiting to reclaim our band of adventurers. Popping and knocking in a wood-reinforced mine, or even the occasional support that is cracked down the middle or falling, can be one way of signalling this. All of this creates a sense of danger, and creates a time / action pressure on the PCs. It's one way to create what I've called in the past "fast" levels - levels that don't encourage in-depth exploration.

Of course there's also the possibility of PCs being trapped in the underworld. This can be a really railroad-ish tactic for a referee, although it can also really turn the "exploration" aspect of the game up if it leads to a new or different area of the dungeon. In a way this is a deeply Gygaxian trap, something like his teleporters that create a sudden need-to-find-the-exit situation in the underworld. This works particularly well if the cave-in happens to be a sinkhole down to a new level.

Last, a military solution would be an interesting cave-in situation: it is possible that some monsters or enemies have built a counter-mine intersecting with the "normal" dungeon that causes a collapse and then reveals a waiting ambush. For instance, undermining most of the way through a floor might make a normal person fall through it and become vulnerable to the attack below. In general counter-mines are a really fascinating way to fundamentally change a dungeon from beneath the ground.

As usual with my "idea" posts, I'll yield the floor with the question: what have you done with disasters, large-scale fires, cave-ins and other catastrophes in the dungeon?