In the Matryoshka Search Technique, I described a method for resolving search checks that keeps players actively engaged with the actions of their characters so that they feel ownership of what their characters discover. You may want to read that post before reading this one, but the short version is:

• On a successful search check, instead of immediately discovering the point of interest, the character discovers an indicator pointing them in the direction of the target of the search.
• This requires the player to draw a conclusion and then act on that conclusion.

For example, instead of saying, “You find a hidden switch on the frame of the painting that opens a secret door,” you would instead say, “You notice that the frame of the painting is quite dusty, except for one spot in the lower right corner.”

Because it is the player’s declared action (“I examine that corner of the painting” and then “I push on the raised portion of the heron engraving”) that result in the switch being discovered and the secret door being opened – rather than the dice roll – the player is empowered and it feels as if THEY were the one who found it. (Because it was.)

I refer to this as a matryoshka technique because it’s like a Russian nesting doll: Instead of showing the players the innermost doll, you instead hand them the full stack and let them open each one.

But it’s not the only matryoshka technique.

If you’re running a hexcrawl, for example, you can use a matryoshka technique even within a single hex.

For example, even if you’re using an extremely simplistic hexcrawl structure in which the PCs automatically encounter the keyed content of each hex when they enter the hex, you can – instead of having them just run directly into it – give them an indicator.

So instead of saying, “You see a goblin village,” you describe the goblin tracks they stumble across. Or describe a plume of smoke on the horizon.

Do they follow the tracks? And, if they do, are they successful?

Do they head towards the plume of smoke? Or avoid it?

Instead of stumbling directly onto an ancient pyramid, the PCs spot a patch of ancient road almost completely swallowed up by the jungle.

Are they able to figure out which direction the road originally ran? Do they divert to see where it went?

And, like any matryoshka technique, this doesn’t have to be just one level deep. It can be nested to several levels.

It can also be interesting if they get two different indicators at the same time. For example, they see a plume of smoke to the west, but the goblin tracks are heading north.

(With a robust structure for running wilderness exploration, multiple indicators like this can be spontaneously generated. For example, you might have both the keyed location in a hex they just entered and the random encounter you just rolled.)

The goal of the technique is to draw players more deeply into the game world and empower them to actually explore the game world (instead of letting the dice do it for them). Not just because it will make them feel cool (although it will), but because it positions them to make meaningful choices. And that’s exciting for everyone at the table, because the consequences of those choices will transform the campaign.

FURTHER READING
Matryoshka Hexes

This is an unusual post.

After writing Part 5: Encounters of the 5e Hexcrawls series, I received feedback from readers that the post wasn’t clear enough and needed several points of clarification. I concluded it would also benefit significantly from some substantive examples of the system in practice.

Which is, of course, what you’ll find below.

This material has been added to the original article (linked above, but I also didn’t want people who had already read the original article missing this update). So this post is being posted… and is also immediately completely redundant.

So if you haven’t read the original article, you should just do that now and skip the rest of this. Otherwise, read on!

EXAMPLE: SAMPLE ENCOUNTER TABLES

Location Check: 1 in 1d6

Encounter Check: 1 in 1d10

Border Encounter: 1 in 1d20

Note: I indicate hexes which are already keyed as potential lairs for this creature type. This can inform the nature of wandering encounters and/or suggest a potential origin/terminus for tracks.

This table uses several advanced rules. When rolling an encounter, I would simultaneously roll a 1d6, 1d10, and 1d20 for each watch.

If the 1d6 result is a 1 (indicating a location encounter), it would indicate that the PCs have found the keyed location in the hex. If I’m not using simultaneous encounters, I would then ignore the other dice rolls (the location check “overrides” them; you could also just roll the 1d6, then the 1d10, then the 1d20, but that’s not necessary and is more time-consuming).

If the 1d10 check indicates an encounter, then you’d check the 1d20 roll to see which encounter table you should be rolling on. (You could also theoretically roll 2d20 of different colors, allowing you to immediately identify what type of encounter.)

With an encounter identified, you would then check % Lair, % Tracks, and # Appearing (although you don’t need to check for tracks if a lair encounter is indicated). Lairs and tracks are also exploration encounters, so if those are indicated when the party is resting, you can treat the encounter check as having no result and the watch passes quietly.

This is, of course, a fairly complicated example featuring a lot of the advanced rules all being used simultaneously. For a much simpler resolution you could just roll 1d12 (1 = wandering encounter, 12 = location encounter), roll 1d20 on the wandering encounter table (if a wandering encounter is indicated), and then the number of creatures appearing.

DESIGN NOTE: PROCEDURAL vs. DESIGNED ENCOUNTERS

A procedural encounter will usually generate one or more general elements. (For example, 1d6 friendly orcs.) As described in Breathing Life Into the Wandering Monster, the expectation is that the DM will contextualize this encounter. In other words, the procedural encounter is an improv prompt for the DM to create the encounter (often combined with a simulationist element of modeling, for example, what kinds of monsters lurk in the Darkovian Woods).

A designed encounter, on the other hand, is far more specific: You’re essentially prepping the material that you would improvise with a procedural encounter.

The Principles of Smart Prep maintain that you generally shouldn’t prep material that can be just as easily improvised at the table, so generally speaking I would describe most designed encounters as being training wheels for DMs who aren’t confident improvising encounters from procedural prompts yet. (There can be a number of exceptions to this, but they’re pretty rare in actual practice, in my experience.)

In other words, designed encounter tables typically result in a lot of wasted prep. They also get used up (a procedural encounter can be used over and over and over again to varying results; a designed encounter is specific and generally can’t be repeated). This creates gaps in your encounter table and a need to frequently restock them.

(Procedural-based encounter tables will also need to be tweaked or restocked from time to time – if the PCs wipe out the goblin village, it may result in no further encounters with goblins – but this is very rare in comparison.)

DESIGN NOTE: SETTING LAIR/TRACK PERCENTAGES

In designing your encounter tables, the % Lair and % Tracks values can be set arbitrarily. For a quick rule of thumb, use Lair 20% (or Nil for animals that don’t really have lairs) and Tracks 40%.

Older editions actually included values for one or both of these stats in their monster entries, so for some creatures you may be able to reference those older resources.

A gamist tip here is to increase the % Tracks value based on difficulty: If there’s a monster that’s a lot more powerful than everything else in the region, crank up the % Tracks so that the PCs are far more likely to become aware that it’s there than they are to run into it blindly.

A simulationist tip is to vary both numbers by a sense of the creature’s behavior. Here’s an easy example: How likely is a flying creature to leave tracks compared to a woolly mammoth? (See Hexcrawl Tool: Tracks for  thoughts on what types of tracks a flying creature would leave.)

A dramatist tip is to think about how interesting each type of encounter is for each creature type. Is a ghoul lair more interesting than running into a pack of ghouls in the wild? If so, crank up the ghoul’s % Lair.

The last thing to consider is that, as noted above, a Lair encounter will generally add a new location to the current hex. The higher you set the % Lair values on your encounter tables, the more often this will happen and the quicker areas of your campaign world will fill up with procedurally generated points of interest.

Conversely, how comfortable are you improvising this type of content? It’s good to stretch your creative muscles, but it may make more sense to keep the % Lair value low until you’ve gotten more comfortable with pulling lairs out of your hat.

Go to 5E Hexcrawls

Go to Table of Contents

With Zariel’s story at the heart of Descent Into Darkness, the legions of Hell play a significant part in the campaign. In designing the Remix, I eventually reached the point where it felt necessary to actually nail down some of the details of how Hell’s military is organized.

My goal with these notes is not to be hyper-detailed or encyclopedic. In reality, any military with have dozens of minute distinctions within and countless exceptions to systems like this. And this is Hell itself — a fiendish bureaucracy with a legacy of aeons. Random Worldbuilding: Creating Noble Titles suggested a three-step process that will probably be useful to remember here:

1. Create a system that makes perfect sense.
2. Create two exceptions: One grandfathered in from some other system. Another that’s newfangled and recent.
3. Whenever the current system doesn’t quite work right for your current adventure… add an exception.

What I’m looking to establish here is a broad framework that will keep us oriented and consistent, while providing clear guideposts for improvising during play.

In my personal canon, the Legions of Hell are logically descended from the Legions of Heaven (because Asmodeus brought that military command with him when he betrayed Heaven to conquer Hell), which is why you’ll see broadly similar ranks in, for example, flashbacks to the Averniad.

Design Note: Our org-chart here takes inspiration from the legions of Rome, primarily because it provides a Latinate patina that feels stylistically appropriate for Heaven and Hell. Rome’s own military was reorganized countless times during its hundreds of years of history, and we’re not trying to accurately model the Roman military in any case. But if you’re looking to expand upon these ideas, you might look there for further inspiration.

If you want to add a lot of byzantine complexity, you might hypothesize that different blood legions use different org-charts (in, say, the fashion of the United States Army and Navy). In which case you could literally go Byzantine, by drawing inspiration from Byzantine military titles and the like.

BLOOD LEGIONS

There are eight blood legions – cruor legio – each commanded by one of the Dark Eight, the generals who serve the Archduke of Avernus.

Each blood legion is made up of an incredibly large number of individual legions. Some scholars cite a specific number (one hundred or one thousand or six hundred and sixty-six are popular choices); others claim there are hundreds, thousands, or even an infinite number of legions. (Thus the elven poet Suntithis famously describes the legions of Hell as the “eight infinities,” which may have inspired Aternicus to describe the apotheosis of Asmodeus as the “infinite betrayal, born eight times in blood.”) Let’s simply say “countless legions” and you can decide how much hyperbole is involved with that, if any.

There is presumably a High Command that serves as an interface between the Dark Eight and the individual legion commanders, but this strata of the military hierarchy is not particularly in our focus. If necessary, we can refer to these officers as Tribunes, with the understanding that there are many different types and gradations of the tribuni. (Members of the tribunate are perhaps further denoted by tiered honorifics that appear after their name, like augusta – e.g., Hastati Betrazalel Augusta is a devil named Betrazalel with the tribuni rank of Hastrati in the class of Augusta.)

Each legion is made up of cohorts, both of which are numbered, with cohorts (particularly cohorts on detached duty) often being identified by their number followed by the number of their legion. (For example, the 9^th Cohort of the 497^th Infantry Legion, also referred to as the 9/497.) Specific legions or cohorts may also have specific titles or nicknames, with varying degrees of “official” recognition. These can denote or be based on elements such as:

• Founding officers or notable historic leaders (e.g., the Belum legions founded by Archduke Bel, although some legions are also known as Belum or Belum Veterana because they are or were directly command by Bel)
• Important historic accomplishments (e.g. the Conquerors of Athalka)
• Awarded honorifics (e.g. triumphantes or perpessio)
• Descriptors of the legion (e.g., the Stygii legions of the fifth layer of Hell or barbazii legions made up entirely of bearded devils)

In other cases, or in addition to these elements, they can also just be adopted as cool names (e.g. Terror Incarnate).

Cohorts are specialized for specific roles or battlefields:

• Infantry
• Aerial
• Cavalry
• Aerial Cavalry
• Aquatic
• Subterrene

Most legions are formed from uniform cohorts and will be referred to similarly (e.g., the 497^th Infantry Legion), although some special legions will have diverse cohorts (for example, Zariel’s 5^th Legion is composed of the 3^rd Aerial Cohort, 7^th Infantry Cohort, and 9^th Cavalry Auxilary.)

Auxiliaries are similar to cohorts, but are either smaller in size, more limited in utility, or both. (Some auxiliaries will be highly specialized, veteran troops with extremely unique skills. Others are essentially trainee cohorts.)

KEY MILITARY RANKS

Optio (pl. optiones): Field officers who command small troop units.

Primus (pl. prima): Roughly equivalent to a lieutenant. They will either be in command of slightly larger troop units and/or have several optios reporting to them.

Triarius (pl. triarii): The commanding officer to which prima report to, and who reports to the leader of the cohort.

Signifier (pl. signifiers): A lesser leader of a cohort or auxiliary; a junior princeps.

Princeps (pl. principia): The leader of a cohort.

Legate (pl. legates): The commander of a legion. This title is sometimes translated as “General” in the Common tongue.

THINKING IN RANKS

As a practical, but completely non-binding, design guideline, I’m going to think of these ranks in the following terms at the table:

• An encounter with a couple of devil soldiers probably doesn’t feature a commanding officer.
• When encountering a squad made up uniformly of one type of devil (e.g., 6 bearded devils), I’ll generally have them led by an optio.
• A primus will generally be a slightly more powerful devil leading a squad of less powerful devils. (For example, you might have a chain devil primus leading a squad of 6 bearded devils.) Alternatively, you might use an alternative stat block from the Enhanced Devils supplement, which are handily designed not to make devils more powerful, but to make them more varied. (For example, you might have a bearded devil squadron led by a bearded devil primus with innate spellcasting.)
• A triarius is usually going to be a significant CR bump above whatever the baseline troops are. If you’ve got an infantry legion of bearded devils, then perhaps the triarius is an erinyes or horned devil. Conversely, if it’s a legion of lemures or imps, then the bearded devils might be triarii.

Legates, Princeps, and Signifiers are probably all significant characters that you would be placing with some care and thought. It would be in no way inappropriate to see pit fiend legates, although less powerful (i.e., slayable) legates are quite possible.

Go to the Avernus Remix

Go to Part 1

Having concluded that mid-dungeon rests should generally be risky and usually require effort in order to pull off successfully, how should we handle spells or other special abilities that make these rests easier to achieve? Some of these – particularly rope trick and Leomund’s tiny hut – seem tailor-made to make risk-free rests effortless!

In handling these spells in play, there are a couple general principles I think are useful to keep in mind:

First, we’ll want to think about how these spells fit into our procedure. Much like our response to a fireball spell shouldn’t be, “Well, I guess this means we don’t need the combat system any more!” our response to PCs using spells to achieve safer rests shouldn’t be to throw out our procedures for handling risky rests.

Second, it will help to remember that spells cast to increase the likelihood of a successful rest are, in fact, resources being spent. (Even if spells can be ritually cast, they’re still chewing up a prepared spell for most classes.) This is particularly true if such strategies require multiple spells to be used. (For example, using pass without trace to conceal the trail leading to an invisible rope trick while using alarm to prevent ambushes.)

In concert, these principles should make us wary of a kneejerk desire to “nerf” such spells. It’s easy to slip into thinking that the goal is to interrupt and disrupt the PCs’ rests, and thus become frustrated with strategies that successfully allow them to rest.

But if we return to thinking about an expedition model, in which players strategically choose how to spend your resources in order to maximize their rewards, this is exactly what we want.

On the other hand, if there’s a strategy which is costing them no resources at all and always results in successful rests, those are the ones we’ll want to interrogate closely and figure out how to handle better. (Or modify mechanically in order to balance them.)

LEOMUND’S TINY HUT

As our first example, let’s look at Leomund’s tiny hut, a 3^rd-level spell which lasts for 8 hours:

A 10-foot-radius immobile dome of force springs into existence around and above you and remains stationary for the duration. The spell ends if you leave its area.

Nine creatures of Medium size or smaller can fit inside the dome. The spell fails if its area includes a larger creature or more than nine creatures.

Creatures and objects within the dome when you cast this spell can move through it freely. All other creatures and objects are barred from passing through it. Spells and other magical effects can’t extend through the dome or be cast through it. The atmosphere inside the space is comfortable and dry, regardless of the weather outside.

Until the spell ends, you can command the interior to become dimly lit or dark. The dome is opaque from the outside, of any color you choose, but it is transparent from the inside.

How does this spell affect our dungeon rest procedure?

By and large, it doesn’t.

The procedure determines whether or not the bad guys locate the PCs. And whether they find the PCs or “just” a mysterious dome of force energy that wasn’t there yesterday, it doesn’t really change anything.

Leomund’s tiny hut makes the PCs perfectly secure… but also perfectly detectable. (You could even argue that an immobile dome of force is probably more difficult to conceal in many circumstances, perhaps granting disadvantage to their Stealth checks.)

What I’ve discovered, in practice, is that this creates a really interesting dilemma for the PCs when the monsters discover their position: Do you remain within to finish your rest, while allowing the monsters to fetch reinforcements and set up preparations to assault you as soon as the spell drops? Or do you disrupt your rest and leap out to deal with the threat?

The challenge for the DM, in my opinion, is to create varied and interesting threats. These will be dependent on the particular circumstances of the dungeon, but can be almost limitless in variety.

For example, in one instance the NPCs in my game pulled a Cask of Amontillado, simply walling up the room that the PCs had placed their dome in. Others have built bonfires over the dome and drenched them in oil. Triggering a cave-in to bury the dome can also be a really interesting option.

Of course, you can almost always default to the bad guys fetching reinforcements, bringing into play the principles of running an active dungeon that we discussed in Part 1. This can easily be the most dangerous option for the PCs to wait out, as a large part of strategically conquering an active dungeon is NOT allowing all the denizens of the dungeon to attack you at the same time.

As a result, in addition to simple fight now / fight later dynamics, I’ve also frequently seen tiny hut dilemmas result in tense negotiations.

My point with all this is that the tiny hut, while it gives the PCs a lot of advantages, doesn’t actually negate the logistics of seeking rest in the dungeon. In fact, the spell creates the opportunity for incredibly cool moments and incredibly difficult choices that otherwise wouldn’t exist.

Nerf Note: If you want to nerf this spell, consider reverting a major change that was made to the spell with 2^nd Edition (which inflated the number of creatures which could fit inside the hut) and 3^rd Edition (which inflated it again to the current total of nine). In 1^st Edition, it could only hold six man-sized creatures.

This makes the spell useful for a PC-only adventuring party, but the minute you add hirelings, allies, or just some mounts, the utility of the spell rapidly degrades.

ROPE TRICK

Next up, let’s consider rope trick, a 2^nd-level spell that lasts for 1 hour:

You touch a length of rope up to 60 feet long. One end of the rope then rises into the air until the whole rope hangs perpendicular to the ground. At the upper end of the rope, an invisible entrance opens to an extradimensional space that lasts until the spell ends.

The extradimensional space can be reached by climbing to the top of the rope. The space can hold as many as eight Medium or smaller creatures. The rope can be pulled up into the space, making the rope disappear from view outside the space.

Attacks and spells can’t cross through the entrance into or out of the extradimensional space, but those inside can see out of it as if through a 3-foot-by-5-foot window centered on the rope. Anything inside the extradimensional space drops out when the spell ends.

Although a lower level spell with a much shorter duration, rope trick, in my opinion, can actually be far more effective in gaining a secure rest than Leomund’s tiny hut.

Once again looking at our dungeon rest procedure, the key thing to note is that the rope trick is a very good way of creating an effective hiding place almost anywhere in the dungeon.

However, let’s consider the key feature of the spell: The dimensional portal created by the rope trick is invisible, but does not actually disappear.

This means that NPCs can enter the portal. (And it only takes a few for the whole thing to collapse.) This might be the result of them tracking the PCs’ to the portal entrance. Particularly tall creatures might also just stumble straight into it if the PCs aren’t careful in where they place the entrance.

The other key thing is that the spell doesn’t block sound. This is particularly significant in 5^th Edition because of how the game handles invisibility: Being invisible makes it possible to hide, but doesn’t even grant advantage on Stealth checks (including the Stealth checks built into our dungeon rest procedure).

In addition to assaulting a rope trick, of course, the NPCs might also choose to besiege it using tactics similar to those described for a tiny hut above.

Nerf Note: A really key limitation of rope trick from previous editions was that you couldn’t put extradimensional spaces inside other extradimensional spaces, preventing anyone carrying a bag of holding from entering a rope trick. Given the ubiquitous utility of a bag of holding, this creates a meaningful strategic dilemma: Be able to carry a lot more loot out of the dungeon OR take advantage of the respite of a rope trick. You can’t do both.

MORDENKAINEN’S MAGNIFICENT MANSION

Mordenkainen’s magnificent mansion is a 7^th level spell that lasts for 24 hours:

You conjure an extradimensional dwelling in range that lasts for the duration. You choose where its one entrance is located. The entrance shimmers faintly and is 5 feet wide and 10 feet tall. You an any creature you designate when you cast the spell can enter the extradimensional dwelling as long as the portal remains open. You can open or close the portal if you are within 30 feet of it. While closed the portal is invisible.

There is more to the description of this spell, but this first paragraph contains the bits relevant to our discussion.

The key thing to note is that Mordenkainen’s magnificent mansion largely bypasses most of the strategies we discussed for tiny huts and rope tricks: Once the portal is closed, it blocks sound and cannot be accidentally passed through. And it’s invisible, so the whole thing will not be easily discovered by patrolling monsters.

Its only Achilles’ heel is that the entrance to the mansion will still be readily apparent to any creatures with truesight (or other means of seeing the invisible).

How much of a problem is this?

Well, probably not that much. This spell is the same level as teleport, which provides an entirely different way of securing a safe rest before resuming your dungeon exploration from the exact point you left off (by, obviously, teleporting out of the dungeon and then teleporting back in).

This isn’t an accident: At the levels where these spells become available, the game is shifting away from dungeon expeditions being the focus of play. (This is not always reflected in published adventures, but nonetheless remains true in the game’s design.) Teleport away? Hide in a magic mansion? Use passwall (albeit nerfed in more recent editions) or transmute rock to reengineer the dungeon? From a design standpoint, that’s fine, because at these higher tiers of play, the game has (or at least should) move beyond the low-key logistics of dungeon delving.

This was an unexpected Part 2. Thanks to the Alexandrites on my Discord, particularly Pooserville, for suggesting it.

D&D was strategically built around the expedition. In its most basic form, your group gathers its resources, journeys into the unknown, and attempts to maximize the treasure they can gain for the resources you’ve invested. If you plan your expedition carefully and execute it well, you’ll return home with riches.

At low levels, these expeditions generally descended into dungeons. At mid-levels, funded by resources from the dungeon, more complicated expeditions would be mounted into the wilderness, with commensurately larger rewards. The resources thus gained, at high level, could be used to clear the wilderness and found baronies or establish churches.

The game has long since evolved to support many more styles of play than this three-tier expedition structure, but its DNA remains deeply embedded in D&D’s mechanics.

A dungeon expedition, for example, is largely about managing a pool of daily resources: Once you’ve expended those resources (spells, hit points, etc.), it’s usually time to withdraw from the dungeon and regroup for your next expedition. Many of these resources, of course, were baked into the class and level of player characters, and most of them are still there today.

The drawback of leaving the dungeon to regroup, of course, is that you’re back at the entrance, and your next daily expedition will have to start over from the beginning. While you were gone, the bad guys will have reoccupied rooms, reset traps, raised new defensive barricades, and generally made it difficult for you to get back to where you were and continue your exploration.

(Or, worse yet, they might just pack up and leave, taking with them whatever you were there to obtain in the first place.)

In other words, it’s going to raise the costs of your next expedition, which will result in it being less profitable.

So even today, whether you’re in the dungeon to liberate treasure or not, it can be very tempting to skip the withdrawal and stay in the dungeon while resting to recover your pool of daily resources. If you can pull it off, you can avoid some or all of the costs of leaving the dungeon and then working your way back to where you are.

Now this is the point where many DMs – particularly new DMs – make a mistake: Their dungeons are static and reactive. In other words, the challenging content of each dungeon room simply waits for the PCs to enter the room. Furthermore, when a dungeon room is emptied by the PCs, it simply remains empty instead of being restocked.

The result is that there is no challenge to resting in the dungeon, and rarely any cost for leaving and then returning.

This breaks the expedition cycle at the heart of D&D’s mechanics. Instead of needing to carefully budget and strategically employ your pool of resources, you can instead simply burn through them as quickly as possible, automatically rest to regain them without consequence, and then do it again. (This is sometimes referred to as the 15-minute adventuring day or the nova cycle.)

In 5^th Edition the cycle of daily resources has been disrupted somewhat through the short rest mechanic, allowing PCs to regain some of what would have previously been daily resources with just an hour of rest instead of eight hours of rest. This can alleviate some of the narrative oddity of the literal 15 minute adventuring day (because after a short rest, the PCs can continue accomplishing things in the same day), but structurally and in terms of mechanical balance, you’re still looking at many of the problems of the nova cycle. In fact, in some cases they can be worse, because it’s far easier to justify being able to catch your breath for an hour than it is to take a break for twenty-four hours.

ACTIVE DUNGEONS

So you’re the Dungeon Master. What can you do about all this?

Well, the first thing is to make sure that there’s a cost to leaving the dungeon to rest. The way to do that is by having the dungeon actively respond to the PCs.

An adversary roster can make it much easier for bad guys to actively respond to PCs. It also makes it very easy to redistribute them into new defensive positions if the PCs retreat and give them time to prepare for their return.

Other restocking procedures, which will refill previously cleared rooms with new adversaries, are also possible.

Beyond that, you really just need to think about what the NPCs would logically do in response to the PCs’ assault and then have them do that. There’s not really a big trick here: Your goal is to make it harder for the PCs to freely return to the dungeon, and that’s conveniently also what the bad guys are going to want.

Note: There are ninety umptillion exceptions to this, because there are any number of dungeon concepts which logically wouldn’t stage an active response to the PCs. That’s just fine. But you’ll probably want to design such dungeons with the nova cycle in mind (or provide some alternative explanation for why the PCs can’t freely rest, like a looming deadline).

RESTING IN A LAIR

A lair-type dungeon is one in which all of the inhabitants are part of the same organization or otherwise closely aligned with each other. (It might be the sewer lair of an organized crime family, a cavern complex swarming with goblins, or a fortified slavers’ compound.) If the PCs attempt a short rest within a lair-type dungeon, here’s a quick procedure you can use:

Is the compound on alert? Do the NPCs know that the PCs have infiltrated the dungeon or been killing them off? Then they’re probably actively looking for the threat. The PCs need to make a Stealth check opposed by the Wisdom (Perception) of the NPCs.

Give advantage and disadvantage for the check appropriately. (If they made some efforts to identify an out-of-the-way portion of the dungeon or disguise their presence, it’s probably a standard check. If they just closed the door to a random room, they should probably have disadvantage. Did they not even bother to close the door? The bad guys find them. You can’t succeed at hiding if you don’t even try.)

If the NPCs don’t find the PCs, they may assume that the PCs left. They’ll spend the rest of the hour calling for reinforcements, raising barricades, or getting ready to pack up and leave (depending on the situation).

If the compound is not on alert, make a random encounter check. 1 in 6 chance if the PCs closed the door, so to speak; 2 in 6 if they didn’t.

If an encounter is triggered and the PCs set a watch, let them make appropriate checks to detect the approaching encounter before they’re spotted. (They may still be able to salvage the short rest if they take clever/decisive action, or if they’ve taken appropriate actions to disguise their presence beforehand.)

If the PCs left evidence of their presence, make an additional random encounter check for each area where they left evidence. If an encounter is indicated, the NPCs have discovered the evidence and the alarm is raised during the short rest. The NPCs will start looking for the PCs, but they’ll make their Wisdom (Perception) checks with disadvantage (since they’re only searching part of the time).

If the PCs took efforts to hide the evidence or clean up after themselves, use an Intelligence (Stealth) check to see if the NPCs actually spot the evidence on an indicated encounter.

Note: Similar techniques can be used if the PCs fully retreat from a dungeon without raising the alarm to determine whether or not evidence of their trespass is discovered after they leave (which could result in pursuit and/or defensive preparations being made).

RESTING IN A MEGADUNGEON

A megadungeon, or any larger dungeon featuring multiple distinct factions, should be broken into zones, with each zone corresponding to territory controlled by one of the factions. These zones can be prepped ahead of time, but this is not strictly necessary: It’s usually pretty easy to eyeball what faction’s territory the PCs are currently in.

Each zone is simply resolved using the procedures for a lair-type dungeon, above.

No Man’s Land: In large dungeons like this, there may be abandoned sections which are not claimed by any faction. It’s also possible that the PCs might clear a zone (by wiping out the faction that lairs there). It’s much safer for the PCs to rest in such locations; even if the alarm is raised elsewhere, factions will generally focus on searching and securing their own areas of the dungeon, rather than venturing out into uninhabited regions. (An exception might be made if the PCs have really pissed somebody off and can be directly tracked to their new location.)

Make a single encounter check at a much reduced rate (1 in 20 if they’ve taken reasonably precautions; 2 in 20 if they haven’t), with a successful check indicating creature(s) from a nearby zone have unluckily entered the area and threatened the PCs’ rest.

It’s not necessary to make checks in such areas to see if evidence left by the PCs has been noticed, although you might make a similar check at a reduced rate (1 in 20) to see if a nearby faction has noticed that the a zone has been recently cleared. (This may also be indicated through general restocking procedures.)

LONG RESTS IN THE DUNGEON

If the PCs attempt a long rest in the dungeon, simply repeat the procedures listed above eight times (once per hour).

This does make it extremely likely that their rest will be interrupted, unless they’ve made a concerted effort to find a truly safe location AND taken precautions to avoid detection even if foes draw near. This is, of course, quite intentional: Dungeons are dangerous places, and if the PCs want to reap the rewards of pulling off a successful long rest in the middle of a delve, they’ll need to earn it.

Go to Part 2