Play Report (AD&D 2e): In Search of the Unknown (Session 7)

After a few weeks off, we finally had the opportunity to play the seventh session of our Dungeon Module B1: In Search of Unknown AD&D 2e game. You can read past play reports here.

The roster for this session was as follows:

• Llombaerth, CG elf thief 2 (Adam)
• Malkara, N human mage 1 (Maya)
• Rory the Small, LN human fighter 1 (David)

In our last session, the party spent some time in Timbershore licking their wounds. They regrouped, adding Robbernar the cleric and Malkara the mage to the team. They also recruited Walgretor, a half-elf porter and torchbearer, and Armstrong and Booker, a pair of human men-at-arms. Llombaerth took some time to gather information about Timbershore's criminal element so that the party might smuggle goods into town on their next return visit. They had an uneventful journey back to Quasqueton, felling some trees along the way to make an impromptu river crossing. 

They found that Grilk's goblins had been attacked by orcs in the time since they'd been gone. The goblins had killed a few orcs and turned the rest to flight, but one goblin, Tobb, had been captured in the ensuing chaos. The remaining goblins had erected barricades at strategic choke points, and had freed a "demon worm" and corralled it into the corridor leading to the orcs' territory. The party investigated the bedchamber in which the worm had previously been trapped, obtaining some books and a nice cloak, and found two chests bursting with treasure, which they concluded was actually magically-disguised rocks.

We picked up from there. The party debated whether they should do anything with the rocks/treasure, such as give it to the goblins as payment in advance for some future service, or have two of their hirelings lug around one of the chests in case it proved useful. Ultimately, they decided to leave it where it was.

They set out to explore more of the dungeon, heading down a long, twisting corridor straight across from the bedchamber. As they approached a T intersection, they heard movement up ahead. The bulk of the party retreated with the lantern, while Llombaerth snuck up ahead in darkness. He succeeded on his Move Silently check, peeked around the corner, and spotted a lone goblin, its head cocked to listen in the party's direction.

Llombaerth pressed himself up against the wall, lying in wait. The goblin began to approach. When it got close enough, Llombaerth leapt out and grabbed it, muffling its cries for help. He dragged his captive back to the rest of the group.

After some questioning, the party learned that the goblin was Pigmug, of the Stonespeaker Clan. Once a member of Grilk's gang, Pigmug had since opened his ears to the Stone That Speaks, and to Drak, the powerful orc leader through which the Stone speaks to its devotees. The party asked Pigmug how many orcs and goblins were part of the Clan, but Pigmug was not good at counting. They tried to convince Pigmug that they were Stonespeakers as well, and that anyone can speak to the Stone if they're brave and strong enough. Pigmug caught this as a lie, and the questioning deteriorated. They knocked out Pigmug and decided to drag him back to Grilk's gang.

Grilk thanked them for bringing Pigmug back into the fold, and the party charged Grilk with deprogramming the goblin and learning what he knew of Clan Stonespeaker. They returned to explore the rest of the corridor.

They went left at the T intersection and found two doors. They listened at the first one and heard orcish voices in idle conversation on the other side. They heard nothing from the other door, so they tried opening it, but found it stuck. Forcing it would make a lot of noise and might attract the orcs, so they left it to explore the other branch of the intersection.

That led to another T intersection with an alcove at the end. To the right, they spotted one of Grilk's goblins, manning the barricade. They concluded correctly that the alcove at the end of the corridor must be the other side of one of the one way secret doors connecting to the entrance hall and spent some time learning how to open it. They returned to the T intersection and this time explored the left corridor.

The corridor went more or less straight a long way, then turned left. They followed it and heard a thumping sound at the end of the hall, beyond the light of their lantern, as if something big was throwing itself at the wall or a door. Llombearth once again snuck off ahead towards the darkness.

Though he passed his Move Silently check again, the thumping suddenly stopped. There was something between a hiss and screech, and suddenly something big was barreling down the corridor towards them. Most of the party was surprised as a horse-sized worm or maggot came wriggling into the light of their lantern, a mass of tentacles writhing from its armored head.

Bronson immediately lost morale and meant to flee, wrestling with Walgretor for control of the lantern. Rory leapt into action, putting himself between Llombaerth and the worm, hacking with his halberd. In the madness, his attack missed completely, his weapon clattering off the nearby wall. 

Armstrong loosed a pair of arrows at the worm, heedless of the possibility that he might strike Rory by mistake. One arrow sailed just high of the thrashing beast, while another plinked harmlessly off of its armored head.

On the following round, Malkara threw a trio of darts. One bounced off Rory's armor, a second off the worm's head, and the third found meaty purchase in the worm's face, dealing maximum damage (which for a dart is 2). 

Llombaerth scrambled back towards the rest of the group, lobbing a flask of alchemist's fire as he fled. He hit an AC of 2...but hit Rory by mistake! Rory burst into flames, but Llombaerth rolled low on the damage (just 3 on 2d6).

Bronson wrestled the lantern away from Walgretor and fled back down the corridor, leaving the party in darkness, were it not for the fact that Rory himself was alight.

Armstrong loosed another pair of arrows, one bouncing off of Rory's armor while another sailed past the worm.

I considered whether the worm would press its attack against Rory, considering he was on fire, until I realized that this type of worm has "Special" morale. According to the Monster Manual, this creature "will continue to attack as long as any of their opponents are unparalyzed." As the players were about to learn, the dreaded "demon worm" of Quasqueton was none other than a carrion crawler!

The carrion crawler pressed the attack, lashing at Rory with its eight tentacles. With THAC0 17, against Rory's AC of 4, it needed a 13 or better to hit. It rolled a 6, a 9, a 10, a 12...and a 20! Fortunately, the carrion crawler's tentacles don't actually do any damage, so the critical hit would have no additional effect, but Rory needed to make a saving throw versus paralyzation or else be paralyzed for 2d6 turns. He needed a 14 or better to succeed. He rolled a 19!

The carrion crawler attacked three more times, rolling a 5, another 5, and a 2! Wow! Needing just a 13 or better to hit, each of the carrion crawler's attacks had a 40% to hit. With eight attacks, it was statistically likely to hit at least three times. The probability of it missing seven times was 2.8%!

Rory, still on fire, stood his ground and pressed the attack. As a specialist with the halberd, he could attack twice this round. His first hack missed wildly, while the second bounced off the rubbery hide of the crawler's body - softer than its armored head, but still tougher than leather.

On the following round, Rory fell back to the rest of the group. Malkara waited for him to be clear of the worm, then loosed more darts. Two more bounced off the armored head, while the third found flesh and scored another 2 damage. Rory stopped, dropped, and rolled on the floor. Walgretor, who was carrying a bucket of orc soup from Grilk's lair, used it to douse the flames. The smell was unbearable, but Rory was no longer on fire. The corridor was plunged into darkness.

Luckily, Llombaerth could see in the dark. He threw another flask of alchemist's fire down the corridor between the party and the crawler. The crawler, heedless of the flames, charged through them to get at its prey. It shrieked as the flames engulfed it, its body twisting as its fatty segmented body bubbled and burst, painting the walls of the corridor with its steaming innards. It crashed to the floor in front of the party, its legs twitching in its death throes.

The crawler had 9 hit points left, and the burning alchemist's fire did 10 damage on a roll of 2d6.

Bronson was long gone, so the party lit another torch and collected themselves. They waited for the flames to burn out, then proceeded down the hall, coming to a pair of door on the left. At the first door they heard nothing. At the second, which had been battered and scratched by the crawler's attempts at entry, a goblin's voice on the other side demanded to know who they were.

The party identified themselves as the slayers of the demon worm, and again claimed to be Stonespeakers. They told the goblin that the Stone demanded it open the door to them. The goblin stalled for time, and the party could hear it dragging furniture to barricade the door. They managed to convince the goblin that the Stone was displeased with it, upon which it opened the door for them.

They found themselves in another bedchamber, this one with a canopy bed, a nightstand, chest of drawers, a full-length mirror, a sculpted basin, and a tapestry on the wall. The goblin was Numbnuts, and he was here searching for any missed treasure which he might offer to Drak in order to gain favor. The party convinced Numbnuts that they knew where a lot of treasure was to be found, but they would only tell him in exchange for information.

Again they asked how many orcs and goblins there were, but Numbnuts was also not good at counting. Slug was the goblin who could count. Lumpknuckles was good at listening. Numbnuts was good at drawing. The party demanded that Numbnuts draw them a map of the Stonespeaker Clan's territory, and Numbnuts obliged.

Numbnuts drew a big rectangle with corridors bordering the top, bottom, and left side. He drew a small rectangle within the larger one in the bottom right corner, which was the room they were in now. To the left of it, across an adjoining hall, was a "junk room" where Lumpknuckles would be. A winding hallway, patrolled by Slug, meandered throughout the larger rectangle. Further to the left of the junk room was the "captain's room". Brubar the orc liked to hide in there and get drunk. 

To the north of that were two more rectangles side by side. The left rectangle was the "stairs room", where Muv the orc stood guard against incursions from the orcs who worshipped the Stone Below, a splinter faction of Drak's orcs who were now enemies of the Stonespeakers. The right rectangle was the "Stone's chamber", where Drak went to speak to the stone and to feed it the essence of heretics. Just above those two was another rectangle, the "treasure room", guarded by the orc Rheev. 

In the upper right corner of the large rectangle was a big area where Drak sat on his throne, guarded by his strongest warriors. The door that party had listened at earlier (where they heard orcish voices on the other side) led directly into the throne room.

With that, we called it for the night. The party has defeated one of Quasqueton's most dangerous monsters and has begun making small forays into hostile orc/goblin territory. I'm looking forward to seeing how it shakes out.

Developing the Sandbox: Dungeons

This post is part of my series on developing the sandbox. You can read other posts in the series here. The series will, when necessary, go into detail on the development of my sandbox for Dungeon Module B1: In Search of the Unknown, but will also discuss sandbox development more generally. You can read play reports from my B1 campaign here.

I'll be honest - these days, I don't actually much enjoy making dungeons. You have to figure out what the dungeon used to be, who built it, what happened to it, who's been using it since, who's using it now, how they get along with their neighbors, where they live, (sometimes) what they eat and drink, what they're doing there, what stuff they have, what stuff is in the dungeon that they don't have, why they haven't gotten it yet, why are there traps, what the traps do and how they work, what a "special" room is, why whoever built the place would put it there, what the layout is, what the rooms used to be and what mundane stuff would be in there to play with, what are the doors like, which ones are secret, which ones are locked, how do the inhabitants get around, how many levels are there and how many rooms are there per level, where are the stairs, what's the percentage chance of encountering such-and-such while wandering around the place - the list goes on and on.

To compound all of this, there is so much advice to be found on the subject, whether it's in rulebooks throughout the many editions of D&D, other TTRPG systems, or online on blogs. It is really overwhelming to take it all in, decide what you like and what you don't, synthesize it all into a coherent methodology, and apply it when designing dungeons for your home game.

It's a lot of level design. There was a time when I really enjoyed that part of the game a lot, but lately, making dungeons has made me very frustrated more often than not. And then you have to write that all up in a key that's comprehensive while also being simple enough to run at the table. 

You have to put in similar work when designing a a stronghold or monster lair, but there's plenty of information about real world strongholds to draw upon, and monster lairs are often smaller, simpler, and easier to imagine. There are dungeons in literature, but few that actually emulate the D&D dungeon-ass dungeon. You're better off just reading old TSR modules to figure out what a dungeon should more or less be like.

I'm more of a big picture guy. I like making big maps with varied terrain, then finding out what all is there, whether it be cities, strongholds, dungeons, or monster lairs. Who lives there? Who rules it? What kind of monsters plague this or that region? How do they all relate? For me, that's where the juice is.

That's why I've basically just started stealing my dungeons. I originally decided to create my B1 sandbox because (1) I thought the dungeon-focused adventure was crying out to be placed within a larger context and (2) I really liked the idea of one of the more tedious parts of the sandbox development process (designing the dungeon) being done for me.

B1, however, does not feature a finished dungeon. The DM is instructed to roll to determine if each of Quasqueton's rooms contains monsters, treasure, or neither. This is relatively easy, however, since the module provides lists of monsters and treasure which the DM can place in each room when a positive result is determined. All that's left is to use the module's thorough room descriptions to figure out where the treasure is stashed, to determine how the monsters relate to one another, and create new wandering monster tables based on the monsters found in my version of Quasqueton.

I used some loosely interpreted opposed reaction rolls to determine monster relationships, eyeballing the map to find out who lives near who (combining groups into factions when appropriate) and who controls what territory. 

What I ended up with was some goblins who control the dungeon's entrance, some orcs deeper within who threaten the goblins and have lured away some of their numbers, and a few smaller groups of intelligent creatures and roaming non-intelligent monsters which are indifferent to the conflict. This has worked out quite well so far, with the players deciding to try to stay on good terms with the gatekeeping goblins, the goblins asking the characters for exchanges of favors, and the orcs launching attacks against the goblins when the party is away from the dungeon.

The DM is also instructed to roll to determine when doors are stuck, which requires more eyeballing to determine who can get into what areas, whose territory they'll need to pass through, and how that affects the dungeon's politics and who is aware of whom. For example, if the bandits must pass through goblin territory to get to the area of the dungeon in which they've been stocked, then the goblins must know about the bandits and have some opinion about them, and there must be some reason why they haven't killed each other.

For the wandering monster tables, I add up all of the monsters in the dungeon to get a total number of monsters. Then, for each monster type, I divide that monster type's total by the larger total to get a percentage. That is the percentage chance of encountering that type of monster wandering the dungeon.

Let's say I have a 20 room dungeon, and there are monster in five rooms: 11 brigands, 6 hobgoblins, 4 gnolls, 2 rot grubs, and 3 troglodytes. I don't think of the rot grubs as "wandering", so excluding those, I have 24 total monsters - 25% hobgoblins, 46% brigands, 17% gnolls, and 12% troglodytes. My wandering monster table would look like this:

d100    Monster
01-46   1d6 Brigands (from Area 8, looking to threaten, plunder, and rob)
47-71   1d3 Hobgoblins (from Area 17, killing any they outmatch)
72-88   1d3 Gnolls (from Area 11, willing to trade information for treasure)
89-100 1d2 Troglodytes (from Area 20, looking for sacrifices for the God of Stink)

(I'm assuming this is a 1st level dungeon, and the number appearing is based on the assumption that there will be 4 to 6 characters of that level, using my cataloguing of AD&D 2e monsters by level to determine roughly how many of each monster would be appropriate. You'll want to make it work for your own assumptions about party size and whatever system you're using.)

However, adapting B1 for my campaign wasn't as simple as stocking the dungeon, making a wandering monster table, and then using the module as written. Like, this isn't really usable:

There's an art to turning this into something that functions at the table. You can't read it as is. You could maybe skim it and put it into your own words on the fly, but you're wasting everyone's time. You could memorize it somehow, but good luck with that. My process has been to rewrite it in my own style, in a way where the interactable elements and the hierarchy of information is plain to see. In my style, this room looks like this:

Descriptive, but not verbose. Concise, but comprehensive. Important stuff is bolded. Bolded items are described in further detail. Additional information is available upon further inspection. I can eyeball the initial description and rephrase it in my own words, then go into further detail when the players spend time examining any one thing. This works for me. Yours might look different.

This is all very specific to my B1 sandbox, or any sandbox in which your dungeons are repurposed from existing published material. That isn't always the case. Though I don't love doing it these days, when I choose to do so (and I'm not wholly opposed to it - my sandboxes for both B2 and B3 include dungeons beyond those detailed in the published modules, after all), how do I design my own dungeons from scratch?

As with other location types, I'll start with a minimalist sketch. In the case of dungeons, this includes the type of structure it once was, who built it, how it fell into ruin, and (broadly) who or what dwells there now. 

In the AD&D 1e DMG, the ruin types are limited to deserted strongholds (with "mazelike dungeons thereunder" and the occasional monster using the surface ruins as a lair), villages (here I'd include the other classes of small settlement like hamlets and thorps as well), cities (or towns), tombs, and shrines. Tombs and shrines are classic dungeon fare. Deserted strongholds are good fodder as well, especially when considering B1 and how it depicts the maddening, hostile subterranean environments which bored, retired adventurers are likely to carve out beneath their seats of power.

Villages and cities are a bit trickier. They're not quite "dungeons" in the traditional sense. I tend to build ruined cities and towns as I would normally build a city or town, then wreck them. The characters can still crawl through the settlement's ruins district by district just as they would crawl through such a place were it not ruined. Perhaps it is now overrun by the doppelgängers, fiends, lycanthropes, and undead which have always prowled its streets. Navigating the settlement itself might be more akin to a point crawl. It's the locations which were once important to the settlement that are the proper dungeons. 

Using my B1 sandbox as an example, if Timbershore were to fall into ruin in the far away future, perhaps the remains of the Evil wizard Lambrecht's castle is one such dungeon, and the ruins of the Evil cleric Blanchefleur's keep another. Add in some crumbling manors or a merchant consortium's sealed treasure vault for variety. A ruined village, hamlet, or thorp could be much the same, albeit smaller and easier to get around. Instead of ruined castles and manors, a dungeon there might be an old mill or granary - a diversionary location with just a handful of rooms.

As for who built the structure originally, it depends. Strongholds are easy for me, since AD&D is quite explicit as to who can build strongholds and when, and the DMG has nifty (if not, at times, wholly consistent with its own rules) tables for determining who a stronghold's ruler is (or in this case, was). 

Shrines are built by clerics, so it's easy to determine who the cleric in question was - just determine their race and alignment and either pick some appropriate classic D&D god and genericize it for your setting or choose some god unique to your setting which you wish to highlight. If there were to be a ruined shrine in my B1 sandbox, for example, I would make it a shrine to the Horned Mother (Blanchefleur's deity), or I would come up with some other deity in opposition to her, its worshipers long since eradicated. I tend to favor either Evil shrines or Good shrines which have been desecrated by Evil powers. Remember that Evil undead are empowered in such places!

The ruins of cities/towns, tombs, and villages/hamlets/thorps are a bit trickier. I quite like the 2014 D&D 5e DMG's tables, which you can roll on to determine who built the dungeon, what type of dungeon it is, and why it was abandoned - so much so that this framework strongly informed my own ideas about creating the initial sketch of such a location. Still, who and what the tables include can be a bit of headscratcher. A dungeon might have been built by dwarves and elves, but not by gnomes or halflings. It might have been built by hobgoblins, but not by orcs. Beholders, giants, mind flayers, and yuan-ti, but not any other monsters.

Instead, it's probably best to look at what's already been established about your setting, and pick or randomly determine some appropriate creator based on that. My B1 sandbox is located in a heavily forested region, so perhaps a civilization of elves once lived there in the ancient past. Rogahn and Zelligar used orc slaves to build Quasqueton, and went to war with barbarians in the region - perhaps the orcs or barbarians had a civilization of their own and built structures which fell into ruin as the two powerful adventurers asserted their dominion over the place. There is an ogre lair in the area - perhaps their ancestors built things as well. Perhaps Timbershore once had outlying villages which were reclaimed by the forest, or perhaps it was once subservient to a larger city which met a similar fate.

The same approach works for dungeon history. Barbarians, ogres, and orcs may have brought ruin to human or elven civilizations and structures, and may have likewise been ruined by the campaigns of Rogahn and Zelligar. Perhaps there was a conflict between worshipers of the Horned Mother and whatever deity's worshipers had a foothold in the region before them. In addition to conflict from war and conquest, you could work in plagues, curses, and magical catastrophes.

All of the preceding information - the type of ruin, who built it, and what happened to it - can inform what lives in the ruin in the present day. A fighter's stronghold which was ruined when a subdued dragon rebelled might still be home to that dragon. A wizard's stronghold might house summoned guardians or magically-altered experiment creatures. Evil and desecrated shrines and tombs will likely have undead or fiends. An enclave of barbarians, elves, ogres, or orcs might still call the ruined city home. A village abandoned due to plague will be home to disease carriers. A curse upon a city may have turned its aristocracy into wererats or vampires.

I try to pick a few different monsters types - intelligent, organized monster groups with lairs and factions inside the dungeon, unintelligent scavengers and predators which roam its rooms and halls, and singular, powerful monsters with or without minions. In AD&D terms, if I'm making a 1st level dungeon, I'll want it to have a mix of 1st, 2nd, and 3rd level monsters, according to the DUNGEON RANDOM MONSTER LEVEL DETERMINATION MATRIX:

Likewise, a 16th level dungeon might include a mix of monsters from levels 1 through 10.

I might go so far as to use reaction rolls at this stage to determine who is friendly, hostile, or neutral towards the party, and which monsters are allied or at odds with one another. The Dungeon Checklist is a great resource - make sure there is fodder for the characters to kill, thinking creatures to parley with, and dangerous creatures to be avoided or carefully neutralized.

This all will give you a pretty good sketch, but dungeons aren't the kind of environment you can easily riff on using minimal information. You need to know specifics. How well-lit is it? How high are the ceilings? How many rooms are there? How many levels? How are the rooms laid out? Are they connected by passages or doors? Are the doors locked or stuck? Are they wood, stone, or metal? What's in each room? How many monsters are there? Where are the tricks and traps? How do they work? What damage do they do? What saving throw is needed? Where is the treasure hidden? How much of it is there?

As I've said, there are so many ways to determine these things that it's nearly impossible to suggest best practices. It will differ for everyone based on their taste, system of choice, and campaign setting.

One of the best tangible tips I can offer is to design dungeons in chunks. This somewhat solves the question of how many rooms and how much of each thing (empty rooms, monsters, tricks/traps, and treasure) to include. I use 20-room chunks along with the AD&D 1e dungeon stocking table:

Thus, if I have a 20-room dungeon, there are 12 empty rooms, two with monsters only, three with monsters with treasure (these are my monster "lairs" in the dungeon), one "special" room, one trick or trap room, and one treasure room (like a vault behind an especially secure or secret door). This to me is a small dungeon (boo, hiss, five-room dungeons stink). If I want a bigger one, I add 20 more rooms. If I need more, I add another 20, and so on. 

Since each 20-room chunk contains a "special" room which might also contain a stairway, I tend to include one stairway per two 20-room chunks. A level can have multiple stairways up or down, of course, so this isn't necessarily a limitation upon how large a single level can be.

If I want to create a big 100-room dungeon, I know I need 60 empty rooms, 10 rooms with just monsters, 15 monsters with lairs and treasure, 5 trapped rooms, 5 treasure rooms, and so on. 

You might use a different dungeon stocking table to inform the size of your dungeon "chunks". For example, from OSE:

Your "chunk" might be 36 rooms: 12 empty (2 with hidden treasure), 12 with monsters (6 with treasure), 6 special, and 6 trapped (2 with treasure). Once again, adjust to taste.

Empty rooms are not truly empty, but contain environmental storytelling, hints at what awaits in other rooms, time-wasting set dressing, or interactive elements which might come into play in the future, such as if a wandering monster shows up there, or if the room is restocked with a monster on a future delve. 

I like using Courtney Campbell's resource on empty rooms to determine what a room's original purpose was, and from there determine what sort of set dressing to include. I do this with rooms that are otherwise occupied as well. The 1d6+6 orcs are not just standing around in some barren chamber, but are instead in an old kitchen scavenging for food, in the wizard's bedchamber scraping the gold leaf from his headboard, or in the lounge gawking at the marble statue of a nude woman.

The specific monsters I use could be determined by rolling on AD&D's dungeon encounter tables, or I might pick a few based on the dungeon's history. Tricks, traps, and treasure would be similarly informed. An elven ruin might have more magical traps, while a dwarven one might have more traps involving shifting masonry. Gnomes use clockwork devices, kobolds use poison and pots of stinging insects, and defiled shrines are warded by curses and negative energy. Elves treasure ornately carved wooden objects, spellbooks, moonstones, and magical bows and swords, while dwarves treasure stone carvings, gemstones (but not pearls, because dwarves HATE pearls), and magical axes and hammers.

The number of monsters appearing, gameplay statistics of traps, and value of treasures will all be informed by your system of choice, of course.

As for layout, you can create your own dungeon and still steal a little bit. You could take an existing map from one module or another and simply restock it to your liking. This is useful if, for example, you know you want to run the hideout for a cult of Elemental Evil, a fire giant stronghold, or a Tomb of Horrors-type thing, and know where to find one. I ran Tomb of Annihilation's Fane of the Night Serpent, Princes of the Apocalypse's Sacred Stone Monastery, and Storm King Thunder's Forge of the Fire Giants this way, using the maps and room descriptions but changing the monsters, treasure, and mechanics of traps and the like to suit my needs, and it worked quite well.

Alternatively, you can steal a more generic map from somewhere online. I've used a lot of Dyson Logos maps in my games. I often find this unsatisfying though, as the types of dungeons I can make are limited by the maps that are available. I have to either compromise my vision to suit a map that wasn't created with that vision in mind, or find a map I like first, then come up with the vision after. Neither is ideal.

So, I often create something from scratch. I don't usually bother with detailed maps these days. I've tried using AD&D's Appendix A to randomly generate dungeons. It's really confusing and it creates a map that kind of stinks. I feel the same way about random dungeon generators found online. I've basically given up on it.

So, if a relativistic flowchart-stye map is good enough for the players, it's good enough for me! I start with a bubble in the center of the page (with stairs leading up to the surface), then roll 1d4 to determine the number of exits. For each room branching off of that, I roll d4 again to determine the total number of exits (that is, if I roll a 1, there is only one way into the room, and no other ways out). I do this until I've exhausted my 20-room chunk or end up with only dead ends (to which I might add secret doors to additional rooms to fill out any remaining rooms from my 20-room chunk). I can add additional notes as needed, like "This room is really big and takes two turns to map," or "This hallway is really long and requires a wandering monster check when passing through it."

I tend to put important rooms like monster lairs (those with treasure), "special" rooms, and treasure rooms in those areas farthest from the entrance, so that the characters will have to endure the longest route through the dungeon to reach those rooms. Rooms with only monsters or only traps will be those at important intersections or choke points, ensuring that the characters either go through them or find a long way around them.

After that, it's a matter of writing the key, which I've demonstrated in my rewriting of the room from B1, above. I like to find a middle ground between writing something descriptive but also usable at the table. Highlight interactable elements. Try to anticipate what the players will want to investigate and provide additional detail as needed. Try to think of all the information you might need. Think of what questions the players might ask, and come up with answers. Include game mechanics whenever possible to avoid looking stuff up during the session. Write it and reread it and refine it.

As you run the dungeon, it will change. Make note of which monsters have been killed, which have had changes in attitude towards the PCs, which traps have been disarmed, which treasures found, which doors unlocked, and the like. Once the players make contact with the monsters, feel free to move them around. They should be doing something, not just hanging out in one room all the time.

When the characters leave the dungeon, I like to change things. Monsters which don't like each other may come into conflict, killing or capturing one another. Those that do like each other may form alliances. Their territory may expand as the characters open doors, disarm traps, and kill other monsters. They may recover treasure themselves, find secret doors of which they were previously unaware, and update their own dungeon maps with scouting missions.

This is also when I restock the dungeon. I go through each room the characters have already explored, checking again for monsters, traps, and treasure. Perhaps a new monster has wandered into the dungeon, or wandered up or down from some other level. Perhaps a trap has been reset, or a new trap placed where there was none before. Perhaps the monsters have moved their treasure to a previously unoccupied room for safe keeping. 

This can be a boon or a bane to the players. Perhaps the friendly goblins have thwarted a counterattack from the orcs, and now the orcs are lesser in number. Perhaps the orcs triumphed and carried off the goblins, and now the characters have lost a valuable ally. The rival adventurers may have perished in their own quest, or maybe they succeeded and carried off some valuable treasure, or simply made progress and disarmed some trap. Discard any results that don't make sense, or stretch your imagination until they do. No two delves should be exactly alike.

As I've explained, this is much too large a topic to possible cover in its entirety in a single post. Hopefully, I've provided a somewhat comprehensive explanation of my approach to getting started and to doing the long, somewhat tedious work of preparing a location like this for gameplay. Everyone's process will be different. This is a only a peek into mine.

Developing the Sandbox: Outdoor Encounter Tables

This post is part of my series on developing the sandbox. You can read other posts in the series here. The series will, when necessary, go into detail on the development of my sandbox for Dungeon Module B1: In Search of the Unknown, but will also discuss sandbox development more generally. You can read play reports from my B1 campaign here.

First, a preamble. The types of encounters I'm referring to in this post are those procedurally generated encounters which occur when the party is traveling overland. These are often colloquially referred to broadly as "random encounters" - which is not specific enough for my purposes - or as "wilderness encounters" - which suggests the party is navigating an untamed, dangerous outdoor environment, and which is too specific for my purposes. Here I will use the term "outdoor encounters" because they are neither city/town encounters - which I describe in my previous post on developing the town - nor are they necessarily wilderness encounters. These encounters can occur anywhere outdoors, and may take place in either wilderness or inhabited areas.

You can create outdoor encounter tables for your sandbox at various points throughout the development process. In many traditional "old school" campaigns, the dungeon which is the initial locus of play is situated very near to the "home base", so it is unlikely that the characters will have many outdoor encounters when traveling to and from that destination.

Indeed, it is even suggested in the AD&D DMG that the local municipality actually maintains a road leading to the dungeon. In that case, your initial outdoor encounter table probably needs little more than bands of fellow/rival adventurers, patrols from the nearby settlement, or groups of bandits and brigands which might prey upon such travelers.

However, Dungeon Module B1 relies upon the premise that the dungeon, Quasqueton, is remote, its location lost to history and local knowledge. Few if any know of its whereabouts, which serves as justification as to why it hasn't yet been plundered. It wouldn't make sense for it to be right next to town, and certainly not for a road to lead right to it. The player characters have the fortune of coming into possession of a map pinpointing the dungeon's location, but that doesn't mean the trek should be easy.

This meshes well with my own preference for remote dungeons and low level wilderness exploration. In older editions of D&D, the wilderness is a wildly variable and dangerous place, unsuited for prolonged expeditions by low level characters or those traveling without formidable armies. That doesn't mean low level characters can't adventure there. In fact, I think they should! They just have to be careful. They might encounter an orc warband, see a giant prowling on a far away mountainside, or spot a dragon flying overhead.

These encounters not only inject some variety and unpredictability into your game, but present the players with dangerous situations they will have to carefully navigate, reminding them that the world is not carefully curated to accommodate their characters - that they are small fish in a big pond and will need to act accordingly until they acquire some power and have more agency.

In the first session of my last AD&D campaign, the party had seven random encounters before finally arriving at the first dungeon. Many of them were potentially deadly, but a combination of mitigating factors like distance and reaction rolls plus the players' own savvy and my wise and benevolent restraint allowed them to reach their destination unscathed. Spotting a lone hill giant wandering in the distance filled them with fear. A dozen or so sessions later, they encountered another such wandering giant and laid a trap for it, slaying it with ease. The characters' growth in power over that span of time was self evident.

In my sandbox for B1, Quasqueton is located about a day and half's travel from town over settled farmland, across a river, through the forest, and atop a high, wooded black crag. It's not a super deadly trek, but it provides a decent chance that a 1st level party on their initial expedition may get wrecked by a band of hobgoblins or worse. That may not be to your or your players' taste, but I find the possibility tantalizing.

Since reaching the obvious first adventure site for the campaign would require some overland travel (both through inhabited lands and the wilderness), creating my outdoor encounter tables would need to be a priority. How would I go about it?

There are more ways to write outdoor encounter tables than I could possibly summarize here. Here are some of my favorite blog posts on the subject. If you want more, Google "D&D encounter tables blog" and go on a voyage of your own making. It's not my job to educate you.

The primary conflict when deciding how to structure my outdoor encounter tables was whether to draw my encounters from a preexisting, generic, all-encompassing table (such as those found in the AD&D DMG) or to populate my tables only with those creatures that would reasonably be found outdoors in my sandbox, based on the settlements, strongholds, and monster lairs which dotted the landscape.

For example, the cultivated farmlands and wild grasslands surrounding Timbershore would be populated mostly by farmers, patrols, merchants, and the like. Brigands would prowl the lands surrounding Fayette's Hold. The forest would be home to ogres, owlbears, giant eagles flying overhead, and the occasional chance encounter with the leprechaun, Jinglepuff.

While this makes plenty of sense, it's also rather dull. It's not super exciting if the only people found in brigand territory are brigands. It forces me to think about what sorts of mundane animals are found in which environments, and to think of which factions have overlapping territory or which might venture into one another's spheres of influence. Most importantly, it eliminates any possibility of being surprised myself.

Using a more generic outdoor encounter table is certainly good for producing results which surprise even the DM: "Huh, I didn't expect there to be a green dragon living in the woods. That's exciting. What does it portend for my campaign?" But it also produces incongruous results: "Uh, so why is there a sphinx hanging out a couple of miles away from the brigands' castle?"

The solution I eventually came up with was to use two encounter tables - a generic one which I used almost like a spark table (in this case, the outdoor encounter tables from the AD&D DMG) to then populate my own curated encounter tables for the region. Basically, I would roll first on the AD&D encounter tables, try to come up with some rationale for the result, add it to my curated encounter table if it made sense, and discard it if it did not.

This was easy enough with many of the most common encounters - merchants in the grasslands were probably traveling to town, pilgrim encounters were with devotees of the Evil cleric Blanchefleur, brigands were from Fayette's Hold, orcs were from Quasqueton, and the like. 

But I also ended up with some unexpected results, like the sphinx just outside of Fayette's Hold (which I discarded, but now think I could have actually used - it would be cool if the brigands had made an alliance with a Chaotic Evil hieracosphinx), or some truly strange things hanging out around Quasqueton (one of which turned my preconceived notions about one of the region's important NPCs on its head, which I won't spoil here).

My sandbox wound up needing seven encounter tables: inhabited/patrolled forest (the hexes bordering Timbershore, Ebongrove, and Fayette's Hold), uninhabited forest, inhabited/patrolled forested hills (Zerelda's Camp), uninhabited forested hills (Quasqueton), inhabited/patrolled barren hills (Fayette's Hold), inhabited/patrolled grasslands (which was all of the grassland on the map), and rivers and lakes.

For each table, I generated four results. If the result was a creature which could be encountered only during the day or night, I generated another "mirror" result for the opposite time of day. To keep things interesting, I rerolled duplicates, so "brigands" was only ever one of the entries on the table. At most, the table for any given terrain type would have eight entries, so I would have to generate a maximum of 56 encounters. I ended up with 39.

Since this is a lot of encounters, I tried to keep them light, but I also wanted to have most of the work done for me ahead of time so that I wouldn't have to roll much at the table. If the encounter is with people or creatures who can speak, what is their leader's name? How do they react to the party, why, and what do they do or say? If their reaction is indifferent, how can I make that into an encounter which is still engaging? What treasure are they carrying? Where are they from? Where are they going? What are they doing? If the encounter includes spellcasters, what spells do they have prepared?

Some encounters are easier than others. Animals and unintelligent monsters are easy. Mounted patrols with mixed arms and multiple character-type NPCs among their number are more challenging. But it's worth it to put in the time to prepare them properly.

Throughout the process, I tried to keep in mind a few things: logic, but also fun, and variety. There's no proper orc lair in the region, so it didn't make much sense to have a band of 100 orcs wandering the forest. However, a group of a dozen or so orcs venturing into the region from just off the map would suit my purposes. There would be no bands of roaming giants, but a single wanderer passing through the region seemed within the realm of possibility.

While I do enjoy the idea of the wilderness being an unpredictable and dangerous place with huge groups of monsters prowling about, I also had a mind for balance - again, because it makes more sense to me. It seems appropriate that the party would be more likely to encounter a handful of brigands rather than a giant army of them. 

Thus, I loosely "balanced" these encounters based on a d100 roll: a result of 01 to 65 would be an encounter for 1st to 3rd level characters (1d3), a result of 66 to 85 would be for 4th to 7th level characters (1d4+3), 86 to 95 would be for 7th to 12th level characters (1d6+6), and 96 to 100 would be for characters 9th to 20th level (1d12+8). Most encounters will be with small groups, but there is the chance of running into a large force which must be evaded or negotiated with. 

This roll was made only after I had determined the general type of creature encountered, so if I rolled up a dragon and could make sense of it (easy to do, since dragons fly all over the place, whether they live there or not), then there would be a dragon, balance be damned. The roll only determined whether it was a singular dragon or a gang of teenagers.

As we play out this campaign, I replace those encounters I've already rolled as makes sense. For example, the hobgoblins encountered in the forest during our first session were merely frightened away, not defeated, so they are still out there. The giant tick which was slain is gone from the encounter table until I roll up another giant tick to replace it. There are certainly more giant ticks in the woods, so I could simply reuse this encounter, but again, two of my priorities when creating these tables are fun and variety, and encountering a giant tick 25% of the time in the forest is neither.

Hopefully, as the campaign continues, the party will have more outdoor encounters, and I'll be able to showcase some of the work I've put into them in my forthcoming play reports.

Play Report (AD&D 2e): In Search of the Unknown (Session 6)

We played the sixth session of our Dungeon Module B1: In Search of Unknown AD&D 2e game last night. You can read past play reports here.

The roster for this session was as follows:

• Llombaerth, CG elf thief 2 (Adam)
• Malkara, N human mage 1 (Maya)
• Robbernar, CG elf cleric 1 (Nael)
• Rory the Small, LN human fighter 1 (David)

As you can see, we had a new player joining us. Maya is playing Malkara, a human mage who, like the rest of the party, has an exceedingly average spread of ability scores. She did manage to roll a decent selection of 1st level spells: comprehend languages, magic missile, and spider climb. She should be an adequate replacement for the recently deceased Pommernar.

Last time, the party returned to town from the dungeon. They encountered another giant tick while camping in the wilderness, and it slayed Pommernar. Back in town, they were taxed heavily for their recovered treasure, but earned enough gold for Llombaerth to reach 2nd level. The party spent some time brainstorming the best way to use the few days they would need to rest and recuperate in civilization. They met Robbernar, Pommernar's father and a cleric who agreed to adventure with them.

After some number crunching, I told the players why it might be unwise to buy a boat. I reasoned that whether they bought the boat or not, they would still have to pay the crew. And if they owned the boat, they would be responsible for all its upkeep expenses. The main benefit of owning a boat would be that the party would not need to pay the regular fee for passage - 66 silver per head for a trip downriver to the dungeon and back. But the boat itself costs 500 gold, so to break even the players would need to go back and forth to the dungeon about 25 times (assuming an average party size of three - with one hireling and a donkey it's closer to 15 times, which is still pretty excessive). 

So, they abandoned that notion for now. They set off for the River Ward nonetheless, as they wished to recruit more hirelings and to learn more about smuggler activity. Robbernar managed to hire a half-elf porter/torchbearer named Walgretor as well as a pair of human men-at-arms named Armstrong and Booker. Rory also decided to retain their original porter and torchbearer, Bronson. Llombaerth got a lead on a seedy tavern named the Drunken Pike, where the ward's criminal element liked to congregate.

At the Drunken Pike, Llombaerth met Malric Two Fingers, a grizzled old sailor with just one hand, and that hand with just two fingers. Llombaerth initially offended him by making thief signs with his hands - flaunting all his fingers in Malric's face - but eventually (after some bribery) the two got on the same page. Llombaerth asked if Malric knew how someone might bring goods into town unnoticed by the constabulary.

Malric told him about the Rusted Anchor, an old barge tethered to the wharf but long out of commission. Malric recounted how, back in the day, sailors would board the barge and give a special knock at the door to indicate that they were ready for work. He had heard that sometimes, out of respect for the old ship, people would board it at night and give the same knock - since Malric did not have the proper number of fingers, he had his bar boy demonstrate. Llombaerth took note. Though the barge was out of service now, Malric had heard that there are somehow still sailors unloading its cargo to this day.

With that, the party regrouped and set off into the wilderness. They were nine now - Llombaerth, Malkara, Robbernar, Rory, the men-at-arms Armstrong and Booker, the porters Bronson and Walgretor, and the mule, Hasselback. After crossing the river halfway through the first day of travel, they set Armstrong and Booker to work felling some trees to create an impromptu crossing for any subsequent trips. I told the players that this should last at least until there are some heavy rains, which are likely to wash the logs away.

At the end of the first day, they made camp at the bottom of the wooded black crag upon which sat Quasqueton. The night passed without incident. The following morning, they had a three-hour hike up the hill to the ruin, which was also uneventful. They plunged back inside the dungeon.

As they moved down the familiar long entrance hall, a goblin voice sounded an alarm. They saw that the goblins had fortified the landing at the top of the short staircase at the end of the hall with a makeshift wooden barricade, a bloody orc's head skewered atop a spear with an arrow through its eye.

Grilk emerged, recognized the party as friends, and told his goblins to stand down, welcoming his friends back inside. The party learned that a band of orcs had attacked the goblins in the night, with three groups of two attacking from different directions. The goblins had heard them coming, but not quickly enough to save Tobb, who had been wounded with arrows and carried off as a prisoner. Three orcs were slain in the counterattack, at which point their morale broke, and they retreated.

This was part of my restocking/updating the dungeon process. I made opposed reaction rolls for those dungeon factions which might be in contact with one another and determined that the orcs would launch an attack on the goblins. I looked at my map and figured out how they might go about that strategically and what numbers they could bring to bear, then roughly simulated it with surprise rolls, attack rolls, movement, morale checks, and the like. It was a fast and bloody conflict which did not go very well for the orcs, but they managed to take a captive.

Grilk informed the party that earlier, the goblins had managed to trap a horse-sized "demon worm" inside the closet in Zelligar's bedchamber, which was nearby. After the orc attack, the goblins had freed the worm and corralled it into the corridor which led to the orcs' territory. The party decided to explore that room.

The goblins had put lots of orc meat into their stew, in which the party declined to partake, but Robbernar did instruct Walgretor to fill a bucket with the slop in case the worm showed up. They entered Zelligar's bedchamber and found it in disarray, furniture overturned and broken down for the barricades outside - including a large bed (its frame dismantled) and some smaller tables and chairs.

The party had their men-at-arms bust open a nightstand which they thought might be trapped, but found nothing inside. The far wall was decorated with a long stone carving of a mighty wizard casting a fearsome spell. A door in the room opened onto a 30' long corridor, at the end of which was a pair of jeweled chests overflowing with coins, gems, and jewelry.

This was obviously a trap, so Robbernar used detect magic and determined that the chests and their contents did indeed radiate magic. The party was very careful. They threw a chair leg at the chests, then poked them with their 10' poles, then used a fishing hook and line to fish out a piece of jewelry, then lassoed a whole chest and dragged it out into the main room. Robbernar finally touched some of the treasure with his hands and noticed that everything felt and sounded like rocks, not precious metal and gemstones. It was a trick, but one the party might be able to leverage against their enemies.

Lastly, the party checked the closet. Inside, they found a small pile of slimy giant rat bones, partially crushed. Malkara claimed a cloak studded with circular bits of pewter. They looked over some papers detailing stronghold inventories, expenses, and construction, and found four books: a history of the local area, an Elvish encyclopedia of herbology, a handwritten notebook in an indecipherable language of runes and esoteric markings, and an illustrated tome describing meteorological phenomena, with similarly indecipherable scrawls in the margins. They pocketed the books.

That's where we left off. The party has explored a good amount of the dungeon, but there's still some ground to cover. I'm looking forward to seeing what they do next.

Developing the Sandbox: The Town

This post is part of my series on developing the sandbox. You can read other posts in the series here. The series will, when necessary, go into detail on the development of my sandbox for Dungeon Module B1: In Search of the Unknown, but will also discuss sandbox development more generally. You can read play reports from my B1 campaign here.

(It's actually surprisingly hard to find a map from a classic D&D adventure of the sort of town I'm referring to here, since so many famous modules deal in either villages or cities - very little in between.) 

What I'll be discussing in this post are settlements which are big enough that it's not worth it to spend time detailing each individual location, where it's instead better to break the whole thing down into a few thematic "chunks" (for example, "Herald's Market" and "Alchemist's Ward" in the image above, which is taken from donjon's Fantasy Town Generator).

What I won't be discussing here is smaller settlements - thorps, hamlets, and even villages - where the whole place is essentially one big area. Development of those locations will look different, and I'll discuss that in another post.

If you refer back to my post on Prioritizing Points of Interest, you'll find why I decided to detail the town of Timbershore first and foremost when developing my B1 sandbox. The road into the region leads straight into town. Unless the players decide to go around it, it's the first location they'll interact with in their adventure.

However, a town like Timbershore is a complex environment, and it will require a great deal of development. It's one of those locations, like a tentpole megadungeon, which a DM could theoretically sink limitless hours into as the campaign progresses, fleshing it out as demanded by in-game events and the players' interests.

You can read a lot about how best to utilize settlements like this in play. There was a whole blog bandwagon about it (technically this was about cities, but any advice on that subject is equally applicable to towns as well). Of course, the most important writing on this subject is my own. Or, at least, being familiar with my writing on the subject will likely aid you in understanding my perspective and this post in particular. Check out my posts on settlements in general and the implied setting of towns and cities in AD&D specifically.

So, where to start? If you refer first to my post on minimalist location sketches for sandboxes, you'll find that, knowing that this first location is a town, I'll next want to know its population, any character-type NPCs who live there, who rules the town, and how the powers that be respond to the player characters when they first arrive there.

Let's start with population. This will determine the number of character-type NPCs in the town, which will in turn determine their level distribution, which will in turn determine who are the town's "bosses". I determine population using AD&D's INHABITATION table:

A town's population is 1d6 x 1,000 + 500. For Timbershore, I rolled a 5, so its population is 5,500. You are also free to look at your hex map and determine the hectares of airable land in the region and do some realistic calculations based off of that. I hope you have fun. I'm going to go with the dice roll.

My towns and cities and such are primarily populated by humans. Large groups of dwarves and elves and halflings don't live in such places - at least not at this scale (though perhaps they used to). They are "monsters" with wilderness "lairs". But, some of them do live in human towns and cities.

To determine racial demographics, you could use this table:

Or even this table:

I prefer the first. What's the difference? The first one is 80% human, 5% dwarf, 5% elf, 2% gnome, 5% half-elf, 2% halfling, and 1% half-orc. The second is 70% human, 8% dwarf, 5% elf, 2% gnome, 8% half-elf, 2% halfling, and 5% half-orc. Do you want fewer humans and more dwarves, half-elves, and half-orcs? Use the second table. Or make up your own arbitrary percentages. It doesn't really matter.

Timbershore has a population of 5,500. That's 4,400 humans, 275 each of dwarves, elves, and half-elves, 110 each of gnomes and halflings, and 55 half-orcs. This isn't really important, except:

While this excerpt applies specifically to the recruitment of henchmen, I extrapolate from it that 1 in 100 of the human and half-orc populations and 1 in 50 of the remainder are character-type NPCs. This is, incidentally, why I prefer the demographic breakdown which yields more humans and fewer of everyone else - it's fewer character-type NPCs that I have to deal with, which is less work.

Using our demographics and this rule of thumb, Timbershore has:

• 44 human character-type NPCs
• 5 to 6 of each: dwarf, elf, and half-elf character-type NPCs
• 2 to 3 of each: gnome and halfling character-type NPCs
• At most 1 half-orc character-type NPCs

That's 63 to 69 NPCs with character levels. That's a lot, and if we were detailing a city, it'd be even more. Surely we're not going to detail them all (not yet at least - you can do it, but it's totally optional). Instead, we're going to figure out who are the important ones, and detail those first. How do we determine who's important? 

First, let's go back to that excerpt on Number of Prospective Henchmen. 1 in 1,000 people in the general population will be interested in employment as henchmen. I generalize this a bit by saying 1 in 10 of those identified character-type NPCs will be interested in employment as henchmen. For Timbershore, that's 4 to 5 humans and, at most, 1 of any other given race. 

You can roll a d10 to determine this, with the number of total character-type NPCs of each race serving as a percentage chance that one is interested in employment as a henchmen. For example, if there are 5 dwarves in Timbershore, then there's a 50% chance that one is recruitable as a henchman.

Henchman are locked in at levels 1 through 3:

For these NPCs, I'll just determine their level (1d3), and give just those of 1st level (because 2nd and 3rd level henchmen can only be recruited at higher levels) a name, a class, and an alignment, then put them aside. Nonhumans will have a chance to be multi-classed (see Race and Multi-Class above). Their alignment can be anything. I will usually have demihumans tend towards their Monster Manual alignment (dwarves will often be Lawful and/or Good, elves will often be Chaotic and/or Good, and so on), whereas for humans I'll roll 2d3 to determine it. I determine class by rolling on the Character Subtable ("Used For Encounters On All Dungeon Levels" - but ignore that!):

(Note that just because this is all AD&D stuff does not mean you must be playing AD&D. You can change out the races for whatever you like. You could rule that a result of "Ranger" is actually a 50/50 chance of being a ranger or barbarian. A result of "Illusionist" could instead be a sorcerer or warlock. These are generally very useful worldbuilding tools regardless of your prefer system. I am merely attempting to demonstrate their usefulness. Adjust to taste!)

Henchmen are one of the most important NPCs to detail first because the players have a good chance of interacting with them early, if for example they feel the need to bolster their numbers before beginning their adventure. The next most important will be the high level NPCs - those who run thieves' and assassins' guilds, rule strongholds, and command factions within the town.

These high level NPCs will be drawn from those remaining - let's say 40 humans, 5 dwarves, 5 elves, 5 half-elves, 2 gnomes, 2 halflings, and 1 half-orc. What to do with these? I start with breaking them down based on a loose notion of "tiers" comparable to D&D 5e's "tier of play". 65% are Tier 1 (levels 1 to 3, or 1d3, but not interested in employment as henchmen), 20% are Tier 2 (levels 4 to 7, or 1d4+3), 10% are Tier 3 (levels 7 to 12, or 1d6+6), and 5% are Tier 4 (levels 9 to 20, or 1d12+8).

That gives me 26 Tier 1 humans, 8 Tier 2 humans, 4 Tier 3 humans, 2 Tier 4 humans, and so on. Start with the highest level NPCs and work your way down, detailing them the same way you did henchmen. Don't generate all of them. For now, you only need to worry about those NPCs which have strongholds. They are the ones who will have large amounts of henchmen and followers to command, which means they'll be factions.

These high level NPCs and the factions they command will provide the most basic level of color to your large settlement - a town ruled by an evil wizard is much different than one ruled by a paladin, and a city with half a dozen thieves' guilds is much different than one with a like number of temples.

So then, who can rule a stronghold? Luckily, I wrote a whole post about this. At 8th level, clerics can build places of worship, and monks can build moat houses and friaries. At higher levels, fighters, paladins, rangers, thieves, magic-users, illusionists, assassins, and bards can build strongholds as well (note that druids do not build strongholds in towns and cities - you might rule that rangers, monks, and bards do not, either, according to taste).

You might prefer a system in which a character can build a stronghold at any level provided they have the means and the desire to do so, but I find that woefully unhelpful when trying to develop a location like this. I for one really appreciate AD&D's black and white approach to this. "X class can rule a stronghold at Y level" is a very useful rule of thumb for a DM trying to build a setting.

So for now, we need be on the lookout for character-type NPCs of 8th level or higher only - that's our Tier 3 and Tier 4 NPCs. When I rolled up these NPCs, I got a LN human fighter 7, a N human magic-user 7, a LG human ranger 8, a LN human magic-user 11, a CE human cleric 13, and a NE human illusionist 16 (since AD&D 2e has specialists of all eight schools of magic, I treat the illusionist instead as a specialist and roll d8 for their specialization, determining that this one is instead a diviner). 

Surprisingly, that gives me just two NPCs who rule strongholds (fighters and rangers get their strongholds at 9th level, and magic-users get theirs at 12th level). I determine that the human cleric 13 rules a large shell keep and that the human diviner 16 rules a large walled castle with a keep, using this table:

I make some reaction rolls and determine that the cleric is unfriendly and the diviner is friendly. This suggests to me at least that the cleric will be uncooperative with the party and that the diviner is willing to work with them - he is NE, so this doesn't mean he's altruistic, only that he thinks he stands to gain something by working with the party.

Remember that while the powers that be may be willing to work with the party to accomplish their goals, high level NPCs view adventurers as a threat to the established order, and none will be permitted to grow too powerful without being checked by their overlords.

This bit tells me that NPCs with strongholds will have 1d4+1 henchmen and several men-at-arms:

And this bit tells me what level these NPCs' henchmen ought to be:

The 13th level cleric will have 2-5 henchmen of at most 8th level, and the 16th level diviner will have a like number of henchmen of at most 10th level. Some of those lower level characters I rolled up (the LN fighter and N magic-user) can fill those rolls (the LG ranger's alignment is incompatible with both). Since the LN magic-user 11 is too high level to be a henchman to either and also falls into that "Tier 4" 9th to 20th level range, I'll make her an important NPC too, even if she isn't high enough level to rule a stronghold.

With my three most important NPCs identified, I'll make some contested reaction rolls to determine how they feel about one another. I determine that the diviner is aligned with the magic-user (friendly) and has an uneasy truce with the cleric (indifferent/neutral), and that the cleric wants the magic-user dead (hostile). 

I mill this over a bit and determine that the diviner and the cleric have some sort of agreement wherein the cleric is allowed to collect tithes and is exempt from taxation. Although the diviner is higher level, AD&D clerics get a small army and tons of fanatical followers when they build a stronghold, whereas magic-users must rely on paid men-at-arms and henchmen, so the cleric actually wields the more threatening force in town. The place is probably crawling with the cleric's agents. 

At the same time, the diviner possesses powerful scrying magic and is aware of everything that goes on in his domain, rarely needing to leave his stronghold at all. The cleric is certainly not a good influence in town, but her agents sow chaos and keep the people divided and frightened, which makes them easier for the diviner to control.

The magic-user is, in turn, the diviner's boots on the ground agent, carrying out his laws and orders. She is on the cusp of being able to build a stronghold, so perhaps she is dutifully serving the town's overlord in hopes that he will grant her permission to begin construction when the time comes. Since the magic-user is an agent of law and order, the cleric wants her eliminated, which may push the town over the brink into all-out chaos. She's not bold enough to take out the magic-user herself, but would perhaps align herself with some outside agents to get the job done.

I think that's a pretty good initial sketch for my town. After this, I would move on to sketching out my other locations, then circle back around for another pass at the town.

Somewhere in all of this, you're going to have to stat out these NPCs. You can wing it to a certain extent. For example, if you know an NPC is a 7th level human fighter, it's easy enough to run them at the table by simply looking up the relevant statistics when needed. The tricky part is in the finer details, like what magic items the NPC possesses, or, if they're a wizard, what's in their spellbook, and what spells they usually have prepared. 

I don't particularly like figuring all that out at the table. And if the high level wizard happens to have a ring of three wishes or some other powerful magic item, I like to know that as soon as possible, as that is potentially game-altering information. Once I'm done with my initial location sketches, I will usually do a second pass through all of my locations, during which I jot down stats for monsters and NPCs and roll up their treasure hoards. 

This can be really tedious when dealing with a place like a town where there's a high density of high level NPCs. There's not really an easy way to do it - I just put my head down and power through it. The players are unlikely to decide to try and confront some high level NPC in their first session, but players are unpredictable, so I like to be prepared in case they do.

The next most important step - really the last one to get your town ready for your first session - is to break it down into districts. This is an idea I took from Brave, and I find it really brings settlements like this together in a simple, gameable way. There's this nifty little table:

Personally, I'm only really looking at the Districts row, though you may find the rest of the information illuminating. A town has 1d4 districts. Instead of rolling, I base it on population size. Timbershore has a population of 5,500, which is about 85% of the maximum population for an AD&D town (6,500). Since the maximum number of districts is 4, I multiply that by .85 and get 3.4. Thus, Timbershore will have three districts, with a 40% chance of a fourth. I check for a fourth and determine there are indeed four districts.

I want one of these districts to be where the diviner lives and one to be where the cleric lives, so I create the High Ward - a wealthy administrative district at the town's center, with the diviner's castle atop a high hill - and the Chapel Ward - filled with hostels, almshouses, and minor shrines (remember there are no other clerics of high enough level to have actual temples here), with the cleric's religious stronghold at its center. The town is on a river and right next to the forest, so I also have a River Ward with docks and warehouses and a Timber Ward with lumber yards and sawmills.

I draw up a little flowchart map to show how the districts connect and create a brief description for each, including what general sorts of locations and people are found there:

• Chapel Ward (NE): Blanchefleur’s Keep looms over this district. Features minor shrines, cult safehouses, “charity” kitchens, mortuaries, hostels, graveyards, and catacombs. Flagellant processions make their way through the streets daily. Devotees of Blanchefleur, desperate wretches, cultists, and hospice workers live here. Connections: Timber Ward (S), High Ward (SW)
• High Ward (C): Situated atop a hill. Surrounds Lambrecht’s Castle and observatory on an even higher hill ("the High Hill"). Features markets, spell licensing offices, the courthouse, elite guard barracks, and residences of wealthy merchants and officials. Senior officials, licensed magi, wealthy traders, and administrators live here. Connections: Chapel Ward (NE), Timber Ward (SE), River Ward (S)
• River Ward (S): Noisy, profitable, and corrupt, where outsiders enter town when sailing downriver. Features docks, warehouses, guildhalls, caravan yards, river shrines, taverns and flop houses, smuggler dens, and toll houses. Boatmen, dock workers, traders, caravan guards, smugglers, and transients live here. Connections: High Ward (N), Timber Ward (NE)
• Timber Ward (SE): The working population center, closest to the forest. Features woodcutters’ lodges, charcoal burners, sawmills, tanneries, cheap inns, labor markets, militia mustering yards, fence yards and animal pens, repair shops, and the hunters’ guildhall. Laborers, trappers, foresters, hunters, and militia families live here. Connections: Chapel Ward (N), River Ward (SW), High Ward (NW)

From here, it's pretty easy to improvise where different things might be located in town. Need healing or want to visit a shrine to a deity? Go to the Chapel Ward. Want to yuck it up with the elite or auction off a captive dragon? Go to the High Ward. Want to secure passage downriver or fence some stolen goods? Go to the River Ward. Want to sell some giant beaver furs or hire a guide to help you navigate the forest? Go to the Timber Ward.

From there I can flesh out a few additional locations, like an inn in each district, a graveyard in the Chapel Ward, a bath house in the High Ward, a smuggler's den in the River Ward, a boxing yard in the Timber Ward, or whatever. I try to think of what the players might want to do in town (looking over your list of downtime activities is a good place to start), then create a few locations that make sense for each, with maybe a sentence or two of description. For example, my list of locations for the High Ward looks like this:

For now, there's just one thing left to do, and that's encounter tables. You can pretty much get by simply riffing on AD&D's CITY/TOWN ENCOUNTERS MATRIX, but I like to be a little more prepared. I generate four encounters each for day and night in each district (that's eight encounters per district, so 32 total for Timbershore). This is probably overkill - I didn't roll up any encounters in either of our sessions which took place in town, so you can probably get by with like, one prepared encounter per district per time of day, improvising as needed.

Since the encounters are by district, the district in which they occur should color the encounter. A demon or devil encountered in the Chapel Ward was probably summoned by the high level cleric herself. A high level magic-user encountered in the High Ward could be the diviner (or a projected image of the diviner) wandering the streets. A press gang makes a lot of sense in the River Ward, but less so elsewhere. Laborers in the Timber Ward might be woodcutters, but in the River Ward they would be dockworkers, and so on.

These encounters can also aid you in fleshing out the town. Some encounters will include character-type NPCs with levels. You can use these to "backfill" those other character-type NPCs which were too low-level to detail up front, adding them to your roster as you go. By doing this, I was able to determine the identities of all my remaining character-type NPCs. For example, the River Ward's roster looks like this:

One can also encounter monsters like fiends, lycanthropes, and undead wandering the streets. Because these are unusual, they're worth rolling up ahead of time. These can further flesh out the town by allowing you to add monster lairs as locations. If I roll up doppelgangers in the High Ward, I may want to add a location to that district where a wealthy family has been replaced by imposters. If there's a vampire wandering the Chapel Ward, it probably lairs in some old mausoleum where the locals fear to tread. If there are wererats in the River Ward, there's probably a sewer or an abandoned ship where they make their lair. This is all worth adding to your rumor tables as well.

When rolling up these encounters, remember to disguise them using vagueness and similarity, and that dealing with NPCs should be expensive and irritating.

We're not quite done with this town. We will need to come back and fill in any monster lairs as determined by our random encounters, and we will need to key our high level NPCs' strongholds as well. Since the party is likely too weak to confront either of these locations at the start of the campaign, we can circle back around to that later. For now, the town is probably quite ready for play.