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.

Where are AD&D's High-Level NPCs?

This is a question I've pondered quite a bit in my exploration of AD&D. Theoretically, my assumption would be that there's nothing to restrict a character-type NPC from reaching the same level of experience as a player character of the same type. That is, a human character can become a cleric of 29th level, a fighter of 11th level or higher, or a magic-user of 18th level or higher (no upper limit is provided in the latter two cases, whereas 29 seems to be the cleric's limit for whatever reason). 

Non-human NPCs would also presumably abide by the same restrictions imposed on player characters of that type, so an NPC elf could advance as far as 7th level as a fighter and 11th level as a magic-user, and no more, just like a player character. Interestingly, NPC dwarves, elves, and gnomes can be clerics, and NPC halflings can be druids, even though these options are not available to player characters:

The PHB, in the section titled THE ADVENTURE, describes three different adventure types, characterized by the setting in which they take place: Town Adventures (which include not just towns but also cities and large villages, and so are probably more accurately described as settlement adventures), Dungeon Expeditions, and Outdoor Exploration (this is, for my money, a better configuration of modern editions' "Three Pillars of Play" - which are Combat, Exploration, and Social Interaction).

The DMG provides some guidance as to what level character-type NPCs encountered in each of these scenarios will be, so I'll analyze the subject within this framework.

I'll begin with Dungeon Expeditions, because NPCs encountered in these scenarios will tend to be of lower levels. Character-type NPCs are encountered in dungeons as early as the 1st dungeon level:

They can also be encountered all the way in the deepest levels of the dungeon, and on every level in between:

Here is the relevant section describing the level of these NPCs:

Character-type NPCs encountered on the 1st through 4th levels of the dungeon are of the corresponding level in their class. Such NPCs encountered on the 5th through 15th levels are level 6 to 12. NPCs encountered on the 16th level or deeper can be as high as 13th level. 5th level NPCs are oddly never encountered in dungeons, nor are those higher than 13th level.

This suggests that 13th level is the maximum level at which player characters should continue to embark on dungeon expeditions - that's before clerics or magic-users gain access to 7th level spells, for context.

Next, let's look at Outdoor Exploration. For whatever reason, character-type NPCs are encountered in temperate and sub-tropical uninhabited/wilderness areas and nowhere else - not in arctic, sub-arctic, sub-tropical, or tropical areas, and not in waterborne environments. Not even in inhabited areas! The occasional high-level NPC is still encountered in groups of men (bandits, berserkers, brigands, etc.) and demihumans (dwarves, elves, gnomes, etc.), but full parties of fellow/rival adventurers are not.

According to the special footnote to these encounter tables, character-type NPCs in the outdoors are always of 7th through 10th level:

But random encounters are not the only way to encounter character-type NPCs outdoors. We must also account for those NPCs who rule strongholds:

Excepting the bard, which is an outlier by design, there is a good deal of overlap with those character-type NPCs found in the lower levels of dungeons and in random outdoor encounters. Only magic-users, thieves, and assassins exceed 13th level - and, as I mention in my post on Thieves and Assassins Guilds, this table is not totally reliable when used to stock an outdoor area, since the PHB states that thieves can only build small-type castles within or close to towns and cities, and only Grandfather/Grandmother assassins (15th level) rule wilderness strongholds (14th level assassins' strongholds are always in a town or city).

Where are all the higher-level character-type NPCs, then? According to the rules presented by the DMG, characters adventuring in dungeons and outdoors will never encounter a cleric capable of casting 7th level spells or a magic-user capable of casting 8th level spells (16th level in both cases), an illusionist capable of casting 7th level spells (14th level), a Great Druid (14th level), a Grandfather/Grandmother assassin (15th level), or a monk of higher than 13th level.

This has gameplay significance because players may need to be able to find such NPCs. They may need a cleric to cast resurrection, a magic-user to cast wish, or an illusionist to cast vision. Druids, assassins, and monks will need to find higher-ranked NPCs of the same type and defeat them to advance in their own class. Where are they?

As I've already mentioned, the Grandfather/Grandmother assassin should be in a wilderness stronghold. Druids, likewise, according to the PHB, "do not dwell permanently in castles, or even in cities or towns", but "in sacred groves...set in woodlands and similar natural surroundings":

While no such restriction is imposed upon higher-level monks, it seems fitting to me that these too would be (literally) cloistered away in some remote wilderness monastery. It seems to me that the highest-level druids, assassins, and monks are left off the castle sub-table merely because they are unique individuals who should be deliberately placed in the play area with careful consideration by the DM - though this rationale again is not applied to monks in the same way, who appear on the sub-table despite the fact that those of 9th level or higher are, according to the PHB, one-of-a-kind:

That still leaves the higher-level spellcasters, which brings us to Town Adventures.

In my post on henchmen, I mention that 1-in-100 humans and 1-in-50 demihumans are "suitable for level advancement". Of those, generally, 1-in-10 are "interested in offers of employment as a henchman". Since newly recruited henchmen can only be 1st through 3rd level, this means that 10% of the character-type population in a settlement is of 1st through 3rd level, but no guidance is provided for the level distribution of the remaining 90% of character-type NPCs.

The city/town encounters detailed in the DMG shed a bit of light on the level of NPCs found in these environments:

• Assassin encounters are with assassins of an indeterminate level (my guess is that one is to refer to the Thief encounter below to determine the level of the assassins).
• Bandit and Brigand encounters at night will include "1 or more leaders" (based on the Monster Manual there could be a fighter from 8th to 10th level, a magic-user of 7th to 10th level, or a cleric of 5th to 6th level, but it's unclear how to determine which leaders and of what level will be present because the Monster Manual deals with much larger groups of bandits).
• City guard encounters include fighters of 2nd to 5th level and indentured magic-users of 1st to 4th level.
• City official encounters include fighters of 1st to 4th level.
• City watchman encounters include fighters of 1st to 5th level and indentured clerics of 2nd to 5th level.
• Cleric encounters are with a cleric of 6th to 11th level and can include lesser clerics of 1st to 4th level.
• Druid encounters are with a druid of 6th to 11th level and include either lesser druids of 1st to 4th level or fighters of 1st to 6th level.
• Drunk encounters can include individuals from many of the other encounter types, presumably at similar levels.
• Fighter encounters are with a fighter of 6th to 12th level and include henchmen of 1st to 4th level.
• Gentleman encounters can be with fighters of 7th to 10th level.
• Harlot encounters may be with a harlot who is a thief or is working with one (presumably of the same level as those described under Thief, below).
• Illusionist encounters are with an illusionist of 7th to 10th level and include either illusionists of 1st to 4th level or fighters of 1st to 6th level.
• Magic-user encounters are with a magic-user of 7th to 12th level and henchman magic-users or fighters of 1st to 6th level or 4th to 7th level, respectively.
• Mercenary encounters include 1st level fighters and may include fighters of 2nd to 5th level.
• Merchant encounters include a fighter of 1st to 4th level.
• Monk encounters are with a monk of 7th to 10th level.
• Noble encounters include fighters of 1st to 4th level, and the noble may even be a fighter or cleric of 5th to 12th level.
• Paladin encounters are with a paladin of 6th to 9th level.
• Pilgrim encounters may include a cleric (or druid) of 2nd to 8th level, a monk of 5th to 6th level, a fighter (or paladin or ranger) of 1st to 8th level, a thief (or assassin) of 2nd to 7th level, or a magic-user of 6th to 9th level (as with Bandit and Brigand encounters, it's not entirely clear how to determine which of these character-types will be with the group, as the Monster Manual details much larger groups).
• Press gang encounters will be with 1st level fighters with a leader of 2nd to 5th level.
• Rake encounters will be with fighters of 5th to 10th level.
• Ranger encounters will be with rangers of 7th to 10th level.
• Ruffian encounters will be with 2nd level fighters and may include an assassin of 5th to 8th level.
• Thief encounters will be with a thief of 8th to 11th level and may include lesser thieves of 1st to 4th level.

So, that's a lot of information that I probably didn't need to type out. Do you notice anything? The highest-level character-type NPC which can be randomly encountered in a city or town is of 12th level (and that's only individual fighters and magic-users and noble clerics). Again, where are the higher-level NPCs?

Well, they must be somewhere. The DMG contains the following list of prices for hiring NPC clerics to cast certain spells:

(Interestingly, there is no such list for druid, magic-user, or illusionist spells, suggesting that while clerics are probably happy to accept tithes to their faith in exchange for service - if the character requesting such is "of similar alignment and religion as the cleric" - those other classes will probably only provide their services in exchange for magic items or service in kind.)

This list includes 7th level cleric spells like astral spell, control weather, earthquake, gate, regenerate, and restoration. Higher-level NPCs are present in the world (and specifically in large enough settlements), but for some reason are never randomly encountered in dungeons, outdoors, or in settlements.

This suggests that such NPCs are essentially retired from adventuring (they do not go into dungeons or the wilderness) and insular (they do not leave the place in which they reside to wander the streets of towns and cities). They have to be deliberately sought out where they live, and they probably won't be happy to see you.

So where do they live? Clearly not in dungeons, and apparently not in wilderness strongholds. However, while there are certainly domain-less high-level NPCs in D&D, I have a hard time believing that retired, insular, very high-level NPCs would be without them. 

Rather, I suspect that these individuals have ruled their strongholds for a long time - long enough that they are well a part of the established order. Strongholds maintain order in the borderlands, attracting people to settle around them in tax-paying single dwellings, thorps, hamlets, and the like. Eventually, I imagine, those communities grow into villages, towns, and even cities, with those old strongholds at their core. Is this where the highest-level NPCs live?

I think so. I think cities especially probably have lots of strongholds within them - those may be ruled by NPCs like those found on the castle sub-table in the DMG, but I imagine others will be ruled by even higher-level ones. Unfortunately, there's no guidance as to what the distribution should be.

In my post on settlements (which leans towards 5e), I stated that the character-type NPCs in my settlements are broken into groups corresponding to the "Tiers of Play" - 65% will be Tier 1, 20% tier 2, 10% Tier 3, and 5% Tier 4. In a later post, I wrote about how a similar "Tiers of Play" framework might be applied to AD&D (specifically 2e). While in that post I found that the "Tiers of Play" framework did not clearly map onto AD&D, I do still find it helpful to consider here.

In my post on henchmen, I provided an example AD&D city of 35,000 people with the following character-type NPC demographics:

• 280 humans
• 35 dwarves
• 35 elves
• 14 gnomes
• 35 half-elves
• 14 halflings
• 3 or 4 half-orcs

Since the total is 416 or 417, and 10% of these will be henchmen of 1st to 3rd level in accordance with the AD&D DMG's rule of thumb, that leaves 374 to 376 character-type NPCs to be distributed as follows:

• 243 to 245 low level (level 1 to 3 but not interested in work as henchmen)
• 74 to 76 medium level (level 4 to 7)
• 37 to 38 high level (level 7 to 12)
• 18 to 19 very high level (level 12 to 20 - or 20+ in AD&D 1e's case)

We can use AD&D's Character Subtable for dungeon encounters to get some idea as to what classes those very high level characters are:

Since this is for dungeon encounters, it may not apply to cities exactly, but it's what we've got. It's also worth noting that I'm not taking race into account, which would affect the distribution somewhat. Dwarves can't be magic-users, for example, so allowing all of these NPCs to be of any class will result in skewed numbers - but this is just a broad overview.

We can expect the average AD&D city to have the following very high level character-type NPCs in residence: 3 or 4 clerics, 7 or 8 fighters, 0 or 1 paladins, 0 or 1 rangers, 3 or 4 magic-users, 0 to 1 illusionists, 1 or 2 thieves (each of which might have their own guilds or gangs at war with one another), 0 to 1 assassins (but no higher than 14th level), and 0 to 1 monks. There will be no druids, because druids do not live in cities, and you might opt to rule the same for rangers and monks depending on taste (this is how I'm leaning).

While the demographics I'm using to divide the NPCs into "tiers" are entirely of my own invention and not based on anything in particular, I think they work reasonably well for the setting AD&D is depicting. The assumption seems to be that player characters will be able to find very high level NPCs in large settlements if they go looking for them, and these numbers support that. It also makes sense that player characters will be bumping into NPCs of up to 12th level in the streets, since there are a decent number of them living in the settlement.

Hopefully this is helpful to someone, and hopefully I've demonstrated that this is not just worldbuilding and number crunching for its own sake, but that these numbers actually have gameplay implications and are important to consider when running a game. What are the chances that a high level cleric of the player character's alignment lives in the city? What is the political landscape among the city's fighter lords? If the party has earned the ire of one powerful magic-user, is there another they can call upon for powerful magical aid? How many Thieves Guilds are active in the place, what is the status of their internecine turf war, and which would it be best to align with?

The player characters are not the most important or powerful individuals in the setting, and they may never be. I for one think it's important to know who those individuals are, how many of them there are, where they can be found, what they want, and most importantly, how the player characters might have to deal with them.

Alignment Languages are Socially Repulsive

In my last post on languages in AD&D, I teased that alignment languages were a topic worthy of their own post. Was I right? Let's find out.

It's fitting, for context, to start with original D&D to understand the intended function of alignment languages:

Here we see that alignment languages are treated much the same as other languages in terms of their function as a tool of communication, except that creatures who do not speak a particular alignment language will nonetheless "recognize a hostile one and attack". This I take to mean that if you speak Law to a creature aligned with Chaos, the creature will not understand you, but will recognize the language and recognize you as an enemy. If you instead speak Neutrality to the same creature, it will neither understand nor attack (and may not even recognize the language at all).

Now let's switch over to AD&D (specifically the PHB, in the section on CHARACTER LANGUAGES), where, in keeping with that edition's extremely thorough list of languages (and many other things), the number of alignment languages has exploded to nine:

There are two things to come back to here - the parenthetical about The Assassin and the fact that a character who changes alignment no longer knows the language of their previous alignment. For now, we'll leave these be.

Continuing on in that same section of the PHB, we have one of the more interesting tidbits:

"Open alignment speech will be frowned upon as a serious breach of social etiquette." Huh. That's kind of similar to the OD&D thing where a creature attacks if you speak a "hostile" alignment language in its presence, except here that specific rule of thumb is made a lot more general. Any open alignment speech, regardless of the audience, is a breach of social etiquette. Why is that?

Later in the PHB, in the section on THE ADVENTURE, the use of alignment languages is described as "socially repulsive":

Furthermore, "questions about rank, profession, god and alignment are perilous". In my post on henchmen, I mentioned that, during the recruitment process, "direct questions about alignment and religion are usually taken poorly" or "amiss". Interesting.

First, let's try to grasp what alignment languages actually are. Here we will move on to the AD&D DMG:

There's a lot here, but the most important part - the actual definition of alignment language in AD&D - is highlighted, namely that they are sets of "signs, signals, gestures, and words which intelligent creatures use to inform other intelligent creatures of the same alignment of their fellowship and common ethos". Well, that makes sense. 

But...speaking alignment language openly is "socially repulsive" and "a serious breach of social etiquette." This same paragraph says that "alignment languages are NEVER flaunted in public" and "are not used as salutations or interrogatives if the speaker is uncertain of the alignment of those addressed." How then, if it is so taboo to speak openly, is alignment language meant to be used to inform others of "fellowship and common ethos"?

It is worth noting that Gygax compares alignment language to Thieves' Cant (more on that later), Latin (as used by the Catholic Church), and the languages of secret organizations and societies. These example languages include "recognition signs" and "recognition phrases" and are used "as a common recognition and communication base". Basically, you use alignment language to identify people who are of the same faction as you. So why is the use of alignment language so taboo unless you're already certain of the recipient's alignment?

Gygax goes on to clarify that "alignment languages are of limited vocabulary and deal with the ethos of the alignment in general, so lengthy discussion of varying subjects cannot be conducted in such tongues". The next paragraph expands on this:

Again, alignment language is identified as a means to recognize like-aligned creatures. Again, Gygax says that the language is used "to discuss the precepts of the alignment in detail" and that all other communications in the language will be rudimentary at best. You can discuss the ethos of Lawful Good with a dwarf, but you can't use the language to plan a counteroffensive against the orcs.

Here too Thieves' Cant (and Druidic) come up. It's interesting that these are lumped in with alignment languages. This is because they are "specialty tongues" which, like alignment languages, are used exclusively to discuss those topics which are within their sphere of interest - crime and nature-y stuff, respectively. Presumably, like alignment language, it is considered rude to speak them openly, but they can be used to recognize members of their respective factions. Also presumably, if a druid converts to some non-Neutral deity or a thief forsakes skullduggery, they would likewise somehow suddenly lose the ability to communicate in those tongues.

The next paragraph describes the consequences of open alignment speak more thoroughly:

As in OD&D, a creature which is hostile to the alignment which is spoken will recognize the speaker as an enemy (although, unlike in OD&D, they will not necessarily immediately attack). However, even creatures of the same alignment will refuse to associate with the speaker out of embarrassment, because speaking the language openly is "unmannerly, rude, boorish, and stupid."

The final paragraph in the section clarifies the situation in which alignment language is to be used:

The discussion about which creatures can actually speak their corresponding alignment language is not particularly germane to this post. Suffice it to say that a creature of a given alignment can only speak its alignment language if its Monster Manual entry explicitly says that it can. Thus, just as blink dogs do not speak Lawful Good, neither do animals speak Neutral, nor do undead speak Chaotic Evil.

What picture is being painted by all this information? Alignment language is a set of signs, signals, gestures, and words which allow creatures of the same alignment to discuss the concepts pertaining to that alignment and little else. It is used as a tool of recognition, but never publicly, and only after communication has been established between the two parties. 

One does not simply walk into a tavern and ask in the language of demons, "Anyone else Chaotic Evil?" Rather, you scope out the place first, recognize the Evil High Priest in the dark corner bearing the subtle indicators of a follower of Orcus, pull up a chair and introduce yourself, and, when you're absolutely certain that no one else can overhear, offer them a Sign of some sort. You might discuss the latest schism in the faith presented by the anti-pope's recent declaration, but you wouldn't use that language to negotiate the cost of a raise dead spell or to recruit the priest as a henchman.

Think of alignment language as ideological jargon. If you're well-versed in Marxist political philosophy (Lawful Good), you can probably carry out a conversation with someone similarly educated at a party, but a third person who is not fluent may have a hard time following along. Even that same well-versed person might be put off if you launch right into theory without introducing yourself first. If you start loudly talking about it to anyone and everyone who will listen, you will put people off. Someone who otherwise agrees with your ideology may nonetheless be hesitant to associate with you out of embarrassment. If you do it at Thanksgiving with your conservative family, they will probably become hostile towards you.

But it's not exactly a 1:1 comparison, because you can only speak an alignment language if you are of that alignment. You can't speak it by merely being familiar with its ideas, and you lose the ability to speak it if your alignment changes. But in real life, a fascist can learn Marxist buzzwords, and a Marxist can become a fascist and still speak the language of Marx.

Or can they? Speaking alignment language is basically a means of determining if someone is truly of a given alignment. It's a test. The fascist might be fluent in Marxism, but if you have a long enough conversation with them, you'll probably figure out that they don't really believe it, or perhaps don't understand it. They don't embody it. They're not a true Marxist. They're a fascist.

Perhaps you are Lawful Good, trying to pass yourself off as a follower of Orcus to learn what the Evil High Priest is up to. Since you're both Chaotic Evil, the priest switches to speaking in that tongue. He's not speaking an unintelligible Black Speech - he's talking (maybe in what we would call the common tongue) in a ideological jargony way about eating babies and having sex with zombies (or whatever it is they do). You're familiar with the ideas of Chaos and Evil so you're like "Yeah, haha, I love doing that stuff too", but you can't really hang in the conversation for long. The priest will know you're not what you're pretending to be. He's testing you to see if you know the Signs.

In that sense, alignment language makes a lot of sense to me. It's off-putting to most people, but when employed with tact, it can be used to identify your true allies and reveal hidden enemies. This is maybe best represented when we dig into the assassin (I said we would get back to it), specifically the assassin's ability to learn alignment languages other than their own:

High level assassins are the only ones capable of speaking the language of an alignment to which they don't belong (including Druidic and Thieves' Cant, which are again lumped in with alignment languages), and thus the only ones able to pass themselves off as belonging to an alignment other than their own when tested in this way. Since assassins are the masters of disguise and subterfuge, this makes a lot of sense. They're kind of like the COINTELPRO/CHAOS agents of AD&D - the rats who will infiltrate your movement, give your plans away to your enemies, and kill your ideological paragons. And just like those agents in real life, their alignment can only be Evil.

How you feel about alignment language will of course be dependent on how you feel about alignment. It is a famously difficult concept to grasp. Is alignment a cosmic faction of which your character is a part? Is it a set of ideological guiding principles which influence their behavior? Is it their religious beliefs? Is it simply the distillation of who they are? Is it what they do? Is it their position on a graph?

It's kind of all those things! From the DMG:

Here alignment is described broadly as an "ethos of thinking". It does not "dictate religious persuasion", but "religious beliefs will dictate alignment". For NPCs, alignment determines behavior, but for PCs, behavior determines alignment. Alignment not only defines/is defined by behavior and religious beliefs, but also groups creatures basically into factions which are (usually) friendly, unfriendly, or neutral based on their alignments' relation to one another, similar to OD&D, where alignment is simply a "stance" the character takes:

All of this, of course, you're somehow supposed to graph:

I'm not graphing shit, Gary!

Fortunately, despite providing definitions and examples for each of the nine alignments, Gygax gives us permission (not that we need it) to decide for ourselves what exactly is Lawful, Chaotic, Good, or Evil:

So yes, you can simply ignore that one forum post where Gygax says it's Lawful Good to kill orc babies or whatever. Or not. It's your game.

Even after you're done grappling with alignment, you may still struggle with alignment language. Why does the Lawful Good paladin who kills an orc baby and becomes a Lawful Neutral or Lawful Evil fighter suddenly lose the ability to speak of the Lawful Good ethos with his former contemporaries? Can the truly Lawful Good somehow recognize that the fallen paladin no longer truly believes that killing children is wrong? This is exactly what AD&D suggests.

And why is the Lawful Good paladin unable to converse in the tongue of Lawful Neutral or Neutral Good? Can the paladin not at least discuss the precepts of Law and Good, if not those of Neutrality? If we view alignment language as less a full language and more a set of codes for recognizing what someone actually embodies - which is again what Gygax suggests - then I would argue that the Lawful Good paladin can converse with a Lawful Neutral or Neutral Good creature in the language of those beliefs which they have in common. The two would simply come to recognize that they do not fully agree with one another and thus are not of the same alignment (although not necessarily enemies).

To simplify this, and because I don't like the nine alignment languages, I would reduce the number to five - Law, Chaos, Good, Evil, and Neutrality. Most creatures can converse in two of these languages, True Neutral creatures being the exception (not sure how I feel about this - should there be an Order-Neutral language and a Moral-Neutral language?). Lawful Good creatures can converse with Lawful Evil creatures about the ideology of Law, or with Chaotic Good creatures about Goodness.

I see no reason why a character shouldn't also be able to speak in the language of a former alignment. The Lawful Neutral fighter who was once a paladin still knows the language of Good, but actual Lawful Good creatures would recognize him as not being truly Lawful Good. If you conceive of alignment language as something bestowed upon the character by some higher power - which is kinda sorta what AD&D suggests - then this probably seems wrong, but it preserves the originally intended gameplay function of alignment language while reducing cognitive dissonance in the fiction.

If there's a serious drawback to the use of alignment language, it's that deception is much more difficult. You might be able to convince the Evil High Priest that you too are Chaotic Evil, but if he is wise enough to test you to see if you know the correct phrases (and he is a high level cleric, so he is likely wise), you have no hope of tricking him (unless you're an assassin, in which case you are Evil by necessity, if not Chaotic, although this is treating the example too literally). This reduces player agency, as subterfuge is suddenly off the table.

Then again, detect evil/good and know alignment are spells in AD&D, and they are available at low levels, so this kind of deception is tricky anyway (the availability of those spells also makes this use of alignment language somewhat redundant, but then again not every creature is a spellcasting cleric). I would err on the side of allowing characters to bluff proficiency in alignment languages, rather than treating them as a totally binary detection system. 

Just as I allow characters to more or less "get by" in languages with which they're not proficient, I imagine a Good character would have some familiarity with the beliefs of Evil and might know some of the correct phrases. This would work well with a skill-based system (to determine how well the character passes as the desired alignment), or in an old school system you could adjudicate it using reaction rolls.

Another option is to forego alignment language entirely as presented in AD&D - that is, as a language an intelligent creature automatically knows by nature of their alignment. Instead, the languages of Law, Chaos, Good, Evil, and Neutrality are simply additional languages characters can learn. Gandalf is not Evil, but he knows the language of Mordor all the same. This is similar to how modern editions of D&D handle it. There are no languages named Law or Chaos, Good or Evil, but there is Celestial, Infernal, and Abyssal - languages generally spoken by creatures of a particular alignment but which anyone can learn.

Hopefully I've demonstrated an interesting use case for alignment language and have shed some light on the way in which it was originally conceived. Like alignment itself, I feel that alignment language has gotten a bad reputation as a weird AD&D thing and has been doomed to be the butt of jokes made by people who learn about it from memes and never really try to understand what its purpose originally was. I've come to find that alignment language is actually quite interesting, and my gears are churning trying to figure out how I'd use it in my games.

Warring Guilds and Rogue Assassins Plague the Streets of AD&D

Thieves guilds and assassins guilds are ubiquitous in AD&D cities and towns. Both assassins and thieves are found on the DMG's city/town encounters matrix:

The assassin entry tells us that not only are cities and towns assumed to have a Thieves' Quarter, but also that (many) assassins may be encountered "at the guild":

The thief entry tells us that "other thieves encountered will be on guild business":

Thieves can "build a tower or fortified building of the small castle type for their own safety", but only within or very close to a town or city:

This calls into question why thieves are listed as potential masters of castles - including medium and large ones - in Appendix C within the procedure for generating fortresses in uninhabited areas:

...but this is certainly not the only contradiction between the PHB and DMG, and isn't particularly relevant to this post, so I won't dwell on it.

What is relevant is that if a thief does build a stronghold, they may use it as a headquarters for a gang, and in so doing attract thief followers. However, if they do so they will "bring the enmity of the local Thieves Guild". This will result in a war which ends "only when and if all the Master Thieves on either or both sides are dead, or if the thief character removes to another locale" (where, presumably, there is simply another Thieves Guild to contend with):

Similarly, the PHB tells us that "most towns and cities" are home to an Assassins Guild which "controls an area of from 10 to 100 miles radius around the headquarters town or city", and that "all non-player assassin characters are members of such guilds". A player character assassin who is discovered in a guild area will be invited to join the guild. They may decline, but will be sentenced to death if they perform an assassination while not a member of the guild:

Later, we are given more information about the nature of the Assassins Guild headquarters, which is "always within a large town or big city", but must be "nondescript" and safeguarded so as to avoid "attention and unwanted notoriety":

Compare that to this paragraph from the section of the DMG on poison, which describes the Thieves Guild as "an accepted part of communal society", so long as they stick to doing things that thieves do. The Thieves Guild can provoke the enmity of the Assassins Guild by engaging in acts which bring unwanted attention (or unwarranted blame) to the latter, or which are seen to step on the latter's toes:

This suggests that while the Thieves Guild may be at war with an upstart gang outside of their organization, and the Assassins Guild may condemn to death an outsider carrying out unsanctioned killings, the two organizations may also be in conflict with each other.

Unlike the thief, who can simply start a gang but must then go to war with the local guild, an assassin must assassinate or otherwise kill the local Guildmaster Assassin to attain 14th level (and later do the same to the Grandfather of Assassins to attain 15th level):

The headquarters of the Grandfather/Grandmother of Assassins can take a variety of forms but, unlike the local guild headquarters, is obvious and "must be located well away from all communities":

(Here we have another inconsistency with Castle Sub-Table II.B. That table lists the assassin as a possible master of the castle, but they can only be 14th level. 14th level assassins are Guildmasters, though, and their headquarters "is always within a large town or big city". As Grandfathers/Grandmothers - 15th level assassins - are the only assassins who build headquarters in uninhabited areas, it would be more accurate for the assassin on Castle Sub-Table II.B. to be of 15th level. But I digress yet again.)

The local Guildmaster Assassin commands 7-28 lower-level assassins. If the Guildmaster is killed, 75% "leave the area" (presumably to defect to a Guildmaster in another city or town) and new (1st level) assassins eventually arrive to fill the ranks:

The DMG goes on to state that the Grandfather/Grandmother of Assassins always has 28 followers of 2nd through 8th level, plus 4-16 of 1st level, and that when they are overthrown, as with the Guildmaster, 75% of these followers desert:

Since Guildmaster Assassins are referred to as "the local Guildmaster Assassin", whereas the Grandfather/Grandmother of Assassins is referred to as "the Grandfather/Grandmother", I'm going to conclude that there is only one such Grandfather/Grandmother of Assassins in the entire world, similar to The Great Druid:

It is my conclusion then that those assassins who desert when the Grandfather/Grandmother is supplanted do not return to the guild or defect to another but instead become masterless rogue assassins who wander the world, employing their skills in unsanctioned killings (thus making themselves persona non grata within the newly established order). 

They may even plot to overthrow the new Grandfather/Grandmother (the player character) once they've accumulated sufficient power. This would require reengaging with the guild structure which, since they've already deserted, may not be the intention. In any case, the player character assassin would be wise to closely monitor any new Guildmasters within their domain!

The assumption that there is only one Grandfather/Grandmother of Assassins and that assassins who desert in the wake of one's thus killing leave the guild forever is at odds with the previous declaration that "all non-player assassin characters are members of such guilds", but I've already demonstrated Gygax's tendency towards contradicting himself. Gygax may also simply be describing the status quo at the time when campaign play begins, not accounting for every possible action the player characters might take during the course of a campaign.

Whether this is another contradiction or I'm simply over-interpreting or making a false assumption, the idea of rogue assassins roaming the world is at least a fun one with interesting gameplay implications, so I'm willing to accept it at face value.

All of this, in fact, is highly gameable information. A player character thief may start a gang and provoke a war with the Thieves Guild. A player character assassin may decline to join the Assassins Guild and be condemned to death for unsanctioned activities, or join the guild and climb the ranks to Guildmaster or even Grandfather/Grandmother and so have rival (and even guildless) assassins plotting to kill them. But also keep in mind that anything the player characters can do, NPCs can do too.

Whether or not there's a player character thief or assassin in play, these organizations are assumed to be present in all AD&D towns and cities and will be operating in the world on their own terms. The player characters may be unlucky and run into thieves or assassins in the streets of the town or city, but perhaps they can also help out the local guild by toppling the upstart gang or capturing or slaying the rogue assassin. Perhaps the players help the upstart gang of thieves or mistakenly hire a rogue assassin, thus earning the enmity of the respective guilds. Perhaps they employ poison in some high-profile skirmish in the city, and now the Assassins Guild is out for their blood.

In summary, pit your guilds against each other for control of the streets, and let your player characters get mixed up in the conflict or caught in the crossfire. Remember that all dealings with thieves and assassins are wrapped up in layers of intrigue, subterfuge, and esoteric rules. The guilds are everywhere, and something as simple as helping or hiring the wrong person or putting a bit of poison on your blade may be seen as the opening salvo in a protracted war with AD&D most dangerous organizations of burglars and cutthroats.